The Mulebox is an open source, build-it-yourself hardware guitar processing unit based on the JohnH attenuator, Electrosmith Daisy Seed DSP module, and Cleveland Audio Hothouse platform. It works as a reactive load box for silent playing, with cabinet simulation via Impulse Response files providing a line level, stereo output to e.g. speakers or a mixer.

Check out this video for a full description and audio examples, including a brief comparison with the UA Ox Box.

This document provides guidance for how to build your own. Both software and hardware are open source and contained in this repository. You can build it as described here, or use this as a starting point for your own build of a similar device.

WARNING #1: The basic function of this box is to replace the speaker in an amplifier. A tube amplifier, in particular, will be damaged if used without a suitable load. Do not turn the amp on without a speaker or a load (like the one inside the Mulebox). If connections are made incorrectly, become loose, or are poorly soldered inside the Mulebox, the load could be disconnected. In this case, you could damage your amplifier.

WARNING #2: The components described here are for an 8 Ohm speaker output of an amp rated 50W or less. Only plug into an 8 Ohm speaker output on your amplifier, and never use it with an amp more powerful than 50W. If you need a different speaker load or power rating, you will need to build it with larger resistors of different values.

WARNING #3: In the suggested enclosure, this is a very tight build. You need to take care selecting components that can handle the required power, including wires. You need to be familiar with soldering, guitar electronics (e.g. you have built guitar amps or pedals before), and be capable of building the code for this software project using basic development tools. There are detailed instructions below, but some experience is assumed and probably required to be successful.

You will also need to supply your own IR files, e.g. ones you have sourced or recorded yourself. A few free IRs are included, and they sound decent, but professionally recorded ones will be even better. There are plenty of places you can buy good quality IRs, and plenty of free ones available from Tone 3000.

This is the assembled Mulebox:

To use it, connect a 1/4" TS speaker cable (not an instrument cable like the one you might use to plug your guitar into an amp or connect two effects pedals) from an 8 Ohm (only!) speaker output on your amplifier, into the amp input on the Mulebox rear.

Connect a pair of 1/4" TS audio cables from the left and right output jacks to a mixer or other recording device. If you want to connect to a pair of speakers, make sure they are either powered or that you use a suitable stereo amplifier. Note that the outputs are not balanced, and there is no headphone output, so you will need to plug the Mulebox into something cabable of amplifying a stereo signal to hear its output.

Turn the rotary switch on the front to position 1. Connect a 9V DC, centre-negative (only!) "pedal" power supply to the power socket at the back. Turn the other knobs to noon.

The front knobs can be used to control the output level and the reverb mix. The reverb is a lush Dattorro plate algorithm.

The trim knob at the rear is used to control the input level between the amp and Hothouse. In general it should be as high as possible without causing clipping.

When powered on, the red LED will blink several times, indicating the number of IRs currently loaded (up to 12 times), and then stay on whilst the Mulebox is active. The blue LED will then blink to indicate the current position of the rotary knob, i.e. which IR is currently selected.

The blue LED will stay on if an IR is loaded, and blink when the selected IR is being changed by turning the rotary knob, the number of blinks indicating the selected position. If there are fewer than 12 valid IRs, some of the slots may be empty. In this case, the blue LED will remain off, and the raw sound of the amp will be passed through, which will probably not sound very good.

It is important to avoid clipping in use. Clipping is indicated by the LEDs blinking:

• The red LED will blink if the input signal starts to clip. In this case, turn down the trim knob on the rear, or reduce the output volume of your amplifier.
• The blue LED will blink if the output signal starts to clip. In this case, turn down the output level knob on the front.

By default, the Mulebox includes the following basic, royalty-free IRs. They sound pretty good, but you will definitely be able to get even better sounds if you buy and load more professional IRs.

1. Blackpanel 1x12
2. Greenback 4x12
3. V30 4x12
4. Chime 2x12
5. Tweed 1x12
6. Boutique 2x12

Think of the Mulebox as two separate devices hardwired together:

1. An attenuator, using a variant of the "JohnH" Attenuator design. This provides a safe 8 Ohm reactive load with an impedance curve that is similar to a speaker cabinet, which will make a tube amp react more realistically.

The JohnH design can be used to build an attenuator box that quietens an amp whilst passing output to a real speaker in a guitar cabinet, with selectable levels of attenuation. The Mulebox does not support pass-through to a guitar cabinet. Instead it takes a "line out" signal and sends it to the DSP module.

The attenuator absorbs all the energy of the audio signal (up to 50W) and turns it into heat. Therefore, the attenuator gets hot. An aluminium chassis acts as a heat sink, and is cut with plenty of ventilation. There is also a small fan to provide airflow. This is important not only to ensure the passive attenuator components don't fail, but also since we are running a DSP module in the same box and this has a narrower operating temperature range.

2. A DSP processor that takes the raw amp sound and applies a selectable Impulse Response (IR) to it, emulating a guitar speaker cabinet, before passing the audio out as a stereo signal (really, dual mono). This is based on the Electrosmith Daisy Seed module, which can be programmed to perform a variety of audio-related functions. The source code in this repository is the Mulebox firmware, written in C++ for the Daisy platform.

The Daisy Seed needs to be connected to knobs and audio I/O. For that, we use the Cleveland Music Co Hothouse. This is a PCB that provides up to six knobs, three switches, two footswitches, two LEDs, and stereo audio input and output, originally intended in a guitar pedal form factor. Mulebox uses the Hothouse PCB, but not the breakout boards for the jacks and footswitches and LEDs. This runs on 9V "pedal power".

Note that the attenuator is entirely passive. This means that if you accidentally or temporarily turn your amp on but leave the Mulebox unpowered (there is no on/off switch), your amp is still safe. However, the cooling fan will not run, and there will be no audio passing through to the output, so you are not recommended to stay in this state for very long.

As described here, the Mulebox lives inside a "Hammond 1590DD" size enclosure – mainly because this is the largest enclosure one can order from Tayda Electronics and have them paint and drill. It is of course possible to use a different box, so long as it has the required cutouts and enough ventilation, but we provide Tayda drill and UV print templates below, which make it much easier to get a professional-looking enclosure.

Here is the schematic of the full circuit:

There are a number of ways the Mulebox could be extended or improved, given more time, budget for components, and – perhaps most importantly – a larger enclosure. Here are some ideas:

• If you want to use a 4 Ohm or 16 Ohm speaker tap, or a higher-wattage amp, you will need to modify the attenuator part of the circuit. The Marshall Forum thread where the attenuator is described, contains the information you need to do this, but there is a lot there to unpick. Almost certainly, you'd need to change the large resistors and possibly the inductor coil, which in turn might make it difficult to fit everything in the suggested enclosure. But it is certainly possible to do this. The Hothouse, Daisy Seed, and software parts of the circuit should not need to be changed, so long as the "line out" part of the attenuator that is hardwired to the Hothouse PCB is of a safe level.

• The JohnH attenuator design can be used, with one or several switchable steps, to reduce the output power of an amp whilst still playing through a guitar cabinet with a real speaker, as an alternative to or in parallel with speaker simulation via IRs. The Mulebox does not support any attenuated "pass-through" and only passes signal to the DSP unit. With more space and the addition of some switches and an extra output jack, it would be possible to pass an attenuated (i.e. quieter) signal to a speaker like a normal reactive load box, instead of or in paralell with the simulated speaker output.

• With additional active components, the Mulebox could gain balanced outputs, and/or a headphone output (you would need to add a headphone amplifier).

• Commercial units like the UA Ox Box or Two Notes Torpedo Captor X provide not just cabinet simulation, but other digital effects. The Daisy Seed could be programmed with additional effects, though this might also necessitate the installation of further knobs (the Hothouse has support for three further knobs and three toggle switches that are not used in the Mulebox). Note that there would likely be some tradeoffs in both storage space and processing power on the Daisy Seed. Already, adding reverb has forced us to restrict max IR lengths to 85ms (down from 170ms without the reverb – still fine for cabinet IRs, but not suitable for reverbs).

• To save space and cost – the attenuator circuit does not include JohnH's optional second inductor coil that simulates the bass "hump" found in many real world cabinets. The Mulebox firmware applies a slight, hardcoded DSP low-end boost to roughly emulate this resonance. With more knobs or toggle switches, the Mulebox could gain an adjustable two or three-band EQ for additional tone shaping.

When selecting hardware components to use, it is important to consider not only their values (e.g. resistance and power rating/wattage for the resistors; inductance for the inductor coil), but also the physical dimensions of each. The Tayda drill template provided will need to be modified for any components with a different footprint from the ones suggested below. This is most important when it comes to the large, high-powered resistors. Since these are clamped to the bottom (technically the "lid") of the enclosure, it might be preferable to drill the mounting holes for these manually.

First, buy the Daisy Seed – the brains of the operation.

The Hothouse does not come with it. You can buy it from Electrosmith directly, or a third party seller close to you. It can be reprogrammed as many times as you like and used for other projects (e.g. using the Hothouse platform in its default pedal-sized enclosure).

Next, you need the Hothouse.

This is what the Daisy Seed will mount to to gain access to audio I/O, knobs (pots), and LEDs. Ideally, buy the Hothouse from Cleveland Music Co. They sell a version without an enclosure, which will save you a bit of money, whilst still supporting the developers of the Hothouse.

However, the fine folks at Cleveland Music Co have made the PCB available as open hardware which means that you can also manufacture your own from their "Gerber" files – probably using JLCPCB. In this case, you want the main PCB only (not the footswitch or I/O breakout boards), and you probably want to get JLCPCB (or whoever you use) to provide and solder the SMD components for you.

There are several online tutorials that explain how to do this, but the basic idea is that you upload the Gerber (.zip) and Bill of Materials (.csv) file for the main PCB to the JLCPCB website, and choose their PCB assembly service. It's a very slick process. You will probably have to order at least five boards, but even then it's pretty affordable (plus you now have spares).

If you go this route, you will also need to buy two single-row, 20-pin, 2.54mm pitch female square-pin headers and a 100uF low-ESR electrolytic capacitor, which are otherwise included in the Hothouse kit. See the shoppling list below for specific examples.

You will also need pots, jacks, a power socket, and LEDs, but we'll detail those below, as you'll need to get them regardless of whether you buy the Hothouse kit or manufacture your own PCBs. The components that come with the Hothouse are different from what we will mount in the Mulebox.

The Hothouse PCB is meant to hang from its pots and switches, but for the Mulebox we'll need to mount these to the front and back of the chassis and wire them to the board. That means we need another way to mount the PCB, without having it touch the chassis itself. Unfortunately, there are no mounting holes on the Hothouse either.

Instead, we can use four Essentra TCEHCBS-4-01 PCB mounts and some #6 screws (or M3.5 size metric ones). These snap around the corners of the board. The drill template below is spaced for these. They are available from many places, including (with screws) The Pi Hut.

Next, you need an inductor coil: literally a large coil of thick wire wound into a coil. This is what makes the load "reactive" and helps ensure the amp acts and feels like a real speaker is connected.

For an 8 Ohm, 50W build, you want an "air core", 18AWG coil with an inductance of 0.9mH, such as the Dayton Audio LW18-90. These will usually be stocked by specialist audio hardware or repair shops, rather than general purpose electronics retailers.

The coil needs to be mounted in such a way that it does not touch any other metal components, including the enclosure. It is also pretty large, and so a tight fit. The Dayton Audio coil, at least, comes with a pair of plastic zip ties to hold the coil together, and you can use the thicker fastener on the zip ties as a spacer, keeping it off the bottom of the enclosure. The suggested layout/drill template has two holes drilled to run additional zip-ties through, for fixing the coil in place.

The circuit calls for three large, wirewound resistors: one rated at 100W and two at 25W. Again, these take up a lot of room and they get hot.

In contrast with the inductor coil, it is very important to bolt these securely to the metal enclosure, with a small amount of thermal conduction paste smeared on the bottom. This helps turn the enclosure itself into a heatsink, dissipating heat more effectively.

The suggested layout/drill template is based on resistors from the TE Connectivity HSA series (see links below). This choice is somewhat arbitrary, and there are other manufacturers. However, you will need to either drill your own, precise mounting holes, or adjust the drill template by referring to the relevant technical drawing, as the hole spacing is not standard across manufacturers.

Speaking of heat dissipation, the design includes a small fan at the rear, intended to pull cool air through slats at the front and sides of the enclosure across the resistors, expelling hot air at the back. A large circular cutout exposes the fan, which is protected with a metal finger grille.

The drill template has space for a 30mm square extractor fan with standard hole spacing. This is very tight and might require some creativity when mounting and closing the lid. You may need to add a bit of extra spacing with a nut or small washer on the inside of the enclosure, to allow the lid to close.

The fan needs to run off the 9V power supply. You can get fans that run natively on 9V, or e.g. 12V fans that will just run a bit slower at lower voltages. A lower-voltage fan (e.g. a 5V fan) will require additional components (not described herein) to step the 9V voltage down to 5V.

Cheap fans are noisy. Quiet fans don't move a lot of air. The quality of the fan will make a difference. Look for 'vapo' magnetic bearings, and noise ratings below 20dB, noting that a 12V fan run at 9V, for example, will be slower and thus quieter than the stated rating.

Sunon makes good fans and the Sunon MF30101V2 12V 30mmx30mmx10mm fan is a good, relatively quiet fan that can run at 9V, but there are many other options.

Apart from needing to fit all the components, the most important facet of the enclosure is that it needs to help manage heat effectively. A 50W amplifier at full volume generates enough power to create serious heat, which could cause the Daisy Seed to shut down or impact the performance of the circuit in unpredictable ways. An aluminium enclosure can act as a heatsink in its own right if the large resistors are clamped securely to it, but we also need to cut ventilation holes.

Since the components (just!) fit inside a Hammond 1590DD enclosure, we can use Tayda Electronics to paint, drill, and UV print an enclosure for us. See this guide for more details about that process, including how to set up the UV print templates.

You can use this drill template as your starting point, which looks like this:

We will use the "lid" of the enclosure as the base, drilling several mounting holes for the large resistors, induction coil, and the Hothouse PCB standoffs.

Note that Tayda drill the lid from the outside but we are mounting components from the inside, which means the drill holes need to be translated accordingly. The template matches the screw spacing of the components listed in the shopping list below, but these will likely need adjusting if you use other brands. Alternatively, remove these holes from the drill template and drill them manually once you have all your components laid out – the base will not be visible in any case.

"Side E" will be used as the front, mounting two LEDs in bezel holders, the rotary switch, and the output level and reverb mix control pots. It also contains crucial slat cutouts that sit in front of the large resistors, directly opposite the fan, allowing it to draw cool air across the top of the resistors. The potentiometer and switch cutouts match the UV print template for the front.

"Side C" will be used as the back. It contains cutouts for the jack sockets, the "trim" knob that is used to adjust the volume of the attenuated signal going into the Hothouse, and a large cutout for the fan. Again, the position of these holes match the UV print template for the back.

The top and sides contain a large number of slats – so many in fact, that you need to order "extra holes" from Tayda for them to be willing to cut them. See below.

Please review the Tayda drill and UV print guide, and see the list below to understand which components and services to buy to manufacture the enclosure. Alternatively, if you wish to drill your own enclosure, see the Layout directory in this repository, which contains SVG files with precise hole positions and dimensions that match the Tayda template. You can even use this tool to extract the positions and dimensions of each cutout.

Resources:

• Tayda Electronics Drill template
• Tayda Electronics UV templates in PDF format for the front (side E) and back (side C) with white, gloss and colour layers
• Affinity Designer files for each side, including drawings of the cutouts in the drill template

You need the components below. We suggest ordering everything that is available from Tayda Electronics if you are going to also have them drill and print the enclosure, because their components are cheap and their shipping is better done in bulk. Other components are listed from Farnell. You can of course choose whatever supplier you want, and use equivalent components, though you may then need to adjust some of the enclosure cutouts and other aspects of the layout.

(SKU numbers in brackets)

• 1 x red 5mm LED (A-1554)
• 1 x blue 5mm LED (A-3005)
• 2 x 5mm LED bezels (A-660)
• 2 x B10K mini pots with solder lugs (A-1961)
• 1 x B5K mini pot with solder lugs (A-1970)
• 3 x dust seals for pots, unless supplied with pots (A-5527)
• 3 x small knobs for 6mm splined pots, ~16mm (e.g. A-6071)
• 1 x 100uF electrolytic capacitor (A-959) – unless supplied with a full Hothouse Kit
• 2 x 20pin female headers (A-1310) – unless supplied with a full Hothouse kit
• 11 x 1K 1/4W metal film resistors (A-2200) – used for the rotary switch resistor ladder
• 1 x 1K 1W metal film resistor (A-2196) – higher power rating, used for the attenuator circuit
• 3 x isolated jack sockets (A-6042)
• 1 x small isolated DC power socket (A-991)
• 4 x 10mm M3 bolts (A-1252) and 4 x nuts (A-1247) – for mounting 25W resistors
• 4 x 10mm M4 bolts (A-4351) and 4 x nuts (A-4357) – for mounting 100W resistor
• 4 x 25mm M3 bolts (A-1257) and 8 x nuts (A-1247) – for mounting cooling fan and finger grille

And for the enclosure:

(skip all these if using a different/custom-drilled enclosure)

• 1 x 1590DD enclosure in a suitable colour (e.g. A-5906)
• 1 x 1590DD custom drill service (A-5141-CST-DR1)
• 1 x 1590DD lid custom drill service (A-5141-CST-DRL) – you can skip this if you prefer to place and drill your own mounting holes for the large resistors, induction coil, and PCB
• 64 x additional hole drilling service (A-99999-CST-DR-H1) – adjust this so that the total number of cutouts in the enclosure is covered (40 holes are included in the default drilling service, so 40 + 64 = 104 total holes)
• 1 x 1590DD side C UV print service (A-5141-CST-UVC)
• 1 x 1590DD side E UV print service (A-5141-CST-UVE)
• 2 x gloss layer UV print service (A-99999-CST-UV-GL) – optional, but recommended to protect the printed labels
• 2 x "print twice" white layer UV print service (A-99999-CST-UV-WH) – optional, but recommended if printing white on black

(Order codes in brackets)

• 4 x TCEHCBS-6-01 PCB standoffs (4691573) – plus suitable screws (#6 or M3.5, self-tapping). Alternatively, this kit from PiHut contains both the standoffs and the screws.
• 1 x 12-position make-before-break rotary switch – Lorlin CK1034 (1123690) – note this has an imperial sized 6.35mm (1/4 inch), D-shaped spindle and so requires a suitable knob. There is a metric version of the same switch (CK1024) which can take a 6mm D-shaft knob, but the front cutouts will then need to be adjusted.
• 1 x 3K3 2W metal film resistor - used for the attenuator circuit (1738645)
• 1 x Sunon MF30101V2 30mm fan or similar (4382050)
• 1 x Fan finger guard for 30mm fan, e.g. Multicomp Pro MCSC30-W1B (1781195)
• 1 x 10 Ohm, 100W resistor, e.g. TE Connectivity HSC10010RJ (1174288)
• 1 x 100 Ohm, 25W resistor, e.g. TE Connectivity HSA25100RJ (2009323)
• 1 x 39 Ohm, 25W resistor, e.g. TE Connectivity HSA2539RJ (2805189)

• 1 x medium knob for 6.35mm D-shaft rotary switch, ~19mm, with the pointer towards the flat edge (assuming you are using the Lorlin CK1034 rotary switch, which is their imperial-dimensioned one). These can be a bit hard to source, but the Cliff K21 D-shaft 1/4" knobs with a suitable cap works. Note that many D-shaft knobs have a pointer 180 degrees away from the flat edge of the D, which will be wrong. A 6.35mm round shaft knob with a set screw might also work.
• 1 x Daisy Seed – see above
• 1 x Hothouse main PCB – see above
• 22 AWG or similar (i.e. thick) wires, preferably in a few different colours, for wiring the attenuator.
• Thinner wire in multiple colours for wiring up the Hothouse to the output jacks, LEDs, and potentiometers.
• Optional: For wiring the front pots, LEDs, and rotary switch, you can use sets of wires with 3-pin JST 1.25mm (or larger) connectors (you'll need both male and female connectors). The advantage of this is that you can easily disconnect the front components when opening the enclosure.
• Plastic zip ties for mounting the inductor coil
• Thermal paste (only a tiny amount) for mounting the large resistors

You will need:

• A good quality soldering iron with both a pencil tip for PCB soldering and a larger tip for the large resistors, jacks, and so on.
• Other soldering tools – high quality solder, an extractor fan and/or mask to avoid breathing in fumes, a solder sucker for rework, helping hands or clamps, etc.
• Screwdrivers, pliers, wirecutters, wire strippers, spanners in various sizes to affix pots to the enclosure, etc.
• A multimeter that can measure resistance and continuity (beep mode)

Before the Mulebox will work, you need to build flash the firmware from this source code repository onto the Daisy Seed. This also transfers the prepared IRs (see above) to its QSPI flash memory. For this to work, you need to connect the Daisy Seed to a computer via USB – either directly or using an STM32 debug probe if you have one.

1. ARM GCC Toolchain

   # macOS
   brew install --cask gcc-arm-embedded
   
   # Linux (Debian/Ubuntu)
   sudo apt-get install gcc-arm-none-eabi
   
   # Or download from: https://developer.arm.com/tools-and-software/open-source-software/developer-tools/gnu-toolchain/gnu-rm

2. dfu-util (for flashing via USB)

   # macOS
   brew install dfu-util
   
   # Linux
   sudo apt-get install dfu-util

3. Clone with submodules

   git clone --recurse-submodules <repository-url>
   cd Mulebox

   If already cloned without submodules:

   git submodule update --init --recursive

make                # Build the project
make update-irs     # Rebuild the IR data header from WAV files in irs/
make clean          # Clean build files
make clean-all      # Clean including library builds
make program        # Flash firmware via debug probe (QSPI direct)
make program-dfu    # Flash firmware via USB DFU (or use: make flash)
make help           # Show all available targets

There are two ways to flash firmware: via USB or via an STLINK debug probe. The debug probe method is simpler (no buttons or timing required) but you need an STLINK-V3MINIE or similar SWD debug probe, and to have separate power running to the Hothouse/Daisy Seed.

Both methods require a one-time bootloader installation step, after which the firmware can be updated as often as needed.

This step is only needed once per Daisy Seed, or when updating the bootloader.

With USB:

1. Connect the Daisy Seed to your computer via USB.
2. Enter STM DFU mode:
   • Hold down the BOOT button on the Daisy Seed
   • While holding BOOT, press and release the RESET button
   • Release the BOOT button
3. Flash the bootloader:

   make program-boot

With STLINK debug probe:

1. Connect the STLINK debug probe to the Daisy Seed. Power the Daisy via 9V or USB.
2. Flash the bootloader:

   make program-boot-probe

After flashing, the Daisy Seed will restart. Its onboard LED should pulse for about 2.5 seconds (the bootloader's grace period), then stay off since no application is loaded yet. If no application is found, the bootloader continues to wait for a DFU connection indefinitely.

After the bootloader is installed, use one of these methods for all firmware updates.

With USB only:

1. Connect the Daisy Seed to your computer via USB.

2. Enter the Daisy bootloader by pressing RESET on the Daisy Seed.

   Important: Press RESET only — do not hold BOOT. Holding BOOT enters the STM DFU mode (for bootloader installation), which is different from the Daisy bootloader (for firmware updates).

   The onboard LED will pulse for about 2.5 seconds.

3. Flash the firmware within the 2.5 second grace period:

   make program-dfu

   If you miss the window, press RESET again and retry. On the very first flash (when no application is loaded yet), the bootloader waits indefinitely so there is no time pressure.

The firmware loads and begins running automatically after flashing.

See the Daisy Seed C++ SDK documentation for more details.

With STLINK debug probe:

1. Connect the STLINK debug probe to the Daisy Seed. Power the Daisy via 9V or USB.
2. Build and flash:

   make
   make program

The firmware is written directly to QSPI flash via the debug probe and begins running automatically.

This project ships with six royalty-free IRs as a starting point, but you should load your own IRs that you have either sourced or recorded yourself. One free source of IRs is the Tone 3000 website, though there are many others.

Requirements:

• WAV format, 48kHz sample rate, mono or stereo
• Up to 12 files – if fewer, some positions on the rotary switch will be blank
• Each IR should be up to ~85.3ms (4096 samples) – longer IRs will be truncated, but you can still use them
• You need Python 3 installed and available via the python3 command

Note that files are assigned to rotary switch positions 1–12 in alphabetical order, so name them accordingly (e.g. 01_dark.wav, 02_bright.wav) to minimise confusion about which one is which.

Steps:

1. Place your IR WAV files in the irs/ folder.

2. Run the make target to regenerate the IR data header:

   make update-irs

3. Rebuild and flash to the Daisy Seed:

   make
   make program        # with STLINK debug probe
   # or
   make program-dfu    # with USB (press RESET first)

Note that the update-irs step is run automatically before each new build of the software, so if you perform a clean build (make clean all) then it will probably work without the explicit make update-irs step.

Tip: If you have a debug probe, it connects to the Daisy Seed via a 14-pin connector cable. While testing, you can leave this connected and feed it out through the top slats on the left side of the Mulebox, which are cut slightly taller than the others for this very reason. If you leave the Mulebox powered on and connect the probe, you can quickly load new IRs by putting them in the relevant directory and re-running the build per the above instructions. However, if the debug probe is left connected to your computer via USB while playing, it will likely introduce additional noise – a whine or buzz that will be audible through your speakers.

You've ordered the parts. Waited for the delivery driver. Laid everything out. It's time to put it all together!

The Mulebox in a 1590DD enclosure is a tight fit. You want to think carefully about cable length and dress, and avoid any components unintentionally touching, which could lead to grounding issues or shorts.

Here is an illustration of the final component layout, as if looking down from the top of the inside of the lid (i.e. the base of the unit), with the front- and back-mounted components seen from above.

You will mount the largest components to the removable "lid" (in our case the base), but the other components (jacks and pots) are on the back and front long sides. This creates a problem, because you need wires running from the base to the front and back. If these are too short, you will be unable to connect them. If they are too long, they will get in the way of the internal components or block the fan (and increase the risk of shorts) when the box is closed.

We recommend that you assemble the unit with the lid and box opened like a clamshell, lying as close together as possible along the rear long edge. In this position, run cables that are as short as possible from the rear jacks and pots to the relevant points on the base. You can have a bit more leeway with the front components, which use thinner wires, though it's even more elegant to use small plastic JST connectors that can be done up with the lid partially closed. Care and planning is everything.

(Note: The build in this photo is not identical to the suggested layout above – the inductor coil and 25W resistors have been swapped around due to a drilling error. It is however electronically identical. This uses wires with JST connectors for the front-mounted components.)

It can also be useful to fit only three of the four PCB corner mounts to begin with. This makes it easy to take the PCB out so it can float inside the enclosure with the lid open. All the wires to the front of the unit run directly to the PCB.

Here is one feasible layout that aims to minimise cable runs. Note that the wires that attach to the resistors, inductor coil, input jack should be 22AWG (thicker, handling the higher current of the amp signal). The ones connecting to the PCB should be thinner (handling only the electronics and line level audio signal).

You can largely build the Mulebox in any order you want, but here is a suggested approach.

The cooling fan can be one of the most fiddly components to install, because if there is any pressure on its housing, it will likely either stop or run noisily. Unfortunately, there is only just enough space for a 30mm fan inside a 1590DD enclosure, and the sides also slant outwards slightly, making it difficult to mount the fan cleanly. We therefore recommend getting the fan in place first and testing it with a 9V battery.

1. Mount the finger grille from the outside of the enclosure and run 25mm M3 bolts through at least two screw holes. Depending on your model of fan, you may only be able to use two or three bolts in any case!

2. Put a nut on the screw and tighten on the inside of the enclosure. This will keep the finger grille in place, but also create some spacing at the edge, which will likely be necessary to allow the lid to close without compressing the fan housing.

3. Slide the fan onto the bolts, with the direction of air flow towards the outside of the enclosure (there should be a small arrow indicating the direction, but you can also power the fan up and confirm that air is blowing out through the finger grille). Do not use too much force when installing the fan and check for any bending or warping of its plastic housing. If necessary, carefully enlarge either the holes on the fan housing or the enclosure or both, to make it possible to angle the bolts appropriately. You want the fan to be scurely in place so it does not vibrate, but not under any bending tension.

4. Whilst doing this, frequently test that the fan runs, by connecting it to a 9V battery or temporarily wiring it to the power socket. If at any point it stops or makes a grinding noise, you have probably bent the housing. Free the fan and confirm that it works still, then attempt to seat it again.

5. Install the lid and confirm that the fan still runs (it can be helpful to pull the fan wires out of a slat to connect to a battery). If not, check that the inner lip of the enclosure lid has not bent the fan housing.

6. Fasten the enclosure lid screws and confirm that the fan still runs and that air is flowing freely out of the finger grille at the rear of the unit. If not, again look for signs of bending. It may be necessary to put a thin spacer between the lid and the enclosure before tightening the screws, to gain half a millimetre of extra headroom.

Later on, be careful that no wires or other components end up pushing or warping the fan housing.

Once you are confident the fan is properly installed, we recommend tackling the bulky resistors and inductor coil, with their thicker wires.

1. Begin by placing the large resistors (1 x 100W, 2 x 25W) as shown in the layout. Smear a very thin layer of thermal paste on the bottom and bolt them securely to the lid, with the nuts on the inside (i.e. the side where the resistors are mounted) and the screw heads on the outside.

2. Fix the inductor coil in place using zip ties. Be careful that it does not wobble around, and that no part of it touches any metal components, including the enclosure.

3. Screw in place the rear and front-left PCB standoffs, but leave the front-right standoff off for now.

4. Insert the amp-in jack at the right of open (upside-down) enclosure, making sure to use an isolated jack as specified so that no metal part of the jack touches the enclosure.

5. Mount the 9V power socket, the B5K pot – remember to snap off the key pin that will stop it from sitting flush with the enclosure, and to add the plastic dust seal to the rear – and the two output jack sockets. This completes the rear of the enclosure (which will be lying, upside-down, close to the rear of the lid if you have placed them like an open clamshell).

6. Run 22AWG (thick) wires between the large resistors, inductor coil, and rear jack as shown in the wiring diagram. If you have the enclosure open and upside down, realise that the diagram is a bird's eye view onto the inside of the lid, so the rear-mounted components will be upside down! Keep wires as short and neat as possible, and strip enough insulation off to wrap securely into the solder lugs, but no more. Double check your wires, then solder everything to the large resistors and inductor coil (i.e. the components mounted to the lid). Don't solder the input jack socket yet. Use your multimeter to double check continuity where it is expected.

7. Run an extra length of 22AWG wire from the sleeve (ground, closest to the enclosure wall) of the jack socket to the pin of the B5K pot at the rear that is closest to the input jack socket if you mounted the socket with the lugs facing out of the enclosure (i.e. "up" when the enclosure is upside down) – pin 1. You can now solder the ground wire from the large resistors and this new wire to the input jack socket.

8. Solder one side of the 2W, 3K3 "small" resistor to a wire, and protect the exposed leg with some heatshrink tubing. Do not put heatshrink over the resistor body itself! Make sure the wire is long enough to reach from the tip (hot) of input jack socket to pin 3 of the B5K pot on the rear, which means the pin that is furthest from the input jack socket, without getting in the way of the fan when the lid is closed. Solder the wire end to the tip lug of the input jack socket alongside the wire that runs to the large resistors on the lid.

9. Place the other "small" resistor – 1K, 1W – between pins 1 and 3 of the B5K pot. Insert the ground wire from the sleeve of the input jack into the lug for pin 1, and the loose end of the 2W resistor that connects to the tip of the input jack int the lug for pin 3. Solder a further short wire to the same lug as the ground – we will use this later to connect to the "ground bus". Keep the exposed resistor legs as short as possible, and solder both lugs, snipping off any excess. Leave the wiper (middle) lug of the B5K pot unsoldered for now – this is our line out to the Hothouse PCB.

At this point, you should have all three large resistors, the inductor coil, and three of the PCB standoffs securely mounted to the lid, with thermal paste under the large resistors. The fan will be in place, and the input jack socket will be connected, with two wires running across towards the B5K pot. The wire from the tip of the input jack socket will end in a 3K3 2W resistor. A 1K 1W resistor will be soldered across the outer lugs of the pot, one side shared with the 3K3 2W resistor and the other side shared with the wire to the amp in sleeve (ground).

We will soon come to the DSP part of the circuit, with the front and rear-mounted pots and sockets, but first we need to prepare the resistor ladder on the rotary switch, which is much easier to do outside the enclosure.

1. Identify pin 1 and pin 12 on the switch. Insert a thin wire and one end of a 1K 1/4W resistor in the lug for pin 1 and solder in place. Insert the other end of this resistor in lug 2, alongside one leg of a second 1K resistor, and solder these too. Continue this up to pin 12, where you should also insert a second thin wire. The idea is to create a "ladder", where there is a 1k resistor between each adjacent pair of pins, and also a wire at the first and last pins. Make sure to cut away any excess resistor legs, and double check each solder joint.

2. The common pin should be in the middle. Solder a third short wire to it.

3. Mount the rotary switch to the front panel of the enclosure.

You can test the rotary switch with a multimeter. It should measure close to zero resistance between the pin 1 wire and common in position 1, 1k between those same wires in position 2, 2k in position 3, and so on. In this regard, it will act like an 11k potentiometer where the measured value between pin 1 and the common moves in steps of 1K as the switch is turned. The value between pin 12 and the common will be 11k in position 1, 10k in position 2, and so on, and 0 in position 12. The resistance between pin 1 and pin 12 will always be 11k.

If you get unreliable readings, double-check your solder joints. The rotary switch is not complicated, but it can be tight with 11 resistors laddered, and so there is a possibility of shorts or poor connections.

Here is a close-up of a completed rotary switch with the 1k resistor ladder.

We will start by installing the PCB and mounting to the components on the rear, which, if you leave the enclosure open as suggested, turns out to be easier than the components on the front.

1. Complete the Hothouse PCB, outside the enclosure. The surface-mounted components should already be in place if you ordered it from Cleveland Music Co. or used the PCB assembly service, but you will need to solder in the female headers that seat the Daisy Seed itself, as well as the 100uF electrolytic capacitor. Remember that these are polarised! Do not solder any pots, switches, or ribbon cables to the Hothouse PCB.

2. Place the PCB into the three installed standoffs. This helps you gauge the length of wires required. You can solder wires to the top of the PCB, though be careful not to touch any other components while you do!

3. At the rear of the PCB, solder a wire from the "LIN" (left input) pad to the middle lug (wiper) of the B5K pot on the rear. This is the audio input, and since it's mono, we'll use only the left channel. The "RIN" (right input) pad will remain empty.

4. Solder a wire from the "ROUT" (right output) pad to the tip lug of the right output jack socket, and from the "LOUT" (left output) pad to the tip of the left output jack socket.

5. We will now create a ground bus. The goal is to connect the GND pad at the rear of the PCB to the sleeve lug of one of the nearest output jack socket, and then run that to the sleeve lug of the adjacent output jack socket, the ground pin of the B5K socket (which itself is connected to the sleeve of the input jack socket), and eventually to the ground pin of the 9V power socket. There are a few ways you can do this, but one way is to use a short piece of solid core wire that has been fully or partially stripped, placing this through the ground lugs of the jack sockets as a bar that you can then wrap other wires around before soldering them together. You want to avoid ground loops (multiple parallel ground paths), and ensure that any exposed wire cannot come into contact with any other components and accidentally create a short.

At this point, the audio connections to the rear of the PCB will be in place, and there will be four trios of wires that need to connect to the front of the enclosure – three for the LEDs, three for the rotary switch, and three for each of the two B10K pots (reverb mix and output level knobs).

If you have left the enclosure open like a clamshell along the rear long edge, you have been able to run short wires to the PCB and large resistors that will fold relatively easily into the unit when the enclosure is closed onto the lid. However, with the enclosure opened like this, the components that will be at the front will now be upside down and far to the back. Ideally, we want to avoid very long wires that run almost double the depth of the unit. There are a few different strategies you could employ:

• Run long, thin wires, and try to fold them in as neatly as possible. This can be messy and increases the risk of the wires disturbing airflow or getting in the way of other components.

• Measure an appropriate length of wire and solder the wires in place before mounting the components to the front panel. Then "unsnap" the PCB from the standoffs (which is most feasible if you only installed three standoffs to start with) and carefully fold the PCB back over the open enclosure, so it floats upside down whilst you mount the front components. The problem then will be re-seating the PCB securely as you close the unit up. The longer the wires, the easier it will be to get your finger into the enclosure to do so.

• Solder the wires with the front-mounted components loose, and only install them as you close the box up. This is most tricky with the LEDs, since these need to be pushed into the bezels from the inside.

• Use plastic triple JST connectors crimped to wires. Solder the wires with the female connectors to the PCB and the wires with the male connectors to the pots, switch, and LEDs. Cut them to a length where they can be connected with the enclosure only slightly open at the front. This creates neat wiring, but you have to be careful when opening the enclosure not to snap the connectors or wires.

Some trial and error may be required, and it might be a good idea to do a single pot first as a test.

The images above illustrate using JST connectors for the front components. Here is a photo of the "flying PCB" approach:

Regardless of your mounting strategy, the connections need to be made as follows:

1. Solder trios of wires to the pads for pots 4, 5, 6 on the Hothouse PCB. The pads are quite large, and it is important that you get the wires to touch the metal edges of the pad while you solder. Do not allow the wires to float in the middle of the pad in a pool of solder, as this can lead to a poor connection. Leave the other pot and switch pads empty - we won't use them.

2. Solder another trio of wires (or a trio of wires with triple female JST connector) to the LED1, LED2, and GND pads at the front of the PCB. These will run to the LEDs. Leave the other pads on the front (FSW1, FSW2, PIN1) empty – we will not use them.

3. Install the LED bezels in the enclosure. Temporarily push the two LEDs (red for the top light, which will be at the bottom of the enclosure if opened, and blue for the bottom light) into the bezels to help gauge the length of wire required.

4. Identify the negative (short) legs of the two LEDs, angle them so they touch securely, and solder them together. Make short loops for the positive (long) legs. Solder the LED1 wire to the red LED's positive leg, the LED2 wire to the blue LED's positive leg, and the GND wire to the two negative legs that are soldered together. Use heat shrink or careful positioning to ensure that the positive legs can't touch the metal bezel or any other components, creating a short. The close-up photo above of the front-mounted components shows the wiring of the LEDs.

5. Wire the rotary switch to the POT4 pads. Pin 1 (the rightmost pin when viewed from above) is marked on each trio of pot pads, and should connect to pin 1 of the rotary switch.

6. Wire the two B10K pots to POT5 and POT6, respectively, on the Hothouse. You need to take special care to ensure you do not wire them backwards. Again, wire the Pin 1 pad to pin 1 of the pot. If you look at the Hothouse build guide you can see how it expects the pots to be mounted directly to the rear of the PCB in a certain orientation, with the body of the pots facing up towards the audio and power pads (RIN, ROUT, LIN, LOUT, GND, 9V) that in the Mulebox are found at the rear. You can use this to visualise which pin of the pot is meant be in the Pin 1 pad on the PCB. If you get it wrong, the pots will work backwards.

7. Depending on your mounting strategy (especially if you are using wires with JST connectors), you can now mount the final PCB corner standoff to secure the PCB in place.

We are almost done! The only installation that remains is the power.

1. Connect the negative (usually black) wire from the fan to the negative (centre pin) lug of the 9V power socket. (If you are unsure which lug is negative, use a multimeter to check continuity with the centre pin of the socket.)

2. Also connect this to the ground bus at the output jacks, and solder in place. All the ground points – the sleeves of all three jacks, the GND pad on the PCB, the ground sides of the large resistors, etc. – should now be connected.

3. Connect the positive terminal of the 9V socket to the 9V input of the Hothouse PCB, and the positive (usually red) wire from the fan, and solder in place.

1. Flash the Daisy Seed with the latest firmware if you haven't already done so (see above).
2. Carefully insert it into the header pins, so that the USB socket faces to the left side of the unit (i.e. are on the side of the PCB closest to the edge of the lid, furthest away from the large resistors). This may need a bit of pressure, but you should be careful not to bend any of the pins.
3. Close the lid, connecting the front components or their wires as required.
4. Insert the lid screws.
5. Screw in or stick on rubber feet to each corner if desired, so that the bolts in the lid are clear of the ground. Lifting the enclosure a bit can also help heat management.

Before connecting the Mulebox to your amplifier, plug in the speaker cable and use a multimeter to measure the DC resistance across the tip and sleeve of the other end (i.e. the end that would go into your amp). It should measure about 8 Ohms, give or take. If you get an "open loop" (no connection) or very low (e.g. less than 7) or high (e.g. more than 10) resistance rating, do not plug your amp in. Open the enclosure and look for incorrect connections.

Then, plug in the 9V DC power supply – make sure it is centre-negative – the same type of power commonly used for effects pedals. The fan should start to spin and you should see a small light on the Daisy Seed through the top slats in the enclosure. If not, there is a problem with the power connections.

The red and blue LEDs on the front panel should also go on, initially blinking. This means the firmware is running and an IR is loaded correctly. If not, something may be wrong with the software build.

Connect the stereo outputs of the Mulebox to a mixer or interface that you can monitor with speakers or headphones. Listen for any noise before connecting the amp. If it sounds unduly noisy, look for faulty or dry solder joints or damaged wires.

If all is good, plug the speaker cable into the 8 Ohm speaker output of your amp and turn it on. Keep the volume low on both the amp and the Mulebox (both the trim knob at the rear and the line out knob at the front). Try to play something and turn up the volume knobs slowly.

Look for clipping (the red LED will blink if the input is clipping; the blue LED will blink if the output is) and turn down the trim pot at the rear or the line out pot at the front if required.

Before jumping straight to the main firmware, you can verify that your LEDs, front-panel knobs, and rotary switch are wired correctly using the included test programs.

1. Knobs test: Build and flash the knobs test using make TEST=knobs program if using a debug probe, or make TEST=knobs program-dfu if using USB. When running, turning the two front knobs (reverb mix and output level) will independently adjust the brightness of the red and blue LEDs. If an LED doesn't light up or the knobs don't smoothly alter the brightness, check your wiring to POT5, POT4, and the LED pads on the PCB.

2. Rotary switch test: Build and flash the rotary switch test using make TEST=rotary program if using a debug probe, or make TEST=rotary program-dfu if using USB. In this test, turning the rotary switch will change the brightness of the LEDs. The brightness should jump in distinct steps as you turn the switch. If it acts erratically or only works in some positions, double check the resistor ladder on the rotary switch and its connection to POT6 on the PCB.

Once you have verified the hardware works as expected, flash the main firmware by simply running make and make program (or program-dfu).

• Project license: GPL-3.0 (see LICENSE)
• Main application code: GPL-3.0, unless otherwise noted
• Hothouse abstraction layer: GPL-3.0 (Copyright 2024 Cleveland Music Co.)
• libDaisy: MIT License
• DaisySP: MIT License

• Daisy Documentation
• DaisySP Documentation
• Cleveland Music Co. Hothouse
• Hothouse Examples