Lifting the feedback return is forcing the amplifier to run at unity gain (gain of 1) and in that situation its pretty much guaranteed that any amp of this kind of topology would oscillate.
Increasing the combined value of C11/13 might also provoke oscillation.
Increasing the combined value of C11/13 might also provoke oscillation.
Hi everyone. Sorry I missed these last couple of posts. Things have been going fine for past month. Even had a pair of Pioneer HPM-100s cranked up half way Friday night. All good. Then last night power up and within 10 seconds I hear relay click and left channel cut out. Same as before. So I'll have a look at the ground you mentioned. Start taking measurements and report back. It may be a couple of days before I can get to it though.
A lot happens in a week on here 😉
It really does sound like there is some physical issue somewhere. I think you are going to have to start prodding, poking and bending to try and provoke it.
It really does sound like there is some physical issue somewhere. I think you are going to have to start prodding, poking and bending to try and provoke it.
Well it's provoked now. Still up over 800 mV offset on left speaker outs. Also smell of overheating in the area of C39/C41 and R113/R115. Haven't popped the cover yet.
Is that the zobel network, the C and R in series at the output ? If so then its 100% certain that it is oscillation that is causing the resistor to fry. The cause can still be a physical problem though.
Finally have some time to look. And naturally, it is working fine now. 7 mV on left, -4 mV on right speaker offset. I've jiggled some of the leads to the Voltage Amp Assy and did not see a change in left output w meter attached.
So the zoebel network. R113 and R115. Should I see exact same voltage at all points for those resistors? On both left and right, I see same voltage at these resistors as I do at speaker outs. 7 mV left, -4 mV right. Note, both channels fluctuate +- .1 to .4 mV. Neither is steady.
No DC current flows through those resistors as they are part of a series R/C network which means each end of a given resistor will have the same voltage as the other end. The actual voltage is the same as the speaker output because that is where they connect.
The fluctuation is normal and will be caused by a couple of things such as slight noise/hash which 'confuses' many meters. The DC offset will also vary very slightly (and quickly as opposed to slow thermal drift) as the mains varies which in turn causes the rails vary.
The fluctuation is normal and will be caused by a couple of things such as slight noise/hash which 'confuses' many meters. The DC offset will also vary very slightly (and quickly as opposed to slow thermal drift) as the mains varies which in turn causes the rails vary.
So next time it acts up, where should I look first? Mooly, you seem to think its something physical. Do I jiggle connections and wires to see if that corrects things? The fact that it has run fine for weeks, could it be a failing component still?
Well based on the evidence, such as you mentioning earlier that moving the amp caused a problem etc. You really need to catch it in the act.
Lift it up and slam it into the table and see if something happens. You've been dying to do this all along anyway, right? Percussive maintenance is a thing, but here it can be mostly seen as troubleshooting.
🙂 Dare is say that these are the kind of faults that separate the really first class techs from the also rans 😀
You need to get creative in recreating the issue. Solder leads to the amp output and connect those to a scope (there can be no room for the slightest intermitent flaky connection to test gear). Do the same to any leads to the DVM. Gently poke prod and tap around the board, and do this on each component if need be. Also waggle leads one at a time. If you think its heat related then use freezer one drip at a time on suspect board areas/components all the time watching that scope trace. You need football eyes 😀 home and away. One eye on what you are doing, the other on the scope. Did it blink, let alone flicker. Don't give up, be like a dog with a bone... its personal now between you and the amp. And its not going to get the better of you. If all that is inconclusive (and it happens) then you need to go over the fault mentally again. Can you rule out certain areas ? If it still played up with the volume on minimum then its a fair bet the preamp can be overlooked at this moment. Go over every single joint with the iron, does the solder take to the leads or have any oxidised badly ? Are all the leads actually pushed through the board and making it through to the other side. Give any plug/sockets a thorough going over. Crimped connections can be intermittent, the plug/socket interface poor due to contacts losing spring tension.
Don't give up though. Once this point is reached it really does become a personal challenge.
You need to get creative in recreating the issue. Solder leads to the amp output and connect those to a scope (there can be no room for the slightest intermitent flaky connection to test gear). Do the same to any leads to the DVM. Gently poke prod and tap around the board, and do this on each component if need be. Also waggle leads one at a time. If you think its heat related then use freezer one drip at a time on suspect board areas/components all the time watching that scope trace. You need football eyes 😀 home and away. One eye on what you are doing, the other on the scope. Did it blink, let alone flicker. Don't give up, be like a dog with a bone... its personal now between you and the amp. And its not going to get the better of you. If all that is inconclusive (and it happens) then you need to go over the fault mentally again. Can you rule out certain areas ? If it still played up with the volume on minimum then its a fair bet the preamp can be overlooked at this moment. Go over every single joint with the iron, does the solder take to the leads or have any oxidised badly ? Are all the leads actually pushed through the board and making it through to the other side. Give any plug/sockets a thorough going over. Crimped connections can be intermittent, the plug/socket interface poor due to contacts losing spring tension.
Don't give up though. Once this point is reached it really does become a personal challenge.
Well, I haven't tried slamming it on my bench, but I did finally look over things some more last night. With the amp on and meter leads connected to left channel speaker out I was seeing 7.4mV. No issues. I proceeded to tap around on all boards and where various leads connect to each board, trying to get a response using an insulated probe. Tried tapping just about every solder point (except on the main power supply board and ps caps), protection relays, RCA inputs, everything. No response.
I don't have a scope, still looking to borrow one from a friend who has one tucked away somewhere in his garage. Using the freeze spray is a bit tricky with this amp as all of the boards have bottom sides facing outward, components are on internal sides that are not easily accessible.
I don't have a scope, still looking to borrow one from a friend who has one tucked away somewhere in his garage. Using the freeze spray is a bit tricky with this amp as all of the boards have bottom sides facing outward, components are on internal sides that are not easily accessible.
That's not necessarily a disadvantage as long as you know the component layout. A few drips of freezer chilling the joints will conduct almost instantly into the device. So worth trying still.
The tricky part there is that the leads for transistors (and the 2 ICs) are so close together, I'd likely cause a short across leads and I understand that is a big no-no.
- Status
- Not open for further replies.