Vibrobot distortion

[Diynot]

Calling @DGWVI and @Mentaltossflycoon and anyone else who might have thoughts. Mkay, not the reason I thought I was going to open this thread, but while my vibrobot is trem-ing, it has some distortion/fuzziness on the signal. After chasing rabbits before lifting pin 6 I am 99% sure of parts values. During the no trem witch hunt I also have reflowed joints, cleaned the board, the voltages look good, the ICs aside from the 8038, that I got from Robert, came from Mouser, I have messed with the trimpot, but can’t get the signal clean. Obligatory board pics in case I missed something and also a vid for listening purposes. My initial suspicion would have been the TLC2262 except with the depth all the way down the clean is clean. I guess that leaves the 741, 8038, or the 2N7000, but are those actually seeing/producing an audio signal? The vid was made using my p-90 loaded Reverend in the bridge position, but I also tried with my tele and although there is less, the distortion is still present.

 

[Diynot]

Aha, now I know - sincerely oops...

All good. I appreciate you putting eyes on it

 

[Jachiro]

Hello. I don’t know if it’s too late to post on this thread but I just went through the exact same thing on my vibrobot build I just completed. Perhaps I can provide some insight. I had a nasty distortion that I could not dial out at any position of the trimpot.

I traced my signal and found really good clean gain at both outputs of the TLC2262. One output is the clean signal— call it output 1 and the other gets modulated by the ICL8038 circuit— call it output 2.

I found output 2 was getting the negative part of the wave clipped right at the drain of the 2N7000 mosfet. The LM741 chip had no issue and the bias point was right at about half of the supply making it pretty much optimal. I was pretty sure it had to do with the signal generator chip itself.

I found a video on YouTube where Fran herself troubleshoots an original vibutron and talks about its design. She mentions that the ICL8038 was not designed to operate at the voltage that it’s at in the vibutron circuit and that she rejects over half of the ICL8038 chips due to not working well enough in the overall circuit.

When I bought my vibrobot pcb, i purchased 3 ICL8038 chips just in case something went wrong during the build. I swapped out the chip twice during my troubleshooting of this build and when I tried the last 8038 chip I had, that’s when I had my circuit work. It still had a slight hint of distortion, but I was able to almost completely dial it out. Mine is workable now.

I would suggest trying another ICL8038 chip. I’ve built a lot of pedals in the past, so this isn’t my first rodeo. I have to say that this build sucked - almost as much of a troubleshooting nightmare as my low tide build, but at least I learned something.

 

[Diynot]

I would suggest trying another ICL8038 chip. I’ve built a lot of pedals in the past, so this isn’t my first rodeo. I have to say that this build sucked - almost as much of a troubleshooting nightmare as my low tide build, but at least I learned something.

Thanks for the heads up. The build was banished to the pedal oubliette and to be honest will probably stay there. Not going to spend $$$ (relatively speaking) chasing obsolete chips to find one that gains me no greater functionality than I can get from a basic TR-2.

 

[swelchy]

I'm pretty sure I remember Fran talking about the main chip for that pedal was absolute crap and remember He said he had a long vetting process to go through chips to see if they distorted or not...

 

[Robert]

I feel like the circuit design itself is what needs attention... not so much the ICL8038.

The ICL8038 is being used as an LFO generator. If the circuit were designed properly there is no reason the ICL8038 should cause distortion, it's not in the signal path!

The issue, I assume, is that the 2N7000 is clipping to ground (possibly because of how it's biased due to some DC offset from the ICL8038 which may be variable between ICs)...

Maybe the ICL8038 needs to be decoupled from the 2N7000 so the trim pot is the only factor in setting the DC bias? Or maybe there's a better suited component for that particular task.

 

[Jachiro]

Actually, after revisiting, and after playing through my vibrobot for a while I saw that it was still distorted in sine wave mode. I looked more into it and I saw that R21 (91K) was enough out of tolerance to cause distortion. I replaced it with a 100K trimpot with one leg removed, making it a variable resistor. R21 is tied to pin 12 on the 8038 chip and is labeled distortion on the datasheet. It can be as high as 100k. Through a combination of tweaking the value of R21 between 89K and 100K, and tweaking the value of the 100K trimpot tied to the mosfet, I was able to tame the distortion. If you want to use a fixed resistor, 93K, if available, might be better. Or you can make sure to use a 91K resistor with an extremely tight tolerance. They exist, they just don’t come with your standard set-of-multiple-values resistor kit. You may have to order from mouser or digikey.