Diagnosis of a blown transistor

clintrubber said:

Remains puzzling... if this topology is flawed then it wouldn't be widely used for this kind of synth-application... reported problems for other synths would have been more widespread...  Possibly the dimensioning then, as suggested here above?

Click to expand...

Agreed that it's puzzling, and is likely not fundamentally flawed so much as subject to idiosyncracies in the parts. For starters, the whole approach of generating noise through reverse-biased transistors always has the "transistor selected for best noise" caveat.

I also wonder if it's a technique that worked passably well with through-hole parts, but that has some issues when moved to the scale of SMT parts. No idea... (is that what you meant by an issue with "the dimensioning"?).

As for other common noise generation methods, Arp supposedly moved from using transistors to zener diodes for the noise source in the Odyssey... if I rebuild the whole sub-circuit, I'll give this a try as well:

leigh said:

Well, you'll be able to hear if yours sounds as bad as the audio examples I posted! Hopefully yours doesn't...

Click to expand...

While not 'listening' at the same unfiltered node as you did, mine definitely sounds most like the earliest part.
I bought it second hand, and haven't used it 24/7 I must add, but I'll sure be monitoring it from now.

While the riddle is unsolved so far, are you meanwhile continuing to enjoy proper noise for instance by injecting an alternative (pseudo-random kind) noise source into the external audio? In other words: keep making music and/or solve tech-problem?

leigh said:

(is that what you meant by an issue with "the dimensioning"?)

Click to expand...

Meant the component values.

But different packages can/will result in different mechanical stress on the Silicon-die. How that translates to component behaviour depends on the situation. I know this is addressed for complete integrated circuits, but don't know to which extend this matters for 'single devices'. Can well imagine though the mechanism is simply still there, unless properly dealt with.

But hey, we're talking abouit changing behaviour over time, so degradation by applied biasing being far more likely than (an assumed constant) mechanical stress.

clintrubber said:

While the riddle is unsolved so far, are you meanwhile continuing to enjoy proper noise for instance by injecting an alternative (pseudo-random kind) noise source into the external audio? In other words: keep making music and/or solve tech-problem?

Click to expand...

Sure, there's other instruments around I can use, and I do have one other noise source in my synth setup, but the Minibrute is opened up on the bench, so it's currently out of commission.

I opened a support ticket with Arturia for kicks, to see if they could tell me anything. The initial response from them was less than inspiring - a total boilerplate "sorry for the trouble, please send us a video of the Minibrute set up with this test patch"... but the test patch did not include any noise!

leigh said:

Sure, there's other instruments around I can use, and I do have one other noise source in my synth setup, but the Minibrute is opened up on the bench, so it's currently out of commission.

Click to expand...

I see, that's a pain, need to solve it before closing the box again. No idea about available space inside , perhaps replacing the whole section by a small add-on PCB?

I opened a support ticket with Arturia for kicks, to see if they could tell me anything. The initial response from them was less than inspiring - a total boilerplate "sorry for the trouble, please send us a video of the Minibrute set up with this test patch"... but the test patch did not include any noise!

Click to expand...

Think to understand that this is more a test of their support-capabilities than that you expect info & aid towards a solution, right?  8)
But who knows...

Also curious to what the thoughts of Monsieur Usson would be about this topic, for more kicks, did you try him as well?

clintrubber said:

I see, that's a pain, need to solve it before closing the box again. No idea about available space inside , perhaps replacing the whole section by a small add-on PCB?

Click to expand...

Yeah, that's probably what I'll do, rather than try to rework multiple SMT parts. Given the small size of a noise sub-board, it won't be too hard to fit that in there.

clintrubber said:

Think to understand that this is more a test of their support-capabilities than that you expect info & aid towards a solution, right?  8)
But who knows...

Click to expand...

Well, contacting Arturia support was in part to see if they were going to suggest replacing the digital PCB, as had been the reported fix years back (on Arturia forums) – which I think is a red herring. And, then to see what the official word on the noise repair would be, if anything. After another round of emails with them, with someone who actually read my email this time, they sent the part number for Q28/Q29, and said I could try replacing that myself... so that's the extent of the official word I guess.

clintrubber said:

Also curious to what the thoughts of Monsieur Usson would be about this topic, for more kicks, did you try him as well?

Click to expand...

I haven't tried Yves yet, as I suspect he gets a mighty volume of help emails already, so it may be hard to get his attention about this. Besides, although it's his design, it's Arturia's manufacturing, and he might either not have any input about their choice of SMT parts, or be wary about badmouthing them if in fact he thinks they made a bad call with their particular part selection for this application.

Got back to the Minibrute this morning, and decided to try replacing Q28 (the reverse-biased transistor noise source) with a Zener diode. I had some 10V Zeners around, and the Minibrute has 12V rails, so started with that.  After listening to the noise from the Zener on a breadboarded circuit, thought it sounded (and it looked on the scope) like a good white noise source.

Pulled out Q28 and just replaced it with the Zener, onto the original board. No other parts replaced. Seems to work great. Attached is a screenshot of PT129 (the output of Q29) - this is zoomed in to a 20ms slice, like before. Looking a lot better!

So the Q29 gain stage that I thought perhaps was damaged, turned out to be OK. Thanks everyone who helped take a look at that circuit.

leigh said:

I haven't tried Yves yet, as I suspect he gets a mighty volume of help emails already, so it may be hard to get his attention about this.

Click to expand...

IIRC I've contacted him once, some years ago and did get a (good) reply, but can well imagine that's not always possible if he gets flooded by questions.

leigh said:

Got back to the Minibrute this morning, and decided to try replacing Q28 (the reverse-biased transistor noise source) with a Zener diode.

Click to expand...

Good to hear it's solved! And when it might develop again that the fix can be simple.

FWIW, a few days ago I thought to hear the noise-source of my MiniBrute 'sputtering' as well, but it went away again. It could be caused by this thread and might not have caught my attention otherwise.

Now I'm curious whether upside-BJTs-for-noise can have occasional 'sputtering', or will keep doing so once they've gone bad - but that's for another time.