FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 

DUSG Help!
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2, 3  Next [all]
Author DUSG Help!
I've been banging my head against the wall on this for a while now on this DSUSG. I have built up one CGS114 board with the modified values as in the build notes (1nf & 100K) on the TRIG IN. I have used 2n3904 and 3906's placed in backwards to the silkscreen and a BC547 for the LED placed in correct to the silkscreen. I dont really seem to get any output. I started trying to trace in with the scope but basically stop seeing my signal after the 33K. very frustrating

I tried the patch described in the notes, patching the TRIG and END together, at some point the LED lights up and the output goes high, but no triangle wave??
Are both USGs showing the same errors? I recently debugged this PCB, turned out to be a few solder bridges, but I had a lot of help comparing both sections. One of them did actually work so I had a good reference point.

Also this thread with debugging tips from zthee are gold since this module can be a bit complex and daunting.
Ya both sides do pretty much the same thing. I decided to try and troubleshoot one side first. Replaced all transistors and the lm3900, but it still does the same?
Ok, I removed all the panel wires from one half and followed Zthee's instructions. Pin M is high when nothing is connected, when A & M are connected together the output goes high, but there isnt any triangle

Measure +-12v no problem, no short to power either
7 links? You tried the trimmers?
OK, I'm getting a triangle now, my trimmers was all the way down and what I am finding now is that I can only get about 3.5KHz from the trimmers before the output "crashes" and I get nothing, when this happens I need to "reset" it by unplugging A from M before I get any triangle wave back. The trimmers are both almost fully CCW and when the go full CCW this is when the output "crashes"
+1 for reflowing your solder joints.

The next thing I would do is double and triple check you have all your diodes soldered in the right way.

I had a similar problem with my VCS where the output was stuck high and it wouldn't trigger. It was the caused by a couple of dodgy diodes that were not blocking current properly. Once they were replaced it worked fine.

Keep at it mate, you'll get there.
If it is cycling I would reconnect the rise and fall pots, trim it per instructions, and see if it is good.
I seem to have taken a few steps backwards. I ripped out all of the wiring to help troubleshoot. I have left the trig and end connected so i can patch them together. Now one half doesnt output when i patch tirg and end. I have replaced all the transistors, chips and diodes and reflowed everything on that half and i still get nothing. The other half seems to output fine when the trig and end and patched together. I did notice if i place a trig in from a sequncer i get a single triangle wave on either half and the leds very dimly blink on either half....

Im giving up for tonight, it has to be sonething stupid. For heavens sake the board is so simple. very frustrating very frustrating
yan - when debugging this board, the first usual suspects are the trimmer and the led. not sure if you've read this elsewhere, but this thing is crazy sensitive to the current draw on the led. too much and it crashes. if you're not using a low current led (2ma max is as high as i care to go with a dusg) that may be worth considering. for reasons also related to pulling to much current, the trimmer will also bottom the module out if set incorrectly. with an led that's only marginally in spec that margin on the trimmer can be razor thin. (i have one that works perfectly at 5v spec, but crashes at 5.1v) especially with a single turn trimmer.

that said, it sounds like you've gotten that squared away, so now t's probably time to time to break out the loupe - this board is notoriously fragile, and deteriorates quickly with rework when troubleshooting even when handled gently. with all your desoldering - tracing the schematic with a meter in continuity mode and a highlighter may be of some help. i've found lifted pads/traces in areas that showed almost no sign of molestation - it can be crazy. pay particular attention to the negative rail traces - for some reason unknown to me, i've found those dislodged in some manner or another on a half dozen boards i've debugged.

the first one of these i ever built was a real bastard - even after building the circuit six more times successfully i was unable to get it running right (problems similar to yours), and into the fuck-it-bucket it went to be parted out at a later time. after helping a buddy debug a dusg i felt inspired and pulled it out of the junk drawer for a second (or better six thousandth) look. after all the time rattling around in the garbage the suspect pad further dislodged and was easy to spot. 5 mins later, with a jumper properly installed, the bastard came alive - happily sloping along...
As a last ditch effort to throwing it in the fuck it drawer I used a solder vacuum and desoldered every pad on the half that wasnt working. I checked each pad under a microscope and resoldered everything under the microscope. Lo and behold it worked very frustrating

I was able to get both sides working and calibrated them

One final question; in the first paragraph of the build notes it says to patch the END and TRIG together and you should get 5KHz, the most I can get through the trimmer is 3.6KHz, after I set it to 100Hz and trim to 5V Pk-Pk, I get 2.9KHz with both the rise and fall knob turned up fully.

Any comments on this LED, I'm not sure if it has some cause and effect on not being able to hit the 5KHz described above.

Digikey 160-1087-ND
right on! i'd say 3.6k is well within acceptable range - particularly if you've implemented any of the anti-lock-up mods detailed here and on the cgs page (dropping the timing cap to 1n for example) which limit the circuit's range (hi and lo) in favor of the increased reliability. in the interest of science i put my merry band of cgs slopes on the frequency counter along with their leader, the Random Source 4x4 offering, and got the following:

4x cgs boards (w. mods) appx. 2.5k max
2z cgs boards (unmodded) 4.5k max - these do on occasion latch up when cycling
1x random source (smd w. 10n 1% styrene caps) 5k on the nose...

several of the cgs slopes have fancy styrene caps too, though they are 1n - they might go a bit slower, but that could be fancy cap voodoo...

i wouldn't use that led according to its 10ma spec, but with that said, given your measurements i'd leave well enough alone! there are lots of other component variables going on in this thing beyond the blinkenlights. congratulations on bringing it back from the dead thumbs up
I have one of these boards waiting to be soldered up.. any tips on how to minimise soldering problems?
I think just be very cautious is key. I would consider my self an expert solderer with years of almost dailey experience in both smd and through hole. This may have been my downfall by working quickly and too confidentaly. While resolderimg under the microscope i noticed on several occassions solder joints that would form a bubble of flux and then burst leaving a hollow spot on the joint. After finding one under the microscope i tried looking at it without and it appeared like a normal solder joint, the hole looked like a reflection from the light. You had to tilt it in a few directions to see it clearly. I have never come accross this issue; and it may not necessarily be this pcb, I also may have contaminated my solder tip somehow casusing this. Anyways work slow and check your progess often thumbs up

Thanks eveyone for the help and livefreela for the frequency testing. I will try to dig aroundfor a lower draw led and will post a part number if i find one
i concur with yan on all counts - a couple of additional things though:

-whilst i usually love washable solder, i avoid it on this board in lieu of low residue no clean. i love the look of 331 of course, but the combination of the scrubbing and the corrosive flux can make for an unhappy dusg.

-populate sparsely. if you're a jam everything in via panavise and solder in one pass kind of guy like me, you're asking for trouble. many of the pads are tight, so lifts and shorts are easy to come by if you're trying to navigate through a nest of component leads with the iron - particularly in the transistor cores. a component bender can be helpful here to alleviate strain on the pads.

-be gentle with the strain relief holes - many are dangerously close to traces. on most cgs boards these are tight and i open them up a little bit with an awl. try not to do this with the dusg. the nte24 hookup i favor even varies slightly in dimension by color (red being the largest) so i eschew my usual color scheme for what fits.

-to that point, allow extra slack when fitting wiring to the board. good rule of thumb is to avoid desoldering whenever possible, so if you use a test jig for bring up prior to to fitting to a panel you can simply clip the leads and rewire the remainder for final fitting.

-test your transistors first! desoldering that core because you have a funny batch with flipped leads is a bastard and these are more common then we'd care to think...

-and of course, just be slow and gentle - extra deliberate in your placement. this board tends to bite experienced guys on account of its simple appearance. given how rework can compound problems think of any mods you'd like to include beforehand rather than planning on desoldering at a later point in time...

good luck!

also loudest warning's wiring guide can be quite helpful:
I'm bumping this, I'd love to get some help with this build, having some trouble. I've already built a couple of these without problem, but this one is acting up.

One of them is working flawlessly, on the other unit Rise (or Attack) times behave weird. In loop mode they reset to +3V and peak at around +9V, so I'd say it has a +3V offset compared to how it should work. If Rise time is set short and Fall time is longer the envelope works right (drops from +6V to 0V). If triggered from an external source longer Falls with short Rise times work fine as well, but as above, longer Rise times with short Falls peak too high (climbs from 0V to ~8V).

What can cause this? I've already checked soldering/traces around suspected area (4 transistors, inputs from pots,etc.). I've also read this article by Tim Stinchcombe and I have a general understanding of how this circuit works, but still can't figure out why the hell it happens.
In the meantime I've figured out, that when using an external trigger, everything works fine. In cycle mode however, attack phase ranges from +3V to 9V but only with short falls.

Edit: also changed all transistors, diodes, ICs, the LED and the trimmer.
I have a quad slope in which one quarter isn't functioning properly. When cycling, turning the fall pot only works in the first 10% of the pot. After that the fall time falls off too fast and the cycling stops. Turning back reengages the cycling. I know it's not the pot itself, because I swapped the rise and fall pots (I used connectors instead of soldering everything to the board). hmmm.....
search64 wrote:
I have a quad slope in which one quarter isn't functioning properly. When cycling, turning the fall pot only works in the first 10% of the pot. After that the fall time falls off too fast and the cycling stops. Turning back reengages the cycling. I know it's not the pot itself, because I swapped the rise and fall pots (I used connectors instead of soldering everything to the board). hmmm.....

Does it work right with non-cyclic triggering?
Not sure, will check.
Figured I'd bump this instead of making a new thread. I'm having an issue with my DUSG (aside from killing LED's by mistake). I connected everything which I probably shouldn't have as now I have to undo it to check the voltages. Nothing seems to work, the LEDs turn on but nothing else worked. with the multimeter I can't seem to find any continuity issues. I'm going to try and rip out the wires then check the board and then rewire everything slowly.

Any advice on what check because this board has been tough
I would wire in just one half of the board and start there. Once you get half of it running the other one should be the same and you can reference the working half. Once you rewire it start with the basics like power on all the chips make sure all the jumpers are in place. If you have a scope start by giving it a trigger and look for output on pin 9 and then 10 of the 3900
So I took out all the wires then I took the advice on another post regarding doing particular things such as connecting pad M to A and checking DCO. When I connect it to hear what's going on I just get a dull drone which sounds kinda buzzy not at all like a high frequency triangle I should be getting. I get readings but I probably have to check every resistor and stuff.

I'm also confused regarding the PNP transistors. I have the 2n3904/2n3906 replacements but I checked on my multimeter and I'm getting readings between the each lead so I assumed that I don't have to try t out.
peachfuzzmcgee wrote:

I'm also confused regarding the PNP transistors. I have the 2n3904/2n3906 replacements but I checked on my multimeter and I'm getting readings between the each lead so I assumed that I don't have to try t out.

I am not at all sure what you are saying here. I'm sure you know the following, but if you don't it will be useful so here it is:

2N3904 is an NPN transistor and 2N3906 is PNP. You can't substitute them for one another.

One wrinkle with the DUSG board is that its silkscreen for the transistors is "backwards" if you are using 2N3904 and 2N3906 transistors. As long as you make sure you have PNP and NPN in the right places, and that the emitters are in the correct pads, you'll be fine. With 2N3904/6, getting the pins right means orienting the transistor so that its curved side is where the silkscreen has a flat side. But check the pinout of your transistors, just in case!
Well that's what I'm confused about, then just rotate those transistors because I put them in the way that the silk screens are orientated. I just need to turn around all my transistors? annoying but not the worst. Maybe that's what's wrong.

I was just confused as I was told that if my multimeter gave me reading between the pins somewhere between 300-500 all was good Or maybe to tell me that I needed to flip them. I don't know as the transistor thing was confusing so I'll just spin them around and try again.

Maybe that's the only issue.
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2, 3  Next [all]
Page 1 of 3
Powered by phpBB © phpBB Group