I'm doing some work on a 30+ year old Parasound HCA-1200 amplifier. I bought this in the USA, but now live in Germany, and decided to get rid of the bulky/ugly external step-down transformer and convert the amplifier for a 230VAC mains. Fortunately this is easy since the power transformer has two parallel primary windings, which can be put into series for a 230V input. See annotated schematic below (bottom left corner is the relevant bit).
Anyway, I have a question regarding the correct bias current for the output transistors. I found a table online listing the correct bias current settings for all Parasound amps apart from the HCA-1200. Does anyone know the correct value? In other words, what's the correct DC voltage across the 0.33 ohm emitter resistors on the output stage?
I wrote to Parasound to ask this, and got a quick reply saying "we don't know". I then asked if John Curl could be reached to find out, and they replied that he had lost the information in a fire. Perhaps someone here knows the answer...
Cheers,
Bruce
Anyway, I have a question regarding the correct bias current for the output transistors. I found a table online listing the correct bias current settings for all Parasound amps apart from the HCA-1200. Does anyone know the correct value? In other words, what's the correct DC voltage across the 0.33 ohm emitter resistors on the output stage?
I wrote to Parasound to ask this, and got a quick reply saying "we don't know". I then asked if John Curl could be reached to find out, and they replied that he had lost the information in a fire. Perhaps someone here knows the answer...
Cheers,
Bruce
Hi Bruce,
Just look at the others that are similar. The current will not be far off. For most power amps, 20 ~ 30 mA per device is sufficient. The manufacturer may specify something higher, and some designs (like a Stasis) require higher bias current per device.
Just look at the others that are similar. The current will not be far off. For most power amps, 20 ~ 30 mA per device is sufficient. The manufacturer may specify something higher, and some designs (like a Stasis) require higher bias current per device.
Anatech, thanks for your reply. As you know, the distortion levels are dependent on this bias, so I'd like to get it right. I'm attaching the table that I found online, you can see that there is quite a bit of variation between different models. So if possible, I'd like to know what the target is for the HCA-1200. (An educated guess from the table below would be 10mV, since this is the correct value for similar HCA models.) Cheers, Bruce
Hi Bruce,
It isn't that critical for distortion. If you monitor THD looking for the crossover spike, you'll see it disappears on most amplifiers before you reach the bias set point. The manufacturer just wants to make sure the amp meets spec, that's all. To see the crossover spike, in case you haven't, use a THD meter (with a pointer) and look at the monitor output. Sync the scope to the signal and look at the monitor output with Y2. You'll see the spike if you lower the bias current at some point. That's a great indicator of correct bias conditions in most amplifiers. I use a 4R load at 10 KHz, 1 watt.
For some, bias does determine distortion and you set it more to the published THD reading. The problem with them is the design isn't good, so dissipation is a problem if you adjust for minimum distortion. Amplifiers with poor bias regulation tend to have more complicated bias setup procedures (Adcom and others).
For variable bias (optical bias is another type), they increase bias current with signal present trying to control dissipation. So really the bias is varied by a signal detection network, sometimes with an optocoupler. Pain in the rear.
The folks currently at Parasound are a very helpful bunch. A nice change.
-Chris
It isn't that critical for distortion. If you monitor THD looking for the crossover spike, you'll see it disappears on most amplifiers before you reach the bias set point. The manufacturer just wants to make sure the amp meets spec, that's all. To see the crossover spike, in case you haven't, use a THD meter (with a pointer) and look at the monitor output. Sync the scope to the signal and look at the monitor output with Y2. You'll see the spike if you lower the bias current at some point. That's a great indicator of correct bias conditions in most amplifiers. I use a 4R load at 10 KHz, 1 watt.
For some, bias does determine distortion and you set it more to the published THD reading. The problem with them is the design isn't good, so dissipation is a problem if you adjust for minimum distortion. Amplifiers with poor bias regulation tend to have more complicated bias setup procedures (Adcom and others).
For variable bias (optical bias is another type), they increase bias current with signal present trying to control dissipation. So really the bias is varied by a signal detection network, sometimes with an optocoupler. Pain in the rear.
The folks currently at Parasound are a very helpful bunch. A nice change.
-Chris
Hi Chris,
thanks, that's a good suggestion. I don't have a THD meter but do have a signal generator and a scope. I'll follow your advice with a 4 ohm load at 1W, 10kHz sine wave, and adjust the bias until the crossover spike is absent.
BTW, if 10mV across the 0.33 ohm resistors is correct, that corresponds exactly to 30mA of bias, which is at the top end of your 20-30mA typical range.
Cheers,
Bruce
thanks, that's a good suggestion. I don't have a THD meter but do have a signal generator and a scope. I'll follow your advice with a 4 ohm load at 1W, 10kHz sine wave, and adjust the bias until the crossover spike is absent.
BTW, if 10mV across the 0.33 ohm resistors is correct, that corresponds exactly to 30mA of bias, which is at the top end of your 20-30mA typical range.
Cheers,
Bruce
Hi Bruce,
I'm sorry, but to work on any audio, especially amplifiers, you do absolutely need a distortion meter. For this, the minimum would be a darned good notch filter with auto-nulling. That's what a THD meter is.
I use an audio analyser for audio spectrum measurements. I also use classic THD meters because for some things they are the best tool for the job. A good THD meter has a very low distortion analyser built in (HP 339A for example). You need a low distortion oscillator to do this as well. Note the if a THD meter gas a "C" designation, it may not have a built in oscillator.
I'm sorry, but to work on any audio, especially amplifiers, you do absolutely need a distortion meter. For this, the minimum would be a darned good notch filter with auto-nulling. That's what a THD meter is.
I use an audio analyser for audio spectrum measurements. I also use classic THD meters because for some things they are the best tool for the job. A good THD meter has a very low distortion analyser built in (HP 339A for example). You need a low distortion oscillator to do this as well. Note the if a THD meter gas a "C" designation, it may not have a built in oscillator.
Hi Chris, OK, let me see if I can borrow a THD meter. I was going to ask if I could just use the FFT function on my scope and sum the power in the higher harmonics, but I suspect that the input stage and ADC will add more distortion than the amplifier itself. Just for fun I'll try measuring the purity of my 3 different waveform generators that way (one is an ancient HP with a light bulb as part of the AGC feedback loop!). Cheers, Bruce
Hi Bruce,
Yes, if you can borrow a THD meter, you'll see how necessary it is. The crossover notch is tiny in comparison to the test signal (2.83V into 8R for 1 watt). It will also illustrate why you can't just use a scope to judge the quality of any audio stage. Problems have to be extremely gross before you can se them.
Digital scopes don't have the fine signal resolution either, you are further ahead with a decent analogue scope. I'm saying this as an owner of a Keysight MSOX3104T, and often grab my older analogue scopes. The Philips PM3070 or PM3365A being my favorites. The traces on Tek scopes are often too fat. Also, the FFT function on my very expensive DSO isn't that useful where you are looking. I don't even use it.
There is the truth about modern DSO products. Not good enough for analogue work. They do other cool things that are useful, but they do not replace instruments made for the purpose, like spectrum analysers and distortion analysers.
Yes, if you can borrow a THD meter, you'll see how necessary it is. The crossover notch is tiny in comparison to the test signal (2.83V into 8R for 1 watt). It will also illustrate why you can't just use a scope to judge the quality of any audio stage. Problems have to be extremely gross before you can se them.
Digital scopes don't have the fine signal resolution either, you are further ahead with a decent analogue scope. I'm saying this as an owner of a Keysight MSOX3104T, and often grab my older analogue scopes. The Philips PM3070 or PM3365A being my favorites. The traces on Tek scopes are often too fat. Also, the FFT function on my very expensive DSO isn't that useful where you are looking. I don't even use it.
There is the truth about modern DSO products. Not good enough for analogue work. They do other cool things that are useful, but they do not replace instruments made for the purpose, like spectrum analysers and distortion analysers.
For the record, I found some notes in my files. On April 13th 1993 I had a telephone conversation with Andy at Parasound products in San Francisco. He suggested using the central trimpot to set the bias in the range 12-15mV across the emitter resistors. The other trimpot is for adjusting DC bias.
Andy also sent me a full service manual for the Parasound HCA-1200, including Specs, Operating instructions, Parts location, Adjustment and Test point locations, Block Diagram, Schematics, PC board patterns and a BOM/Parts list. This document shows that the "idle current" should be set to have 6-10mV across the emitter resistors, and the DC bias set below 30mV at the output.
I would like to scan the service manual and upload it to some archive for such materials. Is there an archive like that here?
Andy also sent me a full service manual for the Parasound HCA-1200, including Specs, Operating instructions, Parts location, Adjustment and Test point locations, Block Diagram, Schematics, PC board patterns and a BOM/Parts list. This document shows that the "idle current" should be set to have 6-10mV across the emitter resistors, and the DC bias set below 30mV at the output.
I would like to scan the service manual and upload it to some archive for such materials. Is there an archive like that here?
one thing in the schematic puzzles me: one of the 4 power transistors in paralell has a 100 ohm emitter resistor while the others have 0.33ohm. is this a preparation for a higher power version ?
There is a typo in my message #9 above. It should read "The other trimpot is for adjusting DC offset."
@basreflex
The 100 ohm emitter resistors on Q311 and Q313 are typos on the schematic. The parts list and bill of materials show that all 8 emitter resistors, including this pair, are 5W cement resistors 0.33 ohms +- 5%.
However, I was wrong about what the service manual instructions for bias current setting. I thought the instructions were to set the voltage across the emitter resistors (0.33 ohms) to 6-10mV. In fact one should set this voltage on the 100 ohm base resistors, for example R331. Weird!
Sharp eyes!
The 100 ohm emitter resistors on Q311 and Q313 are typos on the schematic. The parts list and bill of materials show that all 8 emitter resistors, including this pair, are 5W cement resistors 0.33 ohms +- 5%.
However, I was wrong about what the service manual instructions for bias current setting. I thought the instructions were to set the voltage across the emitter resistors (0.33 ohms) to 6-10mV. In fact one should set this voltage on the 100 ohm base resistors, for example R331. Weird!
Does the diyAudio forum have a place where one can post PDF documents? If so, I'll add the scans of my three HCA-1200 documents (owners manual, service and repair manual, schematics).
Nice to know! Can you provide a link? Does one have to pay for it?
EDIT: I just checked, and only found the service manual for the PARASOUND HCA1200II, which is a different amplifier. Also, in spite of the name, what is there on electrotanya is only a schematic, not the full service and repair manual.
Second EDIT: I've uploaded the three files to electrotanya, it seems like a good place to put them. Hopefully that will help others in the future who might need them.
Last edited:
The values 20-22mV across the Re resistor as shown for HCA-3500 will lead to lowest distortion. Your amp seems to have 0R33 Re resistors, so 60mA per pair approx. If you have a sufficient heatsink.
Hi PMA,
Are you working from the Oliver criterion?
Anyway, bias current per device in service is probably what the goal should be, and of course operating temperature. Even if the heatsinks are large, running outputs at higher temperatures isn't a good plan. That and a amplifier design that drops distortion continuously past a certain point probably isn't the best design.
I'm looking at this from the viewpoint of long term ownership. So a purely theoretical discussion may not fit the real life goal of a long lasting, reliable amplifier. As you've said, depends on the heat sinks and individual device dissipation.
Are you working from the Oliver criterion?
Anyway, bias current per device in service is probably what the goal should be, and of course operating temperature. Even if the heatsinks are large, running outputs at higher temperatures isn't a good plan. That and a amplifier design that drops distortion continuously past a certain point probably isn't the best design.
I'm looking at this from the viewpoint of long term ownership. So a purely theoretical discussion may not fit the real life goal of a long lasting, reliable amplifier. As you've said, depends on the heat sinks and individual device dissipation.
I am writing based on Douglas Self analysis of EF class AB output stages, Bob Cordell analysis, John Curl experience and my own simulation analysis and real measurements of EF and 2EF class AB NPN/PNP output stages. All these resources are saying the same, coming to the same conclusion. Keep voltage drop across Re resistor near 20 - 25mV for lowest output stage crossover distortion. Of course, many commercial amps have insufficient heatsinks and the idle current of 1 pair must be lower. Parasound would not be the case, in my opinion. One of the sane designs.
Interesting. I've seen crossover disappear completely around 5 mA on some amplifiers, and never disappear on others.
I use a THD meter and monitor the output of that on an oscilloscope in order to see the notch clearly. Triggered off the output waveform of the amplifier.
I use a THD meter and monitor the output of that on an oscilloscope in order to see the notch clearly. Triggered off the output waveform of the amplifier.
I hope you have understood I spoke about output stage in isolation. Of course, negative feedback and feedback frequency compensation take a role in resulting distortion. But, the lower the crossover distortion of the output stage in isolation, the lower the distortion with the feedback loop. And, the optimum idle current in case of the CFP output stage is much lower than for 2EF.
I think you might like to study the classical article by Douglas Self
http://douglas-self.com/ampins/dipa/dipa.htm
https://www.eetimes.com/distortion-in-power-amplifiers-part-iv-the-power-amplifier-stages/
It is good to add a theoretical background to the huge workshop experience, both is important.
I think you might like to study the classical article by Douglas Self
http://douglas-self.com/ampins/dipa/dipa.htm
https://www.eetimes.com/distortion-in-power-amplifiers-part-iv-the-power-amplifier-stages/
It is good to add a theoretical background to the huge workshop experience, both is important.