or maybe not so strange to someone who has encountered the problem before.
Anyway I've built four Slone amps (to his deisgn in the high power amp book design 11.4). After a bit of trouble I've got all of them working and sounding brilliant. Each amp is fed a signal from an active xover with two amps driving bass drivers and the other two the mid and tweet combo.
Now when everything is switched on and the system is left alone, for a period of time the system will warm up to idling temp and stay there. However if im unlucky, sometime after 1st switch on (can be 10 mins or can never happen) there comes a noise which sounds like the brown noise created by cool edit. With this noise also comes a rapid heat up of all the output devices in all four amps, this is not thermal runaway as the amp can idle when this happens, without suffering any damage, but everything gets really hot literally too hot to touch.
If the amp is turned off and on again (this sometimes needs to be done more then once) the noise stops and everything cools down. Only five minutes later for the problem to re appear. When I turn the amp off it doesnt go quietly either, almost as soon as the power is cut there are two thumps produced in the speakers. This is absent when the amp has no problems. Normally after a period of time (power supply caps discharging) a small thump is produced in the speakers but this is gentle and very benign sound wise. When the amp is brown noising the thumps are loud and have an unpleasent electric snap to them.
When the problem occurs it does absolutely nothing to degrade the sound quality appart from raising the noise floor (created by the brown noise) which is not detectable most of the time (only when a quiet passage appears in the music is it noticeable). I guess that the distortion of the amp changes, but that too is not audible.
What also solves the problem is removing one of the inputs from the four amps (it doesnt matter which one) if plugged back in the problem presents itself again but will vanish again if the plug is removed. The problem is appart of the signal not the earthing of the input. Adjusting the volume control (which is before the active xover) affects nothing. If I place my finger on the input to the amp when its brown noising (with all inputs connected) the sound vanishes, I remove my finger, it reappears. Also maybe this is not relavent but when the brown noise is produced the amps become VERY VERY sensitive to mains thumps. I have an electric blower heater in my listening room, when running normally turning this on and off has no affect on the sound. But when I get the brown noise, switching the heater on produces large thumps in the speakers and can make the brown noise worse.
The four amps are all arranged within the same case etc with a single outboard powersupply powering them all. The powersupply sits on the floor and supplies DC to the amps.
I dont have a clue whats going on here and would appreciate any help cheers Matt.
Anyway I've built four Slone amps (to his deisgn in the high power amp book design 11.4). After a bit of trouble I've got all of them working and sounding brilliant. Each amp is fed a signal from an active xover with two amps driving bass drivers and the other two the mid and tweet combo.
Now when everything is switched on and the system is left alone, for a period of time the system will warm up to idling temp and stay there. However if im unlucky, sometime after 1st switch on (can be 10 mins or can never happen) there comes a noise which sounds like the brown noise created by cool edit. With this noise also comes a rapid heat up of all the output devices in all four amps, this is not thermal runaway as the amp can idle when this happens, without suffering any damage, but everything gets really hot literally too hot to touch.
If the amp is turned off and on again (this sometimes needs to be done more then once) the noise stops and everything cools down. Only five minutes later for the problem to re appear. When I turn the amp off it doesnt go quietly either, almost as soon as the power is cut there are two thumps produced in the speakers. This is absent when the amp has no problems. Normally after a period of time (power supply caps discharging) a small thump is produced in the speakers but this is gentle and very benign sound wise. When the amp is brown noising the thumps are loud and have an unpleasent electric snap to them.
When the problem occurs it does absolutely nothing to degrade the sound quality appart from raising the noise floor (created by the brown noise) which is not detectable most of the time (only when a quiet passage appears in the music is it noticeable). I guess that the distortion of the amp changes, but that too is not audible.
What also solves the problem is removing one of the inputs from the four amps (it doesnt matter which one) if plugged back in the problem presents itself again but will vanish again if the plug is removed. The problem is appart of the signal not the earthing of the input. Adjusting the volume control (which is before the active xover) affects nothing. If I place my finger on the input to the amp when its brown noising (with all inputs connected) the sound vanishes, I remove my finger, it reappears. Also maybe this is not relavent but when the brown noise is produced the amps become VERY VERY sensitive to mains thumps. I have an electric blower heater in my listening room, when running normally turning this on and off has no affect on the sound. But when I get the brown noise, switching the heater on produces large thumps in the speakers and can make the brown noise worse.
The four amps are all arranged within the same case etc with a single outboard powersupply powering them all. The powersupply sits on the floor and supplies DC to the amps.
I dont have a clue whats going on here and would appreciate any help cheers Matt.
if the reason ive got no responses is because you need more info or nobody knows whats gone wrong, id still like to hear from you even if its just to say something mad like try attatching your output trannies to the cealing of the room!
Im at a complete loss and would appreciate any ideas no matter how odd they might be.
Im at a complete loss and would appreciate any ideas no matter how odd they might be.
Asl Slone
I assume you mean the 11.4 amp in the High Power Amp book. If I recall, Slone says something about this amp pushing the limits with regard to Nyquist stability. That would seem consistant with the suggestion from the prior post. If it were my problem, I would start by double checking everything point by point in the NFB network. That's just my guess. ( I havn't built this one but it is on my future list. )
Did you yous his PCBs or at least PCB art work or did you do your own PCB layout? If you used his you can focus on the components. It's easy to stick in a wrong value resistor somewhere. (Its easy to do it on all PCBs if you build them all at once. Suppose you disconnect 3 of the four amps - does the remaining one still behave as you describe? If it only happens when all are connected I would look for something external to the PCB. These are just my thoughts.
Anyway, try e-mailing Randy. He is usually very willing to help but you have to keep in mind that diagnosis by "remote control" is not easy.
By the way, I have built one of his other amps and it worked flawlessly once all the dumb mistakes on my part were resolved.
Good luck and be sure to post the solution when you work it out.
I assume you mean the 11.4 amp in the High Power Amp book. If I recall, Slone says something about this amp pushing the limits with regard to Nyquist stability. That would seem consistant with the suggestion from the prior post. If it were my problem, I would start by double checking everything point by point in the NFB network. That's just my guess. ( I havn't built this one but it is on my future list. )
Did you yous his PCBs or at least PCB art work or did you do your own PCB layout? If you used his you can focus on the components. It's easy to stick in a wrong value resistor somewhere. (Its easy to do it on all PCBs if you build them all at once. Suppose you disconnect 3 of the four amps - does the remaining one still behave as you describe? If it only happens when all are connected I would look for something external to the PCB. These are just my thoughts.
Anyway, try e-mailing Randy. He is usually very willing to help but you have to keep in mind that diagnosis by "remote control" is not easy.
By the way, I have built one of his other amps and it worked flawlessly once all the dumb mistakes on my part were resolved.
Good luck and be sure to post the solution when you work it out.
Since I don't know the topology of your amp, ie no schematic, I will throw out a bunch of ignorant things to look at and possibly do..... First, do the driver transistors have any kind of filtering like a 100 Ohm resistor in series and then on that junction between the transistor and resistor, a 100pF cap to its corresponding rail? If not you may want to experiment with it. Turn-on thump is usually done by slowly bringing up the diff-pair constant current source, I think...... 
It sounds as if your amp needs a slower turn-on time... Do you have a bypass cap in the input? 2uF should serve well. I just built a pre-amp for my shop and copied it from an old Peavey design. I had a grounding problem as I just used a portion of the preamp so I put a 100K in parallel with the input as I had a DC voltage on the input. I know... Very weird but it solved my problem and I had no sign of sound degradation either. You may have a floating ground in a preamp somewhere. Also, the fact that all of your amps use the same power supply is a problem. There should be some filtering, balancing or suedo isolation or your amp needs great CMRR. You will notice in any quality preamp that ALL of the opamps have a .1 and a .01uF cap on the rails and those caps are almost soldered to the chip itself. I don't know about how much power it puts out but if all of the transistors are in a long line, those .1 and .01uF caps from the rail to ground, every couple of output transistors wouldn't hurt. I would put those same values on the power module and again at the rail input to the amp too. It may be the same as not putting the filtering on an LM317T and then having long DC power lines. YUP, you'll have oscillation or white noise, meaning heat. When you do the trick with your finger, do you do it with only one of the amps or have you done it to all amps individualily and in no certin order? Will it fail with 2 of them connected and in no certin order? will it do it with only one of them? Can you remove the power supply from them individually and look around? You also need, maybe, a grounding resistor from earth to the ground of your DC power. There is a potiential that travels through the AC power line when heavy currents are switched and that doesn't belong in your amp so that's why a ground resistor. What about putting a .1 Cap on the power switch, on each pole? What about an input filter. They use them in vending machines and arcade games. Maybe in an old UPS...... You may just try a .1uF cap from DC ground to earth.......
These are all just thoughts and ramblings and many may not apply to this, but hopefully It got you thinking outside your current box a little bit!!!!!!!
Best of luck and please let us know what you do to fix it!!!!!
Chris
It sounds as if your amp needs a slower turn-on time... Do you have a bypass cap in the input? 2uF should serve well. I just built a pre-amp for my shop and copied it from an old Peavey design. I had a grounding problem as I just used a portion of the preamp so I put a 100K in parallel with the input as I had a DC voltage on the input. I know... Very weird but it solved my problem and I had no sign of sound degradation either. You may have a floating ground in a preamp somewhere. Also, the fact that all of your amps use the same power supply is a problem. There should be some filtering, balancing or suedo isolation or your amp needs great CMRR. You will notice in any quality preamp that ALL of the opamps have a .1 and a .01uF cap on the rails and those caps are almost soldered to the chip itself. I don't know about how much power it puts out but if all of the transistors are in a long line, those .1 and .01uF caps from the rail to ground, every couple of output transistors wouldn't hurt. I would put those same values on the power module and again at the rail input to the amp too. It may be the same as not putting the filtering on an LM317T and then having long DC power lines. YUP, you'll have oscillation or white noise, meaning heat. When you do the trick with your finger, do you do it with only one of the amps or have you done it to all amps individualily and in no certin order? Will it fail with 2 of them connected and in no certin order? will it do it with only one of them? Can you remove the power supply from them individually and look around? You also need, maybe, a grounding resistor from earth to the ground of your DC power. There is a potiential that travels through the AC power line when heavy currents are switched and that doesn't belong in your amp so that's why a ground resistor. What about putting a .1 Cap on the power switch, on each pole? What about an input filter. They use them in vending machines and arcade games. Maybe in an old UPS...... You may just try a .1uF cap from DC ground to earth....... These are all just thoughts and ramblings and many may not apply to this, but hopefully It got you thinking outside your current box a little bit!!!!!!!
Best of luck and please let us know what you do to fix it!!!!!
Chris
Best response Ive ever got cheers guys Ill try and respond to each of your posts.
To Diodes post - the trick with the finger is only on one of the inputs and it doesnt matter which one, the effect is the same a complete extermination of the noise produced. Regarding the noise issue I have used the active xover with two commercially bought amps and had no problems what-so-ever. So im sure the active xover is OK (its battery powered btw). The power supply is built with a cap on the input, i do actually have a filter for the DC supply but not connected to the amp. I originally built one for a Musical Fidelity A1 when I upgraded the power supply. I could try either building another filter or just swapping them over and seing what happens. If by input filter you are refering to a couple of caps and resistors before the amp after smoothing that is. The power supply has a cap on the input, not sure where im not at home so cant go and check. And there is another set of caps before the filter caps to remove high frequency hash.
Sam9 - Yes its the design you mentioned. Ill try checking the amps circuit boards for problems but I dont think I'll find any as I had to do this with two of the boards anyway to get them working, i literally worried every track of copper on one board to find the problem. I used Slones PCB layout. I have not tried disconnecting the amplifiers power supplies and seeing if the problem persists Ill try that (shouldnt be too hard as each amp has rail fuses). I suppose if after shutting one amp down the problem went away then there is something wrong with that amp. The only thing I find strange is that ALL the amps will oscillate (if thats what it is) at the same time and then all stop at othe same time. So either when one amp goes mad it sets them all off I dont know but that to me would seem like the amps are OK and the xover is bust but I can only try and see. Oh and if you do build it you wont be dissapointed it sounds wonderful (when it works that is, I mean it still sounds great when oscillating but just gets HOT)
Tubedude - that seems like the fastest and simplest approach to try Ill have a go when I get home. I think I have a few 47's spare.
Again thanks for the response, Ill try whats been suggested (ive got a two week holiday on thursday to tweak all i want) and post with the results.
Matt.
To Diodes post - the trick with the finger is only on one of the inputs and it doesnt matter which one, the effect is the same a complete extermination of the noise produced. Regarding the noise issue I have used the active xover with two commercially bought amps and had no problems what-so-ever. So im sure the active xover is OK (its battery powered btw). The power supply is built with a cap on the input, i do actually have a filter for the DC supply but not connected to the amp. I originally built one for a Musical Fidelity A1 when I upgraded the power supply. I could try either building another filter or just swapping them over and seing what happens. If by input filter you are refering to a couple of caps and resistors before the amp after smoothing that is. The power supply has a cap on the input, not sure where im not at home so cant go and check. And there is another set of caps before the filter caps to remove high frequency hash.
Sam9 - Yes its the design you mentioned. Ill try checking the amps circuit boards for problems but I dont think I'll find any as I had to do this with two of the boards anyway to get them working, i literally worried every track of copper on one board to find the problem. I used Slones PCB layout. I have not tried disconnecting the amplifiers power supplies and seeing if the problem persists Ill try that (shouldnt be too hard as each amp has rail fuses). I suppose if after shutting one amp down the problem went away then there is something wrong with that amp. The only thing I find strange is that ALL the amps will oscillate (if thats what it is) at the same time and then all stop at othe same time. So either when one amp goes mad it sets them all off I dont know but that to me would seem like the amps are OK and the xover is bust but I can only try and see. Oh and if you do build it you wont be dissapointed it sounds wonderful (when it works that is, I mean it still sounds great when oscillating but just gets HOT)
Tubedude - that seems like the fastest and simplest approach to try Ill have a go when I get home. I think I have a few 47's spare.
Again thanks for the response, Ill try whats been suggested (ive got a two week holiday on thursday to tweak all i want) and post with the results.
Matt.
I tried putting the caps from the input to ground and all the inputs. This caused the mid section to oscillate continuously but silently, so I removed them. The caps stayed to the bass input and seems to have solved the problem but only time will tell. I'll keep y'all posted on any further developments.
Caps on the Bass input?..... What I meant was bypass caps.... .1s and .01s from rail to ground...... Glad you have success even for now.........
Where I come from Y'all is a commonly spoken word. If you say you all or you guys, you are from out of state. At least a person from another country can spell it right!!!
Chris
Where I come from Y'all is a commonly spoken word. If you say you all or you guys, you are from out of state.
At least a person from another country can spell it right!!!Chris
5th Element,
Your amp is oscillating. Seriously, it's a wonder it doesn't blow up the output stage.....
When an audio amplifier oscillates, the feedback loop is out of control. One of the results of this is that for brief portions of the oscillation cycle, both outputs are simulataneously on. This puts one rail in direct contact with the other via the output devices, and very large currents flow. This creates huge heat, directly threatens the life of these devices, and usually, though not always, creates intermodulation artefacts with the audio signal which comes out as randomized noise.
In all audio amplifiers, one, or sometimes two, transistors are used as a voltage amplifier. This stage is normally found directly after the diff input pair, though sometimes it's driven via an emitter follower.
This transistor has its emitter on a rail (or linked to it via a small resistor), and the signal is taken into the base, with the output at the collector. There is invariably a capacitor, usually 33pF to 100pF, between collector and base.
This is the offending component. Increase this capacitor about 20%, and your oscillation problems should disappear.
Of course, it won't sound quite so clear and fast, but it will stabilize it very nicely.
Cheers,
Hugh
Your amp is oscillating. Seriously, it's a wonder it doesn't blow up the output stage.....
When an audio amplifier oscillates, the feedback loop is out of control. One of the results of this is that for brief portions of the oscillation cycle, both outputs are simulataneously on. This puts one rail in direct contact with the other via the output devices, and very large currents flow. This creates huge heat, directly threatens the life of these devices, and usually, though not always, creates intermodulation artefacts with the audio signal which comes out as randomized noise.
In all audio amplifiers, one, or sometimes two, transistors are used as a voltage amplifier. This stage is normally found directly after the diff input pair, though sometimes it's driven via an emitter follower.
This transistor has its emitter on a rail (or linked to it via a small resistor), and the signal is taken into the base, with the output at the collector. There is invariably a capacitor, usually 33pF to 100pF, between collector and base.
This is the offending component. Increase this capacitor about 20%, and your oscillation problems should disappear.
Of course, it won't sound quite so clear and fast, but it will stabilize it very nicely.
Cheers,
Hugh
Part of the problem 5th element described sounds similar to mine. I built the CF OPS (Complementary Feedback Output Stage) Amplifier (figure 11-4, p337) from Randy Slone’s "High-Power Audio Amplifier Construction Manual" from scratch, to fit an existing enclosure, and to fit a heat sink I had on hand. The PCB boards were modified from Randy’s PCB layout to fit the heat sink. They worked during the initial testing phase without problem (though it may have been too early). The rail voltage from a second hand transformer is a little high at 45.4V vs 42V.
With some further testing however, I ran into some problems. I set the bias incorrectly on both boards (a long story), and burnt out the R29 resistor, Q11, and Q12 with sparks and smoke, on both boards. I have replaced all the damaged parts I could find.
There are still some problems however. The amplifier hums and buzzes, but not all of the time (though alsmost all). Noise dvelops when an interconnect and/or source is plugged into the input. It is quiet without any input. When there is noise, I measure a 8 Mhz 1.5V sine wave with a 30 MHz oscilloscope at the output of both channels, and the heat sink heats up quickly. This signal is also evident on the input at a lower level of about 25mV. I wondered if the amp was oscillating at these times due to being set off by some input signal noise. Could this be the cause? Is this then reflected at the input also? What causes the oscillation? Can it be fixed? I hear on the site that this amp is excellent, and I don’t want to give up, but I need some assistance as well, and would be grateful for any offered.
Jorge, what does the 47 pF Ceramic cap do from rail to ground? Sam9, does the layout make a difference to this? My traces are longer and wider than Slone's. Hugh, I will review your suggestion with respect to the schematic, and try your solution.
I have e-mailed Randy Slone also.
With some further testing however, I ran into some problems. I set the bias incorrectly on both boards (a long story), and burnt out the R29 resistor, Q11, and Q12 with sparks and smoke, on both boards. I have replaced all the damaged parts I could find.
There are still some problems however. The amplifier hums and buzzes, but not all of the time (though alsmost all). Noise dvelops when an interconnect and/or source is plugged into the input. It is quiet without any input. When there is noise, I measure a 8 Mhz 1.5V sine wave with a 30 MHz oscilloscope at the output of both channels, and the heat sink heats up quickly. This signal is also evident on the input at a lower level of about 25mV. I wondered if the amp was oscillating at these times due to being set off by some input signal noise. Could this be the cause? Is this then reflected at the input also? What causes the oscillation? Can it be fixed? I hear on the site that this amp is excellent, and I don’t want to give up, but I need some assistance as well, and would be grateful for any offered.
Jorge, what does the 47 pF Ceramic cap do from rail to ground? Sam9, does the layout make a difference to this? My traces are longer and wider than Slone's. Hugh, I will review your suggestion with respect to the schematic, and try your solution.
I have e-mailed Randy Slone also.
Lots of good suggestions here, but before you try any of them you need to check your grounding and psu arrangements. These sort of problems can cause your amps to oscillate and will result in "the finger on the input curing the noise" phenomenon.
There are numerous posts and threads about this in here, search for them.
If you have built the amps correctly then they are unlikely to oscillate because the design is tried and tested. Altering parts and values within the amp circuits should not be necessary and may cause deterioration. So this should be a last resort.
There are numerous posts and threads about this in here, search for them.
If you have built the amps correctly then they are unlikely to oscillate because the design is tried and tested. Altering parts and values within the amp circuits should not be necessary and may cause deterioration. So this should be a last resort.
AKSA said:
this sounds to me like gm doubling than anything else. Oscilating amps in my mind are amps where the negative feedback turns into a positive feedback at certain frequencies, with or without gm-doubling.
If you are concerned about gm-doubling, you would have limited amp bias (to prevent them from conducting more than 180 degrees).
I appreciate the ideas from everyone. Traderbam’s post reminded me that I changed the grounding as well before the amp started to oscillate, so I will try that solution first.
I received an e-mail from Randy Slone. He suspects the problem is from the 2SA1302 / 2SC3281 transistor pairs, due to the oriental forgery problem, and poor imitations of the real Toshiba transistors. He suggests the Toshiba upgraded version of the 2SA1302 / 2SC3281 pairs, which are the 2SA1943 / 2SC5200 pairs. He recommend replacing the output devices in both amplifiers. He also recommend that I replace Q11, Q12, R29, and R32 in both amplifiers, regardless of whether or not they appear to be good. The extreme strain placed on Q11 and Q12 could have caused permanent damage that isn't readily detectable. I will try his suggestions next. I got my latest versions from Digi-key, so I doubt they are forgeries.
If that doesn’t work I will try Hugh’s suggestion about the increase in capacitor, but it is not clear which capacitor would be involved, looking at the schematic.
The noticed the 2SA1302 / 2SC3281 problem elsewhere on the forum.
http://www.diyaudio.com/forums/showthread.php?postid=141025#post141025
I gather there are similar problems with 2SB649 and 2SD669 that is used in this amp.
http://www.diyaudio.com/forums/showthread.php?s=&postid=201905#post201905
I received an e-mail from Randy Slone. He suspects the problem is from the 2SA1302 / 2SC3281 transistor pairs, due to the oriental forgery problem, and poor imitations of the real Toshiba transistors. He suggests the Toshiba upgraded version of the 2SA1302 / 2SC3281 pairs, which are the 2SA1943 / 2SC5200 pairs. He recommend replacing the output devices in both amplifiers. He also recommend that I replace Q11, Q12, R29, and R32 in both amplifiers, regardless of whether or not they appear to be good. The extreme strain placed on Q11 and Q12 could have caused permanent damage that isn't readily detectable. I will try his suggestions next. I got my latest versions from Digi-key, so I doubt they are forgeries.
If that doesn’t work I will try Hugh’s suggestion about the increase in capacitor, but it is not clear which capacitor would be involved, looking at the schematic.
The noticed the 2SA1302 / 2SC3281 problem elsewhere on the forum.
http://www.diyaudio.com/forums/showthread.php?postid=141025#post141025
I gather there are similar problems with 2SB649 and 2SD669 that is used in this amp.
http://www.diyaudio.com/forums/showthread.php?s=&postid=201905#post201905
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.