Works really well and tracks well.
16/08/17 I also bought (dirt cheap) a non-functioning VCO from a forum that I will fix (it was cheaper than the faceplate on its own! One of the transistors was soldered in the wrong orientation. A quick swap and it worked fine.
Looking at the schematic it looks really similar to the Doepfer a110-1 (it sounds really similar).
Great sound. They track well over four octaves and seem stable on long sessions. They also go very low so double up as LFOs.
Worked straight away at power up. Great envelope generator - may have to build another one !
25/10/18 I did build another one ! (Very painless and easy build - no issues).
Had a few issues with this build - one of the jacks was shorting a resistor (which was located underneath the jack !) One of the LFOs was giving a mix of both LFOs. Also the LED wouldnt light up - worked when I reversed it (although that meant inserting it on the PCB opposite to the polarity marked on the PCB !) All working now though.
Really nice sound - very vintage. Love it !
Really impressive module - quite deep and you can do an awful lot with it.
The x-y scope demonstrates the strange attractors. I tried using toner-transfer to make the front panels - needs a little practice!
Simple passive multiple. Bought the panel for £1.75 (!) from the pusherman group. I wired it as either 2x one-to-three or 1x one-to-seven (depending on whether the jack into the fifth socket is inserted or not).
I bought one board from amazingsynths.co.uk and one from the pusherman facebook group (much cheaper !). I wanted to check if there were any quality differences before soldering more complex boards. They both seemed equally good.
The silver panel is a Greyscale panel (bought secondhand from the muffwiggler board) and the black panel is from pusherman.
The boards were surprisingly easy to build. Initially neither worked. The purple board had a cold solder joint on one of the opamps. A quick reflow and it was working perfectly. The black board was a pain to debug. It needed lots of probing with an oscilloscope and multimeter and hours of poring over the schematic ! In the end one of the resistors in the filter cascading circuit was open circuit ! (I hadn't heated it much at all so I have no idea why it was dead). After swapping it out the filter worked perfectly.
The following photos show the build process (the oscilloscope is showing the VCA working correctly).
24/02/18 Ornament and Crime is a multipurpose module for generating CV. It only exists as a DIY module. It uses a more managable size SMT (0805). It is based on a Teensy 3.2 Arduino board and runs firmware inspired and adapted from multiple sources (quantiser, Turing Machine, various Mutable Instruments modules).
It was a fairly straight forward build and worked on power up. The image below shows the AD generator (Piques, adapted from the Peaks firmware) running in the oscilloscope trace.
The encoders I used send a reversed signal compared to the factory encoders (Bourns vs Panasonic), so I needed to modify the source code and recompile the firmware to enable the correct direction of the encoders.
Sounds amazing and surprisingly painless builds !
23/04/18 It's so good, I built another one! (Partly so I could set up drums easily).
Very straight dorward build. Power section had no problems. Flashed immediately within the Mutable dev environment. Fired up fine. The only issue was no audio. A quick trace with the oscilloscope showed signals going from the STM to the DAC but no onward audio. A quick look at the schematic and it showed pins 3 and 4 should be bridged. Tapping both with the scope probe produced audio. A closer inspection showed pin 4 to be slightly raised off the pcb and not making contact. A quick bend with the tweezers and a dab with the solder and it was up and running.
I have another board and panel to build another one at some point in the future!
The only issues with the builds was setting the zero offset in firmware. This is a hidden mode entered by pressing one button on boot up. I just could no enter it at all. I set up the debug client and server and peeked around in gdb. It all looked good, just wasn't registering the button. I think it was the debounce code not registering it. In the end I just set a breakpoint after the check and poked the right bit so that the check passed. Continuing on with the code meant I could easily calibrate both modules fine.
28/05/18 A month later I built another one. I flashed this with a hacked version of Dead Man's Catch alternative firmware, modified to include the latest factory Peaks code, along with software calibration. I needed to use openocd again in order to calibrate the zero offset.
Very easy builds. No issues at all.
It's worth making a note for the future about flashing them. I used an Olimex AVR-ISP-MkII. I had a real pain trying to get it to work on a Mac with avrdude in the Mutable dev environment. I tried it on a PC using Atmel Studio and it worked fine. I know I had to change the firmware for it to work with avrdude so I used Atmel Flip to do that which was straight forward. However avrdude still wouldn't read it. I saw that some people had had problems with avrdude 6 (used in the mutable dev environment), so I downloaded the avrdude 5.9 sources, installed the dependencies (usblib, yacc, etc) and compiled it. Hooray, it worked. Unfortunately avrdude 5 config file didn't have the m88p config. So I copied it from the avrdude 6 config and pasted it into the v5 config file. After that, it all worked trivially easily.
This would have been a nice straight forward build if I hadn't decided to try to program one late at night after going to the pub ! I think I bricked one of the MCUs when programming it from Atmel studio (I decided it would be a good idea to use this software for some reason !) I think I must have set the wrong fuses. I tried reflowing the crystal (removed it to check for bridges/problems) and it was fine. In the end (probably a bad beer-influenced decision) I removed the MCU with hot air. This meant I needed to buy another one on my next Mouser order. This time there were no issues programming it from the Mutable Dev Environment. There were also no soldering issues and it worked first time.
Flashed and built the firmware from within the Mutable Dev Environment. Very painless and straight forward build.
Both were quite fiddly builds with all the passives and ICs quite close together. However, they both worked perfectly first time with no errors.
Above photos: panels and pcbs, large pile of components, power section complete and tested, STM32 MCU and associated decoupling caps installed and connecting the STLink2 programmer, flashing the MCU, jacks and pots inserted but not soldered, testing, completed.
Both worked perfectly first time, with no issues.
The above images show the: bare pcb and panel, the power section built, the microprocessor soldered, flashing the microprocessor, finished soldering the surface mount components, testing all the inputs and outputs before soldering the through hole components, and the completed module.
It was a surprisingly easy build with lots of space and it all worked first time on power up with no issues. It sounds very nice !
Very quick and straight forward builds. One worked perfectly straight away. On the other one the second and third channels weren't working initially. I had a couple of dry joints on two opamps that was a pretty much instant fix. (Initially the iron wasn't heating the joints well - the tip wasn't well-tinned and I think that was the problem with those op-amps).
Very useful modules - offset/attenuator/attenuverter/mixer.
I ordered an 0603 version of Ears with matching panels from www.allpcb.com. (The factory Mutable Instruments version is 0402).
The images above show: the bare boards and panels, surface mount parts soldered, the completed module, testing the gate output with external audio, testing the contact mic and finally the three modules completed.
Straight forward builds. I substituted 1n5819 diodes for PMEG2010AEBs as I didn't have any. I had to stack two 0603 470n caps in parallel as I didn't have any 1uF caps of suitable voltage. Would have worked first time if I wasn't such a numpty and used 6v "special" lm324 opamps instead of general purpose ones !! (Who needs LEDs when ferrite beads glow so pretily). A quick hot air rework and they were swapped out. They then worked fine with no issues.
22/10/18 I thought I'd build one more so I can have one on each channel of a four channel polyphonic voice:
Not much to say. Trivial builds. Makes things quieter !
27/09/18 I built another one. This had an error on the pcb. One of the tracks from the input of channel 2 was shorted to the pot of channel 4. I broke the track and jumpered it with a cut off piece of a through hole passive.
Straight forward build with no issues. Images show: bare pcb and panel, power section built and tested, MCU and associated components soldered, flashing the MCU, module completed apart from control hardware, control hardware placed but not soldered, testing the module, the finished module.
Images show: bare pcbs, 0402 resistor size, populated boards, testing and then the completed builds.
These modules use teeny, tiny 0402 surface mount parts (see photo above of a resistor with my manky thumb for reference ! I was quite nervous before starting these as they are very closely populated boards with tiny parts - the diodes are really fiddly and very tricky to see the stripe which marks polarity. For these builds I still hand-soldered them with a soldering iron and solder. I did consider using solder paste and reflowing with either skillet or hot air but decided to stick with a soldering iron, which I'm glad I did.
Apart from the eye-strain, these were reasonably straight-forward builds in the end. All three had issues in testing: the tl074 opamps all had dry joints on some of the legs (strangely the opa4171s were all ok). These were pretty easy to sort out - just touching up the opamp legs. After that they worked fine.
Images show: bare pcbs, pcbs with the pile of parts, power section of the boards populated, MCU and associated parts installed - front, MCU installed - rear, flashing the MCU, completed build - front with no face plate, completed build - rear, completed build, calibration (more than 1/1000 cent accuracy in this image!).
Straight forward builds. Only one issue - on one of the modules the display didn't fire up initially (I had wondered if I had a questionable short on the MCU but it looked ok visually so I had left it - it was on the display data pins of the MCU so I quickly reflowed that pin and it worked perfectly straight afterwards. I had bought encoders that were reversed (compared to the factory modules) - intentionally as that was what was available. I needed to make some changes to the source code to change the direction of the encoders which was trivial (and indeed I had made those changes before startiung the build).
Images show: bare pcbs, power section populated, MCU and associated parts installed, flashing the MCU, completed build - all surface mount parts placed, completed build - all through hole parts placed as well as the led level metering sub-board, testing the module, completed module.
Fiddly but straight forward build. The main hassle was cutting down the surface mount socket mount to fit the sub-board assembly. The module worked first time with no remedial work needed.
Images show: bare pcbs, power section populated, MCU and associated parts installed, flashing the MCU, completed build - all surface mount parts placed, completed build - all through hole parts placed, testing the module, completed module.
Pretty straight forward build. I had two issues: I missed a bridge on the DAC (that was simple to fix) but there was an issue with the addition of the keyframes - the frame light stayed on a long time and seemed inconsistent. I worked out it was a dry joint on the ADC pin of the frame pot. It took at least an hour to work out if it was not working correctly and then what the problem was! After that it worked perfectly.
PCB and panel bought. However, looking at the schematic, there was an error. The output resistor is included in the feedback loop, leading to the possibility of opamp oscillation in the presence of a capacitive load (eg with long patch cables). For peace of mind, I decided not to build it and used the parts to build two Links modules instead.
I came back to this (28th August 2018) and spent a few hours laying out a circuit and pcb based directly on the buffered multiple part of Links (with normalled connections between groups). I sent the pcb off to allpcb.com (a Chinese pcb manufacturer) and am awaiting 5 test boards. I have enough parts to build three of them if I want.
On the first run of pcbs, I made a mistake in Eagle and didn't create vias to join the ground planes front and rear, so needed to join up three islands of ground. Works perfectly when that is done. (The image showing the module being tested with the oscilloscope has a 'Y' offset on the scope to see both traces clearly - it's not a voltage offset !)
02/10/18 Built another two own-design buffered multiple. I corrected the Eagle files and have bought some more test boards.
Modules work perfectly - very straight-forward builds. The files can be found on my github.
Images show: bare pcbs, power section populated, MCU and associated parts installed, flashing the MCU, completed build - all surface mount parts placed, completed build - all through hole parts placed, testing the module, completed module.
Both worked immediately with no issues.
After testing one of them, I built the other two:
Worked perfectly first time. To compile the firmware I needed to change the source code change as well as change to the avr-gcc flags to enable compatability with latest avr toolchain.
Images show: bare pcbs, power section populated, MCU and associated parts installed, flashing the MCU, completed build - all surface mount parts placed, completed build - all through hole parts placed, testing the module, completed module.
Both worked immediately with no issues.
On the newer modules (plaits, stages, marbles so far as of Oct 2018) MI ow use SWD to program the MCUs rather than jtag. I jumper the pins of the STLink V2 to the SWD pins on the board. With the STLinkV2 pin 1 (target voltage) needs to be jumpered to pin 19 (3.3v *from* the programmer. Here is how I wired the jumpers to flash marbles:
In order to calibrate the dac, edited the source code to output 1v and 3v. I then measured the 4 dac outputs for both 1v and 3v. From knowing the dac scale and offset as well as the calculated dac code to output a desired voltage I derived the scale and offset for each channel. I then hard-coded this into the firmware.
To calibrate the adc, long-press the rate and spread buttons. Feed 1v into rate input, press one of the two buttons, feed in 3v, press button, feed 1v into spread cv, press button and then feed 3v into spread cv, press button and the calibration is done.
No issues with the builds - both worked first time. It's an awesome module!
No issues with the build - worked first time. Wishing I had built one ages ago!
Straight-forward builds, no issues, worked immediately with no problems. I am still awaiting parts to complete the last of the four.
I built my own panel (and didn't do a very neat job ! Trying toner transfer technique - I think I need to try more glossy paper next time). Straight forward build. The only issue was me being a numpty and installed the two power rail filtering electrolytics in reverse - bang! First time I've installed electrolytics the wrong way and first time I've blown up electrolytic caps! Other than that, it worked well!
Straight-forward builds. (Quite long and tedious though). No issues. Worked straight away with no problems.
I bought 10 pcbs from www.pcbway.com and also ordered 10 panels from www.allpcb.com.
Boom!!! Simple build. Just needed to adjust the value of the pull-up resistor on the accent to allow a nice "thump". Now responds nicely to low (eurorack levels) trigger voltage and low voltage accent (with impedance of ~1k) as well as nice and loud when no accent plugged in. I used a tl072 in the end (just used a jrc4558d initially as I had no spare tl072s at the time. I'll probably build a couple more.
