EH Memory Man internal trimpots

[Matt T.]

First post in the shiny new forum.

Anybody know what the 5 each internal trimpots in the EH Memory Man Deluxe do? I have a schematic and can identify the trimpots but I don't know enough to tell what their function is. I can determine something like "it's adjusting the voltage to the 4558" or "it's adjusting the feedback from the output to the negative input of the 4558" but I don't know what that means regarding the sound or operation of the unit. Any help appreciated.

[LooseChange]

You might have better luck here....
http://electroharmonix.ronsound.com/

[Rob S]

From the 3005 datasheet, it looks like a couple are for bias & are adjusted for minimum distortion in conjunction with a gain/drive control in that -ve feedback loop feeding B13.

Another -ve feedback loop trimpot is used to control the level into the filters feeding the compander, again adjusted for max signal without distortion.

That leaves the 5k between the two outputs (6 & 7) on MN3005B on my (awful) diagram. These outputs look to come from consecutive stages of the delay line & probably vary the phase of the signal - wouldn't mind a twiddle of that one because I'm guessing here .

I expect this is of no use at all but never mind.

Rob.

[Matt T.]

Thanks for the link LooseChange. That's a pretty cool site but he was asking $1.50 plus $1.00 shipping just for the schematic (which I already got for free elsewhere). He does mods to these units too. He also has a bio and an interview with Howard Davis who still does work on these things so I assume he either would want to do the work himself or would recommend me sending it to Howard.

Rob S.,
Thanks, that's exactly what I was looking for, or at least a start. So I guess the bias trimmers for the MN3005 are the 100K's connected to pin 14? That's 2 of the 5. Then 2 of the remaining 3 are in the feedback loop of 2 different 4558's. That just leaves the 5K. Regarding that 5K between the 6 and 7 pins of MN3005-B and the phase thing, there is a Chorus/Vibrato in these things, maybe this is affecting that?

My 'problem' is that these things distort/clip the 'echo' signal a bit too much. I actually think that this clipping/distortion is part of the allure and contributes to the slight 'ducking' effect that MM have (IMO) but I wanted to try and reduce it a bit if possible. I've already got TL072's throughout instead of 4558's.

[Matt T.]

For anyone that might find it interesting, I did find this helpful page. That plus the datasheets on the chips and this article and it's getting a bit more clear.

[Rob S]

I did a bit of searching and found a related cached thread (google puretube mn3005) - that odd trimpot is to null the clock bleed from the delay line.

WRT your distortion problem, the THD on that chip is quoted as being 1% typ, 2.5% max at 0.78Vrms - increasing each time that signal is fed back through.

I have no idea how likely it is that yours could be at the high end of that figure though I know they're quite expensive to swap.

Best of luck with it.

Rob.

[Matt T.]

Last night I pulled the circuit card up and traced the circuit enough to identify the trimmers. Then I labeled the 'bottom' side of the card (the side you can see with the bottom cover off) so that I know which trimmer is which. There are little access holes in the circuit card so you can adjust the trimmers from the underside. I then reassembled and measured some voltages.

The supply is 14.2vdc (actually I think it's negative but I forgot to 'notice' that ).

The 5K pot has an unequal voltage from the wiper to the two outer lugs, but it's close.

The 2 each trimmers for the 4558 feedback loops have the same voltage on all 3 lugs; 7.1v on one trimmer and 7.6v on the other.

The trimmer for the voltage to the + input of one of the 4558 (the 7.1v one) has 14.2v/7.1v/0v on the three lugs.

The other trimmer (looks like a feedback for the MN3005) has 14.2v/7.9v/0v.

At that point I stopped and hope to do some listening tests this weekend.

[Rob S]

I used to have a boss (a very good one) with a habit of twiddling every adjustable thing in sight. The lesson learned was to put a stripe of marker pen down through the wiper of the pot before he got his hairy arm in there so we could always take a step back.

What happy days.

Rob.

[Matt T.]

I put a looped signal from my Boomerang through the Memory Man as I tweaked it. Two of the trimmers are voltage adjustments and the pedal seems to work best when I adjust the voltage to 1/2 V-. The 5K pot didn't seem to do much within a range so I just set it to an equal voltage between the wiper and the two outer lugs.

The other two trimmers adjust the gain of two of the 4558 op-amps in the delay circuit. One trimmer seemed to affect the overall gain of the delay circuit while the other seems to affect the feedback more. Between the two you can adjust how the echoes fade out. You can make the first echo really loud and then the rest fade out really quickly or you can make the echoes fade out very gradually. This is different than using the Blend and Feedback controls on the top of the unit but it's hard to describe. It affects the echoes in a slightly different way...hard to describe.

These two trimmers also affect the settings of the Blend and Feedback knobs greatly. You can make it so the thing won't runaway...or you can make it where it will runaway at the drop of a hat. But here's the rub: setting the Blend control very high at all attenuates the original signal. Just a quirk of the way these things are. So it seems to sound best when I set the first trimmer such that I don't have to turn the Blend control up that high, then I set the second one such that the Feedback control gives a good range and the echoes sound about right (regarding the relative volume of the first echo to the second to the third, etc.). This incidentally makes it easy to get it to runaway but I never really do that.

None of the trimmers seemed to affect the distortion/clipping that much, once it's at the 'minimum' (which mine already was) it doesn't seem to get any better from tweeking the trimmers. However, 1/2 V- in my Memory Man is -7.1vdc and at least one of those 4558 is at -7.6vdc (one of the 4558s that I can adjust the feedback/gain of). Don't know how close to 1/2 V- these have to be or how much distortion that little difference makes.

[StephenGiles]

Do you want set up notes?

[StephenGiles]

See if this is any help. I scanned the original and applied OCR, sent scanned immage to Word and I hope that I have corrected all the errors.

DELUXE MEMORY MAN WITH CHORUS AND VIBRATO ALIGNMENT PROCEDURE (8/1/78)
Plug unit into 115V 60Hz. and set feedback and chorus/vibrato fully CCW, blend fully CW, delay fully CW and power on. The pilot light should come on but not the overload light. Connect the scope to pin 2,4,6,10,12, or 15 of the CD4049 and observe the square wave period. It should be between 120 and 140 usec. (add c across 240pf if necessary). Set the delay fully CCW; the period should now be between 8 and 10 usec. Set the delay to the middle of its range. With the switch in the chorus position slowly bring up the chorus/vibrato control, watching the square wave. The period should decrease slightly and become modulated at a rate of slightly less than 1Hz. At max. chorus setting the period should swing approx. 10% of its average value. Set the switch to the vibrato position. The modulation rate should go up to approx. 4Hz. Set the chorus/vibrato control fully CCW. Connect a 250Hz signal of 500mv p-p to the input and adjust the level control for the same level at pin 6 of the NE570. The level at pin 7 should be between .95 and 1.4 V p-p. Watching pin 7, increase the level. The overload light should just begin to glow at about 1.4V p-p and should be almost as bright as the pilot light at about 2.3V. No distortion should be visible in the signal - at this point. Look at pin 7 ot the first MN3005 and turn, down the level for a signal of 1.5V p-p. Slowly increase the input frequency. The response at pin 7 of the MN3005 should be flat up to about 9OOHz, rise to a max. of about 2V p-p at around 2.5KHz at which point the overload light should glow dimly, drop back to 1.5V p-p at_ about 3.8KHz and roll off sharply above this. Set the frequency back to 250Hz, the 2nd trimpot at about center, and connect the scope to pin 7 of the second 4558, which is also the wiper of the first gain trimpot. Increasing the level setting as necessary, set the first bias trim for maximum p-p undipped signal, then at a lower signal level set the gain to unity (output level = input level). The output level before clipping should be 3.8V p-p or greater. Vary the delay over its full range and if clipping becomes very unsymnetrical at either end, trim the bias a bit for a good compromise over the range. It is normal to see alot of clock noise at the highest frequencies. The second MN3005 is aligned in the saw way as the first , except that the clock balance trirs is carefully set for aiin. clock frequency at its wiper at maximum delay (nin clock frequency) setting. This should be done before fine trimming the bias for symmetrical clipping and again after the bias is finally set, with the input signal to the unit disconnected and the scope gain high. Set the 2nd gain trim last looking at pins 14 and 15 of the NE570 and setting this trim so that the signal is equal in level to that at pin 7 of the 570. If this can't be done, go as far as the trimmer permits and re-adjust the first gain trim to achieve it. Set the level so that the overload light is on dimly, looking at the signal at pins 14 and 15 of the NE570. Set delay to max.(CW) and vary the input frequency from 40Hz upward. The signal should be unclipped and clean over the range. The response should show no peaks, but be flat at about 2.5 KHz and -3db,(x.7) at about 3.2KHz, where only a small amount of abiasing ripple of the waveform should be visible. Set the delay to min.(CCW), The response should now have a peak of about +3db (x1.4) around 2.5KHz and roll off sharply above 3.5 KHz with no trace of noise or abiasing.
Set the input frequency to 250Hz and adjust level for 1.0V p-p
at pins 14 and 15 of the NE570. There should now be between .6 and
.8V p-p at pin 10. Set delay fully CW and disconnect the input.
Looking at the unit's output and overload light turn the feedback
fully CW. Touch-the input to introduce a transient signal. Runaway
oscillation should occur with the overload light flashing brightly.
Set feedback fully CCW and blend to center. Reconnect input and observe
output as input frequency is varied slowly. It should show a peak
and null (comb filter)response. Check that the footswitch connects
the direct signal to the output when it is thrown and that this
signal is always present at the direct output and is controlled
with the level control.
The test and alignment is now complete.

[Matt T.]

Thanks Stephen. Unfortunately, I don't have a scope. I'm not sure how I would connect all that stuff anyway, the pins are so close together and there's not much to 'grab'. I played it several times over the weekend and I like the way the echoes sound and it's not noisy so I'll probably just leave it as-is for now. But one day...when I get a scope...maybe I'll revisit this.