Gecho v2 prototyping is done

The final Gecho v2 prototype

Since the previous update about v2 that was posted in August last year, it took 3 more revisions to get here. Although the first prototype of Gecho v2 has been completed, blueprints verified by our skilled hobbyist friends who hand-assembled their own boards, and it even passed CE/FCC testing with flying colours, so we could go ahead with the production - it was just not good enough. It needed a small upgrade or two. And, things tend to turn more complicated than they appear :)

Thanks to everyone for encouraging comments, questions and feedback during the past few months, it has been very helpful and contributed a lot to shaping the new model.

While working on Glo (the whale synth), I came to realize that sticking to STM32 platform looks nice in short-term; in theory, only small modifications are needed to port the firmware over. But there was a risk that when implementing the new features you are waiting for, like MIDI and sync modes, it would be needed to trade off something on the sound side (e.g. level of polyphony) to get the extra power for the additional processing required.

The ESP32 platform allows for a very elegant solution: it is a dual-core processor, and has more power than the one we used previously. Here, one of the cores can run sound engine while the other one handles sensors, LEDs and timings. And even the single core can work at 48kHz with more polyphony than the old CPU at 22.05kHz.

In this context, Glo can be viewed as an intermediary step; it already has the new chipset: MCU and Codec that allows for proper line-level inputs and outputs, two differentially wired MEMS analog microphones, better/faster USB driver chip, 8x more Flash, and (with an expansion board, more info about Glo DIY options is here and here) even a small loudspeaker driver and a micro-SD card interface for expansion of sample memory (however, as it is only using 1-bit mode it is slower than in Gecho).

What's changed

  • Gecho is now a bit larger but thinner, has a front panel instead of acrylic plate, electronics is self-contained and wooden box is optional
  • The buttons were moved a bit, so they are not too close together
  • There is a proper Line-out signal available via one of the HP connector, with a switch to select function, also there are soldering pads for fully differential Line-out
  • The input that used to be called "piezo/line-in" doesn't need to use additional analog OpAmp chip as a pre-amp anymore (the codec chip contains all required analog circuitry) and has wider range in terms of controlling input gain, from passive guitar pickup to Line levels
  • There is charging circuit on board, that can handle Li-Po and Li-Ion cells
  • Mini-USB connector has been swapped for Micro-USB, and uses faster driver now. USB-C was just too complex to fit as it has many more signals that need taking care of
  • Magnetic sensor has been replaced with accelerometer, it does not have full 9 DOF with gyro sensing, only reacts to tilt but is sufficient for what it needs to control
  • There is still a magnetic sensor somewhere in ESP32 bound to one of the ADC channels but I haven't tested it yet, will leave this to those of you who want to hack it :)
  • The Flash memory has been upgraded from 1MB to 8MB... and then, as a last minute decision, to 16MB.
  • There is also more RAM and as the CPU is faster, this allows for 48kHz sampling rate with similar delay echo buffer lengths as in v1
  • A Micro-SD card slot was added, it is wired via fast 4-bit bus to a SDMMC controller and works at about 2MByte/sec r/w with cheap SD cards from eBay. It might be useful for expanding samples memory and the throughput should also be sufficient for storing the generated track in real time, but that hasn't been tested yet
  • Electret/condenser microphones have been replaced with MEMS, and are routed with differential signals, which lowers the noise/interference. They are slightly more sensitive too
  • All units will have MIDI, there is a switch to select in vs. out over one 3,5mm Jack which is also shared with two Sync and CV signals (both directions). There was not enough space to fit one more connector, and you usually don't need everything at once anyway; however in realms of DIY, there are separate soldering pads for MIDI in and out if you want to expand to all 3 connectors
  • The standard powering scheme still accepts 3 or 4 AA/AAA batteries (also Ni-Cd, Ni-MH or Ni-Zn rechargeables), while built-in rechargeable cells will be optional & easy to add. Disposable cells still make sense if you mostly use the synth plugged into the 5V USB source, and need them as a backup when you go outside where there are no sockets
  • The CV inputs via 3.5mm Jack will safely accept up to 10V, but they can only emit up to 3V as that's the power level Gecho is operating on. I don't know yet how precise these outputs are, as the testing hasn't been completed
  • In sync mode, similarly the output can only go up to 3V which is usually sufficient for 5V powered devices too (or those powered with 9V but operating on 5V internally), this is still being tested
  • The extra soldering pads to override what the IR sensors do are still present, they can be used as 4 more CV voltages (within the 0-3V range), but this is very experimental, you would need to modify the firmware. I believe with MIDI now present it makes more sense to override sensors programmatically using CC messages.

Software-wise, the new Gecho will be able to do everything that previous one did, and since there is a bit more memory and CPU power, it allows to pack in some more advanced sound effects, or the existing effects with better resolution.

A few photos from testing

Testing MIDI input, receiving notes from Groovesizer Red


The revision #3 assembled


The revision #2 with Li-Ion cell (the circuit is the same as in rev. #3)


Testing with two Li-Ion cells in parallel, still works fine. Charging current is limited to 500mA so it can be a bit slow if the cells are of large capacity.


Li-Po cells work the best, and they come in various shapes and capacity.


The revision #2 mounted in the v1 box, this one still has connectors on the top side so it was possible to do.


3D printed enclosure for revision #3 that has all connectors moved to bottom side. This is to verify dimensions.


There will be a front panel, this is a Photoshop mockup, still waiting for the test panels to arrive from the factory, they should be here any day now.


Will it also be available as DIY?

We know how much you love DIY! There won't be as many options as previously, since only minimal soldering is required (two or three wires to connect battery holder / rechargeable cells). There will be more work involved in assembling it mechanically though, which would make a nice DIY kit. The "master" level DIY kits (bare board + components) might be available later, as it requires extensive amount of work to pack - perhaps a compromise can be made where the small elements are left out (common R/C values that you can get anywhere).


What's left to do

There is one more (4th) revision of the board with smaller fixes, should be on the way from the factory by now. Meanwhile, we are communicating with suppliers, finishing design of on the enclosure, accessories and arranging production. We should have all quotes here in a week or two, so the pre-orders can open, please stay tuned! If you wish to receive a newsletter or watch for updates, please subscribe / follow us on social media.