Noisy 1978 Memory Man (3 Knob), SAD1024 Chips, 1309B Board

pmkipp · Post by **pmkipp** » 21 Apr 2016, 17:20

I am attaching the schematic from David Morrin's excellent site for this pedal.

Found this for pretty cheap. It works, but it is very noisy - not distortion (like I think can be found with old SAD1024 chips). It is noisy with lots of hissing in the background, affected most by the Blend knob. Noise is the same whether pedal is engaged or not, though this is probably due to it not being true-bypass (which is easily remedied). Using my audio probe, the effect sound stays clean as it comes into the first opamp stage (the 4558 at pin 5) and is boosted, and the same as it goes through the second 4558 chip. The signal is reduced a bit as it heads towards pin 15 on the first 1024 chip, but is still clean there. The signal also sounds fine going into lug 1 of the Blend pot, but gets extremely noisy on lugs 2 and 3, and stays that way through all the caps coming into lug 3. Back at the 1st SAD1024, the signal picks up effect on the other side but then gets very noise going through those two 1uf caps in between the two 1024's. After that it is just noise.

Bad SAD1-24's? Bad electrolytics? Do I have a nice doorstop?

Scruffie · Post by **Scruffie** » 26 Apr 2016, 12:53

Voltages?

Do you actually get delay? These units are pretty noisy in general as they lack companding, might be normal.

pmkipp · Post by **pmkipp** » 27 Apr 2016, 00:19

It was getting delay, and I think I understand noise in old pedals, but this is too much to put in a chain.

Sad update - I was carrying it and stepped on the cord causing the pedal to drop. Now it is nothing but noise. There is faint delay on some settings, but mostly noise and distortion. Not sure why dropping it would cause this change, but it seems pretty shot now.

Sigh...

Scruffie · Post by **Scruffie** » 27 Apr 2016, 00:52

Yeah they're noisy enough not to be chain worthy as stock depending on your tolerance for noise.

Anyway, old EHX pedal solder joints can be pretty weak, you dropping it might have shocked something.

You can do the debug the proper way or by the sounds of it that you don't want to spend much time on it you can just replace all the electrolytics and tantalums (they're nearly 40 years old as it is, even if some are still good, is keeping the originals in there really going to change much other than the time before one does go bad) audio probe through the circuit, re-flow any solder joints as necessary, re-bias and call it good.

Or you can sell it on to someone who knows and loves his old EHX...

mictester · Post by **mictester** » 27 Apr 2016, 12:06

pmkipp wrote:It was getting delay, and I think I understand noise in old pedals, but this is too much to put in a chain.

Sad update - I was carrying it and stepped on the cord causing the pedal to drop. Now it is nothing but noise. There is faint delay on some settings, but mostly noise and distortion. Not sure why dropping it would cause this change, but it seems pretty shot now.

Sigh...

You will just have shaken something loose. Replace the old electrolytics and tantalums. You'll find with your re-work that you'll probably fix the dry joint that's preventing it working now! It's usually worth re-flowing all the joints on the board - the EHX staff used to think that a "good joint" was rolled, not soldered!

These things always did hiss. I used to add pre- and de-emphasis to them (wins you a useful hiss reduction), and then add a compander (an NE571 does the trick) to further improve things. You'll never get it sounding entirely pristine - the BBD chips are too noisy for that - but that's part of the "charm" of these old pedals.

pmkipp · Post by **pmkipp** » 30 Apr 2016, 13:50

Thanks. This is really helpful. Nothing to lose now, so I will go through all those joints and replace the electrolytics and tants. Can or should I use bi-polar caps in their place?

pmkipp · Post by **pmkipp** » 06 May 2016, 11:54

Replaced all the caps, with no change. Cannot find any bad solders or loose components. And when I trace the audio signal it is ok until it hits the input on that first 1024 chip - after that it is an increasingly noisy signal until it is nothing but noise after the 2nd chip. Looks like I have a doorstop...

Scruffie · Post by **Scruffie** » 06 May 2016, 12:13

Possible one of the trimmers shifted out of its position, some of those old ones are quite loose.

Posting all the voltages will help us see if there's some other problem at hand too.

Whatever happens, don't use it as a doorstop!

pmkipp · Post by **pmkipp** » 08 May 2016, 14:19

I won't make it a doorstop. I will need more knowledge about where to check the voltages, and how to read them (I have a multimeter). Aren't those 1024 chips famous for going bad?

Scruffie · Post by **Scruffie** » 16 May 2016, 21:21

No, they're famous for people assuming they've gone bad because they're rare and expensive, i've fixed a lot of vintage EHX delay based effects and clocks, opamps and caps go bad sooner than the SADs or MN chips for the most part.

Reading with a multimeter is easy, set it to the 20V setting, put the black probe to a ground point (the enclosure will do fine) and use the red probe on each pin of the chips.

Dirk_Hendrik · Post by **Dirk_Hendrik** » 17 May 2016, 07:19

Scruffie wrote:No, they're famous for people assuming they've gone bad because they're rare and expensive, i've fixed a lot of vintage EHX delay based effects and clocks, opamps and caps go bad sooner than the SADs or MN chips for the most part.

Let me second that!

pmkipp · Post by **pmkipp** » 19 May 2016, 01:43

Here are the voltages for the 1024 chips (top to bottom as seen on the schematic), the 14 pin clock chip and the two IC's

1024 #1 - pin 01=0v, 02=9.3v, 03=17.6v, 04=11.6v, 05=11.7v, 06=11.7v, 08=17.6v, 09=34.0v, 10=17.6v, 11=11.7v, 12=11.6v, 13=0.0v, 14=17.6v, 15=9.3v, 16=0.0v

1024 #2 - pin 01=0v, 02=10.6v, 03=17.6v, 04=12.4v, 05=13.4v, 06=13.4v, 08=17.5v, 09=34.0v, 10=17.6v, 11=12.3v, 12=12.3v, 13=0.0v, 14=17.7v, 15=10.4v, 16=0.0v

1024 #3 - pin 01=0v, 02=10.3v, 03=17.6v, 04=13.0v, 05=11.0v, 06=10.8v, 08=17.6v, 09=33.9v, 10=17.5v, 11=13.0v, 12=13.0v, 13=0.0v, 14=17.5v, 15=10.4v, 16=0.0v

CD4047 - pin 1=19.8v, pin 2=17.6v, 3=17.3v, 4=35.5v, 5=35.5v, 6=35.5v, 7=0v, 8=0v, 9=0v, 10=17.6v, 11=17.6v, 12=0v, 13=19.9v, 14=35.5v

IC1 - pin 1=9.7v, 2=9.7v, 3=9.6v, 4=0v, 5=12.5v, 6=12.6v, 7=12.6v, 8=35.7v

IC2 - pin 1=10.6v, 2=10.6v, 3=10.3v, 4=ov, 5=6.4v, 6=6.5v, 7=15.9v, 8=35.7v

Let me know what you think.

Scruffie · Post by **Scruffie** » 19 May 2016, 11:44

Well there's something funny going on with your readings, you have a 35V supply, where were you putting the DMM ground probe when you took the readings?

pmkipp · Post by **pmkipp** » 19 May 2016, 12:39

Ground leg of the delay pot. Would that screw up the readings? The voltage where the transformer connects to the board is 15.2 volts.

Scruffie · Post by **Scruffie** » 19 May 2016, 16:01

Well seeing as the delay pot doesn't go to ground and you're reading 35V on a pedal that's only getting 15V supply, yes it would screw up the readings.

Pin 7, 8, 9 or 12 of the 4047 go to ground, use one of them instead, or the enclosure.

pmkipp · Post by **pmkipp** » 19 May 2016, 17:55

Sorry - my bad. It was the ground leg of the Feedback control. Also tried it on the body, and the ground solder joint for the transformer, all with similar results. There are three wires coming from the transformer. Two read at 15 volts, and the third is the ground.

pmkipp · Post by **pmkipp** » 19 May 2016, 18:18

Here's something: It looks like the schematic says the voltage after the diodes at the ends of the positive outputs from the transformer should be 20v, and after the transistor should be 15volts. Right now it is 35 volts after the diodes, and 35 after the transistor. Something up with the transistor? Bad diodes?

pmkipp · Post by **pmkipp** » 20 May 2016, 18:30

I have a DOD 680, which has the same kind of center-tap rectifier setup as the Memory Man. The 680 uses an SAD4096 chip and is also supposed to get +15 volts after the rectifier transistor. Readings are very similar on the 680 as above - 16 volts going into the diodes, 39-40 volts on the other side. At the 4096 chip there are also similar readings as I found in around the chips of the Memory Man, including a pin with a 34.5 volt reading. The 680 pedal works perfectly. Still not sure what is the problem with the Memory Man, but I have learned a lot about rectifiers...

pmkipp · Post by **pmkipp** » 20 May 2016, 21:55

I am an idiot, but am willing to own it. I had been using AC volts on my meter. I will have the DC volts at all the pinouts this evening, and if anyone is still listening I will watch for any advice sheepishly...