Aion L5 Preamp troubleshooting - compression circuit

[ElCapitan]

Howdy folks, first time poster here.   I've built quite a few pedals, particularly from Aion, and this one is the most ambitious, but I'm having a bit of trouble.   Good news is, all of the knobs and switches do actually perform their functions. However, I'm having some issues with the unit sounding horribly splatty/spotty with the limiter engaged. Any amount of gain reduction results in a nasty electric buzz that rapidly pulsates as the gain reduction goes away.

I have some metal can CA3080s, and I initially put them both in backwards. Oops. The pedal didn't work, and the limit light stayed on all the time regardless of where the limit knob was set. Both got extremely hot to the touch. I'm pretty sure I killed them. I put a new one in the IC11 spot, this time oriented correctly, and the limit light issue resolved. I am still having an issue in that IC10 gets VERY hot. I had an NTE902, which is a rebranded CA3080, now I'm sure it's roached as it got really hot and darkened.  So before I sacrifice yet another CA3080 I want to see what I can do.

The voltages on QX1 & QX2 are way off. I've gone through and looked at the resistor color codes around those, just hoping I'd find something. All are correct. Solder joints look good, although someone on facebook pointed out some potential flaws so I reflowed the solder joints. I've double checked all the electrolytic cap, transistor, and diode orientations and they're correct.

What's driving me nuts is I checked every single resistor and capacitor with a multimeter before install, save for the MLCCs, since my meter has trouble reading them (those are all sourced from Mouser). It is still possible something was missed- and I did discover one issue in the comp schematic (some 470Ks got mixed up with my 47K and I misread my meter). Hoping someone can point me in the right direction to look. All of the knobs seem to work, but the sound is problematic. I've used NE5532 in place of the 4558s in all positions. Everything else is as spec'd.


Here's what I've done:

Tested to make sure vim getting 15V from the regulators. Yep.

Isolated the dual gang pot from the PCB using shrink on the tabs. All others have pot covers.

Checked the orientation of all diodes, transistors, and polarized caps. All are correct.

Replaced the 2n5088s. No luck (these are from Tayda).

Double checked every resistor value in the comp/master section: I found that R93 &94 were 470k instead of 47k. Swapped them out, still have the issue.

Tried switching ICs: just cooks whatever is in IC10

Checked for shorts with a multimeter. None detected.

Tried re-flowing all solder joints (no change)

Checked the values of all film caps in that section (no issues)

Used a jeweler's visor to check for shorts. None seen.

So at this point I'm just stumped. I am hoping that one of you may be able to point me toward a possible solution.


Correct voltagesfor IC10 & 11 according to the test chart:

IC10- (CA3080E)
1- 0
2- 0
3- 0
4- -15.2
5- -14.5
6- 0.9
7- 14.73
8- 0

IC11- (CA3080E)
1- 0
2- 0
3- 0
4- -15.2
5- -14.62
6- 0.23
7- 14.73
8- 0

QX1- (2N5088)
1- 0.4
2- 0.9
3- 14.66

QX2- (2N5088)
1- -0.22
2- 0.4
3- 13.21

Actual measurements:

IC10 CA3080

1 -50mv (goes down)
2 -14mv
3 0
4 -15V
5: -14.5v
6: 2.07V
7: 15.26V
8: cycles down to -300mv

IC11:

1 23mv
2 0
3 0
4 -15V
5 -14.40
6 fluctuates but is in the low mv area
7 15.26
8 0

Qx1:
1 12V
2 12.5V
3 15.19V

QX2:
1 10.9V
2 11.4V
3 28mv

[Kevin Mitchell]

Try to test your OTAs in a 9v circuit - rather than fry them out in a 30v circuit before knowing if they are even good!

They fake everything nowadays.

-KM

[ElCapitan]

Good idea!  I have a Ross compressor I can try them out in.   That said, These ones are for sure the real deal.   Got 'em from a reputable source and they are NOS. 

[ElCapitan]

Well- tested the CA3080 that I thought was bad (the one that gets super hot) and it's fine in my Ross style compressor (also an Aion build).   

[aion]

How are the trimmers set? One of them impacts IC10 and one impacts IC11, so those would be crucial, although I don't know what impact they could have if they are way off bias since they are really just dumping excess signal to ground and not biasing anything about the chips themselves.

I'd pull IC10 and IC11 and then use an audio probe on pin 6 of IC9 to see if everything is normal up to that point (volume, bass, mid, treble on both channels). If something sounds off with those (note that there won't be much clipping at this point) then the problem is more than just IC10/11. If they sound OK then definitely keep the focus where you have been.

But yes, the QX1/2 stuff does look very weird. There's not much that can go wrong there since they are just biased directly to +V with standard resistors, but the problem is probably localized around the parts connected directly to them.

[ElCapitan]

OK!   So, per Kevin's suggestion I tested what I thought were roasted OTAs in my Ross style compressor.   They were all functional, despite how hot they got in the preamp.   So I decided to don my jeweler's visor and go 100% over each resistor line by line.   Turns out that I had incorrect resistor values in R6 and R90 (3.3k instead of 33K).   For that to be the case, not only did I have 3.3 in my 33k bin, I also misread my multimeter, AND missed the incorrect value in R6 when I went over the comp section the first time (I didn't see the orange vs red clearly until I used magnification).   Just goes to show you that it's still possible to miss things even when being careful (which I assure you, I was).   

I did cook one of the NE5532s, and it still got hot even after this was fixed, however the voltages on QX1 & 2 looked right, as did the voltages to IC9, so I replaced it and voila!   it works perfectly, and nothing gets hot anymore.   Now to get a better power supply than the Line6, cuz man that thing's a piece of crap.

Anyhoo, I appreciate the help from everyone.   Cheers!

[PRR]

> I didn't see the orange vs red clearly until I used magnification

Lens may help; I've found LIGHT to be critical. Old fluorescent, red/brown/orange all look a lot alike; under incandescent their difference is clearer but blue/green may be tough. CFLs and LEDs come in various color-spreads. MY gold-standard was frosted sky-lights and doing all resistor selection in day-time.
http://www.lightbulbmarket.com/files/1974975/uploaded/Light%20Spectrum%20by%20light%20type.png

And yeah, the decimal point on many meter and calculator displays is totally lame. Every few years TI changes vendors for their TI-30 calc and for about 4 years I needed a lens to find the dot. Recent production is tolerable but not eye-popping.

[tubegeek]

Every few years TI changes vendors for their TI-30 calc

I don't mean to brag, but the TI-30 I keep on my workbench is Pepto-Bismol pink. SO easy to find it in the mess!

I got a few of them on deep back-to-school sale - did you get yours at Ocean State Job Lot, Paul? I think that's where the pink one came from.

[ElCapitan]

Yeah I've been using my incandescent gooseneck lamp for illumination.  I have a halogen one that's insanely bright, but it gets really hot and draws like 300W.