PROBLEM FOUND! Fixed!
I found the problem!
I had been running each amp off a different household circuit, because of the amplifiers' massive power requirements.
I discovered that the difference in grounding between the two circuits was somehow causing the amps to react to the RFI and amplify it.
As soon as I plugged both amps into the SAME circuit (either Circuit 1 or 2), the problem disappeared. Now there is no audible trace of RFI.
What I'm going to do for now is upgrade the breaker and receptacles to 20A-rated versions and just use one circuit to power both amps, while watching power requirements so as not to overload the breaker (i.e. no vacuum cleaner or iron at all during my listening sessions!)
But this is an experience other owners of monobloc amps may be interested in.
BLUF:
In my experience, using two circuits (one for each monobloc amp) caused me massive RFI problems!
I found the problem!
I had been running each amp off a different household circuit, because of the amplifiers' massive power requirements.
I discovered that the difference in grounding between the two circuits was somehow causing the amps to react to the RFI and amplify it.
As soon as I plugged both amps into the SAME circuit (either Circuit 1 or 2), the problem disappeared. Now there is no audible trace of RFI.
What I'm going to do for now is upgrade the breaker and receptacles to 20A-rated versions and just use one circuit to power both amps, while watching power requirements so as not to overload the breaker (i.e. no vacuum cleaner or iron at all during my listening sessions!)
But this is an experience other owners of monobloc amps may be interested in.
BLUF:
In my experience, using two circuits (one for each monobloc amp) caused me massive RFI problems!
Hi mjraudio,
Is it possible your two amps ended up on different phases? That will cause problems for sure. Your entire audio system should be on the same phase.
-Chris
Is it possible your two amps ended up on different phases? That will cause problems for sure. Your entire audio system should be on the same phase.
-Chris
What could cause them to be out of phase?
If the two circuits were out of phase (90-degree offset?), could that be caused by the accidental reversal of the live and neutral lead from the AC line source?
Would reversing the speaker output polarity cause these problems as well?
If the two circuits were out of phase (90-degree offset?), could that be caused by the accidental reversal of the live and neutral lead from the AC line source?
Would reversing the speaker output polarity cause these problems as well?
anatech said:
Chris, I am curious, you mention that amplifiers like to draw current in surges or spikes.. would uprading the filter caps to a higher MFD value decrease these surges in theory?
Joe
barchetta said:
Hmm I'm feeling just a little deja vu here...I had the same question when I was in the 'filter capacitor replacement time' of my amp fix.
So let me add on to barchetta's excellent question:
Assuming that the rectifiers & transformer could safely handle larger filter caps...could larger ones help in this area or does the rule of diminishing returns apply still?
Also, are these surges & spikes synonymous with "ripple" current?
Hi Joe,
Hi mjraudio,
Be extremely careful measuring these voltages.
-Chris
Yes, that will increase the charging spikes and reduce the conduction angle in your rectifiers. This has the effect of raising the peak currents in the diodes and possibly raising the transformer temperature due to higher losses (peak currents). This will aggravate any tendency for the rectifiers to generate spikes as well.
Hi mjraudio,
No, that would cause a 180° reversal. Your house is supplied with two phases to generate the 208 VAC for your clothes dryer and stove. They are physically separate out to the pole transformer. You will generally induce a great deal of hum and other noise in your system if it is supplied across two or more phases compared to the normal noise levels. Measure between your two hots. You should measure 0 VAC. !20 VAC ish and you have one socket reversed in phase. 208 VAC ish and you are running off two phases.
Be extremely careful measuring these voltages.
-Chris
anatech said:
In idiot terms: the answer is yes it will reduce surges to the home power (the power outlet you are connected to)?
And again in idiot terms: it sounds as if you are going to be asking more of your transformer and rectifier???
And if that is the case.. any thoughts on this FRED rectifier I keep hearing rants and raves about.. people say it actually affects the sound of an amp.. something you can hear (assuming you have a crappy rectifier of course).
Sorry to go off topic..
MJR.. stop at your hardware store and pickup an outlet phase tester.. its easy to use and no brain cells are required.
BUT I think what anatech is saying is you may be using one side of your 220 for one amp and the other side for another.. this wont show that.. Just connect them to the same outlet and see if it goes away 😉
Also; please, do NOT pull your hair out.. No matter what they say, women do not like bald men.. I know for a fact. LOL
Hi Joe,
-Chris
Higher capacitance will make surges worse, including steady state current draw (I was talking about that).
Yes, exactly.
My personal viewpoint is that the standard rectifiers are the best provided you don't do silly things with your supply. If you do, all bets are off.
Yup, that works.
-Chris
Thanks for all your comments and help
Gentlemen I have a new problem.
Look at this picture, it shows the DC offset at the speaker terminals for one of the amps.
The right readout is the current temperature of one of the input stage PCB transistor heatsinks! It has actually risen higher than the readout in this picture and has hit 71.2 degrees Celsius, that is 160 degrees Fahrenheit...and it is still rising!
Now I may not be an electrical engineer but I come from a deep IT background and I can tell you that temperature is beyond unhealthy. My amp will surely die soon with transistors operating at that temperature!!!!
Are temps like this normal? If they are, I will not tolerate them at all. I've assembled servers and PCs for years and the temps in the GFA-565 are a disgrace. You wonder why those input stage electrolytic capacitors leak electrolyte on these amps, it's probably because they overheat given the close vicinity to the transistors. The transistors are cooking the electrolytic capacitors.
I am going to design a custom active cooling solution for the input stage transistors.
Gentlemen I have a new problem.
Look at this picture, it shows the DC offset at the speaker terminals for one of the amps.
The right readout is the current temperature of one of the input stage PCB transistor heatsinks! It has actually risen higher than the readout in this picture and has hit 71.2 degrees Celsius, that is 160 degrees Fahrenheit...and it is still rising!
Now I may not be an electrical engineer but I come from a deep IT background and I can tell you that temperature is beyond unhealthy. My amp will surely die soon with transistors operating at that temperature!!!!
Are temps like this normal? If they are, I will not tolerate them at all. I've assembled servers and PCs for years and the temps in the GFA-565 are a disgrace. You wonder why those input stage electrolytic capacitors leak electrolyte on these amps, it's probably because they overheat given the close vicinity to the transistors. The transistors are cooking the electrolytic capacitors.
I am going to design a custom active cooling solution for the input stage transistors.
GFA-565 after 2 minutes of active cooling in this test setup.
DC offset has dropped substantially and the temperature is less than half.
PIC HERE
I'm not a genious but I would say that achieving lower temperatures would substantially increase the life of my amplifier, wouldn't you?
I'm going to start designing a small and reasonable cooling solution (and quiet).
DC offset has dropped substantially and the temperature is less than half.
PIC HERE
I'm not a genious but I would say that achieving lower temperatures would substantially increase the life of my amplifier, wouldn't you?
I'm going to start designing a small and reasonable cooling solution (and quiet).
Hi mjraudio,
I have never measured the temperature of these transistors, but that is unusually high. I would definetly notice that. Your DC offset supports the idea that there is something rotten in the state of Denmark (so to speak). Custom cooling is not required normally and would probably generate LF noise.
The original caps were a bad manufacturing lot.
So find the problem.
-Chris
I have never measured the temperature of these transistors, but that is unusually high. I would definetly notice that. Your DC offset supports the idea that there is something rotten in the state of Denmark (so to speak). Custom cooling is not required normally and would probably generate LF noise.
The original caps were a bad manufacturing lot.
I agree.
So find the problem.
-Chris
Hi mjraudio,
-Chris
Within reason - yes. Fix the fault, not the symptom.
The other amps I've seen didn't even have the optional cooling fan, and there were fine. You already have more cooling capacity that you should need.
-Chris
anatech said:
Excellent advice, yes! I shall do it! 🙂
The thing about the fans Chris, is that they have never turned on. My speakers are 8 ohms, and I have pushed these amps beyond 50% capacity, and still they have never turned on. The reason is, the thermistors which regulate the fan are on the heatsink at the rear of the unit. When the thermistors at that location go beyond a preset value, the fan presumably kicks in. The designers decided to ignore the transistors as there are no thermistors near it to actuate the fan.
This amp has always run hot, even the rear heatsinks put out significant heat when the unit is idling. The other unit that I fixed runs cool to the touch. However the transistors on both amps are the same temperature, which leads me to believe that 70 degrees is about normal for 1-hour of power-on and usage. I still personally find this unacceptable.
I adjusted the bias in this amp and it is at a perfect 24mV. I'm going to order new op-amps and rip up the input boards next week, clean them and try to reduce the offset in addition to attempting to fix this heat issue. Then I'll either upgrade the heatsinks on those transistors and/or design some sort of active cooling solution. I have no idea what else could be wrong, since BOTH amps have the same temps on the input-stage transistors.
Hi mjraudio,
The bias adjust procedure involves running them until they get hot and allowing them to cool off. The top covers must remain on for this and the adjustment performed quickly. This takes an hour or so to do depending on how far off you are to start.
To the best of my knowledge, the only transistors with a heatsink are the predrivers. I have not seen any discoloured PCB's yet. The 2SA1210 and 2SC2912's should be the ones on a heatsink. They are the cascode transistors and do get warm. Yuo may find that cleaning the input PCB settles your DC offset issues nicely. Replace those caps if you haven't already.
The amp that does not get warm is biased far too cold.
-Chris
The bias adjust procedure involves running them until they get hot and allowing them to cool off. The top covers must remain on for this and the adjustment performed quickly. This takes an hour or so to do depending on how far off you are to start.
To the best of my knowledge, the only transistors with a heatsink are the predrivers. I have not seen any discoloured PCB's yet. The 2SA1210 and 2SC2912's should be the ones on a heatsink. They are the cascode transistors and do get warm. Yuo may find that cleaning the input PCB settles your DC offset issues nicely. Replace those caps if you haven't already.
The amp that does not get warm is biased far too cold.
-Chris
Correct me if I am wrong here.. messing with the bias is messing with the wrong issue... its not going to change the heat of the input board.
Do we have a schematic?
How old is this amp? If its 15yr plus just do ALL the lytics NOW.. also have you compared the temp of both amps at this location? Please excuse me if you mentioned this.
Can I see a pic of this board that has a hot sink on it? I reread.. check your other amp for similar temps please.
Is this the board?
http://www.flickr.com/photos/49493821@N00/375230449/
A little research provided a little info regarding your amp..
Check the reference diodes (6.9 volt temp compensated Zeners).. bad caps can take these out.. and you may have flakey bias control transistors (MPS-A63 and MPS-A13)...
Just basically quoting someone who sounds like he knows your amp...
Do we have a schematic?
How old is this amp? If its 15yr plus just do ALL the lytics NOW.. also have you compared the temp of both amps at this location? Please excuse me if you mentioned this.
Can I see a pic of this board that has a hot sink on it? I reread.. check your other amp for similar temps please.
Is this the board?
http://www.flickr.com/photos/49493821@N00/375230449/
A little research provided a little info regarding your amp..
Check the reference diodes (6.9 volt temp compensated Zeners).. bad caps can take these out.. and you may have flakey bias control transistors (MPS-A63 and MPS-A13)...
Just basically quoting someone who sounds like he knows your amp...
anatech said:
Chris, I just followed that procedure and adjusted the bias on both amps to a perfect 24mV (verified and reverified after about an hour of load with the cover on), the amp that was running cool had a bias setting of 9mV, you were right on! 🙂
The predriver transistors 2SA121 and 2SC2912: are they matched? The service manual does not say so...I would like to replace them, seeing as how they have been operating all this time at 71 degrees (160 degrees Fahrenheit!).
barchetta said:
Yes, that occurred to me as well. Chris was just helping me get everything working right. But, since the other amp had been working at a very low bias (running cooler), that did not affect the temperature of the predriver transistors as they still operated at 70 degrees Celsius (same as the other unit).
I have a 4-page 8.5x11" schematic taped together but with no way to scan it at the moment, maybe I'll try to do it next week and get it online.
The filter capacitor PCB boards have ADCOM stamped on them with the year 1987, so it's 20 years old.
Yes, that picture shows the input stage PCB with the two transistors attached to the black heatsinks. The placed the thermistor right at the heatsink at it measures 71 degrees Celsius (160 degrees Fahrenheit) on BOTH amplifiers after functioning for about 30 minutes.
Barchetta, you requested pics of the new capacitors a while ago.
Also, to confirm: both amplifiers input stage predriver transistors measure temps of 71 degrees Celsius. Bias is steady at 24mV on both amps after warmup under load for 1 hour.
Pic of the fixed amp with new capacitors
Closeup of the new filter capacitors
Hmm, interesting note about the bias control transistors Barchetta, that may be part of the problem. I suspect that a sheer lack of users' operating temps of those predriver transistors is leaving me in the dark as to what is an expected temperature range for those. But on my personal terms, those temps are beyond any acceptable range. If that was the temperature of your CPU, there would be thousands of processing errors.
One more note, I have brand new Mallory electrolytic capacitors to replace all the electrolytic capacitors on the input stage boards in both amps, but I want to get headers (pins that solder onto a PCB) and little connectors so I can make quick-disconnects/headers/MTE/rectangular connectors for all the wire going into that board, so in the future I will not have to mess with soldering of wire onto those boards! This is a pain because I have to order the precise sizes of the headers and wire connectors that have to match the right guage, etc... I really will only be able to make my list once I have the boards out...plus I need to order new op-amp ICs, the Analog Devices OP97FS chips I bought are way too small, I was too impatient to look at the schematic, stupid me, bye by $8.
mjraudio said:
I would strongly urge you not to add the quick connects.. this is an audiophile amp not a computer.. every connection adds additional signal loss that is critical.. people spend big bucks on silver wire and you are going the opposite direction..
Thanks for the pics.
So if both amps are at 70c I am guessing this is normal.. but maybe the lytics are bad.. if you want to close this amp for good replace all the lytics.. also; look up the spec sheets on those transitors. maybe they list acceptable operation range.. you might find they are fine..
I am a amateur.. so anatech.. let us know if you agree with what I said here.. you are the resident expert..
Hi Guys,
Well, Arif and a few other people know these amps fairly well. Mr. Walt Jung also has a good idea what is going on here.
Hi mjraudio,
The big difference between a computer CPU core and those predrivers (yes, they do run warm) is that your predrivers are probably dissipating a couple watts. A CPU dissipates a much greater amount of energy. Yes, the die in a CPU does not lke heat at all. The density and trace widths are much finer than these transistors.
If you can find these parts, buy a few. I like to try to get the NPN's matched up with the PNP's the best you can. Otherwise, clean 'em and regrease 'em. Try not to overtighten them. You may be able to find a larger heatsink for them if you want.
Joe is right on the connector thing. Don't bother. Once you clean these boards up, that should be the last time they have to come out. Connectors can be a trouble point.
I'm glad you got the bias right now. I'll bet the amp will sound smoother.
-Chris
Well, Arif and a few other people know these amps fairly well. Mr. Walt Jung also has a good idea what is going on here.
Hi mjraudio,
The big difference between a computer CPU core and those predrivers (yes, they do run warm) is that your predrivers are probably dissipating a couple watts. A CPU dissipates a much greater amount of energy. Yes, the die in a CPU does not lke heat at all. The density and trace widths are much finer than these transistors.
If you can find these parts, buy a few. I like to try to get the NPN's matched up with the PNP's the best you can. Otherwise, clean 'em and regrease 'em. Try not to overtighten them. You may be able to find a larger heatsink for them if you want.
Joe is right on the connector thing. Don't bother. Once you clean these boards up, that should be the last time they have to come out. Connectors can be a trouble point.
I'm glad you got the bias right now. I'll bet the amp will sound smoother.
-Chris
Chris... I am on the road Mon-thur... just came back home from the Airport so the brain is shot...
70C for the OP stage (or any tranny on the main heatsink) is too high uness you are running 1kHz sinewaves into a 2 ohm load at 33% power. 😉
However on the 555's I routinely measure 70C on the 1210's on the PCB's (you know the ones without heatsinks that scorch the PCB) and these are 20 years old and running fine...
I remember Nelson talking about fixing the Thermal breakers on the heatsinks at 85c so that should be the near absolute max.
The transistors are happy in the 30C to 60C range. (The latter in Class-A amps) measured on the heatsink 1-2" from the device.... this is a guesstimated guideline I use for "normal" amps.
Not having read the long thread, I can say that with an 8 ohm speaker that he has, 70C smells fishy. HF Oscillation is possible... that would also manifest in RF pick up.
What are the chances that his grounding resistor is flaky or even open?
70C for the OP stage (or any tranny on the main heatsink) is too high uness you are running 1kHz sinewaves into a 2 ohm load at 33% power. 😉
However on the 555's I routinely measure 70C on the 1210's on the PCB's (you know the ones without heatsinks that scorch the PCB) and these are 20 years old and running fine...
I remember Nelson talking about fixing the Thermal breakers on the heatsinks at 85c so that should be the near absolute max.
The transistors are happy in the 30C to 60C range. (The latter in Class-A amps) measured on the heatsink 1-2" from the device.... this is a guesstimated guideline I use for "normal" amps.
Not having read the long thread, I can say that with an 8 ohm speaker that he has, 70C smells fishy. HF Oscillation is possible... that would also manifest in RF pick up.
What are the chances that his grounding resistor is flaky or even open?
- Status
- Not open for further replies.