A while back Marc Debroey (SEBASTOPOLmusic) asked me to take a look at his Meng Qi Karp, which didn’t trigger well. It was only after I fixed the problem that I discovered that it’s a known issue. However, the solution suggested by Meng Qi stops the trigger LED from working. In this post I will discuss what causes this issue, and share how to fix it without sacrificing the trigger LED.
A mystery unfolds
As Marc handed over his module, he told me that it was a second hand module and the package had gotten damaged during shipping. He suspected this might be causing the triggering issue as the components on the board looked quite skewed. I found that some modules were able to trigger the Karp (although not reliably), while others were not. Beyond this point I checked many of the things that typically cause issues, such as bad solder points and fried active components. Unfortunately, this did not yield any results. Hmmm, strange…
With all the easy fixes ruled out, I decided to dig deeper and so I traced the input circuitry, from input jack to the trigger LED and the trigger input pin of the microcontroller. When you do some calculations on this schematic, you find that for a trigger input voltage of 5V, about 2.9V will be present at the input pin of the Karp’s ATMEGA 328F microcontroller. In the datasheet of this chip, I find that the minimum Input High Voltage (V_IH) is 3V. This explains why triggering the Karp is a hit-and-miss kind of thing.
So far, I had not yet taken into account the 1K protection resistor that is present on the outputs of many modules. This resistor causes the voltage at the trigger pin of the Karp’s microcontoller to drop to 2V. This indeed results in the module never being triggered. This also confirms the Karp’s behaviour I observed in my rack. Output protected modules are not able to trigger it, while non-protected modules are able to trigger it, but not very reliably.
In summary, the problem with this input circuit is that there is no buffer for the trigger input, and the trigger LED needs to be driven directly from the source of the trigger. The way the input circuit is conceived, the LED saps away too much current from the trigger source, resulting in a significant voltage drop in the trigger signal. This problem is made even worse if the trigger source has a 1K protection resistor
Meng Qi’s fix
The solution proposed by Meng Qi is to remove R3 from the board. This increases the voltage at the input of the microcontroller to rise to 4.5V. This indeed results in a 100% trigger success rate, but comes at the cost of a non-operational trigger LED.
In the schematic below I show how I have fixed the issue without killing the trigger LED. The values of resistors R1 and R2 are increased to 4K7 and 47K respectively, to reduce their current draw from the trigger input. R3 is removed from the circuit. Finally, an external 4K7 resistor is added so that the LED can be driven directly from the trigger input. As a result, the voltage divider R1-R2 will be less affected by the current draw of the trigger LED. For a trigger input of 5V, this results in 3.5V at the input of the microcontroller, resulting in reliable triggering of the module.
Doing it yourself
In the images below, I show how to patch the Karp board. For those interested I’ve also uploaded my analysis notes and calculations in a pdf. This pdf also contains the instructions on how to implement this fix yourself.
In october 2016 Meng Qi has released a new core PCB for the Karp and DU-KRPLS. Although they don’t mention the trigger issue, I do hope they fixed it in these new PCBs. In any case, I’m sure there are still some pre-2016 Karps out there. There being no schematic available makes the trigger issue tricky to figure out. For this reason, and for my own personal documentation I decided to share my analysis of the problem and how to fix it.