Would anyone be interested in modifying my Doepfer A-143-1?

[MrTea]

Hello DIYers,

I have a Doepfer A-143-1 Quad AD that I'd really like to get modified. I'm fairly useless at such things and woudn't really know where to begin and to be honest, don't really have the time at the moment to learn.

What I'd like is a range switch for envelopes 2,3 & 4 (drilled and mounted to the panel) that switches the range with 3 settings from:

Fast - very snappy, decent audio rate when cycling for FM

Normal - the stock settings

Slow - down to say 5 min cycles.

I've seen a few threads in the past about this so I believe it's possible to do it without too much trouble if you have some reasonable DIY skill.

I'd obviously be more than happy to pay for parts, labour & shipping to whoever might be interested.

If you fancy taking on a little paid project then let me know. UK/EU would be preferred.

Thanks in advance,

Matt

[Umcorps]

Same here

[EATyourGUITAR]

well the switches are either SPDT or DPDT ON-OFF-ON. caps add up when they are in parallel. the smallest cap would need to be on the doepfer board and the stock and long timing caps need to be on the switch. so all you need to do is replace one cap on the board and run two wires from the switch to the PCB. the other two caps are soldered right to the back of the switch. total cost is less than $2 if you already have the parts in stock. most people do. I could do the mod but I don't have a way to test it out. so you would be better served if someone who owns a euro system does it for you.

[Protocol]

Same here

Uk

[MrTea]

well the switches are either SPDT or DPDT ON-OFF-ON. caps add up when they are in parallel. the smallest cap would need to be on the doepfer board and the stock and long timing caps need to be on the switch. so all you need to do is replace one cap on the board and run two wires from the switch to the PCB. the other two caps are soldered right to the back of the switch. total cost is less than $2 if you already have the parts in stock. most people do. I could do the mod but I don't have a way to test it out. so you would be better served if someone who owns a euro system does it for you.

That does sound easy, but still probably beyond my current abilty

Well, perhaps not beyond my ability per say but out of my comfort zone. My biggest concern would be damaging the module.

[Tim Stinchcombe]

The mod itself is relatively easy - the trickiest part is drilling the faceplate. For those in the Doepfer Yahoo group, Peter Grenader suggested doing this back in message #10644. List member 'loopcycle' then posted some photos of his mod in the photos section here.

Switches need to be on-off-on types, and there is space for them above the one jack of sections 2, 3 & 4 at right. (Note the cap concerned is C1 according to the Doepfer cap page, and not as in the one photo.) Rough calcs for the values quoted suggest the standard 2.2uF gives approx 5ms to 15secs envelopes; a 47uF would be approx 100ms to 5min 20s; so a guestimate of a 100nF might be well into the audio, perhaps around 5kHz (assuming there is no other limiting factor in the circuit), down to approx 0.7secs. The cap on the board would need to be the smallest, with the other two being added in parallel - in turn - by the switch.

I could probably offer to do this, but without seeing a module 'in the flesh' it would be difficult to say how long/how much it would take/cost.

Tim

[Tim Stinchcombe]

I've just been reminded that I was going to add some details of what I did to MrTea's module when I modified it, so here goes!

We decided on cap values of 100nF for fast envelopes (there is a slight issue with going smaller, as outlined below), 2.2uF (the original fit) for standard length, and 47uF for longer ones. In LFO mode these give:

100nF:
shortest attack/decay: 270us/1.37ms 609Hz
longest att/dec: 132ms/408ms 1.9Hz

2.2uF:
shortest att/dec: 5.4ms/17.5ms 43Hz
longest att/dec: 2.6s/8s 0.09Hz

47uF:
shortest att/dec: 144ms/712ms period 0.86s
longest att/dec: 74s/228s period 5mins

Faster speeds can be obtained by patching the 'Cp' out back to the trigger in, and winding the threshold up (so you get a lot of DC too) - with the smallest cap, several kHz or more is obtainable this way.

I have plots of fastest and slowest envelopes for all three cap values, but they are pretty boring really, and the envelope shape at these extremes looks remarkably similar, albeit with different timebases - this leads nicely on to the issue with making the cap any smaller. In LFO mode the envelope output is fed back to the same point as the trigger input, where there is an additional capacitor as part of the trigger circuit. With the larger timing caps the effect of this other cap is fairly transparent, it charges/discharges so quickly we don't notice it. However if the timing cap is reduced beyond a certain point this cap makes itself felt, in two ways: it takes a finite amount of time to charge/discharge, and this makes the decay segment proportionately longer; in addition the leakage through the cap adds to this effect, making it longer still. This plot shows the difference in shape of the envelope at the fastest 2u2 setting (blue), compared to the fastest using a 10nF timing capacitor (red):


(If one felt inclined to use such a smaller timing cap, one could obviously spend some time experimenting with different values for the trigger cap value to alleviate this, and also not impinge too much upon the main trigger operation.)

The blue trace is typical of the envelope shape at the max/min extremes (obviously in between one can elongate/shorten one segment over the other), and note that the relative shortness of the attack over the decay is due to the way the circuit is designed, around a 555 chip - the timing cap charges from effectively the positive rail, and switches from attack to decay at two-thirds the rail voltage, so there is always about 4V or more across the resistor feeding the cap, whereas in decay the resistor is straight to ground, leaving the end of attack with a much steeper slope than at the end of decay, as in attack it is still charging at a faster rate than it is discharging at the end of decay.

There is also potential for at least one other simple mod, were there the front panel space. There is actually a footprint for a jumper on the 'EG' out, to allow for either the main envelope, or the polarized version, to be output. Unfortunately the position on the PCB didn't allow for the nearness of the pot bodies, so there literally isn't space to fit a jumper/option link, and so a link is hardwired to simply give the main envelope. Space permitting, it would be a simple matter to remove the link and add a switch to allow for either to be output.

To mount the caps, I used a small piece of stripboard: the original 2u2s I pulled off their respective boards; the 'default' cap is the new 100n one, with the 2u2 and 47u switched in parallel, in turn, by the on-off-on switches. The stripboard was carefully placed so there should be no danger of shorting to any leads sticking through the PCB, nor the 'bus' wire connecting all 4 boards (to which it is held in position with some thread):





Tim

[Navs]

There is also potential for at least one other simple mod, were there the front panel space. There is actually a footprint for a jumper on the 'EG' out, to allow for either the main envelope, or the polarized version, to be output. Unfortunately the position on the PCB didn't allow for the nearness of the pot bodies, so there literally isn't space to fit a jumper/option link, and so a link is hardwired to simply give the main envelope. Space permitting, it would be a simple matter to remove the link and add a switch to allow for either to be output.

Love the A-143-1 and never had an issue with the times (other than the fact that I'd like VC control), but this is great news!

Have I understood you right - a 'permanent' mod is possible? This would be fantastic as I often have to run the individual envs through Fonitronik's attenuverter.

[J]

^

interesting stuff...
thinking about it, similar to Navs, i'm ok with the times for the env. maybe longer times for the lfo section might be nice.

much more interested in outputting polarized versions of the env. instead of the standard.
a permanent mod would be cool, gonna take a look

[Tim Stinchcombe]

Have I understood you right - a 'permanent' mod is possible?

Yes, it would simply be a matter of desoldering the link currently made, from the centre pin to the one side, and replacing with a new solder link from the centre to the other side. Not sure if I have a suitable photo of that part of the PCB, but I should be able to identify which jumper it is, as no doubt I noted it on the tracing of the circuit I made - I'll check when I get home tonight!

Tim

[mikecameron]

I'm inspired! Going to mod mine as soon as parts arrive. Thanks for the info guys!

[Tim Stinchcombe]

Not sure if I have a suitable photo of that part of the PCB, but I should be able to identify which jumper it is

No photo I'm afraid, but it is 'JP6', and you'll find it between the threshold pot and the 8-pin chip immediately opposite (and you will notice that as it is so far under the pot, there is no headroom to fit a proper jumper/option link!). If I remember correctly, on MrTea's module it wasn't really a proper link of wire soldered through the holes, just a short piece on top of the pads, bridging across them. Thus to do the mod to permanently have the polarized envelope output from that section, remove the link, and short across the other pair of holes (I would suggest clearing the holes by desoldering properly, then actually inserting a link through the holes and re-soldering - much less fiddlier than trying to get a small sliver of wire to solder at both ends at once!).

And if you want it switchable (and can find space!), then use an on-on switch (single-pole change-over), with the middle wired to the middle of the JP6 footprint, then each side to one side of JP6.

Tim

[Tim Stinchcombe]

No photo I'm afraid...

Started thinking properly, and found the component layout at Doepfer's site - JP6 highlighted in red!:

[J]

thanks Tim

[Navs]

Thanks from me too, Tim! You've just given my A-143-1 a new lease of life.

Can you see a simple way to convert the decay comparator out from a gate to trigger, or at least shorten it?

I want to be able to mix the comparator logic for timing duties, but the sustained 'highs' make this difficult without resorting to external modules.

[tommi]

I think a capacitor in series and then a resistor to ground (1st order hipass filter) will do the job!

You 'll have to tweak the values of resistor and capacitor in order to find the right cutoff frequency to obtain the right time for trigger.
The formula to know cutoff frequency is:
Fc=1/6,28xR1xC1
Where R1 is the resistor value and is expressed in Kohms (in case of a 100K resistor you 'll put 100) and C1 is the capacitor and the value is expressed in Microfarads (in case of a 470pF cap you 'll write 0,00047).

[Navs]

Thanks, Tommi.

I wondered if there was something already 'built-in' that could be modified, but passive HPFs should do the trick.

[Tim Stinchcombe]

I wondered if there was something already 'built-in' that could be modified, but passive HPFs should do the trick.

Well I guess the answer to that is both 'yes' and 'no'! The op amp in the top section which sums all the polarized outputs for the 'mix' out, is still present and fully populated for the lower three sections, but doing absolutely nothing - it could easily be made into an inverter by simply adding a 100k input resistor and an output jack. However I can't see an easy way to use it to generate a pulse - I've tried a few quick simulations using it as a differentiator (i.e. sticking a cap in front of it, similar to tommi's idea), but it looks like there would be big stability problems. Which is not to say that major carving-up of the tracks to re-configure it for positive feedback couldn't do the trick!

As for tommi's idea of the 'cheap and cheerful' highpass, there might already be a suitable place to cut a track and simply insert a cap instead (and another diode to prevent negative spikes getting output might be a good idea too) - I can't be certain as my tracing of the circuit may not accurately reflect the actual traces on the PCB. But if I describe the circuit, you might be able to work it out from the PCB.

So, referring to the components as in the Doepfer layout above:

- the comparator generating 'Cpx' is pin 1 on IC4
- this goes to D5 anode
- D5 cathode goes to: LED D6 anode; 15k R16; output jack; section 'x-1' trig input jack

If there is a single track between D5 cathode and (R15 + output + trig in), then cutting there and inserting a cap would probably work.

To get rid of the negative spikes, a diode across R16 might do, but another resistor might be needed to give a complete path for the cap to discharge - depending on where the cap is in relation to D6 (I can't be sure from my drawing), that diode might provide this path, so no extra resistor would be needed...!?

I can run some more simulations to get an idea of cap value etc., if you are seriously considering doing this!

Tim

[Navs]

Thanks for the offer, Tim! I've just unmounted it and had a look - bloody hell, it's tightly-packed. As luck would have it, my camera battery just died, so no pics.

There is a cluster of four pins in a square arrangement. Two pins are joined (diagonally) with a small piece of wire (ex resistor leg?). Is that what I'm looking for for the output mod?

On the comparators, what I would like to achieve is a new output that would sum (diode OR gate?) the four modified gates/ triggers. I've been using a min/ max for this, but it involves extra patching and doesn't always work if, for example, there is gate overlap due to the relative thresholds.

I'll play around with the patch again to see what might be necessary e.g. if the following module is also an AD envelope which ignores the gate/ duration/ off, then maybe I just need to sum the comparators.

[Tim Stinchcombe]

There is a cluster of four pins in a square arrangement. Two pins are joined (diagonally) with a small piece of wire (ex resistor leg?). Is that what I'm looking for for the output mod?

Well the 'ex resistor leg' sounds right, but the 'four pins in a square' sounds like it might be different from the one I saw (?). From the top of the PCB, the white silkscreen markings for JP6 should be clear enough - three holes in a line, and on the solder-side the 'ex resistor leg' should be bridging two of the holes. This link needs moving to the other pair of pins. (Kicking myself for not taking a photo of it now!)

On the comparators, what I would like to achieve is a new output that would sum (diode OR gate?) the four modified gates/ triggers. I've been using a min/ max for this, but it involves extra patching and doesn't always work if, for example, there is gate overlap due to the relative thresholds.

Well quite coincidentally, from the other issue I'm trying to solve (on the A-119), I discovered the way get around the stability problem of using the unused op amp sections as differentiators by sticking a cap in front of them - only another resistor would be involved. Thus you could easily use these unused OAs on the lower three sections to generate a pulse, or you could sum the Cps of all four into one section which would give a pulse out whenever any of the envelopes dropped below its threshold.

And it wouldn't be too difficult to add a switch so that the mix output on the top section was used either for the mix (as intended), or so that it generated a pulse for the top section Cp out (i.e. so that this plus the three other unused sections now all did the same thing) - there are lots of options!

Tim

[Navs]

Is this what we're talking about for the output mod?



Ok, not four .. maybe six pins

Where do I start looking for the comparator gate > trigger mod?

[Tim Stinchcombe]

Is this what we're talking about for the output mod?

...

Ok, not four .. maybe six pins

Ah, makes sense now! The 'diagonal' link is simply because it is easier to anchor the small piece of wire to the IC pin - you'll note that it is the same connection as the end of the jumper, as there is a small trace joining the two!

To make the output always the polarized one, remove the link as it is now (red cross), and add a link the other side of the 'jumper' (green line):

Where do I start looking for the comparator gate > trigger mod?

I'll have a think about how best to describe it (the photo may help)... I'll be back!

Tim

[Tim Stinchcombe]

Aaargh! I had just finished annotating the photo with what to add to give a pulse out, and thought I should double-check the pulse polarity with respect to the 'Cp' polarity - I think I have got it round the wrong way. So to be sure, is this what you are looking for:

so you get a pulse output when (in this case) envelope 2 decays past the threshold set? (If 'yes', then I do have it the wrong way, so back to the drawing board - tommi's idea might be easiest...).

Tim

[Navs]

Thanks for the trace-cut confirmation, Tim. I'll get on to that!

Yes, I'd want a trigger when the CMP goes high. However, given the fact that I want to generate wonky rhythms, it might not matter! What was your solution and at which point is the trigger generated?

If it's too fiddly, I might have another source anyway: the A-137-2 provides four comparators and is DC-coupled so can be used as a quad gate delay.

I'll start a new thread if you think that might be an altogether easier mod.

[tommi]

I 've found a better solution wich is less ignorant, but more difficult to hack than the simple hipass filter. It is the Ken Stone Gate to Trigger circuit. With a dual opamp you've got two gate to trigger, plus it has a buffer wich is always better. Anyway i want to give a try to my cheap solution.. i don't have an A143-1 but i can try with gates coming from other modules.

The op amp in the top section which sums all the polarized outputs for the 'mix' out, is still present and fully populated for the lower three sections, but doing absolutely nothing

I wonder why they mounted on board to do nothing, but if these opamps are really doing nothing they are perfect to become part of that circuit.

Navs, do you want to drill holes on the panel and add 4 jacks, one for every stage, with the pulse output or you just want to modify every 'cp' out to a pulse leaving the panel as it is?