Hello Out There
I have built myself a Power Amplifier using Hitachi mosfets, 2SK176 and 2SJ56, two per channel which run from supply rails up to +/- 100V and will deliver a maximum output of 68V rms. The quiescent current was set at 155mA and this is the question I would like answered. I have recently seen on the net people suggesting that the bias current can be increased to give better sound quality. If this is a viable action could anyone tell me to what level I could increase the quiescent current on my amp. I also have a recent problem of the imaging drifting to the left. I have eliminated everything from the equation and it is the power amp that is at fault. I was wondering if the standing current could be the possible cause, that it might have drifted, as you can see I am a novice when it comes down to electronics. I would appreciate any help, veiws or answers to the above.
Cheers!!
I have built myself a Power Amplifier using Hitachi mosfets, 2SK176 and 2SJ56, two per channel which run from supply rails up to +/- 100V and will deliver a maximum output of 68V rms. The quiescent current was set at 155mA and this is the question I would like answered. I have recently seen on the net people suggesting that the bias current can be increased to give better sound quality. If this is a viable action could anyone tell me to what level I could increase the quiescent current on my amp. I also have a recent problem of the imaging drifting to the left. I have eliminated everything from the equation and it is the power amp that is at fault. I was wondering if the standing current could be the possible cause, that it might have drifted, as you can see I am a novice when it comes down to electronics. I would appreciate any help, veiws or answers to the above.
Cheers!!
John,
The more current you use, the closer to class A you will get. This will entail more heat. You'll need to consult the SOA charts for your devices and keep in mind that your heatsinks will get pretty toasty if you start jacking up the bias. The power supply will need to be able to supply the necessary current without coming unstuck.
That said, class A operation is one ingredient of good sound.
Grey
The more current you use, the closer to class A you will get. This will entail more heat. You'll need to consult the SOA charts for your devices and keep in mind that your heatsinks will get pretty toasty if you start jacking up the bias. The power supply will need to be able to supply the necessary current without coming unstuck.
That said, class A operation is one ingredient of good sound.
Grey
Thanks Grey
I am using heatsinks fom an industrial robot no less and at the recommended bias setting of 155mA they barely get warm so as far as that is concerned I should be OK. The power supply is based around a 1Kva transformer and 88000 mfd of capacitors which have been bypassed with high quality polypropylenes so there should be no problem there.
Where and what are the SOA charts that you refer to and what information will they relate to. If all looks good I might try increasing the quiescent current in 25mA steps and monitor what happens and also see if the overall performance improves. What do you think???
John
I am using heatsinks fom an industrial robot no less and at the recommended bias setting of 155mA they barely get warm so as far as that is concerned I should be OK. The power supply is based around a 1Kva transformer and 88000 mfd of capacitors which have been bypassed with high quality polypropylenes so there should be no problem there.
Where and what are the SOA charts that you refer to and what information will they relate to. If all looks good I might try increasing the quiescent current in 25mA steps and monitor what happens and also see if the overall performance improves. What do you think???
John
You can find SOA curves on device data sheets. SOA means safe operation area. This is limited by power (y=1/X shape curve), max operating voltage, max operating current etc. Then there is the substrate thermal, usually max 150-175degC but you need to stay well below that.
It is unlikely that your drift is a class A issue, more likely a drift situation, perhaps in the driver stage given your usage of industrial components
Petter
It is unlikely that your drift is a class A issue, more likely a drift situation, perhaps in the driver stage given your usage of industrial components
Petter
Thank you Petter
I have had a look at the available information that I have to hand and this tells me that these particular mosfets have the following specification: (a) Id max/A = 8.00 (b) RDS max/ohms = 1.5 (c) Pd max/W = 125 (d) Vds max/V = 200
As I stated previously I am no electronics buff so I find it difficult to keep up with you guys. Based on the specs could you tell me if the quiescent current could be increased without a major catastophy. Can anyone explain when and how an amp is said to be class A and is this a simple matter of increasing the bias to a certain point or is there more to it than this.
Cheers John
I have had a look at the available information that I have to hand and this tells me that these particular mosfets have the following specification: (a) Id max/A = 8.00 (b) RDS max/ohms = 1.5 (c) Pd max/W = 125 (d) Vds max/V = 200
As I stated previously I am no electronics buff so I find it difficult to keep up with you guys. Based on the specs could you tell me if the quiescent current could be increased without a major catastophy. Can anyone explain when and how an amp is said to be class A and is this a simple matter of increasing the bias to a certain point or is there more to it than this.
Cheers John
John,
The specs you mention are, unfortunately, in isolation--meaning that you can have the 8A, but not at the same time as the 200V (note that 8A * 200V would yield 1600W--well beyond the 125W case limit). The SOA chart (aka SOAR on some datasheets, the R at the end standing for Region) should be somewhere in the data sheets. I don't keep Hitachi data sheets around, as I can't get the parts reliably.
Does their website not provide complete information on their devices?
Grey
The specs you mention are, unfortunately, in isolation--meaning that you can have the 8A, but not at the same time as the 200V (note that 8A * 200V would yield 1600W--well beyond the 125W case limit). The SOA chart (aka SOAR on some datasheets, the R at the end standing for Region) should be somewhere in the data sheets. I don't keep Hitachi data sheets around, as I can't get the parts reliably.
Does their website not provide complete information on their devices?
Grey
John,
first of all, from what I understand not all class AB amps really benefit from jacking up the bias current in search of the audio holy grail known as class A. There are other types of distortion besides xover that may actually increase the total distortion of the amp if the amplifer is not designed to operate in class A to begin with.
I don't know what kind of design you are using but from simple transistor considerations a 125 w device operating at 100V will be very ill-disposed to tolerate much more than 0.5 amps. For that matter, I think that one pair of these mosfet would be undersized in an amp operating at 1/2 of the voltage you are using right now!
first of all, from what I understand not all class AB amps really benefit from jacking up the bias current in search of the audio holy grail known as class A. There are other types of distortion besides xover that may actually increase the total distortion of the amp if the amplifer is not designed to operate in class A to begin with.
I don't know what kind of design you are using but from simple transistor considerations a 125 w device operating at 100V will be very ill-disposed to tolerate much more than 0.5 amps. For that matter, I think that one pair of these mosfet would be undersized in an amp operating at 1/2 of the voltage you are using right now!
The suggested bias for Hitachi lateral MOSFETs is 100mA per device.If you are running two pair per channel that would be 200mA plus the current from your driver circuit, say another 50mA, for a total of 250mA per channel.With +/- 100V and only two pair of outputs only 8 ohms may be driven.As far as the image drifting the first thing I would look at would be the cap in the feedback loop to ground.Usually an elecrolytic due to the size, this part is critical.If the image drifts to the left, the cap in the right channel is usually at fault.Replace with as high a quality non-polar as you can afford, and bypass with a small film type.To prevent damage to this cap a pair of reverse paralleled glass diodes should be used (1N4148 or similar).
Thank you all
I am really pleased at the response I've had from all you guys I didn't expect all the help that I have received. I shall try and digest everything that you have imparted and will let you know at a later date what, if anything, I have done and what the overall benefits were.
Thanks to everyone
John
I am really pleased at the response I've had from all you guys I didn't expect all the help that I have received. I shall try and digest everything that you have imparted and will let you know at a later date what, if anything, I have done and what the overall benefits were.
Thanks to everyone
John
While Bipolar AB amps often do not benefit all that
much from higher bias current, the Mosfets definitely
do, and I would suggest that most of the complaints
about Mosfets stem from low bias performance.
I would suggest that you crank the bias up until you
are running 25-30 watts each, which is about 250-300 mA
per device, assuming that you can still touch them after
a half hour or so.
much from higher bias current, the Mosfets definitely
do, and I would suggest that most of the complaints
about Mosfets stem from low bias performance.
I would suggest that you crank the bias up until you
are running 25-30 watts each, which is about 250-300 mA
per device, assuming that you can still touch them after
a half hour or so.
Hello John,
my experience with the Hitachi Fets is that you can get big gains changing the bias from 100 to about 200-225 mA per Fet. After that the changes are quite subtle. Do you really need 100V supply rails? This will give over 500W/8 ohms with no power gain into 4 Ohms cause max current of 2 paralled Fets is only 16 amps. My first amps had about 75V supply rails but I´m down to 42V now and therefore able to use better (noisewise)transistors/capacitors in the driver stage. The amp is still perfectly capable of driving large speakers (Thiel CS6)without clipping (in a normal living room).
In my experience 2 paralled fets (Hitachi that is) are not enough when you´re using "difficult" speakers, with a 100V supply I would use at least 4.
william
my experience with the Hitachi Fets is that you can get big gains changing the bias from 100 to about 200-225 mA per Fet. After that the changes are quite subtle. Do you really need 100V supply rails? This will give over 500W/8 ohms with no power gain into 4 Ohms cause max current of 2 paralled Fets is only 16 amps. My first amps had about 75V supply rails but I´m down to 42V now and therefore able to use better (noisewise)transistors/capacitors in the driver stage. The amp is still perfectly capable of driving large speakers (Thiel CS6)without clipping (in a normal living room).
In my experience 2 paralled fets (Hitachi that is) are not enough when you´re using "difficult" speakers, with a 100V supply I would use at least 4.
william
Clarification
Hi Guys
I have been reading through your responses again and was wondering if I have caused a little confusionon my initial posting. I quoted "I have built myself a Power Amplifier using Hitachi mosfets, 2SK176 and 2SJ56, two per channel...." To clarify this matter what I should have stated was that there are two of each per channel ie: four transitors per channel. I don't know whether this will make a difference to your initial comments but my apologies to all.
John
Hi Guys
I have been reading through your responses again and was wondering if I have caused a little confusionon my initial posting. I quoted "I have built myself a Power Amplifier using Hitachi mosfets, 2SK176 and 2SJ56, two per channel...." To clarify this matter what I should have stated was that there are two of each per channel ie: four transitors per channel. I don't know whether this will make a difference to your initial comments but my apologies to all.
John
Well Done !!!!!
Hi All
I said I would get back to you after I had played around with the bias current. Well I decided not to pfaff around with 25mA increments but went staight for 250mA. I set this up and let it settle for approx. an hour for each channel, no load connected. It took about that time for the heatsinks to reach what seemed to be the stable temperature with no load. This was scientifically done by placing ones hand rather gingerly on said heatsink and leaving it in that position until you figure that you haven't actually turned your hand into a major burns injury. Yep they were hot but not too hot to hold on to. what puzzled me was when I built this amp some four years ago I set the quiescent current to 155mA as recommended but when I connected my meter up to set the 250mA it was only reading 85/87mA I don't know why it had dropped so much.
Now for the best part a) my amp is still working, the heatsinks get mighty hot when driving my Magneplaner MG11c speakers. b) The performance of the amp sonically is a revelation. The bass is much deeper and much better defined, The sound stage is deeper front to back, there is more space around the individual musicians, There is lots of low level detail that I have never heared before. It's JUST GREAT!!!
Thank you to everybody for your fine advice it has been very much appreciated. I still have the drifting image problem but I don't think I am capable of sorting that so I must find someone who knows what he is doing.
Best Regards
John
Hi All
I said I would get back to you after I had played around with the bias current. Well I decided not to pfaff around with 25mA increments but went staight for 250mA. I set this up and let it settle for approx. an hour for each channel, no load connected. It took about that time for the heatsinks to reach what seemed to be the stable temperature with no load. This was scientifically done by placing ones hand rather gingerly on said heatsink and leaving it in that position until you figure that you haven't actually turned your hand into a major burns injury. Yep they were hot but not too hot to hold on to. what puzzled me was when I built this amp some four years ago I set the quiescent current to 155mA as recommended but when I connected my meter up to set the 250mA it was only reading 85/87mA I don't know why it had dropped so much.
Now for the best part a) my amp is still working, the heatsinks get mighty hot when driving my Magneplaner MG11c speakers. b) The performance of the amp sonically is a revelation. The bass is much deeper and much better defined, The sound stage is deeper front to back, there is more space around the individual musicians, There is lots of low level detail that I have never heared before. It's JUST GREAT!!!
Thank you to everybody for your fine advice it has been very much appreciated. I still have the drifting image problem but I don't think I am capable of sorting that so I must find someone who knows what he is doing.
Best Regards
John
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.