Help with Compander NE571?

[Erich F]

I built the PT80 and after some intial debugging, it worked fine. Unfortunately, I screwed the board up when I cut it back a tidge to get it boxed, then it stoped working (no delay,just the dry singal).
Obviously something is broken or cracked but I need some help isolating the problem. I've replaced some caps, resistors and the NE571.

Using a probe I got audio (with delay and one repeat) to the input of the NE571 (pins 2and 3) but only a very faint signal from the output (Pin 7) . As an experiment and jumped a hot signal from before the compander to Pin 7 and got a loud distorted but repeating response. So my theory is that the output from Pin 7 is too low.
What sets the output for pin 7? FYI, the output of the other opamp in the NE571 (Pin 10) is fine.
Voltages are fine everywhere. Any help is appreciated. Thanks.

[Scott Swartz]

You are checking the right things.

The input at pins 2,3 is the compressed delay signal, and the output on 7 is expanded by the 1/2 NE571 that is made up of pins 1-8.

None of the components that connect to pins 1-8 have anything to do with the "volume" leaving pin 7, the chip automatically does a 2 to 1 expansion, assuming pin 6 is connected to pin 7 (as the PCB layout does).

It sounds like a broken or intermittent trace, or maybe a short in the socket for for the NE571.  I would carefully check the traces and replace the socket.

Since shorting audio to pin 7 gave loud audio, I don't think that pin is shorted to ground somehow, but maybe the audio that is on pins 2,3 is not actually getting into the chip.

Did you probe right at pin 7 or after the 1 uF coupling cap?  Perhaps the cap is bad.

The .47 cap sets the time constant for the expander, the 10K sets the DC level of pin 7, and the 100 pf oes to THD trim, none of which will affect volume.

[afranks]

Scott-

What's the lastest on your BBD analog delay project?

thanks-
-alan

[Scott Swartz]

On the BBD circuit, PCB artwork is complete, working on schematic, JD Sleep is working on parts layout, I would guess it will hit JD's site in 2-4 weeks.

My prototype is complete and working perfectly.

[Erich F]

Thanks for your reply, Scott. I put in a new IC socket (boy, it 's a lot tougher when the board is already populated), changed 3 caps and still the same problem. I need to go out in get a new NE571 as I accidently put it in backwards first( DOH! I assume that's not good or the chip) Maybe that'll be the solution.

When you say:None of the components that connect to pins 1-8 have anything to do with the "volume" leaving pin 7, the chip automatically does a 2 to 1 expansion, assuming pin 6 is connected to pin 7 (as the PCB layout does).
That means the volume at Pin 7 should be loud, right?

Anyway, I 'll post when I try the new IC.

[Rob Strand]

I haven't got a PT80 schematic for reference but you should measure the voltage on each an every pin of the compander, even unused ones.  These chips have well defined voltages and a wrong one can be used to track down the problem (if it's that, it could be elsewhere).

[Erich F]

I haven't exchanged the NE571 yet, here are the voltages:

Pin 1   0.46    Pin 16   1.33
Pin 2   2.09    Pin 15   2.36
Pin 3   2.09    Pin 14   2.36
Pin 4   0    Pin 13   12.1
Pin 5   2.15    Pin 12   2.38
Pin 6   8.79    Pin 11   2.33
Pin 7   8.79      Pin 10 8.26
Pin 8   2.36      Pin 9 2.37

Given that Pins 1- 8 and Pins 16 to 9 serve the same purpose for each half of the IC, Pin 1 looks suspicious. I have changed the 0.47 uF cap and still no difference. Any other ideas? Thanks in advance.

[Scott Swartz]

I know from memory that pin 6,7 should be about 6 volts, this is set by the 10k resistor.

Also, pin 10 should be about 6 volts, the two 36k resistors set this.

The NE571 must be fried.

[Rob Strand]

1) The pins which are around the 2V mark are usually 1.8V.  The spec says this can be up to 2V but I;'ve never seen it  that high.  The fact pin 8 and 9 are off is a sign something is wrong.  Also, with no signal all off these should be identical.

Pin 2 2.09 Pin 15 2.36
Pin 3 2.09 Pin 14 2.36
Pin 5 2.15 Pin 12 2.38
Pin 8 2.36 Pin 9 2.37
Pin 11 2.33

2) These two should be < 0.5V the 1.33V looks bad.
Pin 1 0.46 Pin 16 1.33

3) Like Scott mentioned these should be around 6 to 7V.  The percentage they are off is roughly the same pecentage that pin 8 and 9 deviate from the nominal 1.8V - that means the internal opamps are probably OK

Pin 6 8.79
Pin 7 8.79 Pin 10 8.26

Perhaps the most worrying sign is pins 8 and 9, it's hard for external problems to affect these two voltages.  Now it's possible that these are manufacturing tolerances.  However, pin 16 looks off, now if that was caused by external problems I'd expect the voltage on pins 14 and 15 to be different from the pin 9 voltage - but they aren't different!

Quite a few things don't add-up, so it's possible the chip is dead.  You could prove it once and for all by lifting the 1uF caps which connect the chip to the surrounding circuit,  after doing that if the measured voltages are still the same (off) then it's highly likely to be the chip.  If pin 4 wasn't gounded you would expect some weird behaviour but your pin looks well grounded.

[Erich F]

Thank you Rob and Scott. I really appreciate your help on this.
I have spent A LOT of hours on debugging this great effect and I am determined to get it to work again.
I'm embarrassed to admit that since all the voltages read high, I thought I better change the battery on my digital voltage meter...that helped. It started to make more sense. Also, I got a new NE 571 and Pin 16 is still above 0.5 and the effect still doesn't work:

Pin 1 - 0.11 Pin  9 - 1.82
Pin 2 - 1.82 Pin 10 - 6.33
Pin 3 - 1.82 Pin  11 - 1.79
Pin 4 - 0.00 Pin 12 - 1.84
Pin 5 - 1.84  Pin 13 - 9.48
Pin 6 -  6.73 Pin 14 - 1.82
Pin 7 - 6.73 Pin 15 - 1.82
Pin 8 - 1.82 Pin 16 - 0.91

When I disconnect the four 1 uF caps, Pins 1 and 16 read zero. Any further ideas?

[Rob Strand]

OK those measurements look much better.

When I disconnect the four 1 uF caps, Pins 1 and 16 read zero. Any further ideas?

OK the 0.1V on pin 1 looks very reasonable, it never goes entirely to zero.  The 0.91V on pin 16 looks a little if'y, since it's < 1V then it could be due to noise.  If you have some noise in the ckt, perhaps just from wiring or lack of shielding because the box is open then perhaps the small voltage is due to noise - the 1st stage opamp boosts the high frequencies which could increase noise at the compander input.  Now you could try shorting the hot line of the input socket to gnd and see if the 0.91V drops a bit.  If it's any comfort, I just whacked that ckt together on a breadboard without shielding and got 0.1V across the caps.

One thing which I'm not sure about is the type of 1uF caps you used.  The original ckt uses non-polars so if you have sub'ed  these for polars you have to get the around the right way.

From what I understand from you original post you are getting proper level audio signals at the 1st opamp, the compander pin 10, the compander pins 2&3, but not at pins 6 & 7.  Now by proper they should be at least as strong as the input signal, in fact you should percieve these as being stronger and more trebly.   It's important you confirm the signal at pint 2 & 3 is a strong signal and not a weak one.  Also when you check pins 6 and 7  check the level before and after the 1uF cap between the compander and the pots.

Now what to make of what we have so far.  Well it's hard to tell at this point a few more tests are in order.  What I would be doing is:
- check your pot wiring.
- turn the repeats off.
- desolder the wire which goes to the wiper of the delay pot and use this as an audio probe.
- connect the "probe" to the output of the 1st opamp.  If the signal is clean an unmodified then the input and output stage opamps are probable OK.
- next connect the problem to the output of the 1st compander.  In this case the unit should sound like a lite compressor, the audio level should be  about the same as the bypassed stage, perhaps a little louder.
- next put the wire back on the delay pot.
- disconnect the transistor end of the cap between the transistor and pins 2 & 3, connect this back to pin 10 with a wire.  In this case the unit should should pretty clean with an unmodified signal.
- put the cap back in
- lift the delay chip side of the 1uF cap between the delay chip and the transistor filter.  Connect this back to pin 10 with a wire.  In this case the unit should sound clean but the highs will be gone, the level should be reasonable.

OK if one of these test fails is should be possible to isolate the fault further.  If all pass but the unit doesn't work then there's something wrong in the region of the delay chip..

[Mike I.]

Haven't worked on a PT-80 but...in normal operation the voltage at NE570 pin 1 should be more than two diode drops (above say 1.1 to 1.2 volts, from Signetics 570 app notes) above ground.... voltages of 0.1 to 0.5 volts (with input signal going into compressor) are suspiciously low. Perhaps input signal isn't making it into chip or perhaps a capacitor is not working properly and allowing a DC offset to be fed into 570. Note that a digital multimeter should be used so the capacitor on pin 1 isn't loaded down. An old fashioned 1000 ohm/volt analog meter not a good idea. (Have investigated this recently as an NE570 makes a nice envelope follower when the voltage at the rectifier cap pin is buffered with an op amp voltage follower. The 1.1 to 1.2 volt offset present in normal operation needs to be removed for best results)
Regards, Mike

[Scott Swartz]

In addition to the good tips in the previous 2 posts, here is something simple to check.

Pin 13 is shown as 9.48 volts.  This should be the regulated 12 VDC from the 78L12, its a direct connection.

Is there a short to ground somewhere?

[Erich F]

Thanks, Rob. I ran all the tests you suggested  and it passed each one (also thanks , Mike - I am using a DMM to test the voltages). I am using non-polar 1 uf caps.
I assumed you meant the wiper of the "delay level" pot when you said "delay pot". I also assumed when you stated "clean" you meant a signal without delay rather than without distortion ('cause I got some slight distortion on Pin 10 and it lost highs when it was supposed to).
I am getting proper clean sound at Pins 1 and 7 of the TL072.
The audio signal at Pins 6 and 7 sounds much, much weaker than at 2 and 3. I really have to crank my amp to hear a faint sound. There is no difference to the sound before or after the 1 uf cap located between the Pins 2 and 3 and the pots.
The audio signal at Pins 2 and 3 sound somewhat weaker than at Pin 10 but much more so than, say, at Pin 14 of the delay chip (not surprisingly, I guess, given the signal runs thru a couple of 10K resisitors after the delay chip, ).
Scott, the wallwart I'm using feeds 11.04 V into the 12 V regulator and reads 9.48 afterwards. Is that a normal drop in voltage when using a regulator?

[Rob Strand]

I assumed you meant the

Yes on both accounts.  The small amount of distortion you get will probably be on the lower strings.

The audio signal at Pins 6 and 7 sounds much, much weaker than at 2 and 3.

Depending on signal levels this can be normal.  That stage is an expander so when you are below the unity gain threshold the input will be lower than the output.  The first NE571 does the opposite the output of that stage is highe than the input (when below the unit gain threshold).  The two effects will roughly cancel out resulting in an overall unity gain.  If you perceived no volume loss during the test where the 1uF cap on pins 2 and 3 was connected to pin then this mechanism is working and the NE571 is working- that was actually the reason for that test.

> The audio signal at Pins 2 and 3 sound somewhat weaker than at Pin 10 but much more so than, say, at Pin 14

This sounds like a possible problem which needs exploring.   If you are losing signal through that transistor stage then perhaps there's something wrong with the transistor or the biasing.  You could check the base and emitter is at about half the supply rail.  It probably worthwhile just replacing that transistor.  The 10k's should  not attenuate the signal they are part of a filter wich removes the high frequencies, which appears to be doing something form your tests.  Measuring the voltage on pin 14 might not be a good thing, the next best thing is to measure the voltage at the 2nF cap to ground, preferably with transistor out of circuit - if the level here is good and level is bad after the transistor then it points to the transistor.

[/quote]

[Rob Strand]

pin 1 should be more than two diode drops (above say 1.1 to 1.2 volts, from Signetics 570 app notes) above ground.... voltages of 0.1 to 0.5 volts (with input signal going into compressor) are suspiciously low.

That's true with a signal but with no signal there is no rectifier current and hence no current to bias the two transistors which appears after the rectifier cap - so you end-up with voltages < 1V with no signal.

[Scott Swartz]

Rob is on the right track with the signal tracing.

On the power, you need a wall wart that is at least 15 VDC output to get a regulated 12 VDC.

If you don't get it figured out, I would take a look at it for you, email me if interested.

[Erich F]

I replaced the transistor (..again). The new 2n5088 transistor reads the same as the old one: E- 0.34,B-0.34,C-9.54. From what I understood, Scott's build notes stated that the transistor limits the highs (rather than amplifies the signal) so I was never sure what the voltages should read.

Scott: here's my e-mail: efreiler@msn.com. I'd appreciate it if you had a look at the unit. BTW: shouldn't this should work at 9V? It's the same wallwart I used when the project worked before.

Rob - If you want, maybe we should take this off line (I'm enjoying working this thru but it may be too boring for people in general)

Thanks. Erich

[Erich F]

I just want to post this to remind anyone who has tried to make a pedal work and couldn't, that with some patience and the help that is available here at the forum, you can make it work.

I spent many, many hours trying to debug this effect before I asked for help. Well, a little perseverence and A LOT of help from Rob Strand did the job..it was a bad capacitor!

A very grateful thank you to Rob for his time and help. I learned a lot.

Also, thanks to Scott Schwarz, too (I didn't need to send the board to you after all, but thank you for the the offer..and the design for a great sounding pedal.)

This place is great.

[Ed Rembold]

Not part of the problem-  but-  You need to feed a regulator with 2-3 volt higher than the reg output.
so in this case , you should have at least 14-15 volts feeding a 12 volt regulator.  the exception would be if you are using an LDO regulator (Low Drop Out).
(I learned this way back from Gus Smalley)
Ed R.