Amplifier Distortion, DC-Offset, and You!
- Thread starter EchoWars
- Start date
I'd put it on Aux, but with the volume at minimum, it really shouldn't matter.
Sounds like a problem with the diff pair at the amp input, or a fluctuating power supply (a lot of diff pair inputs use a zener to drop the voltage and keep tail current constant..a flaky zener could do this).
Sony always does things different, and I'd bet that several stages might have to be shotgunned to eliminate the problem.
Sounds like a problem with the diff pair at the amp input, or a fluctuating power supply (a lot of diff pair inputs use a zener to drop the voltage and keep tail current constant..a flaky zener could do this).
Sony always does things different, and I'd bet that several stages might have to be shotgunned to eliminate the problem.
Newly acquired Kenwood Super Eleven:
Left channel: 0mV (+/-15 mV)
Right Channel: 80mV, (+/-15mV)
I recognize that the right channel is beyond acceptable, but I do not understand the wide swings (+/- 15mV), and the constant shifts in the DC offset values. It rarely shows the same value as the meter monitors it and the changes are not subtle. The unit is labeled as "High Speed DC stereo Receiver". Does this explain all the shifting and wide swings in the DC offset values?
Left channel: 0mV (+/-15 mV)
Right Channel: 80mV, (+/-15mV)
I recognize that the right channel is beyond acceptable, but I do not understand the wide swings (+/- 15mV), and the constant shifts in the DC offset values. It rarely shows the same value as the meter monitors it and the changes are not subtle. The unit is labeled as "High Speed DC stereo Receiver". Does this explain all the shifting and wide swings in the DC offset values?
The 'DC' just means it's DC coupled (no capacitor or transformer between the output devices and the load (speakers)), and the majority of the amps we all have are like this.
The offset swings can mean a differential pair that need replacing, a faulty cap or two, or power supply problems (or a combo of the above). It may be annoying, but not exactly fatal. If the amp has an adjustment for offset you could set for the 'middle of the swings'.
The offset swings can mean a differential pair that need replacing, a faulty cap or two, or power supply problems (or a combo of the above). It may be annoying, but not exactly fatal. If the amp has an adjustment for offset you could set for the 'middle of the swings'.
jasoncauthron
Active Member
Pioneer SA6800 integrated. Clean, but never touched internally, reads:
Left= -55mv
Right= -47mv
Readings didn't drift much from turn-on, until the final check I did about an hour later.
I don't have a service manual for this, but I'm thinking about getting one and going through it. Does anyone know offhand if this one has pots to adjust the offset?
Jason
Left= -55mv
Right= -47mv
Readings didn't drift much from turn-on, until the final check I did about an hour later.
I don't have a service manual for this, but I'm thinking about getting one and going through it. Does anyone know offhand if this one has pots to adjust the offset?
Jason
Raptor
New Member
Re: your test
-Hey Echo....echo.....echo
I am not sure if your aware of my left channel prob or not but i saw your post and decided to give it a try to see what was comng out of the left sie if anything.
Well the right channel of my AU-710 = 3 with the meter on the 2.5 dcvm setting. The left channel was very confusing. It pinned the needle to the left side of the meter as if the probes were reversed. Meaning pos lead on the neg terminal and neg lead on pos terminal. ????????
I'm confused but then again I am ignorant and simply trying to learn. If you might know what this is indicitive of and think i might be able to fix it, please let me know. Otherwise I will be shipping it a guy i met in this forum for him to fix.
Thanks
-Hey Echo....echo.....echo
I am not sure if your aware of my left channel prob or not but i saw your post and decided to give it a try to see what was comng out of the left sie if anything.
Well the right channel of my AU-710 = 3 with the meter on the 2.5 dcvm setting. The left channel was very confusing. It pinned the needle to the left side of the meter as if the probes were reversed. Meaning pos lead on the neg terminal and neg lead on pos terminal. ????????
I'm confused but then again I am ignorant and simply trying to learn. If you might know what this is indicitive of and think i might be able to fix it, please let me know. Otherwise I will be shipping it a guy i met in this forum for him to fix.
Thanks
jasoncauthron
Active Member
Whether the offset is postive or negative with respect to the +/- speaker terminals doesn't matter, but if you're checking it with an analog meter, you'll have to swap the probes accordingly so that the value can be read as a positive number.
You should swap the leads and read it again.
I'm a little confused about the scale you have your meter set to: Is the scale 0-2.5mV DC? You said that the reading you got for the right side was 3?
You might need to switch to a higher scale, if possible. 0-300 mV, perhaps.
Jason
You should swap the leads and read it again.
I'm a little confused about the scale you have your meter set to: Is the scale 0-2.5mV DC? You said that the reading you got for the right side was 3?
You might need to switch to a higher scale, if possible. 0-300 mV, perhaps.Jason
Raptor
New Member
meter setting
Jason
I have no idea really how to use a meter well. The meter has dc and ac settings and ohm. the options are 2.5, 10, 50, 500 for the DC. I used 2.5 as anything higher reduced greatly the response on the meter.
I am still confused about the left channel reading pinning the needle to the left side of the meter.
I tell you what, I'm going to pack that little meter up and give it back to my friend before I hurt myself.
:uzi:
Jason
I have no idea really how to use a meter well. The meter has dc and ac settings and ohm. the options are 2.5, 10, 50, 500 for the DC. I used 2.5 as anything higher reduced greatly the response on the meter.
I am still confused about the left channel reading pinning the needle to the left side of the meter.
I tell you what, I'm going to pack that little meter up and give it back to my friend before I hurt myself.
:uzi:
dr*audio
Fish fingers and custard!
Echowars,
All my amps have adjustments for the DC offset and I adjusted them all to 0V +/- 5mV when I got them. I can see the obvious reasons for wanting a well matched diff pair whether the offset is adjustable or not. Obviously if the betas are matched the distortion will be lower.
I have had good success in replacing packaged diff pairs that are no longer available, using some really excellent low noise Motorola transistors, MPSA18, that I match for Beta. You can put heatshrink over the pair once they're installed to keep them at the same temperature. I haven't needed to replace a PNP diff pair yet, but I have used 2N5087 as a stand alone low noise amp with good success.
The FETS I have used for a diff pair on my Onkyo Grand Integra M508 are no longer available, but I found some 2N5480 that should be suitable. these can be matched by Idss.
What about those Zetex parts you used, are they low noise? What kind of package are they in?
All my amps have adjustments for the DC offset and I adjusted them all to 0V +/- 5mV when I got them. I can see the obvious reasons for wanting a well matched diff pair whether the offset is adjustable or not. Obviously if the betas are matched the distortion will be lower.
I have had good success in replacing packaged diff pairs that are no longer available, using some really excellent low noise Motorola transistors, MPSA18, that I match for Beta. You can put heatshrink over the pair once they're installed to keep them at the same temperature. I haven't needed to replace a PNP diff pair yet, but I have used 2N5087 as a stand alone low noise amp with good success.
The FETS I have used for a diff pair on my Onkyo Grand Integra M508 are no longer available, but I found some 2N5480 that should be suitable. these can be matched by Idss.
What about those Zetex parts you used, are they low noise? What kind of package are they in?
jasoncauthron
Active Member
Raptor,
I'd say you need to find a more sensitive, digital meter, if you want to get a meaningful DC offset reading. I think the "2.5" scale you mentioned is probably Volts, not mV. In that case, a good reading would be around 1% of the scale, and very hard to read.
Jason
I'd say you need to find a more sensitive, digital meter, if you want to get a meaningful DC offset reading. I think the "2.5" scale you mentioned is probably Volts, not mV. In that case, a good reading would be around 1% of the scale, and very hard to read.
Jason
jasoncauthron
Active Member
My guess is that would be the bias current adjustment, NOT DC offset.
If the procedure for adjusting it involves measuring a voltage inside the amp, across the emitter resistors, then it's the bias adjustment.
Bias current might be referred to as idle current or some other name, but I've never heard DC offset called anything else.
Best not to fiddle with it unless you're pretty sure you know what you're doing.
If the procedure for adjusting it involves measuring a voltage inside the amp, across the emitter resistors, then it's the bias adjustment.
Bias current might be referred to as idle current or some other name, but I've never heard DC offset called anything else.
Best not to fiddle with it unless you're pretty sure you know what you're doing.
Last edited:
^
|
|
What he said....
dr*audio...
I like to use Zetex ZTX795A (PNP, 140V, 0.5A, 1W, 100MHz) or ZTX694B (NPN, 120V, 0.5A, 1W, 130MHz) for replacement diff input pairs. Zetex does not give noise figures on these high gain devices, but I have neither seen/heard nor measured any problems when using them (in a phono stage I doubt I'd use them, but anywhere else I wouldn't hesitate). They are a EBC layout 'TO-92' style, and are very rugged little buggers.
Link: http://www.zetex.com/3.0/3-3-1.asp?tcatid=6
|
|
What he said....
dr*audio...
I like to use Zetex ZTX795A (PNP, 140V, 0.5A, 1W, 100MHz) or ZTX694B (NPN, 120V, 0.5A, 1W, 130MHz) for replacement diff input pairs. Zetex does not give noise figures on these high gain devices, but I have neither seen/heard nor measured any problems when using them (in a phono stage I doubt I'd use them, but anywhere else I wouldn't hesitate). They are a EBC layout 'TO-92' style, and are very rugged little buggers.
Link: http://www.zetex.com/3.0/3-3-1.asp?tcatid=6
dc output not causing protect mode fault?
newbie here and electrical klutz.
this has got to be 1 of the coolest threads i've ever read. almost like disclosing insider information to the commoners. just like computers, a little info shared - ("knows just enough to be dangerous").
course, i ran right out and bought the 29 range radio shack multimeter (radioshack.com has it on sale $10 cheaper).
hastily i bought a sony 100watt '82 model ta-f555es amp on ebay without prior research for $91. later found out it has a history of protect mode issues.
with 1 set of brand new speakers -
sure enough, the fault can occur as low as 10 o'clock with heavy bass passages. even occurs with tone defeat at 11 to 12 o'clock. speaker sensitivity 97 db @ 8 ohms. 14 gauge wire.
tried everything from cleaning, to a cooling fan, to every hookup possible (different speakers - ala avid 8" bookshelves), to new speaker wire, to monster interconnects, to all the inputs (only source i'm using is a cd player), to the ohm switch on the back set to 8 & 4.
nothing is plugged into the switched or unswitched outlets and the amp is plugged directly into the wall.
surely i might have a dc output problem but my readings were:
left: -1.15 millivolts right: -.07 millivolts
sony calls this an "audio current transfer" current drive amplifier system which has electrically separated the 4 amplifier sections.
the amp sounds pretty decent to me but just when the music gets good, i'm thinking anytime now and ouch - 95 mph down to 0 in 0 seconds flat... ugghhhh!!!!
i got what i payed for...
anyone know of a cherry au-717 for sale? by the way, noticed the au-717 has twice the slew rate and power consumption as this amp.
newbie here and electrical klutz.
this has got to be 1 of the coolest threads i've ever read. almost like disclosing insider information to the commoners. just like computers, a little info shared - ("knows just enough to be dangerous").
course, i ran right out and bought the 29 range radio shack multimeter (radioshack.com has it on sale $10 cheaper).
hastily i bought a sony 100watt '82 model ta-f555es amp on ebay without prior research for $91. later found out it has a history of protect mode issues.
with 1 set of brand new speakers -
sure enough, the fault can occur as low as 10 o'clock with heavy bass passages. even occurs with tone defeat at 11 to 12 o'clock. speaker sensitivity 97 db @ 8 ohms. 14 gauge wire.
tried everything from cleaning, to a cooling fan, to every hookup possible (different speakers - ala avid 8" bookshelves), to new speaker wire, to monster interconnects, to all the inputs (only source i'm using is a cd player), to the ohm switch on the back set to 8 & 4.
nothing is plugged into the switched or unswitched outlets and the amp is plugged directly into the wall.
surely i might have a dc output problem but my readings were:
left: -1.15 millivolts right: -.07 millivolts
sony calls this an "audio current transfer" current drive amplifier system which has electrically separated the 4 amplifier sections.
the amp sounds pretty decent to me but just when the music gets good, i'm thinking anytime now and ouch - 95 mph down to 0 in 0 seconds flat... ugghhhh!!!!
i got what i payed for...
anyone know of a cherry au-717 for sale? by the way, noticed the au-717 has twice the slew rate and power consumption as this amp.
Protection issues are common on older amps...and usually pretty easy to fix.
As far as your voltage readings, I'm not convinced that you are interpreting the meter correctly. You say 1.15mV and 0.07mV, but...you see...the decimal is all wrong. My Fluke meters (which cost several hundred dollars) have a max resolution of 1mV on the DC scale, so 1.15mV would read '1mV' on it, and forget about the 0.07mV...that is a reading of 70µv (0.00007V, or 70 microvolts). None of my meters would measure such a voltage...I'd have to look at such a low voltage on a scope. I'd suggest that what you actually measured was 1.15V on the one channel, and 70mV on the other. Read the manual for the meter, and double-check these readings. I think you'll find I'm right.
Sounds like the amp needs some attention. Nothing to freak-out over, but with over a volt on the one channel I'd say that it sure ain't living up to its potential.
RE: The AU-717. A fine amp. But so is the Sony. Sansui made a big stink back when about slew rate...as if it was the 'end-all be-all' to amplifier design. It isn't. High slew rates are fine when used with stable loads, but can cause instability if highly reactive loads are used. And a lot of amplifiers with rather pokey slew-rates are quite well regarded. So, don't shop for an amp based on one touted feature or spec., shop with your eyes (aesthetically pleasing) and with your ears.
As far as your voltage readings, I'm not convinced that you are interpreting the meter correctly. You say 1.15mV and 0.07mV, but...you see...the decimal is all wrong. My Fluke meters (which cost several hundred dollars) have a max resolution of 1mV on the DC scale, so 1.15mV would read '1mV' on it, and forget about the 0.07mV...that is a reading of 70µv (0.00007V, or 70 microvolts). None of my meters would measure such a voltage...I'd have to look at such a low voltage on a scope. I'd suggest that what you actually measured was 1.15V on the one channel, and 70mV on the other. Read the manual for the meter, and double-check these readings. I think you'll find I'm right.
Sounds like the amp needs some attention. Nothing to freak-out over, but with over a volt on the one channel I'd say that it sure ain't living up to its potential.
RE: The AU-717. A fine amp. But so is the Sony. Sansui made a big stink back when about slew rate...as if it was the 'end-all be-all' to amplifier design. It isn't. High slew rates are fine when used with stable loads, but can cause instability if highly reactive loads are used. And a lot of amplifiers with rather pokey slew-rates are quite well regarded. So, don't shop for an amp based on one touted feature or spec., shop with your eyes (aesthetically pleasing) and with your ears.
dr*audio
Fish fingers and custard!
Echowars,
I don't believe it's the high slew rate itself that causes instability, but rather the extreme measures the designers took to get it make the amp unstable. The Sansui AU-X1 has these output transistors with an Ft of 60MHz! Even if the amp was stable, which it isn't, it can amplify any rf signal that gets in there, say, from a nearby radio station, and try to pump that out to the speakers. I think at that frequency it just pumps it into the Zoebel network at the output and heats up the output transistors, but you get the idea.
I recently modded an AU-X1 and ended up with a stable amp by limiting the open loop bandwidth to 30KHz. The closed loop bandwidth, then would still far exceed that, and so I limited it at the input to 40KHz. I also replaced the output transistors with Motorolas (MJ21193/94 ) with an Ft of 6MHz or so.
Ended up with a really sweet sounding amp that is stable and still plenty fast.
I don't believe it's the high slew rate itself that causes instability, but rather the extreme measures the designers took to get it make the amp unstable. The Sansui AU-X1 has these output transistors with an Ft of 60MHz! Even if the amp was stable, which it isn't, it can amplify any rf signal that gets in there, say, from a nearby radio station, and try to pump that out to the speakers. I think at that frequency it just pumps it into the Zoebel network at the output and heats up the output transistors, but you get the idea.
I recently modded an AU-X1 and ended up with a stable amp by limiting the open loop bandwidth to 30KHz. The closed loop bandwidth, then would still far exceed that, and so I limited it at the input to 40KHz. I also replaced the output transistors with Motorolas (MJ21193/94 ) with an Ft of 6MHz or so.
Ended up with a really sweet sounding amp that is stable and still plenty fast.