• 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

Oscillation Explosion!

K

KeithTr

AK Subscriber
Subscriber
I thought I had fixed a Technics SU-V5 with a blown channel, had it running for a few days and apparently it went into oscillation (I'm thinking high frequency) and blew up the RC network (10ohm, .047uf) components (a combined package) on both channels. Schematic of that section attached. Most of what I replaced was pretty standard, but I did use film caps 1uf replacement and did substitute a few transistors (outputs were same as originals that blew previously). Also had used a newer opamp (of same class) for feedback circuit, which I'm thinking might have been the cause.

Since I've not dealt with oscillation before, wondering if anyone has a suggestion or two about how to diagnose the culprit(s) and a possible solution. Thanks in advance..K
 

Attachments

  • Technics V5 amplifier section of Schematic_2.jpg
    Technics V5 amplifier section of Schematic_2.jpg
    106.2 KB · Views: 36
  • Technics V5 amplifier section of Schematic_1.jpg
    Technics V5 amplifier section of Schematic_1.jpg
    101.8 KB · Views: 35
Register to hide this ad
Exactly, what did you replace and with what? If you used a higher speed opamp, you need local decoupling to prevent oscillations.
Or put back the original opamp, but local decoupling never hurts. 100nF x7r mlcc across the power pins.
 
redk9258 said:
Maybe list what was original and what you put in it's place.
Click to expand...
most of what i replaced was non-controversial - electrolytics of same value (a few a bit higher voltage due to what was on hand), a couple of low value (1uf) caps with film caps, transistors (mostly outputs or drivers that were replaced with NOS - not modern replacements), and a couple of signal transistors with modern replacements. And I did replace the opamp which I believe is in the protect/dc regulation circuit. That could have been the issue though it was a modern version of the older chip. Will try to get to today or tomorrow and report back.
 
rcs16 said:
Exactly, what did you replace and with what? If you used a higher speed opamp, you need local decoupling to prevent oscillations.
Or put back the original opamp, but local decoupling never hurts. 100nF x7r mlcc across the power pins.
Click to expand...

Thanks for the suggestion - when you say 100 nanfarads x 7(ohms? ceramic?) across the power pins (of the opamp?). Always something new to learn - now i have to go see what local decoupling really means. :thumbsup:
 
100nF = 0.1uF, x7r is the temperature coefficient. Local decoupling mean placing capacitors right at or as close as you can get, to the power pins Vcc(+), Vee(-) or ground.
"mostly outputs or drivers that were replaced with NOS" so how do you know they are original? so many fakes being sold on ebay. You are far better off buying from a authorized distributor.
 
KeithTr said:
most of what i replaced was non-controversial - electrolytics of same value (a few a bit higher voltage due to what was on hand), a couple of low value (1uf) caps with film caps, transistors (mostly outputs or drivers that were replaced with NOS - not modern replacements), and a couple of signal transistors with modern replacements. And I did replace the opamp which I believe is in the protect/dc regulation circuit. That could have been the issue though it was a modern version of the older chip. Will try to get to today or tomorrow and report back.
Click to expand...
We need to know exactly what you replaced - controversial or not. Otherwise it is not possible to help. Please take the schematics of the section you worked on and mark replacement parts used.
 
rcs16 said:
100nF = 0.1uF, x7r is the temperature coefficient. Local decoupling mean placing capacitors right at or as close as you can get, to the power pins Vcc(+), Vee(-) or ground.
"mostly outputs or drivers that were replaced with NOS" so how do you know they are original? so many fakes being sold on ebay. You are far better off buying from a authorized distributor.
Click to expand...
The drivers and outputs were taken from a parts unit with a cracked circuit board, and tested, so they were original. all other parts were mouser/digikey.
 
p4spooky said:
We need to know exactly what you replaced - controversial or not. Otherwise it is not possible to help. Please take the schematics of the section you worked on and mark replacement parts used.
Click to expand...

give me a day or so, and I'll lay it out. suspecting opamp (which was a modern replacement) or capacitors that were much lower esr than originals (e.g. films).
 
The opamp is the DC servo, they never oscillate. An amp's stability can be tested by applying a square wave with a really steep rise time. Think nanoseconds. Adjust for a 10 Vpp or so output. Use dummyload. Any instability will show up as ringing or actual oscillation. Sometimes the compensation/Miller-Cs have failed (black flags!) or have dry joints. Suspect Cs are C301, 303, 363, 309 and 311. So almost all of them. Often there are caps across the output transistors to GND, I don't see those here.

What I do see is the tone control circuit across the power amp. Hate when they do that. Bad engineering. This could be the source of oscillation too.
 
You have to be careful with power amps. There are things designed into them to avoid exactly what happened to you - ultra-sonic oscillation. The amp is just going to go ahead and keep amplifying those frequencies, too, and eventually be destroyed from the heat.

You can't just replace parts with substitutes without consideration. Also, circuit topography must be maintained - wire routing, grounding, and such.

Heat must be monitored after a repair to verify no unwanted oscillation.

Doug
 
Doug G. said:
You have to be careful with power amps. There are things designed into them to avoid exactly what happened to you - ultra-sonic oscillation. The amp is just going to go ahead and keep amplifying those frequencies, too, and eventually be destroyed from the heat.

You can't just replace parts with substitutes without consideration. Also, circuit topography must be maintained - wire routing, grounding, and such.

Heat must be monitored after a repair to verify no unwanted oscillation.

Doug
Click to expand...

I've repaired or restored a dozen or so units and this is the first time I experienced such an event. As I've said many times, I consider myself an amateur in this field and I'm just trying to relearn some of what I used to know a long time ago. Like Wall Street (where I worked for 40 years), tuition always has a price. Fortunately, this one is far far less than my securities tuition! I will try to review exactly what I replaced and see if possibly adding some resistance or checking the miller capacitors fixes the issue (just read about those). Hopefully nothing else was destroyed other than the rc output devices.
 
PE9ZZ said:
The opamp is the DC servo, they never oscillate. An amp's stability can be tested by applying a square wave with a really steep rise time. Think nanoseconds. Adjust for a 10 Vpp or so output. Use dummyload. Any instability will show up as ringing or actual oscillation. Sometimes the compensation/Miller-Cs have failed (black flags!) or have dry joints. Suspect Cs are C301, 303, 363, 309 and 311. So almost all of them. Often there are caps across the output transistors to GND, I don't see those here.

What I do see is the tone control circuit across the power amp. Hate when they do that. Bad engineering. This could be the source of oscillation too.
Click to expand...

UPDATE:

It appears that the substitution of a few modern transistors (not controversial choices - generally signal transistors in the amplifier section that had previously failed (like KSA992 etc) were the culprits as they are much much faster than the original components. After replacing the Zobel components - which appeared to be the only components that smoked - I was still getting oscillation. AI (and a friend here) suggested that I might need to add small picofarad level ceramic capacitors across the base and collector of the transistors that had "Miller" capacitors in their path. The idea is that these capacitors filter out very high frequencies of the type that typically cause oscillation (think 20-30mHz). Adding small value (around 10 or 20 picofarad) caps in this way seems to have eliminated the oscillation. It is interesting to google the "Miller effect" in amplifiers. AI also thinks that the old Zobel components (a resistor in series with a capacitor package) could be failing, and is suggesting I replace them with my own construction of the appropriate resistor and capacitor in series (10 ohm and .047uf) and I will try this next to see if I can extend the life of the unit. But as of this afternoon, it ran for about two hours at 10 watts with zero oscillation. crisp sine and square waves, and was able to make almost 70 watts before clipping (rated at 60).
 
KeithTr said:
UPDATE:

It appears that the substitution of a few modern transistors (not controversial choices - generally signal transistors in the amplifier section that had previously failed (like KSA992 etc) were the culprits as they are much much faster than the original components. After replacing the Zobel components - which appeared to be the only components that smoked - I was still getting oscillation. AI (and a friend here) suggested that I might need to add small picofarad level ceramic capacitors across the base and collector of the transistors that had "Miller" capacitors in their path. The idea is that these capacitors filter out very high frequencies of the type that typically cause oscillation (think 20-30mHz). Adding small value (around 10 or 20 picofarad) caps in this way seems to have eliminated the oscillation. It is interesting to google the "Miller effect" in amplifiers. AI also thinks that the old Zobel components (a resistor in series with a capacitor package) could be failing, and is suggesting I replace them with my own construction of the appropriate resistor and capacitor in series (10 ohm and .047uf) and I will try this next to see if I can extend the life of the unit. But as of this afternoon, it ran for about two hours at 10 watts with zero oscillation. crisp sine and square waves, and was able to make almost 70 watts before clipping (rated at 60).
Click to expand...

One more update. It seems as though everything is functioning properly, but I am concerned that R601 and R602, which appear to be part of the regulated power supply, run very hot - well over 80C. These resistors apparently drop the rail voltage from about 50v down to something closer to 29-33 (varies a bit by channel) and are specced only as 1/2 watt carbon. That's a pretty big voltage drop and I'm thinking that upgrading these to 1 watt metal film is probably a good idea. Any reason not to do this?
 
may want to verify the voltage drop across them to make sure something isn't pulling more than it should, but if its within spec, a larger resistor spaced off the board for cooling isn't a bad idea. The larger body will dissipate heat better and getting it off the board saves it and gives better airflow at the same time.

voltage drop squared divided by resistor value = power dissipation in watts. Sane de-rating is a resistor at least double that value but larger isn't a bad thing.
 
KeithTr said:
UPDATE:

It appears that the substitution of a few modern transistors (not controversial choices - generally signal transistors in the amplifier section that had previously failed (like KSA992 etc) were the culprits as they are much much faster than the original components. After replacing the Zobel components - which appeared to be the only components that smoked - I was still getting oscillation. AI (and a friend here) suggested that I might need to add small picofarad level ceramic capacitors across the base and collector of the transistors that had "Miller" capacitors in their path. The idea is that these capacitors filter out very high frequencies of the type that typically cause oscillation (think 20-30mHz). Adding small value (around 10 or 20 picofarad) caps in this way seems to have eliminated the oscillation. It is interesting to google the "Miller effect" in amplifiers. AI also thinks that the old Zobel components (a resistor in series with a capacitor package) could be failing, and is suggesting I replace them with my own construction of the appropriate resistor and capacitor in series (10 ohm and .047uf) and I will try this next to see if I can extend the life of the unit. But as of this afternoon, it ran for about two hours at 10 watts with zero oscillation. crisp sine and square waves, and was able to make almost 70 watts before clipping (rated at 60).
Click to expand...

UPDATE #2

I really thought I was done here. Made two final changes, upped 2 resistors to 1 watt from 1/2 watt as they were fairly hot and close to wattage limits and redid the Zobel combination of a 10 ohm resistor and .047uf capacitor right before the output. Neither of those should have affected anything and I redid the bias, the impedence adjustment, and checked offset and all looked good. Was doing some normal testing of sine and square waves at different frequencies - all looked good - no ringing - no fuzzy traces, no visible distortion until I was playing with the volume level at low frequency (about 100 Hz) and all of a sudden got some instability. I quickly shut down and tried again. Good sine wave, no issues, squares looked good, and again an instability showed up only after several minutes and appeared on the scope as what I think was high frequency noise. Mind you, this was after I ran the amp playing music with no issues for over an hour this afternoon. I'm going to check if I missed any transistors that were changed. But all instability was on the channel that previously blew up and I can assure you that I had tested basically every component (except perhaps all ceramic caps) in that channel and nothing looked suspect after the blown transistors and electrolytics were replaced. So it looks like now I've got an intermittent issue to hunt down. I should add that when I played music, the amp ran at very low temperature - no heat issues at all.
 
You must log in or register to reply here.
Back
Top Bottom