a broken Memory Toy

[paul.creedy]

A friend's EHX Memory Toy stopped working overnight (literally, he was playing through it one evening and the next morning it refused to play) and I offered to have a look at it for him, but I'm not sure there's going to be much I can do with my limited knowledge, so I thought I'd ask here for advice, if that's ok

With either battery or power supply connected it lights up, and when switched on and the blend set to no effect it allows signal through, but there's no effect.

I've dismantled it and had a quick look, but the sheer number of SMD parts has me wondering where on earth to start.

Rather than break it prodding things at random, where would be the sensible place to begin?

[thomasha]

Check the BBDs, if there is signal at the input and output go to the next one.
If one chip dies you end up without delay. The BBDs are the non-smd chips, so it is a little easier to check them first.

[paul.creedy]

I'll check them and report back - thanks

[j_flanders]

With either battery or power supply connected it lights up, and when switched on and the blend set to no effect it allows signal through, but there's no effect.

That's the most common problem reported on these.

There are versions that have a daughter board (it contains the delay section of the circuit) which is connected by pushing it onto the pins (like a big connector) to the mother board.
The most common fault on these is that the daughter board gets disconnected. Simply push it back on. That's what fixed it for most people.

Later versions have the two boards sodered together at the left side and a soldered ribbon cable on the right side.

Some versions have all the values of the caps, resistors etc printed on the pcb. If you have one of those I would be very grateful if you took some pictures as I have always wanted to trace this circuit.
I have the Toy version with the soldered boards and no values printed.

[paul.creedy]

Update: I tested the BBD chips with an audio probe (which is what I assume I was supposed to do - please correct me if I'm wrong) with the following results.



Because the two boards are soldered together on my one I had to do it from the back, but using that image I was getting an oscillated sound from the amp from all the outputs which rose in pitch from the right hand chip working left.

There was less volume from the inputs and it died away after first contact from the probe, with the exception of the input of the chip furthest left, though there was a noise as the probe touched it. I don't know if this was due to a problem or that the pitch had risen high enough for my amp to struggle to amplify it (which I doubt, but don't know for sure).

I also tested the power in and out of the four chips and all four had identical values.

If that gives enough information for someone to offer a suggestion then great - if there's a better way to test them let me know.

[j_flanders]

You don't need to test from the back side.
There are specific testpoints labeled 'BBD1', 'BBD2' etc. on the front side. In plain view when you unscrew and remove the back plate of the pedal:

Those test points are the output of each bbd chip.
If you play a note you should hear the unfiltered repeat(s) at those testpoints/outputs.
The bbd chips are connected in series, so if there's output at bbd1 but not not at bbd2, there will be no output at bbd 3 or 4 or at the final output.
You can connect your audio probe to those points and see if all of them have output.
As for the trimmers (which you shouldn't touch for now, except for trim 8 ) :
Trim 1: sets bias voltage for bbd1
Trim 2: sets bias voltage for bbd2
Trim 3: sets bias voltage for bbd3
Trim 4: sets the feedback level (too high= early oscillation; too low never oscillation)
Trim 5: sets the bias voltage for bbd4
Trim 6: sets the balance of the clock at the output of bbd4 (correct position= little clock whine, wrong positions= more clock whine, regardless of the delay time, but longer delay times make it worse)
Trim 7: sets the gain (affects the mix of delay and dry, probably not accessible on your version)
Trim 8: sets the modulation depth (turn it to the left or right until you like the modulation on the repeats)

[paul.creedy]

The side in your picture is the side I meant, which is the back of my picture.

I'll check it again tomorrow using those test points.

[paul.creedy]

Update - I finally got round to using those test points to check the BBDs and all are producing a signal.

So whatever it might be, it seems it isn't that.

With a guitar plugged in, the pedal turned on and the mix all the way off (so no effect) it works as before, but whereas I was only getting a signal like that and any attempt to blend the effect in killed it dead, now I can turn the mix control almost all the way up before it cuts out, though there is still no effect working. I wonder if this is because the boards, jacks etc. are out of the enclosure.

So, having (hopefully) eliminated the switch and the BBDs - any suggestions where to look next ?

[StephenGiles]

It also eliminates the 4047 clock circuitry!

[j_flanders]

The Memory Toy is pretty much the same schematic as the Deluxe Memory Man for which schematics can be found online. At least that way you can somewhat go through the circuit.
Because the memory Toy runs on +9V and Memory Man -24 volt they use different chips (mn32005 vs mn3005, or BL3008 vs BL3008, and also the compander chip and clock chip have different names) but the schematic is in principle the same.

After testpoint BBD4 these sections follow:

1) LP filter to further reduce the aliasing/clock whine (it's a dual 4558 or tl072 chip doing the filtering )
then to:
2) The expander part of the compander chip to make the signal half as loud again (the compressor part made it twice as loud before going into the bbd's)
then to:
3) a signal split:
   3a) one way going to the feedback pot (there's a simple RC filter just before or after that)
   3b) one way going to the blend/mix pot (a simple opamp gain section just before that, one half of another 4558 or tl072 chip)

The delay signal exits the daughter board through one of the 4 pins at the utmost left (looking at my picture posted above)
I can't remember which pin exactly.

So, there are 4 places to test and check where the delay signal is still there:

1)daughter board: one of those 4 solder connections to the pins connecting to the mother board

2)mother board: one of those 4 solder connections to the pins connection to the daughterboard
! It might be wise to (not only) probe at the solder points alone. but trace it one step back on the daugher board and one step forward on the motherboard.
On my Memory Toy, although both boards were soldered together via those pins, the connection was faulty: it arrived at the solder point 'sitting on top' of the pin' but the solder was no longer connecting to the trace below that followed. I could wiggle the boards and see the pins move up and down. The solder on the pin obviously tested OK for continuity but it wasn't going anywhere after that, just hanging in the air, so to speak

3)blend/mix pot: it's like a volume pot where the repeat enters at an outher lug and exits at the middle lug, and from there on it goes to the start of the circuit (for having multiple repeats). Test all lugs for repeat signal. (one will be ground)

4) although it's unlikely that the error is right at the end in the blend/mix pot, you could test backwards and probe both outer lugs of the blend pot.
One outer lug is the dry signal, the other outer lug is the delay signal. The middle lug is the mix.
Going even further backwards you could check if there is a signal at the output of the opamp just before the blend pot (not the input opamp but the one that does some gain for the delay signal right before the blend pot)

Here's a schematic to the Memory Man which makes it easier to follow:
http://bee.mif.pg.gda.pl/ciasteczkowypotwor/SM_scena/Electro-Harmonix/Electro-Harmonix-Deluxe-Memory-Man-Delay-Reissue-(Rev.A)-Schematic.pdf

I'll add some annotations later tonight as to where to probe.

[j_flanders]

memory man schematic: http://bee.mif.pg.gda.pl/ciasteczkowypotwor/SM_scena/Electro-Harmonix/Electro-Harmonix-Deluxe-Memory-Man-Delay-Reissue-(Rev.A)-Schematic.pdf

Trim 6 on the photo above is the 'balance trim' in the DMM schematic..
Test point BBD4 is where my mouse pointer is in the screenshot below, and where you still have signal.


See DMM schematic:
-After bbd4 there's some gain, it could that the Toy doesn't have this.
-From R46 to C25 is the post BBD LP filter to reduce aliasing distortion artifacts.
-After C25 it goes to the NE570 compander chip.
-At pin 10 of the compander chip, the delay signal exits the expander half of the compander chip (not an NE570 in the Toy) and splits up:
  -going to the feedback pot. Probe where my mouse pointer is:


  -going to the opamp for some gain, just before the blend/mx pot. Try to probe where my 3 mouse pointers are:


And don't forget to probe the four pins at the left where the daughter board is connected to the mother board.

[paul.creedy]

So sorry, I thought I'd already said thank you for all this help, but just realised I hadn't 

I'll be able to spend time with the circuit next week, so I'll report back on my findings then.

Many (belated) thanks !