Vintage Deluxe Memory Man Repair has me a bit stumped

[Scruffie]

A friend of mine gave me a deluxe memory man he picked up cheaply to see if I could get it running again, it was in a very sorry state, the transformer's been removed, wires were mostly missing, parts had been removed and the PCB looks like some one got a stanley blade and just went to town on it, not to mention the enclosure mutilation (looks like they tried to add an XLR) on an other wise clean '78 memory man.

Sooo, I traced it through and replaced all the parts missing (a pain when there's lots of pads that serve no purpose) got it wired back up, fixed the damaged traces and wired up a little LT1054 inverter+doubler with regulator board to power it (having removed the original discrete device) and fired it up.

One blown MN3005 replacement later and we're back delaying!

BUT while all is fine at delays above... at a guess 75-100mS as soon as you go to lower times, the delay suddenly gets all distorted and the voltage gets pulled down to about 12V, no good.

So, I figure that as current draw increases with higher frequency, perhaps the charge pump is just struggling, tried unhooking the LFO OpAmp and overload LED to see if that would help, no good, I tried paralleling 2 x LT1054 by, as I understand it, connecting pins 3,5 & 8 of the 2 chips and giving the first one it's own 10uF between pins 2 & 4, no good either.

To be sure it wasn't a faulty delay pot or trace, i've tried disconnecting it and adding just a resistor between pins 2 & 3 and disconnecting the LFO too, no good. Replaced the 10uF filter cap to the 4047 and the 4047 its self, nada.

Also, while the charge pump runs fine from 9V providing -15v unhooked from the circuit, as soon as you connect it, it gets pulled down and wont work unless you feed it 12V which is a bit of an annoyance as powering it from 9V would be preferable (these are both 200-300mA supplies so I assume that's not the issue).

So i'm a little unsure as to where to go from here... am I just paralleling the LT1054s incorrectly? Is it possibly drawing even more current than that can provide at lower delay times (the +15V deluxe memory man clone i've built runs fine from just a single LT1054) could the regulator be at fault (it has shorted on the enclosure a few times when the temporary mounting I used... fell off, although most of the time it does its job fine) is it possible a part is drawing excessive current, remembering that most of the time it does work just fine in all respects.

Here's the schematic for the memory man;


And this is the schematic i'm using for the LT1054, minus D4 & 5 and C4 & 6

[Mark Hammer]

The BBD uses a clock.  The charge pump also uses a clock.  Clocks are like male Siamese fighting fish: they just don't like each other if they get to close.

I would seriously consider skipping the charge pump, making a regulator daughter board, and using a 9VAC adaptor.

[theehman]

The BBD uses a clock.  The charge pump also uses a clock.  Clocks are like male Siamese fighting fish: they just don't like each other if they get to close.

I would seriously consider skipping the charge pump, making a regulator daughter board, and using a 9VAC adaptor.

My thoughts exactly. 

[Govmnt_Lacky]

From what I read. this does not seem like a heterodyning issue as Mark eludes to with his solution. It sounds to me like you are over taxing your charge pump circuit. The LT1054, I have found, is ONLY good to about 50-60mA of current draw in almost all configurations. When you are changing the clock you are upping the current draw and this is most likely what is causing your distortion as it is dropping your voltage AND throwing off your BBD bias as well.

I would recommend getting the transformer replaced with a suitable substitute. ehman can point you in the right direction 

[armdnrdy]

If you have a way to test the DMM by disconnecting the charge pump circuit, you can isolate the problem.

This is where a bipolar bench supply comes in handy.

You can use mark's idea for testing as well. Breadboard a bipolar supply.

http://www.musicfromouterspace.com/analogsynth_new/WALLWARTSUPPLY/WALLWARTSUPPLY.php

[Scruffie]

It is not heterodyning that's for sure, i've dealt with heterodyning and never known it to affect a charge pump in such a way and other than the apparent current issue, they're working quite happily together.

I covered the point of exceeding current draw and higher frequency (shorter delays) requiring more current which is why I brought up paralleling two charge pumps for more current, I thought I covered that fairly well...

Let alone dropping the voltage to a level at which the DMM is not suited to work at, mis-biasing it, it also becomes unstable, i'm aware of all these issues.

A transformer would a) be a large expense (they're not as cheap in the EU as the US) and the guy is just a student b) be a pain as the person this is going back to is in a different country who will have a different socket and voltage to me and c) he's quite set on it working with his power supply, which I agree if can be made to happen is far preferable.

9V AC would be an option but if would be far preferable to have this work from 9V or even just 12VDC.

So, if we're certain it's just a current issue and not some other fault, how do I go about putting 2 x LT1054 (or even 3 if that must be the case) in parallel to provide this with enough current to function correctly as I assume I did so incorrectly OR is there a better inductor based switched based power supply I could use that would provide enough current easily (although LT1054s would be preferable).

[Scruffie]

If you have a way to test the DMM by disconnecting the charge pump circuit, you can isolate the problem.

This is where a bipolar bench supply comes in handy.

You can use mark's idea for testing as well. Breadboard a bipolar supply.

http://www.musicfromouterspace.com/analogsynth_new/WALLWARTSUPPLY/WALLWARTSUPPLY.php

I may have a 12v AC adaptor somewhere, bit of a hassle but I suppose it may be a good idea if there is some issue and it's not the charge pump at fault.

A bench supply would be handy but repairs aren't my usual thing, it was just a bud so figured i'd take a shot at it for him as it was quite the state that a tech would likely have charged more than he could afford.

I would prefer to rule out an issue with the paralleled charge pumps before taking that route though.

[theehman]

18vdc wallwart and a daughterboard w/ a regulator.

[armdnrdy]

18vdc wallwart and a daughterboard w/ a regulator.

An 18vdc is not a very common item....I have a DC brick that supplies both 9 and 18vdc, so I've designed many of my builds around 18vdc in.

When a friend wanted me to build a pedal, I searched the net for 18vdc wall warts.....not a very common item, and a bit pricy when you find them.

for one of my last builds, the Mutron Flanger, I designed the power supply to accept both 18vdc and 12vac.
Almost every DIYer has a 12vac adapter laying around that they have saved when discarding some electronic device.

[mth5044]

On the contrary, I don't have any 12Vac power supplies, but two 18Vdc 

Dunlop still makes an 18V wall wart that is avaliable where dunlop stuff is sold at most online places. Danelectro had made an 18V one too, but I'm not sure if that's still around.

[theehman]

EHX also has one. 

[StephenGiles]

I would use 3 9v batteries and regulate the voltage to the required level, which should give you some time to evaluate whether or not the circuit is working properly before they are drained. Alternatively use a car battery - out of the car of course!

[Kipper4]

would a laptop power supply get you somewhere in the parish of 18v?

[LaceSensor]

Strange. I used an lt1054 charge pump in my DMm build and it works a treat.

[mth5044]

Where are you all building your DMM's from? Madbeans layout or is there somewhere else?

[Scruffie]

Strange. I used an lt1054 charge pump in my DMm build and it works a treat.

As I mentioned, this is the case, my madbean DMM works fine too from a LT1054 and this does, to a point... so I figure even if current consumption is improved with modern parts, 2 in parallel should do the trick, but I couldn't get it to work, so I think I did it wrong, especially as there is a lot less info on inverting and doubling out there than just doubling.

Where are you all building your DMM's from? Madbeans layout or is there somewhere else?

Clones, yes, that and rusty pinto did a 4 x MN3008 version some guys might have done.

would a laptop power supply get you somewhere in the parish of 18v?

I need -15v and laptop supplies are pretty badly filtered and switch mode as far as i'm aware, I think this could also cause issues... it wouldn't be my go to, plus finding the right socket!

I would use 3 9v batteries and regulate the voltage to the required level, which should give you some time to evaluate whether or not the circuit is working properly before they are drained. Alternatively use a car battery - out of the car of course!

Now that is a more practical approach in this case (the 3 x 9vs   ) I honestly didn't think of that but it would help rule out issues. Thanks Stephen

Buying a special supply to test it would add cost both he and I can't afford to find out it's not the issue, I could expect to pay £10 at least and then probably never find use for it again and I can't spend that much on non essentials.

[Mark Hammer]

I don't know why we're even talking about 18V.  Th CMOS chips in the pedal do NOT like anything much higher than 15V, which is why the original regulates things down to around 15V.

[theehman]

I don't know why we're even talking about 18V.  Th CMOS chips in the pedal do NOT like anything much higher than 15V, which is why the original regulates things down to around 15V.

I just figure 18v is a nice amount of voltage to put into a 15v regulator.  The regulator doesn't have to drop too much voltage and shouldn't get as hot.

[Govmnt_Lacky]

I see on the LT1054 datasheet that they do have an example circuit for a Bipolar Supply Doubler (Page 11)

The bad thing is that the datasheet is not very friendly with respects to current draw limits for the example circuits 

I have always gone by the 50mA rule. If using the LT1054 as a doubler... 50mA max. I have "heard" that they can go up to 100mA but have never tried. But, as you want a Bipolar supply, I do not know what the impact would be on the current supply limits.

Perhaps a call to Linear tech support is in order! 

[Scruffie]

I tried Stephens idea and used 2 x 9V batteries which gave me -19V and tested it, the upshot is that it works fine like that, but I did check the voltage and it is still being dragged down at short delays through the regulator, it's just not going quite as unstable and pulled down as much as a charge pump.

Doesn't seem quite right to me...

I'm wondering if there's a bad cap(s) causing excessive current draw... does that sound plausible, cap issues isn't really my strong suit.