Aion Comet spiky distortion

[Awrange]

Hello,

I recently built a Comet from AionFx.
The pedal is built to vintage specs without the Muff Mod.

It generally works fine, but it sound as if something is wrong with the distortion.
When the input signal goes over a certain threshold the pedal starts to distort in a spiky way. The less gain you dial in, the better you can hear it. When you turn the guitar down, the distortion sounds as i imagine it should. I never played an original Ds-1 so I can't know for sure, but every demo I've heard did not have that spiky sound.

I already double checked all the components and their values and also reflowed all the connections twice.
I'll check with an audioprobe later to narrow the location of the spikes down.

Here is a short audio example with the distortion turned all the way down.
At the beginning the guitar volume is turned down and you can only hear the spikes sometimes, then it is fully turned up and you hear it all the time. It fades away, as soon as the level drops below a certain point.

https://on.soundcloud.com/ZW9Fi

Here are the voltages I measured
       
Q3
C: 8,9
B: 3,5
E: 3,2

Q1
C: 8,9
B: 3,9
E: 3,5

Q2
C: 3,9
B: 0,6
E: 0

IC
1: 4,4
2: 4,4
3: 4,4
4: 0
5: 4,4
6: 4,4
7: 4,4
8: 8,9

Here are photos of the pcb





Thanks!

[idy]

Welcome to the forum.
Audio probe is a good idea.
But I'll get the ball rolling

Your voltages on the IC look good...almost too good. We usually see the non-inverting inputs (3 and 5) loaded by the meter and slightly lower than the others. Maybe you have a very high impedance meter.

The schematic in the docs is cut off, with Q3 missing. It is another buffer on the original, like Q1 but with a 1M resistor providing bias instead of 470k. I expect to see E and B .6v different, these two voltages are closer than that on Q1 and Q3
.
Your symptoms ( I can't get the sound sample, to work, but your description) sound like "blocking" distortion or misbiased opamp, but your voltages look good on the IC.

I don't see any electrolytic caps in the audio path. Those, reversed, will cause trouble.

I see you used sockets for the transistors. They make dodgy contact and could be your problem. Sockets are used when you want to experiment with different types, or "snowflake" germanium parts, but here these are not likely to make much difference. If you are sure of the pinout you are OK. I check the PCB, trace the parts that attach to the Q,  and make sure that I know which hole corresponds to which leg (E,C,B) and then make sure that the transistor I have matches that. Some boards and build docs lie about this... really.

[antonis]

I see you used sockets for the transistors. They make dodgy contact and could be your problem.

This..!!
(sounds like "cold joint"..)

P.S. Welcome also..
Is there any paricular reason for D2/D3 PCB pads have been soldered..??

[Awrange]

Thanks for your welcome and your suggestions so far!

The schematic in the docs is cut off, with Q3 missing. It is another buffer on the original, like Q1 but with a 1M resistor providing bias instead of 470k. I expect to see E and B .6v different, these two voltages are closer than that on Q1 and Q3

Maybe you have an outdated link, for me the whole schematic is showing.
https://aionfx.com/app/files/docs/comet_documentation.pdf
I will check around q1 and q3 tomorrow, maybe I'll find something.

Your symptoms ( I can't get the sound sample, to work, but your description) sound like "blocking" distortion or misbiased opamp, but your voltages look good on the IC.

I had a chance to go over the circuit with an audio probe and the ic seems to be the problem indeed. The signal starts getting spiky after the first opAmp stage (Pin 7). The bias voltage seems to be fine though and I also tested another 4580D, which made no difference.

I see you used sockets for the transistors. They make dodgy contact and could be your problem.

This..!!
(sounds like "cold joint"..)

The connections seem to be fine, but maybe I'll desolder the sockets tomorrow just in case.

Is there any paricular reason for D2/D3 PCB pads have been soldered..??

Yeah, well I started populating/soldering the components without noticing, that these parts are for the big muff mod. I removed them again but did not bother to suck the solder away .

[idy]

I would, rather than desoldering sockets (might damage board) suggest lightly soldering Qs in to sockets. Check for continuity, Q legs to place on the board they should be attached to.

But you say trouble starts at IC. So that is something. Confirm continuity pins 7 and 6. That stage should be just a buffer. Open that path up between 6 and 7 and you get  "open loop gain."

[m4268588]

The low voltage of Q1,3 B is due to the DMM load.

This circuit is incomplete.
Add resistor in series of C2 to lower GAIN. (470 to 2.2kΩ...?)
Change the D4 placement. The method of this document is incorrect except for the xx358 similar OP-AMP.

[Awrange]

The low voltage of Q1,3 B is due to the DMM load.

This circuit is incomplete.
Add resistor in series of C2 to lower GAIN. (470 to 2.2kΩ...?)
Change the D4 placement. The method of this document is incorrect except for the xx358 similar OP-AMP.

Right now I have just clamped the parts together with alligator clips but I'm pretty sure that this was it!
No mandatory spikes anymore, as far as I can tell. Maybe there is a tiny bit left, when I don't put a resistor before C2 but even then the difference is huge. When I have experimented with the resistor value and properly soldered everything I'll give you another update, but I'm pretty sure this is done!

[Awrange]

After some experimenting I did not find, that a resistor for gain reduction was neccessary before C2.

Although I generally know how to do basic trouble shooting I'm still an electronics noob and don't think I would have found the issue with that diode on my own.
So thanks again for your help, especially to m4268588!

[aion]

Change the D4 placement. The method of this document is incorrect except for the xx358 similar OP-AMP.

I'm reluctant to cast doubt on a successful troubleshooting venture... but I don't think this is correct. I have a NJM3404 DS-1 (2006-2016) and it has this diode. It's possible you were thinking of the 1k resistor to ground after the op-amp, which is used along with the M5223AL (2000-2006) and is specific to that exact op-amp but shouldn't be used otherwise.

Here's what ElectroSmash says about the purpose of this diode: (note that it's called D8 in the original DS-1)

The diode D8 is placed to protect the op-amp input. At the op-amp input, signals will float around 4.5V biased by R10, and in case that signal coming from the previous stage tries to drive the op-amp input below the negative rail, the diode will clip at -0.7V  (-VF) not allowing dangerous voltages to be present at the + pin.

Q2's collector is biased around +5.8V, and when going into saturation, without D8, U1 pin 3 would be driven to -1.3V. When an op-amp input is driven beyond the rails, bad things happen. Some opamps are damaged, some will latch-up, most take time for the input stage to recover from being over-driven, and that screws up the sound.

It is important to mention that by the time this diode is conducting, the opamp is already clipping. Therefore D8 has nothing to do with asymmetrical clipping or any kind of guitar signal modification.

(source: https://www.electrosmash.com/boss-ds1-analysis)

I don't love the solution of removing the diode because the diode shouldn't be causing the issue you describe.

I haven't seen the schematic of a NJM4580 version (post-2016 SMD) so it's possible the 4580 has some differences that make it so it doesn't handle this diode the same way. This could be confirmed by replacing the 4580 with a 3404 and adding the diode back in. If it works, then the 4580 may need the diode to be omitted. But if it still happens with the 3404 and the diode in place, then I would guess there's an error somewhere else in the circuit and that this is just the symptom.

(FWIW, I've never heard any other reports of this in the two years since the project was released in this format.)

[ElectricDruid]

The diode on the Aion schematic posted in this thread above *looks* like it's between the two op-amp inputs, but it also looks like it's inadvertently drifted over there from the gap where it should be up from ground. Which is where it is in the original Aion document, and in the Boss schematic too:

Compare:

https://www.synthxl.com/wp-content/uploads/2019/10/Boss-DS-1-Service-Manual.pdf

and

https://aionfx.com/app/files/docs/comet_documentation.pdf

[m4268588]

I know what I'm doing.

[ElectricDruid]

I know what I'm doing.

So putting the diode across the inputs like that was your mod? Sorry, I hadn't picked up on that.

I don't know much about protecting inputs on op-amps, so I defer to those who do.

[aion]

I know what I'm doing.

I'm not saying your fix won't do what you say it will. Putting a diode across the inputs of an op-amp is pretty common, especially in input stages where it has to protect the circuit against the outside world. But you're editing Boss, not me! In this case the diode to ground is not an error in need of correcting. If it does make an audible difference, then it's part of the sound of the circuit.

[m4268588]

As ElectroSmash describes, this diode is to limit the input signal (negative side).
BA728, M5223, NJM2904, 3404 are based on 358 architecture. (3404 has an improved output stage, but it is irrelevant to this topic)
NJM4580 is based on 4558 architecture. (NJM4560 and TL072 also)
These require more input limit. (You don't have to read Japanese. You should be able to understand just by looking at the images)
A diode to GND is harmless,  but if you use the 4580, you need to add a diode as I instructed.

I'm sorry, but I won't explain it in more detail if other than jp lang. I think it's good that my knowledge and experience are useful to someone, but I am not obligated to educate.

If your test with 4580 shows no problems, please continue your way.

[micmac1]

Hi all,

I have the aion comet on the breadboard since yesterday. When I first fired it up I didn't like it, with Dist low the circuit sounded farty/fuzzy (broken speaker-like). I have a real BOSS DS-1 and the breadboard version really did sound broken compared to it. I checked everything twice, but I couldn't find anything wrong on the breadboard. Then I found this thread and the diode trick seems to fix all this. For TL072 anyway, but I had the issue also with a JRC4580; will try tomorrow if it goes away for that IC, too. (If you don't hear from me it's fixed )

Thanks @m4268588 for sharing!

By the way, aion cinder also puts a diode between the TL072 buffer inputs (Link to PDF).