Random Source / RVG by R*S Fixes and Information

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: luketeaford, lisa, Kent, Joe.

Post Reply
User avatar
dksynth
Wiggling with Experience
Posts: 341
Joined: Wed Jul 01, 2015 12:56 pm

Random Source / RVG by R*S Fixes and Information

Post by dksynth » Fri Aug 02, 2019 3:00 pm

Serge's fantastic random voltage source has been out for a while and sadly I don't see too many working ones! I've had to fix a few of these Random Sources and it seems to be a continuing issue so I will compile my findings for all here.

Use my info at your own risk yada yada you know the drill!

If you have your own tricks for this one please holler, would love to add any and all info that we've got!

here are some photos of circuits involved:

The pins:
Image


the cv drive circuit:

Image

1M5 to GND on the non inverting leg of the LM3900 has been added. There is a similar suggestion in a note from serge about making LM13700 work in this circuit (was originally CA3080). I won't copy his work here but if anyone wants more info about this PM me.

here is that 120k that is usually on the Smooth (top LM3900) but not on the Step (bottom LM3900). I'm not sure about this, seems like it would just bias the CV summing? But I'm not too sharp on the function of LM3900s.
Image

Here is the output stage:
Image

Some general things to get it working:

1. Check that the S/H source is actually making it to the STEP input. Refer to the pin designations above.

if you have a "loop" between S/H UNIPOLAR SOURCE and the unused pull down pin on your panel pcb as pictured here:
Image
(apologies for the condition, I thought there was a short and sanded it down to inspect)
you may have a mismatched panel PCB with your SSG board! It appears at some point in R*S's versions of the circuit that this loop connected the S/H to the Step input. You will have to MANUALLY wire the UNIPOLAR S/H pin to the STEP INPUT PIN as the trace that should be doing it for you is missing. ugly example here:

Image


2. Check offset of S/H UNIPOLAR SOURCE output. Mine all needed to be cranked up as far as possible to ensure no lock-ups. (IMPORTANT - even a board that is otherwise working will lock up if the S/H source offset is wrong. If you don't know which way is up with the trimmers you used, crank it one way and if it doesn't work try the other way!)

3. Check that the LED isn't pulling too much current and locking things up. You want a low current LED if possible. I have rewired mine to be driven off of the unused side of the buffer opamp instead of the LM3900 since it is right there and unused in the standard super-useful buffered-ground config (rhodes chroma service manual reference, :banana: if you get it). This photo shows how to make the mod:

Image

Lift the non-inverting pin of the TL072, wire it to the LM3900 output that runs there, cut the trace to the LED pin and connect the output of the opamp to that pin.

That's all I got, refinements/advice/more info welcome!!
Last edited by dksynth on Sat Sep 14, 2019 1:05 pm, edited 3 times in total.

User avatar
the bad producer
Super Deluxe Wiggler
Posts: 2532
Joined: Tue Mar 17, 2009 8:43 am
Location: Clapton Pond

Post by the bad producer » Fri Aug 02, 2019 3:53 pm

Outstanding! :yay: Beats swapping it out and wiring the CGS! The other thing I noticed was that a low-voltage LED is best to use - ie one that lights up at low-voltages, as well as low current.

User avatar
Moog$FooL$
Broke Dick Dog
Posts: 1720
Joined: Tue Jun 03, 2008 2:54 am
Location: Vancouver, BC.

Post by Moog$FooL$ » Fri Aug 02, 2019 3:59 pm

wow! dksynth.

that's really nice of you..... i will alert the owner.

Reese P. Dublin & Harold Donnelly..... are you guys seeing this???


:woah:
i use to drink & smoke.
that fat cat had to go..... wasn't even mine.

User avatar
dksynth
Wiggling with Experience
Posts: 341
Joined: Wed Jul 01, 2015 12:56 pm

Post by dksynth » Fri Aug 02, 2019 4:44 pm

the bad producer wrote:The other thing I noticed was that a low-voltage LED is best to use - ie one that lights up at low-voltages, as well as low current.
I don’t 100% grok 3900s yet so I don’t know how the numbers stack but I think the LED draw is screwing up a few people’s boards.... everything (input feedback, cycle comparator, coupler) comes from that bus, tacking a normal LED with a current limiting resistor seems risky.

My mod is a bit invasive, you can also pull the LED pin entirely and short it to the output pin on the panel board so that the output buffer is driving both. Not sure if that’s any improvement.

User avatar
dksynth
Wiggling with Experience
Posts: 341
Joined: Wed Jul 01, 2015 12:56 pm

Post by dksynth » Fri Aug 02, 2019 5:40 pm

Moog$FooL$ wrote:wow! dksynth.

that's really nice of you..... i will alert the owner.

Reese P. Dublin & Harold Donnelly..... are you guys seeing this???


:woah:
I read that thread and it sounds like some of you might also have the weird panel PCB ... it was Leturbo's I fixed here actually. If the PCB works properly as an SSG in the SSG slot, definitely check for that loop, that weird panel PCB. I've only seen one in the wild but I have other PCBs that have faults that were "not supposed to be released" but somehow did. So you never know!

harolddonnelly
Veteran Wiggler
Posts: 526
Joined: Fri Mar 13, 2009 3:00 pm
Location: Vancouver, BC

Post by harolddonnelly » Fri Aug 02, 2019 6:43 pm

dksynth wrote:
Moog$FooL$ wrote:wow! dksynth.

that's really nice of you..... i will alert the owner.

Reese P. Dublin & Harold Donnelly..... are you guys seeing this???


:woah:
I read that thread and it sounds like some of you might also have the weird panel PCB ... it was Leturbo's I fixed here actually. If the PCB works properly as an SSG in the SSG slot, definitely check for that loop, that weird panel PCB. I've only seen one in the wild but I have other PCBs that have faults that were "not supposed to be released" but somehow did. So you never know!
Thank you for this very timely post! @Moog$FooL$ and I are working with the SSG in the Carnivore form.

What should we be looking for in terms of being weird on the Panel PCB?

Thank you so much for your effort on this!

User avatar
dksynth
Wiggling with Experience
Posts: 341
Joined: Wed Jul 01, 2015 12:56 pm

Post by dksynth » Fri Aug 02, 2019 8:01 pm

harolddonnelly wrote: What should we be looking for in terms of being weird on the Panel PCB?
Here are some symptoms of a bad/weird panel board that I had:
board works fine in the SSG side of the module
in the Random Source/RVG side SMOOTH and STEP outputs are dead
Timing Pulse is going nuts

this is the gist of the RVG patch:

Image


Start with checking that the S/H SOURCE (unipolar) is properly connected to the input of the STEP side. You can find the proper pins with the labelled pin diagram in the first post. If you have an oscilloscope check that the S/H source is all voltage at or above GND .... although in my case (I have three) I set all of mine all the way up with the S/H trimmer and then they stopped locking on me.

I have also seen some tombstoning issues on R*S boards, never on an SSG though just on NTO and DUSG. They appear to use quite heavy copper (quality!) and this is one of the downsides. Closely inspect the SMD with a loupe if you have one and see if any are resting on TOP of the solder ball instead of being properly soldered. These can be very tricky to spot.

may your :deadbanana: turn into :bananaguitar:

User avatar
lucid
Common Wiggler
Posts: 214
Joined: Mon Jan 23, 2012 4:46 am
Location: sergeland

Post by lucid » Sat Aug 03, 2019 12:07 am

thanks for great job dksynth
I also have some issue with mine, mostly realted to no blinking LED or freezing LED. I am sure that I used low current LED, but maybe not enough low current and low voltage one. Once I was messing with mine I checked output with osciloscope and there is something happening on the output thought LED is visually not "active". I need to take some time take module out and revive it with your instructions, will be back with mine findings. Are you referring to some specific board revisions?

User avatar
the bad producer
Super Deluxe Wiggler
Posts: 2532
Joined: Tue Mar 17, 2009 8:43 am
Location: Clapton Pond

Post by the bad producer » Sat Aug 03, 2019 8:03 am

dksynth wrote:
the bad producer wrote:The other thing I noticed was that a low-voltage LED is best to use - ie one that lights up at low-voltages, as well as low current.
I don’t 100% grok 3900s yet so I don’t know how the numbers stack but I think the LED draw is screwing up a few people’s boards.... everything (input feedback, cycle comparator, coupler) comes from that bus, tacking a normal LED with a current limiting resistor seems risky.

My mod is a bit invasive, you can also pull the LED pin entirely and short it to the output pin on the panel board so that the output buffer is driving both. Not sure if that’s any improvement.
Interesting - I found low forward voltage was the key specification for the LED so that makes sense. I didn't actually consider looking into it at the level you have, but may have to start doing that as more projects come out without schematics (eg the new CGS Serge circuits from Elby appear to have no schematics)...

User avatar
Reese P. Dubin
Super Deluxe Wiggler
Posts: 1359
Joined: Wed Dec 17, 2008 12:33 pm
Location: Absolute Elsewhere

Post by Reese P. Dubin » Sat Aug 03, 2019 2:05 pm

Just skimmed this and it looks promising!

User avatar
dksynth
Wiggling with Experience
Posts: 341
Joined: Wed Jul 01, 2015 12:56 pm

Post by dksynth » Sun Aug 04, 2019 10:42 am

lucid wrote: Are you referring to some specific board revisions?
it appears that there were some different panel component board versions, yes. The part that you solder the bananas to! I have two different ones here and one is faulty. Check that S/H SOURCE UNIPOLAR pin and STEP INPUT pins are connected (buzz em out with a multimeter if it is already soldered to the panel)
the bad producer wrote:
Interesting - I found low forward voltage was the key specification for the LED so that makes sense.
yes, precisely what I found. if it wasn't clear the first time I meant to be agreeing with you!

User avatar
the bad producer
Super Deluxe Wiggler
Posts: 2532
Joined: Tue Mar 17, 2009 8:43 am
Location: Clapton Pond

Post by the bad producer » Sun Aug 04, 2019 11:02 am

LOL - yes, I was just clarifying what I actually meant to say above :hihi:

User avatar
Reese P. Dubin
Super Deluxe Wiggler
Posts: 1359
Joined: Wed Dec 17, 2008 12:33 pm
Location: Absolute Elsewhere

Post by Reese P. Dubin » Sat Sep 07, 2019 11:55 am

THIS WORKS

I can hardly believe it but this panel is finally 100% functional. The mod was easy but a no-go until dialing the Unipolar trimmer all the way CCW, then the glory of a fluctuating LED after a very very long frustrating couple of years.
A full six foot tall karate tournament style trophy to dksynth for this work.
Happily wobbling an E370 right this second.

User avatar
dksynth
Wiggling with Experience
Posts: 341
Joined: Wed Jul 01, 2015 12:56 pm

Post by dksynth » Sun Sep 08, 2019 8:25 am

Reese P. Dubin wrote:THIS WORKS

I can hardly believe it but this panel is finally 100% functional. The mod was easy but a no-go until dialing the Unipolar trimmer all the way CCW
:) awesome. I'll add a note to clarify about the uni-polar trimmer ... I mentioned cranking mine all the way up but it's a bit buried!

User avatar
Reese P. Dubin
Super Deluxe Wiggler
Posts: 1359
Joined: Wed Dec 17, 2008 12:33 pm
Location: Absolute Elsewhere

Post by Reese P. Dubin » Sun Sep 08, 2019 9:39 am

I cranked it in both directions and CCW won the day.
Seriously over the moon with this.

User avatar
drwiener
Common Wiggler
Posts: 244
Joined: Fri Aug 12, 2011 2:32 am
Location: Berlin

Post by drwiener » Tue Nov 05, 2019 3:39 pm

cool - nice hacks!
for a simpler approach about the freezing, I generally get it to work by using a ~3,3k resistor instead of the Trimmer with a properly rated low current LED on the LED. you can test different resistor value to see which one keeps it from freezing with your LED.

User avatar
xonetacular
Super Deluxe Wiggler
Posts: 2570
Joined: Mon May 21, 2012 10:35 pm
Location: Boynton Beach, FL, USA

Post by xonetacular » Wed Nov 06, 2019 12:14 am

what are some good low voltage/current leds you have had success with?


I have one 4x4 module I built that works, but the one in my edelweiss panel is a no go and I havent messed with it in a long time. may be time to revisit

User avatar
drwiener
Common Wiggler
Posts: 244
Joined: Fri Aug 12, 2011 2:32 am
Location: Berlin

Post by drwiener » Wed Nov 06, 2019 5:52 am

xonetacular wrote:what are some good low voltage/current leds you have had success with


I have one 4x4 module I built that works, but the one in my edelweiss panel is a no go and I havent messed with it in a long time. may be time to revisit
from reichelt or banzai (i'm in germany) - it should be rated at 2ma. if you are only having the LED/freeze issue, not the mismatched PCB then try first replacing the trimmer with a fixed resistor like 3,3k... if it still freezes try higher, if you want to get it brighter try a lower resistor... you don't need to solder in the resistor, you can just bend it to make contact after you remove the trimmer.

User avatar
dksynth
Wiggling with Experience
Posts: 341
Joined: Wed Jul 01, 2015 12:56 pm

Re: Random Source / RVG by R*S Fixes and Information

Post by dksynth » Wed Jan 08, 2020 12:12 pm

I've found one more potential quirk that I am willing to bet is the last cause of lockups.

The S/H source which is what is driving the SAMPLE part of the RVG patch goes high about a second after startup.
This has been confirmed across several versions of the circuit so I'm about 90% sure it is a quirk that has existed since day one.
If you power on your system with the RATE knob of your RVG all the way up, there is a coin flip's chance that the spike makes it through the first integrator ... and the second one can't match it, freezing the whole give and take of the RVG patch.

I've added two GIFs below ... .the big spike shows unmodified behavior whereas the small spike shows the behavior after I added a 5.6z zener clipper.
You can solder it right to the banana jack of the S/H source output and solder the other end to GND ... you may have to use a razor blade to expose GND plane somewhere handy.

This has worked for me now but if someone has a better solution I'm all ears!
You do not have the required permissions to view the files attached to this post.

Post Reply

Return to “Music Tech DIY”