The input pair matching I would doubt would cause such a symptom. I'd think more of a VAS issue , the limiter triggering (OPS SOA or VAS protection) or possibly local instability in the OPS. You mentioned it is clean without a load and shows up with a load.
Sorry to have dropped in on this, I'm no expert on this particular design. There is likely a significant difference in some circuit value(s) from one board to the other or one of the semiconductors is flakey.
Sorry to have dropped in on this, I'm no expert on this particular design. There is likely a significant difference in some circuit value(s) from one board to the other or one of the semiconductors is flakey.
Hi,Terry, that's not bad out in to 8 ohms and at 4 ohms is doing 552.25...wow no wonder things are getting hot, try keeping the i/p to 1.4 volts and see how things pan out.
Clip wave would be when top and bottom sine wave starts to flatten off giving even clip so just before you see this happen that's the max power before clip.. Try to isolate your scope connections form the amp's output during testing via connection two capacitors to your speaker dummy load a value of 220nf at 250v or more to each end of the load. hook your test lead as normal to the amp and your scope leads to the caps free heads,or you could use isolation transformers for the job.
you would of noticed that resistor was getting hot during the first test, it's better to try keeping it cool via fan cooling. Most of us who carry out daliy repairs on amps use fan cooled loads plus have them in metal containers with binding post or jack connections for ease of quick set up. I did notice how some amps would whine or oscillate when hooked to my scope so by isolating the connections solved that fuzzy whine sound plus I could see the wave going crazy on the scope.
My old test load was 400w per side and pushed may times beyond that till it an major up graded to 2kw per side at 8 ohms plus power meters.. more than plenty.
Regards A.
Clip wave would be when top and bottom sine wave starts to flatten off giving even clip so just before you see this happen that's the max power before clip.. Try to isolate your scope connections form the amp's output during testing via connection two capacitors to your speaker dummy load a value of 220nf at 250v or more to each end of the load. hook your test lead as normal to the amp and your scope leads to the caps free heads,or you could use isolation transformers for the job.
you would of noticed that resistor was getting hot during the first test, it's better to try keeping it cool via fan cooling. Most of us who carry out daliy repairs on amps use fan cooled loads plus have them in metal containers with binding post or jack connections for ease of quick set up. I did notice how some amps would whine or oscillate when hooked to my scope so by isolating the connections solved that fuzzy whine sound plus I could see the wave going crazy on the scope.
My old test load was 400w per side and pushed may times beyond that till it an major up graded to 2kw per side at 8 ohms plus power meters.. more than plenty.
Regards A.
Has Jkuetemann just said Vas, voltage amplification stage built around Q8-12-22-9-13-23. next the v-limiter is Q10-Q11-D5-6-7-8-9-10 diodes and few resistors, you can see this part within the feedback part of the circuit, its basic job is it to limit the voltage drive to the o/ps stage if a short circuit happens and hopefully save them from blowing up.
Soa means safe operating area of the out put transistors this can be checked in the data sheet.
Regards A.
Soa means safe operating area of the out put transistors this can be checked in the data sheet.
Regards A.
Hi Anthony,
If you remember, I just replaced all of the transistors on the main board. Not saying one could not be bad still but I don't know how to look for it. I poked around with the scope probe and the oscillation was on both sides of the circuit.
The flat tips of the sine wave is what I looked for and that is how I came up with the 47V limit. 47.2V is where they began to flatten out. It happens much higher without any load on it.
This resistor is what I bought for a load. I hung it in a bucket of water to keep it cool.
The oscillation starts at about .8V on the input. It does it with a speaker attached as well. Sounds like a higher harmonic kicking in. It does it at pretty much all frequencies at about the same input voltage.
Are you saying to connect the 220nf caps inline with my scope test leads on the output?
I seem to remember getting a funny reading on D7 or D8. I will check that again when I get home today.
I'm going to replace Q1-Q4 with matched sets and then scrub the trace side of the board again. I'll report back once I've done that. If I can't get rid of it I will make a short video of what is happening and see if that helps.
Blessings, Terry
If you remember, I just replaced all of the transistors on the main board. Not saying one could not be bad still but I don't know how to look for it. I poked around with the scope probe and the oscillation was on both sides of the circuit.
The flat tips of the sine wave is what I looked for and that is how I came up with the 47V limit. 47.2V is where they began to flatten out. It happens much higher without any load on it.
This resistor is what I bought for a load. I hung it in a bucket of water to keep it cool.
The oscillation starts at about .8V on the input. It does it with a speaker attached as well. Sounds like a higher harmonic kicking in. It does it at pretty much all frequencies at about the same input voltage.
Are you saying to connect the 220nf caps inline with my scope test leads on the output?
I seem to remember getting a funny reading on D7 or D8. I will check that again when I get home today.
I'm going to replace Q1-Q4 with matched sets and then scrub the trace side of the board again. I'll report back once I've done that. If I can't get rid of it I will make a short video of what is happening and see if that helps.
Blessings, Terry
Terry,
My apologies, these are pretty standard abbreviations in this arena. Allow me to take a minor diversion that might make interactions a little easier. Generally speaking the Leach amps are basically based on the ages old three stage Lin topology, named for Harry C. Lin who developed the original concept with RCA circa 1956. Hope I'm not too out of place here.
IPS=InPut Stage, generally recognized by the application of the signal we want to amplify to this section and also generally receives the signal from the feedback network. Often this is the two transistor differential amplifier known as a long-tailed pair, or LTP. It can also be a single transistor known simply as a singleton.
VAS=Voltage Amplifier Stage, this is where the signal really increases in amplitude, but at this point isn't capable of really delivering any current. It is also commonly where we find some sort of 'dominant pole' compensation for stability of the amplifier.
OPS=OutPut Stage, this is where the ability to deliver current comes from. It is basically a 'buffer' that outputs the same voltage as the VAS, but is capable of delivering the required current to the load. These are generally either and emitter follower, often noted as EF2 or EF3 to denote the number of steps in the OPS, or a complimentary feedback pair (Sziklai), the CFP.
SOA=Safe Operating Area, refers to the maximum allowable limits on the combination of voltage and current a semiconductor device may handle. This is often enforced with limiters to protect the amplifier from conditions that would be destructive, such as a short circuit or a highly reactive load.
Ok, enough of that. I hope I was even just a little helpful with understanding some of the abbreviation we see here quite a bit and I have no intent on being factious or insulting in any way. Back to the task at hand.
What you have shown looks very much like some kind of parasitic oscillation. These oscillations can often be at a very high frequency and amplitude, so much so that they can sometimes even just appear to fatten or un-focus the oscilloscope trace. The issue is that such oscillations often cause heating and excess current draw to the point of destruction, despite limiter circuits being in place. If one channel does this and the other not then there must be some difference between them. Ensure C10 and C11 are the specified values and of the correct type. These are the capacitors that create the required reduction in gain of the VAS to ensure stability. They can sometimes be increased in value if required, though don't go any bigger than needed for stable operation. This may be needed if device substitutions have been made.
C8 and C9 also affect stability by introducing phase shifts into the feedback network. Verify these too. Ensure the Zobel resistor, R50 hasn't been smoked and has opened or raised in value. The Zobel is often required for 'external stability' and if not there the OPS may oscillate.
Also, a note taken from Leach's web page:
04/17/99 - Added R61 through R64 and the connections between the collectors of Q18 and Q20 and the collectors of Q19 and Q21 to solve a potential oscillation problem between the two sides of the output stage. Added D13 through D16 to regulate the reference voltages that set the diff amp tail currents. Changed R13 and R14 to supply the additional currents to bias the zener diodes. Changed D5 and D6 to higher breakdown voltage diodes. Added R65 and R66. Added new circuit diagram and new circuit board layouts.
Can I assume your version includes these, amongst the other changes mentioned?
Have there been any part substitutions? Particularly the transistors.
Last edited:
This was very helpful...
Thanks for these definitions. It is often assumed that everyone understands amplifier theory and so a lot of nomenclature is taken for granted by many posters. Some of us just enjoying the challenge of building an amp - either from a PCB or assembled module - and leave it there. The learning often comes in bits and pieces when things don't quite work out the way we had planned and the trouble-shooting begins.
Speaking of...I really do need to pick up Doug Self's and/or Bob Cordell's books and start to educate myself, beyond Ohm's Law.
jkuetemann,
Thanks for these definitions. It is often assumed that everyone understands amplifier theory and so a lot of nomenclature is taken for granted by many posters. Some of us just enjoying the challenge of building an amp - either from a PCB or assembled module - and leave it there. The learning often comes in bits and pieces when things don't quite work out the way we had planned and the trouble-shooting begins.
Speaking of...I really do need to pick up Doug Self's and/or Bob Cordell's books and start to educate myself, beyond Ohm's Law.
Hi Jason,
Thanks for the detailed explanation. Couldn't wait to get home and check the boards against each other. Surprisingly, they were very close in measurements. The biggest discrepancy was at the base of Q11 where on the good board I had 2mv and on the troubled board I had 25mv. Unfortunately, while probing the good board I shorted the base and collector of Q10 and seem to have taken out Q12,14,22,&24. Isn't this fun.
Changed out Q2-4 and Q11 on troubled board. Scrubbed and cleaned the foil side. Oscillation is unchanged and the DC offset went for 40mv to 90mv.I also now have 45mv on the base of Q11. I'm getting farther away from where I need to be.
Something that I've noticed. On the schematic, It shows the base of Q11 going through R31, then through D6 to ground and Q10 going through R30, then through D5 to ground. However, on the board layout, the The resistors are between ground and the diodes. Not sure if that matters since they seem to work that way.
Well looks like I have some soldering to do.
Thanks for the detailed explanation. Couldn't wait to get home and check the boards against each other. Surprisingly, they were very close in measurements. The biggest discrepancy was at the base of Q11 where on the good board I had 2mv and on the troubled board I had 25mv. Unfortunately, while probing the good board I shorted the base and collector of Q10 and seem to have taken out Q12,14,22,&24. Isn't this fun.
Changed out Q2-4 and Q11 on troubled board. Scrubbed and cleaned the foil side. Oscillation is unchanged and the DC offset went for 40mv to 90mv.I also now have 45mv on the base of Q11. I'm getting farther away from where I need to be.
Something that I've noticed. On the schematic, It shows the base of Q11 going through R31, then through D6 to ground and Q10 going through R30, then through D5 to ground. However, on the board layout, the The resistors are between ground and the diodes. Not sure if that matters since they seem to work that way.
Well looks like I have some soldering to do.
OK, I have the good channel fixed. I only had to replace Q14 and Q11. I checked the two channels against each other. These are the measurements that were different between the two.
Voltage drop across R13 and R14; good channel 49v each, bad channel 48.5 each
Q22C good channel 1.68V, bad 1.79V
Q23C good channel -1.60V, bad -1.55V
Q16B good channel 1.16V, bad 1.69V
Q17B good channel -1.17V, bad -1.01V
Q16E good channel .632V, bad .702V
Q17E good channel -.561V, Bad -.510
Good channel .17mv offset, bad channel 39mv offset.
All other measurements between the two channels are virtually identical.
I made a short video of the oscillation I am experiencing.
You can see it here.
Voltage drop across R13 and R14; good channel 49v each, bad channel 48.5 each
Q22C good channel 1.68V, bad 1.79V
Q23C good channel -1.60V, bad -1.55V
Q16B good channel 1.16V, bad 1.69V
Q17B good channel -1.17V, bad -1.01V
Q16E good channel .632V, bad .702V
Q17E good channel -.561V, Bad -.510
Good channel .17mv offset, bad channel 39mv offset.
All other measurements between the two channels are virtually identical.
I made a short video of the oscillation I am experiencing.
You can see it here.
Last edited:
Well I'm no closer. Today I unsoldered some of wires connecting the outputs to the main board so I could twist them. I went through the foil side again and made certain there were no bridges of any sort, then cleaned it with alcohol and a tooth brush. I was really hoping I would see some kind of change but everything measures exactly the same as it did before. Still no oscillation with no load and with an 8ohm load it starts singing at about 18Vac out.
So do you guys have any ideas where to look. Looking at the spots where I see voltage differences, they are pretty much all tied into the DC offset I'm seeing. I even went though the top of the board as I was double checking the parts and made sure no parts were touching each other. Did you guys get a chance to look over the video I made? Did that raise any flags for you? I hate to give up on this thing again when I'm so close.
Thanks, Terry
So do you guys have any ideas where to look. Looking at the spots where I see voltage differences, they are pretty much all tied into the DC offset I'm seeing. I even went though the top of the board as I was double checking the parts and made sure no parts were touching each other. Did you guys get a chance to look over the video I made? Did that raise any flags for you? I hate to give up on this thing again when I'm so close.
Thanks, Terry
Last edited:
Sorry to adding post after post. They won't let you edit a post after a bit of time passes. I had a thought this morning. Since the oscillation only happens under load, is it possible it could be one of the output transistors? I can easily change them out since they are in sockets. Would the fact that it oscillates only on the bottom part of the wave give an indication of which rail is affected?
Thanks, Terry
Thanks, Terry
Hi Terry, did you add the caps to the dummy load? Here's one idea if, you can remove the working channel o/ps transistors and swap round to the faulty side and try it to see if you get the same effect.
If an amplifier goes dc we tend to check around the in put circuit then vas stage, devices under load can break down. This amplifier is turning in to a epic battle of trying to solve this nagging fault also is the zoble circuit linked into circuit as with out it things can act up.
If an amplifier goes dc we tend to check around the in put circuit then vas stage, devices under load can break down. This amplifier is turning in to a epic battle of trying to solve this nagging fault also is the zoble circuit linked into circuit as with out it things can act up.
Hi Anthony,
Thanks for checking in. It is becoming quite a battle. I actually swapped out all the O/P's today and no change. I wasn't too worried about the DC offset. It is only 37mv which won't hurt the speakers at all. I mainly mentioned it because it might relate to the voltage variations on Q16 & 17. If the Zoble is the 10R and .1uf across the speaker outlet then yes it is in place. I didn't add the caps to the dummy load yet because I don't have them. However, it does the same thing with a speaker attached and I don't have to have the scope attached to hear it. The good channel doesn't do this no matter how I hook up the scope. There must be a faulty device somewhere but I just can't seem to find it. Everything works fine until it gets to about 17Vac output.
I keep hoping this is going to pop up a red flag in someone's head and they will know what it could be. I don't want to just keep swapping parts willy nilly, my board is worse for the wear.
Blessings, Terry
Thanks for checking in. It is becoming quite a battle. I actually swapped out all the O/P's today and no change. I wasn't too worried about the DC offset. It is only 37mv which won't hurt the speakers at all. I mainly mentioned it because it might relate to the voltage variations on Q16 & 17. If the Zoble is the 10R and .1uf across the speaker outlet then yes it is in place. I didn't add the caps to the dummy load yet because I don't have them. However, it does the same thing with a speaker attached and I don't have to have the scope attached to hear it. The good channel doesn't do this no matter how I hook up the scope. There must be a faulty device somewhere but I just can't seem to find it. Everything works fine until it gets to about 17Vac output.
I keep hoping this is going to pop up a red flag in someone's head and they will know what it could be. I don't want to just keep swapping parts willy nilly, my board is worse for the wear.
Blessings, Terry
That's the thing when countless rework has been carried out to the point of a fresh pcb is made up and all parts are fitted over..17volts ac is less than 50w say 36w! replace some disc ceramic cap within the vas/driver stage as these can also break down.
I've got crown amp on the bench in work with fault indication and dc at it's o/p plus it's going to be a long trek till I get the schematic , so it's down to using the working side for ref pointers. It's also a switch mode power supply with switching fet's..
It can wait till the weekend is over time to reflect and just imply some clear thinking to re-cap over things as going over the same thing can cloud our thinking...
I've got crown amp on the bench in work with fault indication and dc at it's o/p plus it's going to be a long trek till I get the schematic , so it's down to using the working side for ref pointers. It's also a switch mode power supply with switching fet's..
It can wait till the weekend is over time to reflect and just imply some clear thinking to re-cap over things as going over the same thing can cloud our thinking...
I don't have anything other than what is called for in the BOM and schematic. What's weird is I can't find anything different between the two boards yet one works perfectly and one oscillates. The one that oscillates has stopped oscillating a couple of times. There are some slight voltage differences between the two as noted in post #192.
I'm not opposed to adding some things if they will work.
Thanks
I'm not opposed to adding some things if they will work.
Thanks
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.