PPA V2 HELP! - smoking output resistor (long)
Jul 8, 2005 at 12:12 AM Thread Starter Post #1 of 108

n_maher

Resistorous Conflagorous
Joined
Nov 20, 2004
Posts
8,409
Likes
34
Ok, so I'll freely admit that I'm a bit out of my league here. I helped my friend Ian build a PPA V2, he did all the board work and I built him a power supply and helped with the case work and wiring. So, two weeks ago we had a working PPA that wasn't in a case.

Fast forward to the other night when we decide to put the PPA into its home. Everything goes in and the only hiccup is that while checking the switch I accidentally put my meter probe across the positive LED pad and one of the switch pads (I think) causing a spark. Nothing explodes (LED included) and the only casualty appeared to be my nerves. So, we boot the amp up and everything still seems to be going fine so I check DC offset. The right channel has 25mV so we decide to swap out that 637 for a fresh one and chalk it up as the nights only badness. Recheck all the channels and we're back to everything below 3mV so we hook up a set of test cans. Holy hiss! We then futz around checking this and that for god knows how long and decide to see if it's just the test cans. Hook up a set of HD650s and viola, very nice sounding amp. Plug in SR325is and blech, hiss city. So we put the PPA aside (it's after midnight and we've been at this for 5 hours) and figure we'll revisit this another day. It seems to be a grounding issue as when you touch the wires leading to the headphone jack the hiss disappears.

Now jump to last night when Ian decides to try and see if the hiss is at all related to cranky buffer bias and starts checking the buffer bias. Ruh roh rorge, the bias is jumping all over the place and when he finally gets a steady reading poof, the ground output resistor R33G smokes - big time. So Ian brings the amp back to me with his head hung low and asks if I'll replace the resistor. "No big deal my friend", I say.

So tonight I swap out the toasted resistor for a new guy, recheck all the solder joints on the bottom of the board, double check the wiring, etc., etc. Hook everthing back up, press power and blamo, another toasted R33G. Here's the kicker though, tonight I took the opamps out figuring that I don't really want to shell out another $48 or whatever for [2] 637's and 627 (I gave the opamps to Ian as a "get your PPA started gift"). I thought that this meant that the output resistors shouldn't be seeing anything.

So, anyone have any idea what the heck is going on here? I've got some spare parts so if something thinks it'd be a good idea to swap out this or that I'm all ears. I can't even check the voltage at the sockets cause the resistor will go Mt. St. Helens on me. I have checked the STEPS and it's pumping out a healthy 34.7V with no apparent issues.

HELP, please, I'm trying to get this thing up and running for the Boston folks this weekend so they can hear the glory that is the PPA.

Thanks in advance,

Nate
 
Jul 8, 2005 at 12:20 AM Post #2 of 108
Please tell me that 34.7 volts out of the steps is a typo.
wink.gif
 
Jul 8, 2005 at 12:22 AM Post #3 of 108
Nope, that's what it's set at. The way I read the data sheet for the 637/627 they can handle that just fine.

Remember Marshall, I'm a moron, offer advice accordingly.

N
 
Jul 8, 2005 at 12:33 AM Post #4 of 108
I've had something similar wrecking mine for a week or two now -- I was fiddling with the bias on the left channel and shorted the output resistor to the driver transistor next to it. Smoke, and heat enough to scorch the output resistors and melt the C2 behind the output transistors. I'm not sure what exactly died but I've replaced the entire buffer on that channel and didn't get things down to normal temperatures until replacing the output transistors. Still isn't working right, but I think I'm at the point now where the only problems it still has are a result of the damage the board has seen from the soldering and heat.

So, I guess basically I'd say, check the output transistors, and good luck
smily_headphones1.gif
 
Jul 8, 2005 at 12:48 AM Post #5 of 108
Quote:

Originally Posted by n_maher
... Hook everthing back up, press power and blamo, another toasted R33G. Here's the kicker though, tonight I took the opamps out figuring that I don't really want to shell out another $48 or whatever for [2] 637's and 627 (I gave the opamps to Ian as a "get your PPA started gift"). I thought that this meant that the output resistors shouldn't be seeing anything.


I admit I'm not yet familiar with the PPA, but just a quick comment.

Looks like R33s are part of an output buffer circuit. Each one of these circuits has its own power hook-up in addition to signal in and signal out. This would suggest to me that juice has the potential to flow here, without an op-amp.

If your R33g is consistently suicidal but the R and L ones are fine each time, perhaps there is a short in the vicinity of the ground output buffer?

Best of luck!
 
Jul 8, 2005 at 2:13 AM Post #6 of 108
Quote:

Hook everthing back up, press power and blamo, another toasted R33G.


I've seen this before, and it was caused by the same thing: shorting something in the ground channel's buffer with a slip of a probe. You will also find that one of the ground channel's output transistors is also getting hot, and that replacing it doesn't help.

I fought with it for two hours, but it was an amp someone was paying me to build, so I figured the cost of another board and all the small parts was the cheapest way out. In rebuilding the new amp board, I saved the trim pots, the output transistors for the other two channels, all the caps, and the op-amps. All the resistors, and all the small transistors were certainly written off.

This suggests to me that the problem is in one or more of the small transistors. If you're willing to replace all of them in the ground channel, that may work. I don't think replacing just one will work...I recall spending a big chunk of those two hours tracing paths on the schematic to find ones that could pull enough current across R33 to smoke it so quickly, and never was able to solve the problem with a single transistor replacement.

The problem could also have been with the PCB...a short or a break that I didn't see, though I did examine it closely for this.

Quote:

I'm a moron, offer advice accordingly.


Nate, you're a moron.
evil_smiley.gif


What the absolute maximum voltage rating means is, it won't blow up if you run it at that voltage. Maybe it will work fine. But if it doesn't, TI reserves the right to blow you a loud raspberry if you complain to them. Absolutely none of the specs in the data table are guaranteed when running it beyond the recommended maximum voltage.

What do you need all that headroom for, anyway? That's enough to put probably 9V or 10V rms into headphones.
 
Jul 8, 2005 at 2:36 AM Post #7 of 108
All this news isn't unexpected, but it does make me sad...
frown.gif

Well Nate, it appears you're off the hook, and it was my meddling when attempting to adjust the bias last night that did it in. D'oh! I guess it's time to place an order for some parts.

And tangent, Nate tells me the STEPS has been ramped down in the just under 30V range, so when things get squared away, it will be good to go.

I'm always amazed at the knowledge out there. Thanks for all the input everyone... Even if I don't like what I hear!!!
tongue.gif
 
Jul 8, 2005 at 2:40 AM Post #8 of 108
Tangent,

Thanks for the advice, well except that part about me being a moron
tongue.gif
.

In response, well, the 34+ volt power supply was just cause we could and it made the steps run nice and cool. it is still running a full 2V below (at the sockets) the max rating of the 637s which I thought provided a decent cushion. Either way I don't think this played a roll in the issues that this amp is having, but I appreciate the advice and will add it to my knowledge base.

Looks like I'll cook up a Mouser order along with contacting you tomorrow, geez, didn't I just get a package from you today
biggrin.gif
? Oh well, my wallet has no voice left with which to shriek. I guess Ian's getting an early Xmas present now too
600smile.gif
. I haven't given up on this board yet but I agree that it could be a lot of effort with no results trying to track down exactly what I ruined.

[edit] Ian, we can fight about who did what and who orders what tomorrow
smily_headphones1.gif
.

N
 
Jul 8, 2005 at 3:03 AM Post #9 of 108
Quote:

Originally Posted by n_maher
it is still running a full 2V below (at the sockets) the max rating of the 637s which I thought provided a decent cushion.


The 6 V between the recommended and the absolute max is the cushion. You're depleting it.

Quote:

I don't think this played a roll in the issues that this amp is having


Probably not.

Now, you could probably get away with 32V or so, depending on the Idss of your Q3s. Between the D1 drop and the drops across the Q3s, you can easily drop that to 30V at the op-amps.

Quote:

I haven't given up on this board yet but I agree that it could be a lot of effort with no results trying to track down exactly what I ruined.


I'm glad. I didn't tell you this to get a parts order. I hoped, for the sake of science, that you'd try that transistor replacement first. My experience happened at the end of a tight schedule that had already slipped at least once by that point. I didn't feel at liberty to take any more time diagnosing it.
 
Jul 8, 2005 at 3:08 AM Post #10 of 108
Quote:

Originally Posted by tangent
I didn't tell you this to get a parts order.


I hope that I didn't in any way imply that you did, the thought never crossed my mind. Truth be told I've been wanting to build a PPA for a while now, this is just the straw that will get me to order the board. The board requiring the repair will be mine to tinker with at my leasure, to try and learn what the heck went wrong and how to fix it.

Thanks again for taking the time to help.

Nate
 
Jul 8, 2005 at 3:13 AM Post #11 of 108
It seems to me that a safer way to set the bias would be to measure from the outside of one of the output resistors to some point on the output wire itself, instead of the output on the board. That would keep the probe well away from any transistor leads. It'd be pretty simple to just clip it onto the headphone jack once that's wired up, actually.
 
Jul 8, 2005 at 11:01 AM Post #12 of 108
Reel that power down to 30 or 32 volts. I think personally that the difference at the cans between 24 in and 30 in is pretty much indiscernible on most occasions. I often set the PSU so that I get 24v at the sockets, which may mean about 25.5 or so at the PSU.
Even if you wanted to run the op-amps at the reference +/-15V, perhaps 31.5 would be good. You are surely on your own after 30v rail-to-rail.
 
Jul 8, 2005 at 12:05 PM Post #13 of 108
I say reel it down even more so the opamps only see about 26.5v as a max.. that way you can roll ad8610s if you want..

but that's just me (course it depends on what you're driving with the amp.. if I were driving a set of k1000s I'd wana dial 30-32v into the amp..)
 
Jul 8, 2005 at 1:34 PM Post #14 of 108
Quote:

Originally Posted by Jazper
I say reel it down even more so the opamps only see about 26.5v as a max.. that way you can roll ad8610s if you want..


Not a bad idea, but it was decided long ago that this amp will never be running 8610s, way too bright for the owners taste. And reeling it down means swapping resistors, which is a HUGE pain after you've got the heatsink installed.

So 30V it is, with the STEPS actually putting out 31ish to keep regulator temps as low as possible and have the rails as close to 15V.

N
 
Jul 8, 2005 at 2:19 PM Post #15 of 108
Quote:

Originally Posted by n_maher
Not a bad idea, but it was decided long ago that this amp will never be running 8610s, way too bright for the owners taste. And reeling it down means swapping resistors, which is a HUGE pain after you've got the heatsink installed.

So 30V it is, with the STEPS actually putting out 31ish to keep regulator temps as low as possible and have the rails as close to 15V.

N



just remember 8065s also have low voltage maximums.

There are other opamps too. I agree with the 8610 being too bright ..
 

Users who are viewing this thread

Back
Top