Problems with Heathkit IG-18 (Modified) rebuild.

[SoulFetish]

Where to begin...
So, I got a non-working IG-18 pretty much stock which I was able to get up and running for a little while.f But, after a while, signal output would drop out in different switch settings in the Capacitor Multiplier and 10x frequency switches. Eventually, there was no output.
So, since it would be out of commission for a while, I took the opportunity to do some mods to the power supply and circuit to decrease distortion in the sine wave generator.
Since then, I've been chasing demons throughout this unit. Every time I track down some fault, I think I've found the issue only to have problems persist. For instance–at one point, when powering up, I measured a low + voltage (≈+19V) on the power supply rail indicating a heavy load on the power supply. I measured a 39Ω short across the rails, and found a shorted Q3 transistor. Replaced it, and still measured the same low voltage on the power supply. I was like "what the hell?".
I wasn't measuring 39Ω anymore, but I was getting weird resistances at different nodes. There are areas on the board where there seems to be significant heat damage and i began to wonder if the board had become conductive (It is phenolic). So, long story short, here is where I am. I rebuilt and am using different capacitor multiplier switch, and have removed the wave generator board and rebuilt the circuit on a breadboard with just the sinewave portion of the oscillator. Here are some documents:
The original schematic:

Here is the complete assembly manual (I scanned and inserted pages missing from every other available manuscript I found):
Heathkit IG-18 Sine-Square Audio Generator_Assembly-Manual.pdf

and here is a schematic of the circuit as it is built now:

[SoulFetish]

Here are some pics of the construction. The breadboard was a little incomplete at this point, but I can take some shots of what it looks like now if that helps.



Here is the old board, with the heat damage:


In any case, the power supply rail are where they should be. But the voltages at the semiconductor nodes are not at all what I should be getting.
I am using 2N3417s for Q1, Q2, & Q4.
2N3906s for Q3 & Q5

[nosaj]

Here are some pics of the construction. The breadboard was a little incomplete at this point, but I can take some shots of what it looks like now if that helps.[ATTACH=CONFIG]56127[/ATTACH]

[ATTACH=CONFIG]56128[/ATTACH]

Here is the old board, with the heat damage:
[ATTACH=CONFIG]56129[/ATTACH]

In any case, the power supply rail are where they should be. But the voltages at the semiconductor nodes are not at all what I should be getting.
I am using 2N3417s for Q1, Q2, & Q4.
2N3906s for Q3 & Q5

How about some TP data? Specifically at TP1 and from there we can work down. Are you going to put the square wave back in the circuit?
What were original transistors?
nosaj

[SoulFetish]

How about some TP data? Specifically at TP1 and from there we can work down. Are you going to put the square wave back in the circuit?
What were original transistors?
nosaj

Cool, sounds good. Ill try and take some measurements and list them accordingly.
I don’t plan on reinstalling the Schmitt trigger the way it’s drawn. If I reinstall one, Ill probably use op amps and shield it from the rest of the circuit
Also, Ill try and get the original transistor#s. But im pretty sure Q1,2, & 3 were 2N3416s to start with.

[dmeek]

Take a close look at the rotary switch wafers. Where there are terminal lugs and switch contacts connected by rivets, solder the rivets on both sides to make sure the connection is solid.
Be very careful so that the solder does not flow onto the the back of the contacts.
I always found that to be a problem with Heathkit switches.

[SoulFetish]

Take a close look at the rotary switch wafers. Where there are terminal lugs and switch contacts connected by rivets, solder the rivets on both sides to make sure the connection is solid.
Be very careful so that the solder does not flow onto the the back of the contacts.
I always found that to be a problem with Heathkit switches.

Good tip. Ive been really suspicious of the switches. I don’t have any exact replacement and even went so far as to disassemble a sealed mil style multi-deck switch to reprogram it from a 2 position to a 4 position switch.
I thought I nailed it, too. It went back together beautifully. But I don’t think I could have ended up any worse. It was a clinic in precision failure. After reassembly, I couldn’t get continuity on any terminal, in any position.

[g1]

Doesn't seem to be an issue for you, but my unit had problems with the trim pots. There was LF modulation of the output which I was pretty sure must be a bad cap issue, but turned out to be that the trimmers needed a shot of cleaner.

I can verify any voltages or transistor types if you need. (assuming version close enough)

[SoulFetish]

Doesn't seem to be an issue for you, but my unit had problems with the trim pots. There was LF modulation of the output which I was pretty sure must be a bad cap issue, but turned out to be that the trimmers needed a shot of cleaner.

I can verify any voltages or transistor types if you need. (assuming version close enough)

It gives us expected voltages on all the relevant nodes on the schematic. But if you can confirm that yours match up with the schematic we're working with, G, that'd be cool. Don't sweat it though. I just pulled my replacement multiplier switch to confirm that capacitance readings for each "side" and make sure I wired everything correct. The test out fine.
Here is the 'eff'd up thing; So, the voltage reading on the + rail (TP1), is 41.7V. - fine.. But I have +39V on the base of Q5?? at the junction of the diode, base, and 10k resistor.
I'm sure I screwed something up here, but you know when you've been looking at a project for so long, through several edits, and you can't see the glaring error right in front of your face? This is where I think I am.

What are you using for the NPN and PNP transistors, G1?

dmeek, I checked out the switches and couldn't really see what I thought you meant. But I just re-read your post and I understand what part of the terminals you were referring to now. I'll check those out tomorrow.
Some might think that to be too small a target to solder to... But I used to bullseye womp rats in my T-16 back home. They're not much bigger than two meters.

[Helmholtz]

I just quickly checked but I think you need to load the sine output with a 15k resistor replacing the square generator input resistance to allow for ample DC current.

[SoulFetish]

I just quickly checked but I think you need to load the sine output with a 15k resistor replacing the square generator input resistance to allow for ample DC current.

Oh, dude, that makes sense. When I get a chance in a an hour or two, see if that gets the DC levels to settle in.

[dmeek]

It looks like from your photos this is the type of rivet you have. The rivet at arrow A has been soldered already. The important thing is that solder must not flow to point B.
It's a delicate and risky operation but only necessary on rivets that hold separate pieces together, curiously not the case in my photo.

https://music-electronics-forum.com/...1&d=1575233689

[g1]

What are you using for the NPN and PNP transistors, G1?

My transistors are:
Q1,2,4, 2N3416
All others stamped with Heathkit numbers from parts list.
Q3 & 5 plastic TO-99, Q6,7,8, small cans as shown in wave gen. pictorial (#10)
Q10 as shown in pwr. supply pictorial (#5)

[SoulFetish]

okay, so I worked on those rivets; installed a 15k load to ground at the junction of the 15Ω resistors, and there is no change. voltage readings on most all relevant nodes right around 38-39V, except for the bias/feedbback pot terminals. totally weird.
Here are some higher resolution photos of the current build:








ps - oh, if you see random leads sticking out of the board that are cut and go nowhere, those are just reinforcements for some of the leads which are very thin.

[g1]

One thing I'll mention is I was surprised to see so many ECB pinouts in this unit when I'm used to seeing EBC on 2N types.

[Enzo]

I only see the pics in post #2, and the empty small xstr spots look to me to have the B in the center. Are they TO92 or are they round ones?