SSL 9k troubles - fake LM394?

[jeffrey_burr]

I finally finally got my two SSL 9K preamps all put together, after some years of procrastinating.  I used the switched gain described by PRR here:  http://www.groupdiy.com/index.php?topic=4731.0
I'm talking about the idea he floats of having the two stages adding gain at different rates.  It works, in one channel at least.  I thought this by itself might be of interest to people.  That's a vintage thread.

The other channel has problems, and I suspect it's a fake LM394.  It looks different from the one in the working channel, the top surface is shiny and wavy - it looks like someone put a layer of CA glue on there or something.  It sounds to me like something a counterfeiter might do to prevent a cheap marking from rubbing off.  The gain knob gives me a terrific "crack" when I'm switching it (the working channel works quietly) but not much else happens.  Signal passes but not as loud as the other channel.  I think one of the gain stages isn't doing anything but I can figure it out easily because of the staggered gain arrangement.  The first servo is doing something strange I think.  The more I think about it the more it seems like a poorly matched input stage could cause this weirdness.  I've been hanging around here for a long time but I don't really have that much build experience.

Maybe this evening I will have time to take a closer look at how this thing is behaving, which I'll post when I can.  I sure hope someone might be willing to hold my hand with the troubleshooting.  I only have my DMM and my ears, but I might be able to access a oscilloscope and signal generator next weekend.

 

[jeffrey_burr]

I'm guessing that the polarity issue has already been caught sometime in the last few years?  the pdf that I have shows pin 1 of the SSM2142 giving the + signal but that's not what the datasheet says.

 

[maxwall]

you can test the LM394 with a DMM using the hfe socket , test three legs of the LM394 at a time. Make sure orientation is correct, ex. BCE CBE. Then test the other three legs. The LM394 is just two matched transistors
with six legs, each of the three legs form the transistor on the single die. I have seen LM394  where just one side works but the other side is dead. Also have seen both sides dead , resulting in a bad device. You can also check the relative match between the two sides as well.

 

[jeffrey_burr]

Electrochronic, thank you kindly for responding.

I don't like de-soldering but, employing my time-honored procrastination technique, I did check the voltages at all the LM394 pins, comparing the working channel and the non-working channel.  The voltages matched closely on the working one, but were significantly off on the other.  I see now that the two stages are AC coupled though, so this wouldn't be the cause of funnyness with the servo.  It goes all the way negative at the highest couple of gain settings.  I'll take a closer look at what's going on and post more specific observations/questions later.

I also need to search this place thoroughly for known issues with the board and documentation, I keep finding little things but don't want to waste anyone's time with this stuff because it's probably already been found long ago.  I've been sitting on this one for a while.

 

[jeffrey_burr]

Here's the link for the documentation that I have:
http://www.studio21.ch/diy/neeno/ssl9k/DIY_9k_Preamp.pdf

Okay, I measured different voltages at the 2 inputs of "IC22" and figured, technically speaking, the input transistors are insane in the membrane.  An opamp is supposed to force the two inputs to the same voltage, I learned.  I happened to have a small quantity of MAT02 around, which work like a charm here.  It's up!  Still have grounding issues to work with.  I'll chime in again if that starts to look hopeless.