• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

SA-8800 Oscillation?

Some of the guys will have a fan blowing across the heat sink when setting the adjustments also you should have it sitting like it would on all four feet. If you don’t already.
 
also try with top cover on sitting upside down worked for me . it was picking up interference big style with no cover on .
 
basically getting to point of the limit of my knowledge or understand of this design, unless I spend a whole lot more time on it.
Would be nice if I had one to play with myself.
No two implementations are exactly the same, there are a number of variables involved.
Spice represents a lot of perfect cases, like I have not included factors such as wiring parasitics, noise, coupling, real world issues to address.
Whatever you do, I'd match the channels as far as comps go. If we need to over compensation one channel to stablize it, then do the same to the other channel even if it looks as if it does not need it.

Good point Pete, EMI ingress is another factor.


Some of the guys will have a fan blowing across the heat sink when setting the adjustments
That's a new one for me. What is to be gained by doing this method?
 
Last edited:
i must say so far as i remember the bias setting was stable cover off . cant remember board numbers but mine likes to burn up one of the resistors that goes across the coil at the output . i never checked it for oscillation .it always sounds good to me . must pull it out again one day to use for a while . stopped using it as power hungry when idling .
 
Well, channel two is directly underneath the EQ board so maybe that contributes? All the times up until now when adjusting the idle current, it was on its side. When adjusting it now, its just sitting on the bench right way up.

zebulon1 said:
Some of the guys will have a fan blowing across the heat sink when setting the adjustments also you should have it sitting like it would on all four feet. If you don’t already.
Click to expand...

I've read that too but I don't remember why. From what this amplifier is doing so far, if the driver transistors are cooled, the idle current drops off rapidly so I imagine if I put a fan over it, it'll be even harder to set it? I think I do remember MTF saying that it was very sensitive to heat.
 
petehall347 said:
but mine likes to burn up one of the resistors that goes across the coil at the output
Click to expand...
The only thing that i can think of to cause this is high frequency oscillations. At low freq's the R is shorted out by the much lower coil R, or the coil is open which is another hard to believe one.
 
i always thought it was some oscillation and caused by myself before it started doing it . was testing a home made frequency generator . wish i could remember what setting it was on . could well have been triangle wave . or square wave . will get to the bottom of it one day with the help of this thread .
 
zebulon1 said:
Some of the guys will have a fan blowing across the heat sink when setting the adjustments also you should have it sitting like it would on all four feet. If you don’t already.
Click to expand...
I noticed when I adjusted the NSA on a sx-3700 I had the set on its side. The heat from the lower sink was engulfing the top heat sink causing a little frustration. They are sensitive.
 
It adjusted ok on its side with the heat rising through channel two as its all one heatsink. Now when the heat can rise off the heatsink directly it wont set.... I might just try the fan idea and see what happens...
 
Cooling it only made it even hard to set the idle current. Neither channel seems very keen on going over 63mV with or without a fan. I'm wondering if the drivers are getting hot and the thermistor nearby is causing the circuit to shutdown thermally?

I guess the simulation could be set to 70mV? Wondering where the difference lies...
 
Wondering if I'm following here, did this amp work fine, then you recapped it
and the trouble started?
One channel failed first, when did the second channel fail?

I have to wonder why new caps would cause a failure, was one in backwards,
did they have significantly different ESR, ESL?
 
QSilver said:
I'm wondering if the drivers are getting hot and the thermistor nearby is causing the circuit to shutdown thermally?
Click to expand...
That is a possibility. Shield or remove the thermistor. We are assuming it is 10K, we do not know what its characteristics are. slopes NTC, PTC
As an another experiment you can try and reduce the driver bias current back to what it was originally to see how it behaves. That way the drivers are running much cooler.
I stated earlier what the VR9 pot value was to get 70mV of bias.
see if you can figure out ltspice and you can play with this circuit in sim and in real life.

I think just leaving it at 63mV of bias is good enough, unless you want to keep fiddling with this design. I think we accomplished our goals to tame the oscillation and get it bias up to a good enough state. As it is, people report this design runs too hot as it is.

PeteB, the thread started by saying,
QSilver said:
Went to switch it on today and it took longer than usual to come out of protection. Then I heard a slight crackling noise from inside the amp. Turned it off, and smoke came out from the front right corner of the amp. Damn.
Click to expand...
Then he tried to use sub bjts and the design would not bias & oscillated.
 
I might look at dropping the current down and just seeing if the idle current starts to behave differently. My main concern is getting it to run with the same THD figures and to make sure all the sections of the amp will function as intended. I don't completely understand the purpose of the second adjustment yet so I think I'm just defaulting to trying to get the original results back and if the sim biases correctly and the real one doesn't It'd be nice to know where the difference lies.

In reply to PeteB, I think the whole cause of this was down to very tired cooked driver transistors. Post #48 shows some pictures of the old drivers and one in particular is going quite rainbow on it's tab. When the amp took out one of it's outputs, this driver tested ok but on shutdown of the amp, started to leak, breakdown and oscillate. I imagine the rest weren't far behind as they all looked very toasty....

Another thing I think I've mentioned in another thread somewhere to rcs16, is that this amp has no bleed resistors across the main filter caps. Mine stay charged to 45V after weeks of no use at all. I even think the SA-9800 has bleed resistors on it but the 8800 doesn't.... The only filter cap that does discharge is the one that also feeds the regulators IIRC.
 
Info I found on Th1:

Neg Temp Co Thermistor TH Series
SPECIFICATIONS 　 　 　 　 　
Part Number: TH103-2
Zero Power Resistance at 25℃(Ω): 10K
B-constant (25/85): 3435
Rated power at 25℃(mW): 2
Thermal Dissipation Constant (mW/℃): 3
Thermal Time Constant (sec): 10~15
Operating Temperature: -50~150 C
 
But did you recap it due to a problem or did the problems start after recapping.

The drivers may have run hot for a long time but it is possible that they were fine.

As I read your description when smoke came out one channel failed, how did
the second channel fail?
 
QSilver said:
My main concern is getting it to run with the same THD figures
Click to expand...
Are you able to make this measurement?
QSilver said:
. I don't completely understand the purpose of the second adjustment yet so I think I'm just defaulting to trying to get the original results back and if the sim biases correctly and the real one doesn't It'd be nice to know where the difference lies.
Click to expand...
me too but my time is limited on this one.

I would suggest to add a bleeder R and a 1u/100V poly film across the big ecaps.
 
Once you believe that it is fixed, you should drive it to full power with a dummy load,
and even into clipping. Look to see if there are any oscillations as you increase the
power. I would not go above 20KHz with full power testing since cross conduction
can blow the outputs.
 
You must log in or register to reply here.
Back
Top Bottom