IV Infusion Compatibility Auto-Compiler, or Enhanced Compat Mode Revisited.

If you would like to check out IV Infusion Compatibility Auto-Compiler, please click here.

A year ago I built a web app called Enhanced Compat Mode that checked IV infusion compatibilities and came up with the best combinations of 2 to 3 main infusions across as few IV lines as possible and with considerations made to secondary infusions. Nurses could use it to figure out how to organize compatible IV infusions on the fly. Furthermore, I wanted to rename it to something more sensical: IV Infusion Compatibility Auto-Compiler.

The app differentiated between two modes:

• Stacked Mode maximized lumen efficiency by allowing incompatible secondary medications to share a line, provided they were each compatible with the primary infusion and scheduled to run sequentially rather than simultaneously. For example, a patient receiving vasopressin, potassium phosphate, and magnesium sulfate can have all three infusions assigned to the same lumen in Stacked Mode. Although potassium phosphate and magnesium sulfate are incompatible with each other, they are both compatible with vasopressin, and since they will not run at the same time, they can share the lumen.

• Separated Mode prioritized safety by requiring all medications within a lumen to be mutually compatible, isolating incompatible drugs into different lumens entirely. For example, a patient receiving norepinephrine, vancomycin, and piperacillin-tazobactam (Zosyn) can only have compatible medications grouped together. If vancomycin and Zosyn are incompatible, Separated Mode will automatically assign them to separate lumens to prevent any risk of adverse interactions. This approach is particularly useful for patients requiring multiple antibiotics or electrolytes administered simultaneously.

In the original iteration, Stacked mode worked well, however, Separated mode was not reliable. The problem was a greedy algorithm. 

For example, Vasopressin, Magnesium Sulfate, and Potassium Phosphate. Vasopressin is compatible with either of the other two, but Magnesium Sulfate and Potassium Phosphate are incompatible with each other. The optimal output is two lumens with Vasopressin paired with one of them while the other on its own line. But the greedy algorithm consistently separated all three into isolated lines, because an early pairing decision blocked the better arrangement.

Compatibility data was derived from resources like LexiComp, Pepid, and Stabilis. Then the compatibilities were manually recorded in a Google Sheets file with values of 0, 1, or 2. Values of 0 meant incompatible or unknown; 1 meant variable; and 2 meant compatible.Utilizing an AppsScript, I was able to convert the data to JSON format. The resultant JSON file was separate from the code and gave the app scalability.

Google Sheets file.

JSON data.

New data could be added or old data updated without touching the code. The 1s were particularly tricky, since variable compatibility often comes down to concentration or diluent. I handled this with a Strict mode toggle, or a conservative (Risk-Averse) output that treated 1s as incompatible, and a liberal (Standard) output that included them but flagged them for further investigation. Categories in the metadata (vasoactives, sedatives, etc.) were mostly for the UI, making it intuitive for nurses to select medications with the exception of "dedicated" drugs, which were those that could only run alone.

The greedy algorithm problem sat there for a while. After a year of growth in programming, I came back to it and finally saw what was really going on. This is a graph coloring problem. 

Source: Wikipedia

Each medication is a node, edges connect incompatible drugs, and the goal is to assign the fewest "colors" (lumens) possible so that no two incompatible drugs share one. More precisely, it's a clique cover problem, or an issue with partitioning medications into the minimum number of groups where every drug in a group is mutually compatible.

I ditched the guess-and-check approach for a backtracking algorithm. Instead of grabbing the first arrangement that looks okay, the app now explores the entire decision tree. It starts by aiming for the absolute minimum number of lumens and tries to slot each medication in while cross-checking compatibility with everyone else already in that group. When it hits a dead end where no more drugs fit, it backtracks to the last successful step and tries a different path. To make this efficient, I added a heuristic that sorts medications by their number of incompatibilities first. By tackling the troublemakers early, the algorithm kills off impossible combinations immediately rather than wasting time on setups that were never going to work.

I highly recommend Learning Functional Data Structures and Algorithms by Atul S. Khot and Raju Kumar Mishra if you are interested in reading more about this approach.

I kept the strongly suggested maximum of three continuous infusions per line. While the backtracking algorithm is capable of organizing up to five medications per line, the Auto-Compiler is programmed to prioritize a "Rule of Three" safety threshold. This recommendation is rooted in clinical research showing that beyond three concurrent infusions, the risk of unpredictable pH shifts and micro-precipitation increases exponentially. Most binary compatibility data (1:1 testing) simply cannot account for the volatile chemistry of multi-drug cocktails (Keum et al., 2024).

In clinical chemistry, the transitive property does not exist. Just because Drug A is compatible with Drug B, and Drug B is compatible with Drug C, it does not guarantee that A, B, and C can safely coexist in the same line simultaneously. Most data is derived from binary 1:1 testing, which cannot account for the cocktail effect where the cumulative pH shifts of three or more medications may finally overwhelm the solution's buffering capacity and cause a drug to precipitate (Keum et al., 2024).

A critical mechanical constraint is the manifold dead-space volume. This is the stagnant fluid trapped within connectors, valves, and Y-sites that is not actively participating in the primary flow. In a complex manifold, this dead space becomes a reservoir for concentrated, high-alert medications. As you add more infusions to a single lumen, you increase this stagnant volume, which creates two major risks: accidental bolusing and delivery lag. If a carrier rate is adjusted or a line is flushed, that trapped volume of vasopressors is suddenly shunted into the patient, potentially causing harm to the patient (Lovich et al., 2013).

I updated the Stacked Mode algorithm to better reflect clinical reality by treating secondary infusions as sequential rather than concurrent. This change removes the max medications limit for secondaries, allowing an unlimited number to share a lumen provided they are compatible with the primary driver.

I updated the Stacked Mode algorithm to include a context-aware dynamic carrier fluid placeholder, which visually represents the primary saline infusion required for sequential piggyback administration. This feature allows the tool to better align with diverse hospital protocols by automatically injecting a "Normal Saline Primary for Piggybacks" label into any lumen grouping containing a secondary infusion. The UI remains clean by only displaying the toggle when Stacked Mode is active and secondaries are selected, and the system is now fully reactive, instantly updating the results table as the setting is adjusted.

Revisiting this project a year later has been a lesson in how software, much like clinical practice, is rarely a "finished" product. Identifying the flaws in my original logic and implementing a backtracking algorithm wasn't just a bug fix; it was about maturing the tool to match the actual complexity of my job. This evolution highlights that our programs should grow alongside our expertise. Building this was an iterative process that reflected my own growth as both a developer and a clinician.

References

Cassano-Piché, A., Fan, M., Sabovitch, S., Masino, C., Easty, A. C., Health Technology Safety Research Team, & Institute for Safe Medication Practices Canada (2012). Multiple intravenous infusions phase 1b: practice and training scan. Ontario health technology assessment series, 12(16), 1–132.

Lovich, M. A., Wakim, M. G., Wei, A., Parker, M. J., Maslov, M. Y., Pezone, M. J., Tsukada, H., & Peterfreund, R. A. (2013). Drug infusion system manifold dead-volume impacts the delivery response time to changes in infused medication doses in vitro and also in vivo in anesthetized swine. Anesthesia and analgesia, 117(6), 1313–1318. https://doi.org/10.1213/ANE.0b013e3182a76f3b

Keum, N., Yoo, J., Hur, S., Shin, S. Y., Dykes, P. C., Kang, M. J., Lee, Y. S., & Cha, W. C. (2024). The potential for drug incompatibility and its drivers - A hospital wide retrospective descriptive study. International journal of medical informatics, 191, 105584. https://doi.org/10.1016/j.ijmedinf.2024.105584

TI-83 USB Rechargeable Mod

I really like my TI-83 graphing calculator but I am not a fan of the fact it uses batteries so I modified it to run on a rechargeable power source instead of standard batteries. The project utilizes a 3.7V lithium polymer battery which is the standard voltage for single-cell lithium cells. I paired this with an MT3608 DC-DC adjustable boost module and a lithium-ion battery charger module to manage the power flow.

The wiring sequence involved soldering the battery to the charger module and then connecting the charger module to the boost module. I adjusted the boost module to 6V to match the requirements of the calculator before soldering it to the power terminals. This step is critical because the boost converter takes the varying battery voltage and steps it up to the stable level the calculator needs to function.

I used Gorilla Tape double-sided tape to secure the PCBs and the battery to the rear housing of the battery cover. Now my calculator no longer needs batteries because I can just plug in a USB cable and let it charge.

Granted with the exposed wires and taped-on components it currently looks like something that would cause a security incident at an airport but it effectively solves a forty-year-old hardware limitation. The next phase of the project will involve a 3D printed enclosure to hide the wires.

Jazzmaster Rewire: Bypassing the Rhythm Circuit for a 3-Position Selector

Couple days ago, my Made-In-Mexico (MIM) Fender Jazzmaster suddenly went silent. This wasn't entirely a surprise; I had been experiencing intermittent issues with the rhythm circuit switch in the days leading up to the complete failure. Specifically, engaging the rhythm mode resulted in weak-to-zero output, effectively leaving me with only the lead circuit.

Despite my extensive experience and expertise in building pedals and amplifiers, my skills are quite limited when it comes to working on my guitars beyond basic re-stringing. My attempts to learn how to intonate them, for instance, have been unsuccessful. When I took this Jazzmaster in for a professional setup, the technician immediately pointed out that I had installed the bridge backward. Nevertheless, I saw this as a valuable chance to become more familiar with my own gear. I figured that even if I made a major mistake, I could rely on the numerous local guitar shops to correct any errors.

Original wiring.

Upon opening the guitar to troubleshoot the wiring, I made a somewhat impulsive decision: a complete overhaul. Or rather a reduction. During the guitar's downtime, I had been playing my Telecaster more and more, and I slowly came to the realization that I rarely, if ever, fully utilized or even truly understood the whole point of the complex rhythm circuit on the Jazzmaster. It felt like an unnecessary feature for my playing style.

Inspired by the simplicity and functionality of my Telecaster, I decided to re-wire the Jazzmaster to mimic that setup: a straightforward three-way switch offering bridge pickup, bridge and neck in parallel, and neck pickup configurations. This simplification was an upgrade in utility for me.

This re-wiring endeavor provided an opportunity to tackle several persistent issues. The initial installation of my Curtis Novak pickups had been a poor, rushed job, so I took the time to completely redo most of the wiring. For this, I used cloth-wrapped solid core wire leftover from an amplifier project a couple of years ago.

Previous technician had used cardboard as pick-up risers, which is fine. I had previously used bubble wrap.

Despite having the guitar professionally set up last year (intonation, an electronics review, and the addition of neck shims), I was still wary of the quality of the work. The technician, while having a massive online following, came across as somewhat "sketchy." A minor detail that exemplified this was his inelegant use of small pieces of cardboard as makeshift pickup risers. Though functional (I had previously used bubble wrap for the same purpose), it felt like a temporary and unsophisticated fix.

View fullsize

View fullsize

View fullsize

As part of this comprehensive upgrade, I decided to tackle the risers. Leveraging access to a 3D printer, I designed and fabricated custom, solid risers to replace the cardboard pieces, providing a more stable and professional platform for the pickups.

In the end, what started as a simple repair evolved into a significant de-customization project. This final product is a Jazzmaster that now perfectly aligns with my preferences, both in terms of wiring simplicity and component (wiring) quality.

ModuLine Revisited

This entry is in regards to a previously developed IV line management device I produced called ModuLine.

The development of ModuLine presented two primary challenges. The first was achieving a secure fit on the side rail. Given the existence of three distinct bed types within my facility (and likely countless others outside of it) finding a universal fit was extremely difficult. Ultimately, I opted for a design that accommodated two of the three bed types. To enhance securement, I incorporated microsuction tape, a material I had previously experimented with on another project.

Application of microsuction tape/padding on underside of ModuLine.

The second challenge involved the additive printing process itself. Printing the ModuLine as a single piece, in its current geometry, would have resulted in significant overhangs, which are notoriously difficult to manage in additive manufacturing. My solution was to split the component in half and reattach the pieces later. After trying various joining methods, I found that simple furniture dowels provided the most secure and effective solution. Although hammering the dowels in was necessary, it was a rapid, three-second task that guaranteed a firm and safe final assembly.

Use of dowels to create secure and stable attachment of the two halves of the 3D printed product.

Ultimately, the latest round of R&D underscored the core philosophy of the ModuLine project: practical iteration. By embracing the constraints of diverse clinical equipment and the realities of additive manufacturing, the design has once again been significantly improved. The combined use of microsuction tape for a secure bed-rail fit and furniture dowels for a robust assembly highlights a commitment to actual real-world solutions.These final refinements ensure ModuLine is highly functional and also efficiently and safely manufacturable, validating its effectiveness for use at the bedside.

Developing a Low Cost Ultrasound Phantom

Introduction

Ultrasound phantoms can cost upwards of $700 (source). This is a steep price for students, educators, or clinicians who simply want to practice their technique. I wanted to create a low-cost alternative that would let me rehearse ultrasound-guided peripheral IV (PIV) placement without breaking the bank.

While reviewing organic chemistry, I started thinking about polymers and remembered Mary Roach’s Stiff: The Curious Lives of Human Cadavers mentioning ballistic gel. Ballistic gel is a synthetic or gelatin-based material originally developed to replicate the density and viscosity of human tissue for projectile impact testing. That made me wonder if it could mimic the echogenicity of expensive phantoms like the CAE Blue Phantom. 

Come to find, many others came to the same conclusion, so there is a wealth of info online about making these low cost alternatives.

Armed with that insight, I decided to start experimenting. My goal wasn’t perfection, but progress: to build something functional, testable, and cheap. Each attempt became a prototype, teaching me something new about materials, design, and the trade-offs between realism, durability, and ultrasound performance. Here’s a breakdown of what I tried, what worked, and what didn’t.

Prototype 1

For my first attempt, I used 20% NATO ballistic gel produced from a low cost ballistic gel alternative from Fit Lane Nutrition/Amazon and opted for silicone tubing instead of latex. I had come across a tip on 5MinSono noting that latex tubing can shrivel when exposed to hot gel before it set.

I used an ice mold for shaping, which was easy and accessible, but I made the phantom nearly 2 inches (5.08 centimeters) thick. That turned out to be a major misstep and far too deep for realistic PIV insertion practice. In real patients, superficial veins are usually within 0.5 to 1 cm of the skin surface, so this block made even shallow access feel like a central line attempt.

Still, it was a useful baseline to learn from, especially about material behavior and the importance of scale in procedural realism.

Prototype 2

For the second attempt, I went with a 10% FBI ballistic gel mix and poured it into a bread pan, creating a block about 1.5 inches thick. The size and shape worked somewhat better but the depth was barely acceptable for practicing superficial vein access.

Functionally, it was solid for basic training, but the “mock tissue” lacked echogenicity. The gel appeared mostly anechoic, making it difficult to appreciate needle depth or surrounding structures on ultrasound. To improve vessel visualization, I injected saline into the vein cavity. Without it, the air-filled veins showed up like bright fog lights—far too echogenic for realistic image interpretation.

One major drawback is that the block sheared apart after just one or two attempts. The lower gel concentration made it soft and fragile. However, ambient temperature may have also played a role. The room was warm, and the gel felt noticeably softer than when it had first set, suggesting that even small temperature shifts could affect its mechanical stability. This made the block less forgiving to needle passes and more prone to tearing with repeated use.

Prototype 3

For this iteration, I went with a smaller form factor and re-melted my original 20% NATO block. This time, I incorporated cornstarch at a 10% concentration to increase echogenicity in the surrounding “tissue” (source). The goal was to make the phantom appear more realistic under ultrasound, with improved scatter and soft tissue texture. It worked, and the structures looked more like actual tissue, and the needle was easy to track on screen as it pierced and advanced through the gel.

That realism came at a cost. The block lost its translucency. In earlier versions, I could visually track the needle as it approached the vein, which helped correlate hand motion with needle tip position. With the added echogenicity, that visibility disappeared. It’s a trade-off: better imaging fidelity on the screen, but less direct visual feedback.

One additional observation: the added density made it more difficult to fully compress the mock vein with the probe. While this might not perfectly simulate real-world vein behavior, it didn’t seem to hinder needle guidance. Overall, this version struck a better balance between ultrasound realism and practical usability.

Comparisons

Prototype 2 (left) via Mindray versus Prototype 1 (right) via Butterfly.

Left is a $500 US phantom from a medical simulation company. Right is Prototype 2, which comes in at approximately $10.

Future Considerations

The bargain-brand ballistic gel I used for these prototypes is made from organic gelatin, most likely bovine in origin. While affordable and easy to work with, this type of gel isn’t truly shelf-stable. It can degrade, soften, or even grow mold over time when stored at room temperature. For short-term use or small batch experimentation, it works well. But for long-term durability and reusability, I’m looking into synthetic alternatives.

These synthetic gels come with a higher upfront cost for raw materials, but they’re still significantly cheaper than commercial medical training models. They’re also more stable, reusable, and less prone to biological breakdown. In future iterations, I also plan to experiment with graphite powder instead of cornstarch as my echogenicity agent. Graphite offers a more consistent ultrasound appearance and, importantly, it doesn’t support microbial growth, making it a better choice for models intended to last.

Building an Arduino-Based Guitar Pedal Relay True Bypass System: Part 3

In the original demonstration, I used a white noise generator to show the effect going on and off, but in vivo, the bypass didn’t work. Turns out the actual bypass wiring was incorrect and the affected white noise—the effect “on”—was most likely interference from the bypass being activated. The original idea was to replace the 555 timer in the Incandenza Bypass with an Arduino, but in practice, the bypass didn’t function properly. I had to rewire it, ultimately wiring the DPDT relay like a conventional DPDT foot switch.

Other changes I made included switching to an Arduino Nano for better space management. I also reassigned the LED control: instead of being triggered by D2, the LED is now tied to the relay. However, users can easily move it back to D2 if they want to incorporate specialty LED lighting. The LED-related code remains in the original sketch for that reason.

The final working project, along with the schematic, Arduino sketch, and a stripboard layout, is available on my site.

Building an Arduino-Based Guitar Pedal Relay True Bypass System: Part 2

Circuit Design & Prototyping

In Part 1, I outlined the concept behind this relay-based bypass system, drawing inspiration from Incandenza Bypass Demedash Effects, EarthQuaker’s Flexi-Switch, and MAS Effects’ open-source approach. Now, it’s time to dive into circuit design and prototyping, turning the theory into a working system.

As expected, the transition from concept to prototype wasn’t entirely smooth. One issue I ran into was that my initial base resistor value for the transistor was too high, preventing the Arduino from fully switching the relay. I'll cover how I debugged that problem, refined the circuit, and got the relay working reliably.

Circuit Overview How the Relay Bypass Works

The goal of this circuit is to control a DPDT relay using an Arduino, allowing both tap-to-toggle latching mode and hold-to-engage momentary mode functionality from a single footswitch. Since an Arduino’s I/O pins can’t provide enough current to drive a relay directly, I used an NPN transistor 2N2222 as a switch.

When the footswitch is pressed, the Arduino sends a HIGH signal to D7. This activates the transistor, allowing current to flow through the relay coil. The relay switches the guitar signal between bypass and effect mode. An LED indicator is used to show whether the effect is active.

Debugging the Transistor Switching Issue

Initially, I used a 1kΩ resistor between D7 and the base of the transistor, assuming it would provide enough current for the transistor to turn on fully. However, when I pressed the footswitch, the relay barely responded. Instead of getting the expected 5V across the coil, it was only producing 0.4mV, not nearly enough to activate the switch.

The Fix Lowering the Base Resistor Value

After some troubleshooting, I realized the 1kΩ resistor was limiting the base current too much, preventing the transistor from fully saturating. Since an NPN transistor needs sufficient base current Ib to allow full current flow through the collector-emitter junction, I lowered the resistor value to 220Ω, although I have tested it up to 470-480Ω, and the circuit worked.

With 220Ω, the transistor now received enough current from D7, fully switching on and allowing the relay coil to activate properly. The takeaway too high a base resistor can prevent full switching, while too low a value can overload the Arduino pin. A 220-480Ω resistor should provide a good balance.

Prototyping the Bypass Circuit

With the transistor properly switching the relay, I assembled the full prototype. Arduino D7 controls the transistor base via 220Ω resistor. Transistor collector connects to the relay coil. Relay coil is powered from the 5V output of the Arduino, allowing it to switch between bypass and effect mode. However, the schematic does include an alternative method to power the relay if the user wants to use the power from the pedal supply of at least 9V. Flyback diode 1N4007 protects against voltage spikes when the relay turns off. A normally-open, momentary footswitch is connected to D2 allows tap-to-toggle and hold-to-engage modes. LED indicator D6 tracks relay state for visual feedback.

Testing this setup confirmed smooth and silent switching. The Arduino correctly distinguished between short taps and long holds, triggering the relay accordingly. The LED synced perfectly with the effect state.

The up-to-date code, or “sketch,” for the relay bypass can be found here.

Next Steps Refinements & Final Implementation

Now that the basic circuit is functional, I’ll refine the design further. In Part 3, I’ll cover making the stripboard (and later PCB), inserting it into a pedal enclosure, and adapting the design to use an Arduino Nano for a more compact form factor. 

Building an Arduino-Based Guitar Pedal Relay True Bypass System: Part 1

Introduction

For many guitarists, the bypass system in their pedals is an afterthought—just a mechanical switch routing the signal in and out. But for those of us who obsess over the finer details of our rigs, a relay-based bypass system offers a smoother, quieter, and more flexible solution. Inspired by the Incandenza Bypass by Demedash Effects, the Flexi-Switch system by EarthQuaker Devices, and MAS Effects’ open-source work, I set out to develop my own smart bypass system that integrates both momentary and latching functionality using an Arduino-controlled relay.

The Arduino is a versatile microcontroller platform that makes it easy to integrate automation into DIY electronics projects. With its simple programming interface and wide range of available libraries, it allows developers to control components like relays, LEDs, and sensors with minimal effort. In this project, the Arduino acts as the brain of the bypass system, detecting footswitch presses and toggling the relay accordingly. Its ability to differentiate between short taps and long holds enables the dual-function switching behavior, making it an ideal choice for implementing a flexible and responsive bypass system.

This three-part blog will document the why, how, and what of this project—from concept to implementation. In this first part, I’ll cover the design philosophy, why I chose a relay over a mechanical 3PDT switch, and the key challenges I’m working to overcome.

The Problem with Traditional Bypass Systems

Most guitar pedals use one of two common bypass systems:

• True Bypass (Mechanical 3PDT Switch): Simple, passive, and reliable, but suffers from loud switching noise and potential mechanical failure over time.

• Buffered Bypass: Keeps the signal active through a buffer, but can introduce tonal coloration and is always active, even when bypassed.

While a relay-based true bypass system still follows the core principle of true bypass, it eliminates the mechanical noise of a standard switch, offering a quieter and more durable alternative. Additionally, by using a microcontroller like an Arduino, I can introduce smart switching behaviors like tap-to-toggle (latching mode) and hold-to-engage (momentary mode) in a single footswitch.

Designing the Smart Relay-Based Bypass

The main goal was to improve upon traditional bypass switching by combining the best aspects of mechanical and electronic switching. Here’s the basic concept:

• Relay Instead of 3PDT Switch:

  • A DPDT relay routes the guitar signal between bypass and effect modes.

  • Silent switching, no mechanical clicking.

  • Longevity compared to a physical switch.

• Microcontroller-Controlled Switching:

  • Instead of a standard latching switch, a momentary footswitch is used.

  • An Arduino detects short taps vs. long holds, enabling flexible control modes.

• Multi-Mode Operation:

  • Tap once → Latches effect ON/OFF (Latching Mode).

  • Hold → Keeps effect ON, releasing turns it OFF (Momentary Mode).

  • Mimics the feel of EarthQuaker’s Flexi-Switch but implemented in a DIY-friendly way.

Key Challenges & Considerations

As with any project, there are hurdles to overcome. Here are a few key considerations in the development process:

• Power Handling: Most relays require more current than an Arduino can safely supply, so a transistor switch must be used.

• Debouncing the Footswitch: Since the bypass relies on a momentary switch, proper debouncing ensures smooth and reliable switching.

• Fail-Safe Design: If the Arduino fails or resets, the relay should default to bypass mode, ensuring the guitar signal isn’t lost.

• Soft Switching for Audio Purity: Preventing pops or clicks in the signal path by ensuring smooth transitions.

What’s Next?

In the next part of this series, I’ll dive into circuit design and prototyping, breaking down the relay switching circuit, the role of the transistor driver, and how the Arduino code processes footswitch input. I’ll also discuss potential refinements, such as integrating an LED indicator for status feedback.

If you’re interested in smart bypass systems, relay switching, or DIY pedal builds, stay tuned! Part 2 will bring us one step closer to making this versatile bypass system a reality.

Enhanced Compat Mode: Optimizing IV Medication Management and Streamlining the Process of Creating Infusion Compatibility Groupings

Background

(Project is now defunct.)

Patients in the ICU often require multiple simultaneous infusions, making drug compatibility critical. Enhanced Compat Mode operates through a straightforward interface where users select medications from a predefined list. The algorithm evaluates these selections against a compatibility dataset and iteratively forms optimized groupings, ensuring no incompatibilities arise. Results are presented in an intuitive table, clearly showing which medications can be grouped together, with the goal of utilizing the fewest lumens possible. Powered by JavaScript and leveraging an external JSON compatibility database, the web app is fully expandable and can be updated dynamically to accommodate new medications or guidelines, making it a future-proof solution for evolving clinical needs.

Machinations

If you are not interested in the innerworkings of the algorithm, you can skip this part.

Algorithm

1. Input Collection: The user selects all the medications the patient is receiving, categorized dynamically into primary and secondary infusions based on metadata.

2. Behavior Selection: The user chooses between Stacked Mode and Separated Mode for secondary infusions:

   • Stacked Mode: Secondary infusions can share a lumen even if incompatible with each other, as long as they are compatible with the primary infusion.

   • Separated Mode: Secondary infusions are separated into different lumens if they are incompatible with each other.

3. Compatibility Evaluation: Using the compatibility dataset, the algorithm checks:

   • Primary Compatibility: Ensures all infusions in a group are compatible with the primary infusion.

   • Secondary Handling: In Stacked Mode, secondary infusions are only checked for compatibility with the primary. In Separated Mode, all infusions in the same group must be compatible with each other.

4. Strict Compatibility Mode (if enabled): If Strict Mode is active, only medications with well-established compatibility (highest evidence) are grouped, excluding conditional or uncertain pairings.

5. Grouping Process: The algorithm iterates through the selected medications to create groups:

   • Prioritizes grouping medications with the most compatibility.

   • Non-secondary infusions are limited to three per group.

   • Secondary infusions in Stacked Mode can exceed the three-per-lumen rule if compatible with the primary.

6. Output Generation: The algorithm generates optimized groupings, displaying each group with its assigned infusions. It minimizes the number of lumens used while respecting compatibility constraints.

Compatibility information was garnered from Lexicomp and Stabilis. Those combinations with confirmed compatibility are scored 2; those with conditional compatibility are scored 1; and all else (incompatible and untested) are scored 0.

JSON and Scalability

Compatibility data from Lexicomp and Stabilis is converted into a spreadsheet, which is then converted into a data matrices.

The JSON structure in Enhanced Compat Mode transforms raw compatibility datasheets into a dynamic and actionable format. Compatibility data, typically presented as rows and columns in datasheets, is converted into a nested JSON structure, effectively creating a matrix. Each medication becomes a key, with its compatibility with other medications stored as subkeys and corresponding values.

For example:

This format enables the algorithm to quickly evaluate compatibility relationships without manual lookups.

The metadata` section enhances scalability by providing additional context for each medication, such as its category (e.g., Sedatives, Secondaries) and aliases. This allows the app to dynamically group medications and display user-friendly names. Adding a new medication is as simple as including it in the `metadata` and `compatibility` sections, ensuring seamless integration into the UI and algorithm.

The separation of metadata and compatibility makes updates modular and efficient. New medications or changes to compatibility guidelines require only localized updates to the JSON file, without altering the core functionality of the app.

This scalability ensures that Enhanced Compat Mode remains adaptable to evolving clinical practices and datasets.

Philosophical Considerations

The Conditional Compatibility Dilemma

The original compatibility dataset, sourced from Lexicomp and Stabilis, included not only binary outcomes (compatible or incompatible) but also conditional compatibilities—pairings that might work under specific circumstances. I initially introduced a scoring system to account for these conditional interactions, assigning values of 0 (incompatible), 1 (conditionally compatible), and 2 (fully compatible). The algorithm prioritized groupings with the least contention, aiming for configurations that were both compatible and stable.

However, as I progressed, I realized this approach added unnecessary complexity. Clinicians prioritize clarity and actionable results over navigating gray areas. Conditional compatibility often depends on variables like drug concentration, infusion rate, or compounding agents, which are rarely captured in static datasets. Without this context, the scoring system risked creating false confidence and undermining the tool’s practicality. Additionally, the distinction between “conditionally compatible” and “fully compatible” rarely impacted real-world groupings.

This realization led to Strict Compatibility Mode, which focuses exclusively on well-established, evidence-based compatibility. By eliminating ambiguity, this mode simplifies decision-making and ensures the tool remains reliable, intuitive, and aligned with the needs of clinicians.

Secondaries: To Be or Not to Be…Treated as Continuous Infusions?

Secondary infusions, like intravenous piggybacks (IVPB), are typically intermittent rather than continuous, which changes compatibility dynamics. To accommodate this, the tool introduces Stacked and Separated modes:

• Stacked Mode: Secondary infusions can share a lumen even if incompatible with each other, provided they are compatible with the primary infusion. This assumes secondaries won’t run simultaneously, allowing them to “break” the three-per-lumen rule. For example, magnesium sulfate and potassium phosphates, which are incompatible with each other but compatible with vasopressin, can share the same lumen if administered at different times.

• Separated Mode: Secondary infusions expected to run simultaneously are grouped into separate lumens. This ensures incompatible secondaries are never mixed, prioritizing safety. For instance, a patient requiring multiple antibiotics and electrolytes administered at the same time would benefit from this mode.

These modes provide flexibility to reflect real-world practices while maintaining safety.

Shortcomings

While the tool offers significant benefits, limitations remain. It does not account for central venous access requirements or prioritize infusions based on clinical urgency. It also assumes all secondaries pair with a single primary (e.g., Normal Saline), without addressing scenarios involving multiple primaries. Additionally, it cannot enforce institution-specific policies, such as dedicated lumens for TPN or lipids. These limitations underscore the importance of clinical judgment in complex cases.

Future Synergy with DripMap

Enhanced Compat Mode is an offshoot of DripMap, a tool designed to visually map infusion setups. Future integration will combine compatibility optimization with visual mapping to create a comprehensive solution for IV management, dubbed FlowMatrix. Challenges like accounting for medication aliases (e.g., Levophed vs. Norepinephrine) highlight the importance of bridging technical precision with clinical usability. By addressing these gaps, the tool aims to offer clinicians an intuitive, end-to-end solution for managing complex infusion setups.

Flowchart for FlowMatrix.

Conclusion

This project illustrates the delicate balance between technological innovation and practical clinical application. By evolving from a complex scoring system to a streamlined, evidence-based approach, Enhanced Compat Mode prioritizes clarity and usability, making it a valuable tool for optimizing IV management. While limitations remain, the lessons learned lay the foundation for future iterations, ensuring the tool continues to grow alongside the needs of clinicians and patients.

Simple Fick/Thermodilution CO/CI Calculator

Note: An updated version of this calculator can be found here.

Disclaimer: This Fick/Thermodilution cardiac output/index calculator is intended for educational purposes only and should not be used as a diagnostic tool or a substitute for professional medical advice, diagnosis, or treatment. Use of this calculator is at your own discretion and risk.

Body Surface Area (BSA, m²):

Oxygen Saturation of Arterial Blood or Oxygen Saturation as Detected by the Pulse Oximeter (SaO2 or SpO2, %):

Mixed Venous Oxygen Saturation (SvO2, %):

Hemoglobin (Hgb, g/dL):

Calculate
Clear

Results

Cardiac Output (L/min):

Cardiac Index (L/min/m²):

If you have taken the time to utilize this calculator, I kindly ask that you consider filling out a brief survey.

Your input is incredibly valuable to me and will significantly contribute to the development and enhancement of future projects. Thank you for your participation and support.

Kintsugi, revisited: a JFET overdrive based on the SSBS Mini with a dynamic sag circuit based on the SSBS F-k overdrive.

A year ago I became really determined to figure out the dynamic sag portion of the SSBS Fk Overdrive. At the time, no readily available schematic of the dynamic sag circuit existed other than a rudimentary diagram The Tone God had shared. Therefore, I had to deduce what I believed was the circuit based on numerous blog posts by Brian from SSBS in conjunction with tracing photos of the PCBs from the various pedal forums and Instagram. As Dino from DeadEndFX would later inform me, my final result was off by 2-3 parts and one incorrect signal path. Not bad for surmising a circuit with the limited resources available to me.

I stuck that dynamic sag circuit into a SSBS Mini, which is in turn a derivative of the SSBS Fk Overdrive. The final product sat on my pedal board in an unmarked, hammer-finished enclosure for over a year. I never had any intention of releasing this as a commercial product, but I wanted to make something more “official.” I spent a few months figuring out how to do artwork - as in, have the labels align with the knobs, sent the design to AmplifyFun, and reboxed my work.

I don’t have a schematic right now, but I do offer this free project for the Mini, and if you look up the Fk Overdrive schematic on PedalPCB or DeadEndFX, you can finagle the circuit together. Furthermore, my version adds a sensitivity control and a control that determines the magnitude of the voltage drop in the sag, the Threshold knob. I did remove the Heavy/Light control from the Fk OD because the Kintsugi knob made it sonically redundant.

Anyways, I had not written on this pedal in a while, and I just want to share. The past few years building have been great, and it’s always crazy to think I went from literally melting my first pedal kit to being able to create a viable pedal that I put together from an idea to a PCB CAD file to artwork.

Kintsugi, a JFET drive with dynamic sag circuit.

Last year I built my first Smallsound/Bigsound Mini based on a PCB of my own creation. Over the course of the months that followed, I started becoming more curious about the F- Overdrive, which is a JFET based overdrive/fuzz/distortion that features a dynamically modulated starve circuit that is said to mimic the sound of a dying tube amp. To me, the effect is akin to a compressor and tremolo working in tandem. On the most subtle level, the dynamic sag circuit can make a drive pedal sound like Neil Young’s “exploding small amp” from Weld, and on the most extreme level, the pedal will struggle and strive to illicit a sound. For reference, the demo at the bottom of this page is set to about 75% of the dynamic sag’s full capabilities.

What I believed would be a simple project derived from researching a series of Google searches exploded into a full fledged R&D project. As I would later find, the information needed to construct the dynamic sag circuit was either non-functioning and/or incomplete. I solicited advice from the usual forums (FSB, DIYStompBoxes, Reddit), and ultimately rebuilt the dynamic sag circuit from scratch. While I am still not completely sure my dynamic sag circuit is 100% a part for part remake of what’s in the SSBS model, I believe I have come very close and am approximating similar sounds (effects). Furthermore, up until now, the only person I have seen publicly claim to get the dynamic sag circuit to work is Brian from SSBS. Even Andrew/The Tone God who first shared the dynamic sag circuit which later made its way into the SSBS drive has never made a claim to producing a functioning unit.

This project started back in April and it wasn’t until now that I’ve completed the product I set out to create: A JFET drive with a dynamic starve circuit. The primary circuit is the SSBS Mini, but I did make some changes along the way like implementing some snippets from the Fairfield Circuitry Barbershop and Benson Preamp. I also added an additional control to the dynamic sag circuit: “Brokenness” that acts sort of like an attack/decay knob. Just like the SSBS F- Overdrive, the dynamic sag unit is footswitch bypass-able, meaning the pedal can either be run with the normal power supply or be run with the sag unit activated. I opted against making it momentary activated because (1) it would require an additional SPDT toggle switch that I had difficulty fitting on to the PCB layout and (2) for my style of music/playing, I found myself preferring it to be a “latching” effect versus momentary.

Controls

• Volume - Controls audio level output.

• Gain - Controls two gain stages simultaneously and increases distortion levels.

• Bass - Passive EQ control for bass frequencies.

• Treble - Passive EQ control for treble frequencies.

• Threshold - Determines aggressiveness of dynamic sag.

• Sensitivity - Adjust for audio input. Hotter signals and pick-ups may require less sensitivity.

• Range - Determines how much voltage (although technically, current) gets starved but also acts like an attack/decay knob.

• If you would like to play around with the dynamic sag and see its effect on voltage, see this model: https://everycircuit.com/circuit/4638353567514624.

Demo

Updates

• August 5, 2023: Added a “full band” demo. Made changes to some of the wording regarding the controls. Updated with dynamic sag model.

A deep dive into deciphering the elusive dynamic sag circuit in the Smallsound/Bigsound Overdrive.

I originally posted two similar versions of this blog on Reddit and the PedalPCB forum. This version includes changes based on feedback from those posts.

Summary

I spent the past few months figuring out the dynamic sag portion of the Smallsound/Bigsound (SSBS) F- Overdrive, had a breakthrough, and then stuck it into a Fairfield Circuitry Barbershop + SSBS Mini-hybrid drive.

The companion effect, an envelope-controlled voltage starve circuit, mimics a dying tube amp and responds to an individual's playing: As described by Brian/SSBS, the dynamic sag circuit is “dynamically modulated (envelope controlled)…turn up the threshold and the harder you play, the more the overdrive crackles, tries to keep up… and fails miserably.” I call it a practice in kintsugi - or discovering and embracing the musicality in the “brokenness” of the pedal.

In this post I have included video demos, a schematic of the dynamic sag circuit, background information, a schematic of the dynamic sag circuit in practice, and final notes on shortcomings and future things to explore.

The post is quite lengthy, so I implemented headers in case you want to jump ahead to a specific section. Also, instead of posting ongoing updates or progress reports, I will just update this post.

(Last updated: 7-4-2023)

Video Demos

• Diagram of the signal path.

• Rhythm guitar demo.

• Lead guitar demo.

Background

A few months ago, after having built a SSBS Mini clone, I came across a video demo of the F Overdrive. The video detailed a feature within the pedal that was described as an envelope follower that starved a single gain stage. I thought the process would be as simple as a Google search, but whenever I did, I found threads suggesting people just Google the answer which leads to more threads suggesting people just Google the answer which leads to more threads…so on and so forth. In all my hours of Googling and perusing forums like FreeStompBoxes and DIYStompBoxes, I had only found one person who figured out the circuit and produced a functional dynamic sag circuit: Brian from SSBS.

Eventually I happened across a thread on Reddit in which someone used the phrase “dynamic sag,” which lead me to the Tone God Punisher circuit.

The Tone God Punisher circuit is not so much a pedal as a companion piece for a pedal. Utilizing an envelope filter, a NPN BJT, and a JFET, the circuit starves current and mimics a dying tube amp. The core concept of the circuit (envelope follower and voltage controlled resistors) can be used for a myriad of applications in which an input signal dynamically controls another part of the pedal - be it, gain, volume, delay time (like the Retroflect), clipping, etc.

The circuit itself is derived from the Nurse Quacky, which in turn is derived from the Dr. Quack, which in turn is based on the EHX Doctor Q.

The dynamic sag feature is a hallmark of the SSBS F Overdrive. More recently The Shields Blender sported a similar feature. Fender ostensibly claims the dynamic sag - or “reactive sag” is “the first of its kind,” but clearly it isn’t, lending credence to the fact that Brian/SSBS is (was?) quite ahead of his time. To clarify, I don’t know if the sag circuitry in the Shields Blender is the same as the SSBS, but I do take issue that the reactive sag is being purported as being revolutionary when another pedal had the same function years prior.

Research

To say it took me a couple of months to figure out how to get the dynamic sag circuit to work would be an understatement. The original Tone God Punisher schematic was not functional. Through a series of trial, error, soliciting advice, multiple breadboards layouts, meticulously looking at internal of the SSBS OD and unpopulated SSBS OD PCBs, and a random late night epiphany, I managed to get the circuit working.

The big changes I had to implement were (1) the biasing of the op amp and (2) the removal of the JFET. The latter was a last minute decision brought about by a FreeStompBoxes.Com post in which a user felt the JFET and BJT interaction would be too unpredictable and warrant careful (laborious) selection of components. Alas, I got the circuit working and stuck it into a Fairfield Circuitry Barbershop/SSBS Mini hybrid.

Because of the alterations I have made and lack of the heavy/light control in my circuit, I have to state that this is not 100% a replication of the dynamic sag circuit in the SSBS OD; however, it is very similar in topology and produces the same effect.

Schematic of the Dynamic Sag

Note: Since originally sharing this schematic, I have made some changes based on feedback I received. The above schematic also includes a few proposed changes that I have learned about since my initial postings on the forums.

Practical Application in a Pedal Schematic

Based on my analysis of the SSBS Overdrive PCB, my circuit is probably only 80% similar to what's in the SSBS Overdrive - but I do believe it is a fair approximation.Above is a schematic of the Fairfield Circuitry Barbershop with the implementation of the dynamic sag unit. Note this is not the same schematic as the demo videos. I just wanted to share how the schematic would look once all the pieces are put together.

Explanation

(Note that parts referred to in this section is referring to the primary Dynamic Sag Circuit schematic.)

The input of the guitar signal splits off into the actual audio circuit and into the dynamic sag circuit. The signal goes through a buffer into the Sensitivity control. Thereafter, the signal goes to an envelope follower, which is a oddly-biased op amp. The envelope follower triggers an LED. LED charges up a capacitor, which in turn discharges to the BJT and allows current/power to flow to ground. The current/power that flows to ground is current/power that’s meant for the gain stage, effectively starving the stage in the process. LEDs’ forward voltage must be matched. Vout goes to the power source of the gain stage and R10 is whatever the bias resistor is.

The two knobs, Sensitivity and Threshold, control circuit. Sensitivity should be adjusted based on playing style - eg. Picking/strumming strength and Threshold determines how much starve should occur. Mark Hammer proposed that the RC network could be changed in order to increase/decrease the decay or even make the effect less aggressive overall. I postulate that the SSBS Fuck’s compression switch (heavy/low) goes between two different RC networks (R8, C4). In my own personal builds, I usually make C4 10uF. I find that values greater than 22uF drain too much power and take longer for the circuit to recover. Another proposal is to replace R8 with a potentiometer in series with a 50-100 ohm resistor in order to make an Attack control. The two foot switches are Bypass and Dynamic Sag. Bypass activates the drive and Dynamic Sag activates the Dynamic Sag.

The schematic for the dynamic sag footswitch bypass wiring is shared in the schematic under "Pedal Schematic or Practical Application." This will enable you to toggle the dynamic sag on or off. You can also opt for a monetary DPDT switch so that pressing/holding activates the sag and releasing immediately turns it off.

Shortcomings and Proposed Changes

Based on my analysis of the SSBS Overdrive PCB, my circuit is probably only 80% similar to what's in the SSBS Overdrive - but I do believe it is a fair approximation. In the future I would like to find a better way to control the rate of voltage change. I think a slew rate limiter is a plausible solution.

Updates

July 4, 2023

When my Skeptical Buffer schematic came under criticism - which I invite for the sake of correcting my work, I decided to proceed with using Simulation Program with Integrated Circuit Emphasis (SPICE). The program I chose was EveryCircuit. I plugged in the Dynamic Sag circuit and experimented with the different suggestions I posted in my original schematic. To my surprise, none of what I proposed did much. The one change that produced the most change - a more subdued voltage starve - was increasing R9, the resistor from Threshold to the base of the BJT.

Furthermore, as a result of this change, I am inclined to believe that the Compression switch on the SSBS overdrive toggles between two different resistors. The larger the value of R9, the higher the minimum voltage will be when the gain stage starves, and the smaller the value of R9, the lower the minimum voltage will be. Also in future builds, I have considered either omitting the Sensitivity control or making it an internal trimmer because I find that if you can fine tune R9, the circuit may not need a sensitivity knob and the Threshold control will offer more usability.

September 22, 2023

Dino from DEAD END FX has stated that their company plans to release a SSBS F*ck PCB. He has offered the following insight as to my research/development:

Theoretically, there is nothing wrong with the TG (Tone God) schemo. It really depends on the application, as well as the gain structure of what you want to influence with it. Where the F*ck is concerned, there is no Vref scheme in the power structure at all, meaning that the opamp input is not biased, and it works fine. Even with a guitar signal. The sag structure is more or less verbatim to the TG schemo, EXCEPT... there is one resistor to ground added at the back end of the envelope. You're not influencing the main rail voltage here, but rather, the bias voltage of one JFET within the audio gain structure. As such, minor (and I mean MINOR) tweaking might be required to get it to work with the destined application. From my breadboarding, I find that the resistor acts as a bleed control of the sag function (biasing the base of the NPN transistor). My experimenting led me to replace the resistor with a pot, which enhances the sag depending on the gain structure of the circuit being influenced, which is important. I found the F*ck that I had to lack sufficient "fade, crackle, pop" when the gain was turned up high, which using a pot instead of a fixed resistor value (as previously mentioned), helps compensate for.

As for the F*ck, I can tell you from experimentation that even the type of JFET's used as the VCR is important. The 2N5457's gave the smoothest voltage transitions, while others (ex.PF5102) were much more abrupt. At the initial attack, the bias voltage of the JFET used in the audio gain side of the circuit can drop from it's normal 4.5v to less than 1v when being "sagged", adequately providing the "edge of tube disaster" effect that is sought after.

Building a Smallsound/Bigsound Mini.

This was first published in two different places: PedalPCB forum and Reddit.

Disclaimer

Please exercise safety when building pedals. While the voltages being worked with are small, be mindful of electrocution. Wear proper safety gear, take breaks, and solder in a well-ventilated room. I don’t offer any sort of technical assistance and these build logs are meant for educational purposes only. Proceed with caution.

Also, I am not associated with any of the companies or groups/people mentioned in this post. I patronized their businesses and entities so I would like to share/promote them.

Build Log

I used this schematic that I redid from FSB. Gray Bench Electronics also did a teardown of the pedal that showed a few variances like the Bass potentiometer (C500K -> C1M), the Gain potentiometer (B100K dual gang -> A100K dual gang), and JFET selection.

PCB was designed in Eagle. I did do a version to allow for SMT JFETs; however, the SMT versions of the JFETs are not available at this time. I opted for a J201 for the first JFET sourced from PedalPCB and 2N5457 for the other JFETs sourced from StompBoxParts. I also changed all the trimmers to 50K because the 10K on the Bias control didn't allow me to adjust the voltage appropriately. I put some "empty pads" on the PCB so I could easily adjust the voltages with my multimeter probe. Chuck's insight about biasing JFETs and this runoffgrove article helped with the biasing of the JFETs. One of the challenges was making the Bias control (not to be confused with the bias trimmers) to work, and I did need to change the source resistor (R8) of Q3 to 100R from 1K. Furthermore, some of the JFET transistors from my pile did require 100R for R10 on Q4 as well in order for the JFET to bias correctly.

I built this in a 125B Pro from StompBoxParts. I highly recommend it; it's legitimately one of the best looking plain aluminum enclosures I have ever seen. My drilling was a bit off and I had to drill bigger potentiometer holes to allow for more "play" to allow better positioning, but I think it turned out fine. Graphic art is from Daiso, a Japanese dollar store in my area. The overall schematic reminds me of a Big Muff but upon closer inspection, the topology is more like the Fairfield Circuitry Barbershop with passive EQ controls and a dual gain stage.

Updates

• 7-17-2023: Corrected schematic with updated values. Uploaded relevant files to GitHub. Shared PCB on OshPark.

Building a germanium fuzz face.

Disclaimer

Please exercise safety when building pedals. While the voltages being worked with are small, be mindful of electrocution. Wear proper safety gear, take breaks, and solder in a well-ventilated room. I don’t offer any sort of technical assistance and these build logs are meant for educational purposes only. Proceed with caution.

Also, I am not associated with any of the companies or groups/people mentioned in this post. I patronized their businesses and entities so I would like to share/promote them.

Background

For a while, I avoided making any sort of germanium transistor-based pedals because of a series of errors while building some Hudson Broadcast clones. However, I became enamored with revisiting the idea after hearing demos of a fuzz by a local builder, LST Gnome Electronics. Germanium fuzzes consisting of 1-2 transistors may seem simple, but they are like making cacio e pepe: Small quantity of ingredients but require the right execution to get working.

Looking to bypass the steps of germanium transistor selection, I opted for matched transistors from Amplified Parts. From there, I wanted to really mimic the look of vintage fuzzes from the UK, and purchased a slanted fuzz case, large potentiometers, and axial capacitors to mimic that old school look. Also, having forgotten to source 22uF radial capacitors from Amplified Parts, I defaulted to using eBay. (Although during a chance trip to LA, I did come across an electronics store that sold “mojo parts” including radial 22uF capacitors.)

I implemented a couple ideas like the Incandeza Bypass by Demedash Effects relay bypass, SMD components, and a voltage inverter to provide a -9V power supply. I did have plans to insert a pick-up simulator and temperature controls as adapted from the Benson Germanium Fuzz, but the idea proved to be overly ambitious. Furthermore, the slanted structure of the fuzz case I was working with was difficult to design a PCB around.

Making a Stripboard Version

This version features stripboard. The insides are not pretty, and I did have to heavily pad the enclosure because the components/joints kept making contact with the case.

As you’ll read in the following section, I did have plans to make a PCB. While waiting for the PCBs to be delivered, I did make a stripboard version of the pedal. Because the transistor set includes specific values for the resistors, I did have to copy those resistor values by soldering together resistors of differing values. This worked but did not lead to pretty results, and because of the compact stripboard layout for the fuzz portion of the circuit (the charge pump is on a different piece of stripbroad), the resistors had a tendency to touch the germanium transistors, short circuiting the pedal in the process.

Creating a PCB

I used the provided Fuzz Face schematic from Amplified Parts in combination with the Sun Face schematic from PedalPCB. I wanted to go with something that was mountable to the footswitch, and I was going be using lugged potentiometers. Later I would find that the pre-drilled holes on the fuzz case were more fitting for the smaller Alpha potentiometers. Using the larger potentiometers did require me to expand the holes in the case. The metal used for the case is incredibly strong - almost to a fault, and I found that expanding these holes as well as adding an additional hole for a DC jack was quite difficult.

The PCB design as done on EagleCAD. Note the wavy traces - they are meant to mimic vintage fuzz circuit boards. To be clear, this is NOT an optimal way to do traces and only serve an aesthetic purpose.

I did make some changes to the original fuzz face schematic by implementing the popular Bias knob mod. I also put in the original resistor (R9 in the schematic below), which can be used if a person opts against the Bias knob. Furthermore, I implemented jumper pads into the PCB to allow for a silicon version, based on the Analog Man Silicon Sun Face, to be built. Also, the board size allows this pedal to be build into a 1590B enclosure; however, this will require smaller solder-lug potentiometers. If anyone wants to bias their transistors, I recommend substituting R7 and R9 with these boards.

Gutshots

Axial capacitors are used for the actual fuzz circuit. The footprints allow for a full 18mm long component with 5-8mm width. They offer no sonic advantage. They just look cool.

Completed board sans germanium transistors. The footswitch holds the PCB against the enclosure. There is a piece of cardboard between the PCB and the enclosure to prevent the joints of the circuit from shorting against the enclosure. I lifted this idea from Effects Layouts’ Incandenza Bypass layout.

Socketed transistors. There is a LED inside the enclosure. I was going to omit it all together but I needed it while testing to ensure power is going to the circuit.

Completed pedal. The PCB version had way less squealing than the stripboard version.

Schematic

Schematic of the final design. The Test pad is a suggestion by Demedash to check for any power spikes. Note the jumpers. Pad 1 and Pad 2 are an error when I thought I could sub in a LT1054. C3, C4, and C5 appear as resistors in the schematic because I couldn’t find axial capacitor footprints for the PCB.

Parts

If you wish to build a similar pedal, here are the parts I sourced from Amplified Parts:

• Germanium transistor set. I used a T7-H, which is somewhat questionable for testing germanium transistors. I got an hfe of 50-60 and 70-80 for Q1 and Q2 respectively. I did not test for leakage.

• Fuzz case.

• Potentiometer (audio).

• Potentiometers (linear).

• Axial capacitors: Non-polarized, polarized.

• Additional parts like the relay, footswitch, DC jack, audio jacks, and radial capacitors, ICs were sourced from Electronics Warehouse in Riverside County (CA), Jameco, and Love My Switches.

"Closure", a short film treatment and script.

I recently took a class on script writing at San Diego City College. The following are from my fiction script titled Closure. The storyline was inspired by Ubiq by Philip K. Dick and written as if I was doing a project for Wong Fu Productions.

Hook

In the future, hospice patients are given the opportunity to wear a cybernetic device (Reminiscence) that allows them to traverse their memories and explore their pasts. Many use it to revel in a moment that brought them great joy and others use it to rectify an error from ages ago. Unfortunately, the technology is still limited and users are only allotted the ability to travel to a single event.

Isiah, an elderly man dying from pancreatic cancer, has a final wish of revisiting an incident prior to his wife’s (Grace) death. Present at his side is his estranged daughter, Anna. However, things go awry when the machine malfunctions, and Anna must go into the fading mind of her father to help him confront the greatest mistake he ever made and prevent a Compartmentalization Event from occurring.

Files

• Treatment

• Script

Failed Builds: RAT+Muff+HM2.

I first came across the idea of combining a RAT and a Big Muff upon learning of the Drunk Beaver Pedals Disambiguation: The body consists of a mish-mash of a RAT and a Big Muff. Playing off that idea, I designed a pedal with the input stage of the RAT followed by the clipping section of the Big Muff. The initial design was heavy. In many ways “doomy,” and in many other ways, compressed and gated. However, I found the tone to be far too niche for any commercial application. I let the idea gestate for a few months before happening upon various online discussions about using the HM2’s EQ section as the tone stack in various drives. As luck would have it, Drunk Beaver has implemented a similar idea before.

I wish I could say that I invested time and calculation to put together this design, but in the end, I just copy and pasted schematics that I found online and used ElectroSmash to identify the key parts of the RAT and Big Muff. I have managed to get this pedal to work…but I have never managed to make it sound viable.

Additional Sources:

• ElectroSmash Big Muff Pi Analysis

• ElectroSmash Pro Co Rat Analysis

Progress Report: Envelope Follower, Voltage Controlled FETS, and the Elusive "Dynamic Power Supply Sag"

“...the highlight of this pedal is a unique blown amp feature which provides crackles, compression, gating and eq changes similar to that of a tube amp pushed past it’s (sic) limits...”

— Smallsound/Bigsound F- Ovedrive Manual

The Smallsound/Bigsound F- Overdrive has been on the market for some time, and alongside its amazing overdrive pedal, is a highly unique feature: A Crackle/Burst/Threshold system that is an enveloper follower circuit based on the Dr. Quack schematic. The Dr. Quack is an off-shoot of the Dr. Q by Electro-Harmonix. The Dr. Quack was used by The Tone God to create The Punisher, and the Punisher is widely said to be the schematic behind the F- Overdrive’s Crackle/Burst/Threshold (CBT) feature.

Go to 3:35 for demonstration.

“The Punisher is a dynamic power supply sag simulator. The Punisher is not an audio effect but a companion effect that is to be used to control another effect to provide new sounds. You plug your audio signal and power supply into Punisher in series and connect the output audio signal and power supply into the effect to be controlled.”

— The Tone God, "Punisher"

The envelope follower feeds a control voltage to a JFET, which in turn alters its resistance, producing a voltage-controlled resistor in the process. The JFET is run in series with the power supply to a JFET stage, thus mis-biasing the JFET stage in the process. The resulting sounds is said to emulate that of an amp on the verge of explosion whose tubes wax and wane.

For further reading, I suggest visiting this page and scrolling down to “Driving Effects Circuits with the LFO Waveforms.”

Having built the Smallsound/Bigsound (SSBS) Mini, I started to set my eyes on the SSBS F- Oveddrive. However, I had little clue as to the underworkings of the CBT. Initially I thought the circuit was some sort of phaser/tremolo circuit that altered the volume/output of a specific stage of the SSBS Mini. But I wasn't too sure. Also, pictures of the F- Overdrive PCB did not look anything like a tremolo/phaser circuit. Although, oddly enough, the concepts of JFETs as voltage-controlled resistors (VCRs) is closely associated with phasers.

I began researching the most readily available information out there: It's a voltage-controlled envelope follower circuit. I pulled up schematics for the Glowfly Retroflect and Frantone Vibutron, which features envelope followers affecting FETs. I shot off a few Google searches, which immediately turned up the Dr. Q schematic. Unfortunately, there was very little information about how that particular circuit fit in the CBT. I wish I could say that the subsequent research was streamlined once I aligned the CBT circuit with the Dr. Q schematic, but in reality, I spent quite a few days perusing various forums to no avail. A culmination of 2-3 days of digging deep into forums helped me learn that the envelope follower controlled a "resistor," the resistor was associated with the second stage of the SSBS Mini, and the overall result was the JFET being mis-biased.

Simple enough, right?

A few more hours of Google searching lead me to Reddit where someone used the keywords, "dynamic sag." (Coincidently, I had just completed a LM317-based sag unit.) A few more Google searches lead me to the Pusnisher circuit. Furthermore, if you Google the Punisher circuit, the CBT pops-up alongside it!

I breadboarded the circuit, but I could not get it to function. Furthermore, I found several posts of individuals identifying the Punisher as being a key part of the CBT, but only a handful of people who successfully implemented the circuit: The Tone God, author of the circuit; some random guy on DIYStompboxes; and the mind behind SSBS. After a day of troubleshooting the circuit, I crowdsourced Reddit and DIYStompboxes for assistance.

The key piece missing from the original schematic was just a pair of resistors: The op-amp was not biased! The addition of the transistors allowed the circuit to work. I also experimented/ with and implemented some other changes from forum suggestions and the Dr. Q, which allowed the envelope signal to be much stronger. However, one change that helped was found on accident: I was trying out different biasing resistors and left one resistor in. This made the circuit work much better for reasons I am unsure of.

Schematic with many changes.

And like that, I figured it all out...sort of.

See, I call this "Progress Report" because there are still a few of issues to solve:

• How to calculate the resistance that is bestowed upon by the VCR/JFET.

• How to make the circuit smoother as the attack/release is really fast.

• How to elicit more crackling. I actually did manage to do this when I was using two 1M ohm resistors for biasing, but the envelope signal was not as strong.

• What the “Heavy/Light” function changes. I think it either changes the electrolytic capacitor attached to red LED or it may be the Attack function seen in the Nurse Quacky.

As for now, this is it.

Left side is the Crackle/Burst/Threshold function. On the right is a dummy unit to test the circuit. The dummy unit based on the Fairfield Circuitry Barbershop, which is like a minimalist SSBS Mini.

Updates:

• May 8, 2023: Grammar and spelling errors corrected.

Variable Voltage Regulator Pedal, or the Dying Battery Simulator Pedal

A dying battery simulator allows the voltage of an analog pedal to be reduced to a suboptimal level in order to mimic a dying battery effect. At lower voltages, distortion/overdrive/fuzz pedals are said to produce different sonic textures. For example, fuzz pedals may become gated, and high gain distortion pedals may take on a more subtle overdriven sound.

While the design is presented in pedal form, the schematic can be minimized to fit into another pedal in order to allow a user the ability to vary voltage within a pedal. Furthermore, pedal builders can also place the design into their test boxes in order to provide a wide range of voltages. In the lattermost case, building a 35V charge pump with a LM317 would allow a builder to have a wide breadth of voltages from ~32V to 1.25V - although I personally believe, 25V, 18V, 9V, and 4.5V would be satisfactory for many pedal designs.

The LM317 is a 3-terminal adjustable regulator that is capable of outputting a voltage range of 1.25 V to 37 V depending on input voltage. The regulator has three terminals: Input, Output, and Adjust.

Source: Wikipedia

When the JHS Volture first came out, I immediately knew it wasn't simply a potentiometer in series with a power supply. As soon as gut shots for the Volture became available, it was pretty clear the circuit was based on RG Keen's original concept, which in turn is probably from the actual datasheet for the LM317.

For my version, I utilized a video from Volos Projects to corroborate/refine the design. I then added in a LED voltmeter from Amazon and an internal Age control. New batteries differ from old batteries in that, “while new batteries can put out a lot of current, old ones can't - they develop a high internal resistance.” The Age knobs acts a rheostat. Putting “a resistor on the output of our regulator…will fake the internal resistance of the battery" (source). Voltmeter addition was inspired by MAS Effects’ UV Meter Drive and Silktone Fuzz. I did attempt to have a rotary switch for a Battery Simulator (carbon zinc, alkaline, “wall wort”) to play off the JHS April Fool’s joke, but the idea was too ambitious and overly complicated to implement. I also would've loved the Age control to be an actual potentiometer, but I could not find a reasonably priced and appropriately sized 100 ohm potentiometer.

This circuit allows you to go from ~7.5VDC to 1.25VDC which is similar to the Voluture specs on the JHS website ("...adjustable voltage from 7.5VDC down to 1.25VDC..."). The basic circuit is 5-6 different components and takes a matter of minutes to piece together.

Schematic.

Two changes I would've liked to implement were:

• Reverse polarity protection.

• 100uF to 1000uF electrolytic output capacitor for stability.

Demonstration. Note that the LED I used was very cheap and requires calibration. The read-out is not as smooth or accurate as I would like, but I am sure if a better, well-calibrated LED was used, the result would be better.

Files:

• (Pending)