Arif,
these amps were designed to deliver a lot in low impedance loads.
A lot of the amplifiers of yesterday have bias emitter resistor voltages that are now considered to be on the low side.
The Palladium amps had huge heatsinks that run really hot, and were very expensive amplifiers here.
I've seen them a lot, because i really liked the V shape design back then.
these amps were designed to deliver a lot in low impedance loads.
A lot of the amplifiers of yesterday have bias emitter resistor voltages that are now considered to be on the low side.
The Palladium amps had huge heatsinks that run really hot, and were very expensive amplifiers here.
I've seen them a lot, because i really liked the V shape design back then.
Jacco:
I opened the 4004 (Pioneer of the V groove) and the sinks below the V groove were tiny vis a vis the rest of the amp's assets... Caps were just "ok" 10kuF/rail /channel (40 for the entire amp or less.. if memory serves) But the Donuts were 2 and huge... at least 800 to 1000va each. 68v rails... 8 mj15003/4 per channel.
1 more thing, it used 4 bridge rectifiers... 2 per channel
I reckon they could drive 2 ohms... but sinks were as I said.... needing some "extenze".
They had the typical Mj15003/4 sound... Sweet but blurred with no finesse.
Still remember the ad in the "Audio" magazine (boy I miss their technical/ circuit descriptions that stereofilze glosses over) that read across the top, "Sheer Aragons".... it was that ad/review that dubbed it the poor man's krell... and krell being such a big thing in the late 80's... I really wanted to hear it... I guess I was expecting too much...
I opened the 4004 (Pioneer of the V groove) and the sinks below the V groove were tiny vis a vis the rest of the amp's assets... Caps were just "ok" 10kuF/rail /channel (40 for the entire amp or less.. if memory serves) But the Donuts were 2 and huge... at least 800 to 1000va each. 68v rails... 8 mj15003/4 per channel.
1 more thing, it used 4 bridge rectifiers... 2 per channel
I reckon they could drive 2 ohms... but sinks were as I said.... needing some "extenze".
They had the typical Mj15003/4 sound... Sweet but blurred with no finesse.
Still remember the ad in the "Audio" magazine (boy I miss their technical/ circuit descriptions that stereofilze glosses over) that read across the top, "Sheer Aragons".... it was that ad/review that dubbed it the poor man's krell... and krell being such a big thing in the late 80's... I really wanted to hear it... I guess I was expecting too much...
Tiny Tiny correction...
I had set the voltage over the resistor to 30mV (not mA; oops!)
Back to the story:
reset Vre... (???)
Yeek, I had things set correctly the first time (@10mV). Will have to wait a few days before attempting to re-move the amp. Blew my wrist out by just putting on a sweatshirt. Was lugging my monitor/computer around all day (which I shouldn't even be doing; but try and stop me when I have a mission), without a problem, then pop!
Maybe I'll go mid-way ~20mV. Could barely notice any heat at 10 (just over the coldness of the metal itself), and just mildly warm at 30, 20 looks like a good trade-off. 15 if I want to be ultra secure.
I like the heatsinks on the 8002 much better than the v-notch models. The plates on the 8002 are each about 6"x12" facing outward, with the transistors in direct contact (no L bend). You can get a lot more air contact this way than having it just convect through a small gap. Unless it creates some kind of turbo stack-effect. May take some pics when physically able...
Makes me curious what a 50w class-A Pass would be like... Cook pasta on it...
Off to rest, and do the "keep the wife happy" (DIY "now what's that?" prevention) Valentine's prep.
Cheers,
bluesmoke
I had set the voltage over the resistor to 30mV (not mA; oops!)
Back to the story:
reset Vre... (???)
Yeek, I had things set correctly the first time (@10mV). Will have to wait a few days before attempting to re-move the amp. Blew my wrist out by just putting on a sweatshirt. Was lugging my monitor/computer around all day (which I shouldn't even be doing; but try and stop me when I have a mission), without a problem, then pop!
Maybe I'll go mid-way ~20mV. Could barely notice any heat at 10 (just over the coldness of the metal itself), and just mildly warm at 30, 20 looks like a good trade-off. 15 if I want to be ultra secure.
I like the heatsinks on the 8002 much better than the v-notch models. The plates on the 8002 are each about 6"x12" facing outward, with the transistors in direct contact (no L bend). You can get a lot more air contact this way than having it just convect through a small gap. Unless it creates some kind of turbo stack-effect. May take some pics when physically able...
Makes me curious what a 50w class-A Pass would be like... Cook pasta on it...
Off to rest, and do the "keep the wife happy" (DIY "now what's that?" prevention) Valentine's prep.
Cheers,
bluesmoke
Want to hear something funny:
output stages used to be biased by setting the iddling current, not the voltage across the emitter resistors.
A common figure was 25mA per device. A total bias of 100mA for an average power amplifier with 4 output devices gave less than 0.1 watt in class A operation.
One of the threads showed not so long ago that adjusting the voltage difference at a minimal level is the way to minimize crossover distortion.
Even Mr John Curl confirmed the ideal range of the voltage difference across the emitter resistors.
My suggestion would be to give it at least 20mV, the extra 20 watts of dissipation will definitely not kill it.
output stages used to be biased by setting the iddling current, not the voltage across the emitter resistors.
A common figure was 25mA per device. A total bias of 100mA for an average power amplifier with 4 output devices gave less than 0.1 watt in class A operation.
One of the threads showed not so long ago that adjusting the voltage difference at a minimal level is the way to minimize crossover distortion.
Even Mr John Curl confirmed the ideal range of the voltage difference across the emitter resistors.
My suggestion would be to give it at least 20mV, the extra 20 watts of dissipation will definitely not kill it.
Did'nt Self say that the ideal voltage drop across a ER is 47mV for BJT's and (now I forget the number) xx mA's for MOSFET's... he said BJT's are more partial to a specific voltage drop while MOSFET's to a current drop... While one has the highest possible respect for Self and ilk, though one thinks optimal varies by amp to amp...
Hi,
D.Self 2nd ed, A P A Design Handbook p137 says
for EF output
0r10 21.3mV
0r22 23.1mV
0r33 23.8mV
0r47 27.4mV
and for CFP output
0r10 1.53mV
0r22 2.31mV
0r33 2.82mV
0r47 3.59mV
The figures he actually quotes in the table are across both Re in series (that's where the 47mV across 0r33 comes from). I have converted the figures above to Vre as we are used to seeing.
D.Self 2nd ed, A P A Design Handbook p137 says
for EF output
0r10 21.3mV
0r22 23.1mV
0r33 23.8mV
0r47 27.4mV
and for CFP output
0r10 1.53mV
0r22 2.31mV
0r33 2.82mV
0r47 3.59mV
The figures he actually quotes in the table are across both Re in series (that's where the 47mV across 0r33 comes from). I have converted the figures above to Vre as we are used to seeing.
With all this wonderful info, I'm getting a touch dizzy again. However it seems like the original calculation may be in the right range with idling bias at 30mV (if 0r47=27mV). May check to see what the total current draw is next time I can manage opening it. Will still bring it back down to 20mV if the general consensus makes it the recommendation (10 just feels too horribly low; almost freezing).
Completely forgot about testing the idling current right off of the power supply fuse points, the way the Haflers illustrate to test. Yamaha goes by the emitter resistor voltage. Maybe just because it was easier to measure that way.
More about something funny... The mosfet hafler I have uses 2 pair of transistors per channel, biased to an idle current of 200mA per channel for factory spec. Seems double what's being said...
May look into getting an in-ear ir thermometer to more accurately check temps, seems like a handy tool for the toolbox anyway. Strange that computer CPUs and GPUs can dump 100w into a dinky heatsink averaging 60C internal and last pretty much forever, where as power transistors shouldn't see more than 48C at heatsink (118F) going by the palladium spec...? Of course no forced airflow.
Completely forgot about testing the idling current right off of the power supply fuse points, the way the Haflers illustrate to test. Yamaha goes by the emitter resistor voltage. Maybe just because it was easier to measure that way.
More about something funny... The mosfet hafler I have uses 2 pair of transistors per channel, biased to an idle current of 200mA per channel for factory spec. Seems double what's being said...
May look into getting an in-ear ir thermometer to more accurately check temps, seems like a handy tool for the toolbox anyway. Strange that computer CPUs and GPUs can dump 100w into a dinky heatsink averaging 60C internal and last pretty much forever, where as power transistors shouldn't see more than 48C at heatsink (118F) going by the palladium spec...? Of course no forced airflow.
no, just knew where on the shelf it lay.
But more to the point, it does show that designers that occupy different camps can come up with the same answers (I wonder if independantly?).
right in the middle of the recommended range (15 to 25mV) from at least two eminent designers using this Forum.
not necessarily.
Iq=200mA through 2pair = 100mA/output pair. If Re=0r15 then Vre=15mV and if Re=0r25 then Vre=25mV.
It seems that the Hafler could be right on the button if Re is in the range 0r15 to 0r22 and these seem quite possible for a 2pair output stage.
is more down to ensuring that the output and driver devices do not exceed the long term and short term SOAR when driving loads within the specification.
Running the output stage hot increases the likelyhood of exceeding the SOAR when driving high power output or driving a low impedance load or driving a very reactive load and worse, when all three conditions occur at the same time.
Who? me! well you did say a "little" and ""at tmes" so I suppose there is a possibility of salvation.
bluesmoke said:
Wrong assumption.
The 118F/48C is the steady state condition.
Drive the amp hard and the temperature can get much higher.
Saturday Night HS party=> high continuous output level => output stage temp. up=> SOAR goes down=> BiBaBadaBoom !
Regular semiconductor max. temperature is 150 C.
With a heatsink temperature at 50C, there's a 100C margin.
A fraction of this is lost through thermal difference between the transistor and the heatsink=> the (mica) insulator.
At high output levels, the output stage needs the remaining thermal buffer for coping the stress of high output current bass tones.
Ignorance is bliss...
Spent several hours adjusting bias to an even safer 15mV. Temp is even on both channels and a touch lower. Barely drifts (~0.2mV).
Did some other measurements (current at the positive rail fuses) while it was open, now leading me to believe there might be a pre-existing problem I never noticed...
Rails are 65V (64.6V)
Idle @ 20mV bias:
L channel - 278mA
R channel - 230mA (!!)
Idle @ 15mV bias:
L channel - 225mA
R channel - 175mA (!!)
Seeing this, I've got a horrible feeling that maybe a transistor on the right channel could be dead...? If so, I'm feeling well beyond my reach in trying to find and replace it properly at this point (looks like it would be a huge undertaking)... Getting a touch discouraged (but will wait for advice before getting excessively depressed)... Odd that temps feel equal, but would need to measure more accurately...
On the flipside, hooked it up, still sounds great. I don't listen at deafening levels at which point I might notice something...? Possibly early clipping?
Anyway, will give up for now on any more tweaks...
Spent several hours adjusting bias to an even safer 15mV. Temp is even on both channels and a touch lower. Barely drifts (~0.2mV).
Did some other measurements (current at the positive rail fuses) while it was open, now leading me to believe there might be a pre-existing problem I never noticed...
Rails are 65V (64.6V)
Idle @ 20mV bias:
L channel - 278mA
R channel - 230mA (!!)
Idle @ 15mV bias:
L channel - 225mA
R channel - 175mA (!!)
Seeing this, I've got a horrible feeling that maybe a transistor on the right channel could be dead...? If so, I'm feeling well beyond my reach in trying to find and replace it properly at this point (looks like it would be a huge undertaking)... Getting a touch discouraged (but will wait for advice before getting excessively depressed)... Odd that temps feel equal, but would need to measure more accurately...
On the flipside, hooked it up, still sounds great. I don't listen at deafening levels at which point I might notice something...? Possibly early clipping?
Anyway, will give up for now on any more tweaks...
Done for tonight but will re-dabble tomorrow checking the current draw other (neg) rails.
Will triple check again, but the readings off of several emitter resistors were 0.5ohm each. Mondial didn't use the ceramic block high-temp style anywhere, so I'll check the markings/bandings to re-make sure, just in case my metering was somehow flawed. Used 1 each per transistor.
Will also take and post some close-up shots of the board. Insanely happy if it isn't a shot transistor, but would still be a ultra-pain to have to replace most any part due to the clean yet un-service friendly design.
AND will also prep a friendly email to ex-Mondial, hopefully getting the 8002 schematics in return. Looks like it might be more useful than expected now...
Definitely grateful for all this help and input!
Cheers,
bluesmoke
Will triple check again, but the readings off of several emitter resistors were 0.5ohm each. Mondial didn't use the ceramic block high-temp style anywhere, so I'll check the markings/bandings to re-make sure, just in case my metering was somehow flawed. Used 1 each per transistor.
Will also take and post some close-up shots of the board. Insanely happy if it isn't a shot transistor, but would still be a ultra-pain to have to replace most any part due to the clean yet un-service friendly design.
AND will also prep a friendly email to ex-Mondial, hopefully getting the 8002 schematics in return. Looks like it might be more useful than expected now...
Definitely grateful for all this help and input!
Cheers,
bluesmoke
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.