Hey everyone,
I’ve build quite a few Moritz Klein modules so far, but the sequencer I’m currently working on is really driving me nuts. Everything works fine, except for the gate output, which I can’t get working properly. It outputs very weird gate signals, but mainly it seems to pick up the clock signal from somewhere and outputs it as gates. I did all kinds of measuring and experiments, but couldn’t figure out what the problem is. What I can say is, that
I haven’t done any mistakes in the build (build it twice on breadboard and another time on stripboard with the same results)
There are no defective parts. I measured or swapped out all relevant components to test this
The inputs to the comparator measure as they should
The problem is still present, when the clock circuit is entirely disconnected from the rest of the circuit, thus indicating, that the clock signal might get transferred through the power or ground rails
Any idea on what the actual problem might be and how to solve it?
Do you have a scope? Is the clock signal visible on the comparator output (pin 7)?
Of course the comparator and clock op amps are on the same chip, so one could imagine some kind of crosstalk happening inside the chip. But I’d be extremely surprised if that were the explanation. The comparator output is about ±10 V. To get bad gates you’d need crosstalk at the 20 V level! I also have a hard time imagining how power rail noise could do it inside or after the comparator.
And you say the inputs are correct. The threshold voltage is about 4 V, so again, you’d need huge noise at one input or the other to make bad gates.
It’s all very weird. Were all the components sourced from reliable suppliers?
I’d suspect a problem with VD6 but you say it still happens with the clock entirely disconnected from the rest of the circuit, so that can’t be it.
Does it still happen with the comparator input (pin 5) disconnected?
Yes, it’s visible, but not always in pure form, but with additional spikes I presume coming from the actual gates and/or the reset signal.
The breadboard-build (using only one TL074 chip) suddenly works as expected, when I fired it up today… although I didn’t make any changes. The stripboard build still shows the same problem. Here the clock and gate comparators are realized with separate TL072 chips.
Yes, no Aliexpress or similar sources involved.
This was my first guess as well and the very first thing I tested. VD6 is fine.
I’m reluctend to solder it out again, but it definitely still happens, with no input to the diodes VD8 - VD12
Further observation: It stops, when I disconnect the clock from its buffer, but not when I disconnect the clock signal AFTER it is buffered from the rest of the circuit!
It could well be a problem in the stripboard layout. From my experience with guitar pedals that have LFOs, this is a very common problem. Perhaps you could try a different layout and try to keep the grounds of the two sub-circuits separate for the most part… I’d definitely try a different op amp for the clock first.
I think I will try this later this week. For now, after three days of troubleshooting, I’ve had enough of this circuit…
The breadboard build suddenly working bugs me even more, than both builds not working. I spent the last two hours with a different approach, trying to reintroduce the problem on breadboard, but to no avail… it suddenly runs so stable as if there never was any problem. Really drives me crazy. It’s build exactly as the stripboard version. I checked every connection a dozen times, replaced every part over and over. The mystery remains
Ive been (somewhat) keeping up with this topic, and i realize from what I’ve read that you have been very thorough with your check and rechecks and rebuilds. I haven’t noticed (i may have missed, obviously) any mention of your decoupling circuit. I’m guilty of NEVER doing it as per schematic. This is partly from me not really understanding it on a precision level and honest laziness. Cramming extra stuff on a breadboard is annoying and the cap values i sometimes just don’t have available for having used them in the circuit on the breadboard often enough. I almost never go to strip board without a successful breadboard first and even then extra stuff is that much more of a headache when it comes to soldering as close as possible to the ic. I have no better excuse than laziness and indifference when the circuit is working. That being said i KNOW there are circuits I’ve had fail or not work properly because of my general misunderstanding and often misapplication of the decoupling circuit. This includes my first go at a 4017 sequencer. So…how’ve you done the decoupling caps? Lol
I never add them on breadboard, but always on my permanent stripboard builds. Here I kinda like cramming extra stuff in… planning the stripboard layout is to me like doing a puzzle and i always challenge myself to get it as compact as possible.
I try to get them as close as possible to the power pins of the IC, which usually means right next to them. From what I’ve read it’s best to avoid path resistance when connecting them to ground, which of course isn’t that easy on stripboard.
Usually I wouldn’t go to stripboard, before I got something working on breadboard, but this time I took the chance, because I attributed the weird behaviour to the general wonkiness of breadboard builds and I trust the erica x mki schematics.
That said: I seem to have found a solution just now! Adding a resistor to ground after the diodes VD8-VD12 seems to solve the problem. Still don’t understand how and why, but it’s a relieve for sure
That’s strange, because R19 on the schematic is already a resistor to ground after the diodes (as well as being part of a voltage divider with R20). The resistor to ground in combination with the diodes makes a half-wave rectifier. It sounds like your “very weird gate signals” might have been resulting from the negative half of the square wave that the 4017 puts out (I think, just going from memory here, haven’t checked the datasheet to see whether it puts out a unipolar pulse). Have a close look at R19. Is the solder joint dodgy? Does the stripboard layout actually take it from TP6 to ground?
I’m pretty sure you’re right and the main issue was a problem with R19 not being properly connected. This would have left the input to the comparator floating in between the pulses from the 4017, which is probably the reason for the weird behaviour. I build this one into a “baby g8” sequencer (because I didn’t have a blank panel big enough to include 8 pots) and due to all the wiring it’s not that easy to do further testing now. But I’m soon gonna give it another go and build a CV sequencer… then I will pay special attention to it!
I did something similar. I had the extra 4017 and Jack sockets but no pots and I didn’t want to use reset switches. So I made a gate house with gate switches to a single channel using the gate circuit from the Moritz Klein schematic, and an output jack for each step 1-8 coming straight from the 4017 as a separate trigger out and to patch the reset through a jack into the reset pin.
The individual outs for the steps are a nice idea! I have a rotary switch I wanted to use for the reset function but I couldnt fit it behind the panel. Now i think I’m gonna build a small 2hp expander for the individual outs instead.
I don’t really get the infinity/step 9 output though. Since 8 is the maximum step length in this build, you could hardwire step 9 to the reset input and spare one input jack. If you patch a shorter lenght step 9 just won’t trigger…
I just wanted the pattern length to be variable up to 8 steps and I didn’t want to use switches for it. I dont remember if I had enough switches or not. I just thought having a patchable step count from the trigger out of any step was cool.
Yes, I got that, but for the full 8 steps you wouldn’t necessarily need any patching. If the 4017’s 9th output is wired to the reset input, than 8 steps will be the maximum length, while you could still use your patch points to shorten the sequence
Ohhhh, ok. Never occured to me. I’m still quite new at this. It was about a year ago when I made the module. It also happens to be on breadboard. At the time I had just finished the baby8 by following Moritz Klein’s video. It said nothing about how to switch the resets and I had 8 steps so I had to figure out how to wire the switches on my own in a way I could understand. It was way overcomplicated (part of the reason I left the switches out of the gate house for sure). I’ve since studied a bit on @Dud schematics and I understand it a bit differently now. I see what you mean now. Yes that infinity jack is useless, but I do have a module with an infinity out jack…just saying…
I think the problem could have been caused by the clock signal being coupled through the diodes internal capacitance, the extra resistor to ground would steal most of the capacitively coupled current through the diodes effectively eliminating the unwanted clock signal.
Never occured to me. I’m still quite new at this. It was about a year ago when I made the module. It also happens to be on breadboard. At the time I had just finished the baby8 by following Moritz Klein’s video. It said nothing about how to switch the resets and I had 8 steps so I had to figure out how to wire the switches on my own in a way I could understand. It was way overcomplicated (part of the reason I left the switches out of the gate house for sure). I’ve since studied a bit on 
