Debugging EHX Frequency Analyzer

[Cyeos]

Hi guys,
Anyone have an idea what would cause an EHX Frequency Analyzer (in the big old-style box) to hum loudly when 'off' and do nothing when 'on'?
When it is humming, all 3 pots and filter switch cause changes in the sound, as if the oscillator is working just fine, but I can not get the guitar to come through or be modulated.
Turning it on makes all sound go away except a hiss.

I have resoldered all original joints, all original wires appear in place, etc. It doesn't appear anyone ever opened it before.

The only suspicion I have is that the two diodes on the power section get hot when powered on.

It is this board (from the EA site):

[mojokorn]

Do you have a schematic version of the ckt?  If so, it's easier for me (and others) to analyze than the board layout.

Without the schemo my best guess would be that one or both of the the 470uF caps might be leaky, or that the bypass switch has gone belly-up.  Since you seem to be handy with the soldering iron, I'd try removing the bypass switch and hardwiring it to the "on" position to see if that's the issue.  If so, then you can order a new one.  If that's not it, maybe try replacing the 470uF caps.

I can give more detailed analysis with schemo if you have it.

I hope this helps.

[Cyeos]

Thanks!

Here is one of the schems from the same site:
http://experimentalistsanonymous.com/diy/Schematics/Ring%20Modulators%20and%20Frequency%20Shifters/

[mojokorn]

Thanks, so much easier to review with schematics.

Backtracking a bit, the on/off switch is not switching to ground, so I doubt that's the problem.

You noted, that both of the diodes are running hot and since they are both connected to different caps to ground, I don't think they are the issue either.  They probably run a bit warm under normal conditions.

Looking at the schematic, I see the oscillator chip IC4 appears to be controlling the multiplier chip IC2.  If the latter half of IC1 (2nd op amp stage) or IC3 (3rd and 4th op amp stage) was not working, then I suspect you wouldn't hear anything at all.  Hence, I suspect the multiplier chip IC2 or the 100k trim control on IC2 may be bad.  I'm always on the lookout for bad caps (resistors don't tend to blow and if they do you can see it), and since no caps are used in IC2, this further drives my suspicion that IC2 or it's trim may be bad. 

You might try wiggle the trim to see if that' changes anything.  Based on the datasheet for IC2 (http://www.analog.com/media/en/technical-documentation/data-sheets/AD633.pdf), this trim is a voltage divider that drives equal levels into the inverted and non-inverted inputs of the chip.  The 1k resistor from trim to ground is there to avoid an infinite setting.  Assuming it's a linear pot, you can set it half way and it should be fine.  Otherwise probing pins 1 and 4 with a sign wave at the input and adjusting the trim such that input 1 and 4 on IC2 are equal should put it spot-on (if you change the trim setting or have to swap it out).

Of course, playing devil's advocate the oscillator could be affecting the ground plane (noise floor) and one of the op amp chips is bad.

If the chips are socketed any of this is an easy replacement.  If not, maybe installing sockets (if there's space in the z plane) would be a cool idea.  It would allow you to try different, newer, higher performing ops amps on for size.

Of course, we are on the lookout for the most likely candidates and only further hands-on troubleshooting will truly tell the tale. 

I hope this helps.

[Scruffie]

Read this http://www.diystompboxes.com/smfforum/index.php?topic=29816.0

And post voltages readings.

That schematic and layout don't match, does it take a 40V DC supply or have an internal transformer?

Leave the trim alone, it will have been factory set to the optimal position and I assume you don't have a scope.

[mojokorn]

Read this http://www.diystompboxes.com/smfforum/index.php?topic=29816.0

And post voltages readings.

That schematic and layout don't match, does it take a 40V DC supply or have an internal transformer?

Leave the trim alone, it will have been factory set to the optimal position and I assume you don't have a scope.

Good call.

I was laying bed this morning thinking more on this and I realized I should have suggested measuring the output of each op-amp while a signal was injected into the ckt.  This post seems to lay that process out well.  Basically, the stage that doesn't have output is the issue.

Also, for simple voltage dividers, I'm not too afraid to adjust trims.  For other more complex ckts, you'd want to know the desired voltage readings before messing with it.

[Scruffie]

Well the trim is a differential null, i'm sure it can be set by ear but it should be (assuming it's not drifted for whatever reason) where it should be optimally already and fiddling with trims before debugging is something I try and avoid suggesting as it can lead to a lot more issues when someone sees it being stated in a thread, applies it to their completely different pedal and issue and causes a host of problems on those more complex circuits.

The voltages should shed light on the problem, the bypass switching (assuming it's not a later model with true bypass) has the input connected directly to the output so it's possibly a bad first opamp but without voltages and if necessary following that, audio probing and further debugging, that's speculating.

[mojokorn]

Yes, best to hold off messing with the trim for now.

[Cyeos]

Alas, I cannot find a schematic that has the same power section.
Mine indeed has a 40V wall wart. The layout is spot-on, at least.

I DO have a scope - I'm just not very good with it yet.

Will post voltages as soon as I am able!
Thanks guys.

[Scruffie]

I have the schematic, it's a factory one though so I can't post it as it's intellectual property but I can see what's going on.

Ah okay, most people don't, well regardless we'll leave the trim for now until we know more.

[Kipper4]

I have an EH500 that as far as i know still works
if you need voltages for comparaison i could always crack it open.

[Cyeos]

OK! Chip voltages going clockwise on the layout.
Voltages starting at pin 1 going down.

4558D
1.2mV
15.5
15.28
1.8mV
15.32
15.44
15.47
30.79

AD633JN
15.61
15.61
15.41
15.63
2.3mV
15.63
15.63
31.04

ICL8038CCPD
25.09
23.36
23.36
27.44
27.44
31.05
27.51
27.51
15.42
23.36
15.64
21.62
22mV
23mV

4558D
15.61
15.61
15.61
1.7mV
15.61
15.62
15.62
31.02

HA17741
10.4mV
15.52
15.45
1.5mV
10.4mV
15.5
30.84
14mV

[mojokorn]

I have the schematic, it's a factory one though so I can't post it as it's intellectual property but I can see what's going on.

Ah okay, most people don't, well regardless we'll leave the trim for now until we know more.

No soup for you!

[Scruffie]

OK! Chip voltages going clockwise on the layout.
Voltages starting at pin 1 going down.

4558D
1.2mV
15.5
15.28
1.8mV
15.32
15.44
15.47
30.79

Okay this is the input opamp and Pin 1 here should be 15V, looks like it might be bad.

[Cyeos]

Excellent - I'll order some 4558Ds.

Do you think another op amp like a TL072 would work there in the mean time?

[bluebunny]

Do you think another op amp like a TL072 would work there in the mean time?

Yep.

[Cyeos]

So, I now have 15.5V on pin one of that IC, and correct-sounding effect when the pedal is 'on'!

The only remaining problem is I still have the huge oscillator hum when the pedal is off.
It is a loud squareish sounding wave, with frequency controllable by the normal pedal controls(!)
Seems very odd...like the carrier wave is being passed directly to the output when it's supposed to be set to off.

[Cyeos]

So somehow disconnecting SW2a (an original wire) seems to make the oscillator stop when the pedal is off, and still have it function normally when on - i.e. fixed.
How can this be?

Who had the correct schematic for this?