Hello;
I have a faulty Swiss Physics Model 6 power amp for repair. This amp is equipped with 2SJ55/2SK175 VMOS devices in the output stage. I have to replace one complete O/P device set, which means 3 N-type and 3 P-type FET`s. Browsing my parts suppliers, I had to recognize that these devices are out of stock, and I wasn`t able to find a TO-3 substitute. Futheron, I strongly believe these have to be matched, do they? What about using 2SK49/2SJ135 (I know they have a lower Vmax, but as far as I see this won`t be a problem).
Anyone else who has experience with these amps? I can`t find much information in the web; I believe Model 5A (pre) and Model 6 A(power) were the only items manufactured by Mr. Mauricio del Nobile? These amps seem to be rather well-done.
Thanks;
audiomatic
I have a faulty Swiss Physics Model 6 power amp for repair. This amp is equipped with 2SJ55/2SK175 VMOS devices in the output stage. I have to replace one complete O/P device set, which means 3 N-type and 3 P-type FET`s. Browsing my parts suppliers, I had to recognize that these devices are out of stock, and I wasn`t able to find a TO-3 substitute. Futheron, I strongly believe these have to be matched, do they? What about using 2SK49/2SJ135 (I know they have a lower Vmax, but as far as I see this won`t be a problem).
Anyone else who has experience with these amps? I can`t find much information in the web; I believe Model 5A (pre) and Model 6 A(power) were the only items manufactured by Mr. Mauricio del Nobile? These amps seem to be rather well-done.
Thanks;
audiomatic
Uh ? those outputs is like almost impossible to break. replacing them ? but normally the fuses in the psu blows long before ure able to blow those output's. Thei also has drive limmitters. U better check drivers. AND check for chorted caps ( the blue phillips) as thei somtimes short. Output relays etc also blows after switching with signal. never never so maybee anyway.
I havent yet seen a 6A with blow'n output devices (!).
BUZ 900 series (lateral fets) from magna teck there is some complementary among them.
I havent yet seen a 6A with blow'n output devices (!).
BUZ 900 series (lateral fets) from magna teck there is some complementary among them.
The lower voltage devices will work if the total rail-rail voltage is less than the max of one device... rule of thumb.
They are essentially the same devices... I think the higher voltage ones are merely selected from the batches.
I agree these are difficult devices to blow - but I have seen quite a few amps with blown devices, although not the amp you mention. Don't know that one's internals.
If the devices have source resistors for each device, matching is *less* important - most amps do. The larger the resistor the more the "balancing" effect.
Contact Borbely Audio in Germany, he may have some still left around... mention that I saw your post, might help.
If you were in the USA I could likely provide the parts...
_-_-bear
They are essentially the same devices... I think the higher voltage ones are merely selected from the batches.
I agree these are difficult devices to blow - but I have seen quite a few amps with blown devices, although not the amp you mention. Don't know that one's internals.
If the devices have source resistors for each device, matching is *less* important - most amps do. The larger the resistor the more the "balancing" effect.
Contact Borbely Audio in Germany, he may have some still left around... mention that I saw your post, might help.
If you were in the USA I could likely provide the parts...
_-_-bear
All of the Hitachi MOSFETs in TO-3 packages have been out of production for years. Sometimes you can find new old stock (NOS) for silly prices (like $50 each!) A better bet are the Magnatec clones:
http://www.magnatec-uk.co.uk/mosdata.shtml
The ones you need are BUZ901 and BUZ906. These are better than the Hitachi original parts, plus much easier to find and also much cheaper (around $10 each).
http://www.magnatec-uk.co.uk/mosdata.shtml
The ones you need are BUZ901 and BUZ906. These are better than the Hitachi original parts, plus much easier to find and also much cheaper (around $10 each).
2SJ55/2SK175 L-mos are rated at 25% more power than the lower current rated 2SJ50/2SK135, use the lower rated part at your own risk.
Exicon anb Semelab (Magnatec-uk) have TO3 replacements.
Designers using these FETs thought they 'could take it' and just didn't use enough parts.
Inadequate gate protection when driving low impedance loads is another frequent cause of fai;ure.
Exicon anb Semelab (Magnatec-uk) have TO3 replacements.
Designers using these FETs thought they 'could take it' and just didn't use enough parts.
Inadequate gate protection when driving low impedance loads is another frequent cause of fai;ure.
Mr del Nobile encountered much tragedy in his life, i believe the 5 and 6A were his second attempt for audio stardom.
The 6A was very lovingly made, one of the few at the time that used a heatpipe for cooling the outputs.
It is 2SJ50/2SK135 and 2SJ49/2SK134, as reffered in another posting.
I built a number of amplifiers based on the 50/135's, in the 80s the favorite Mosfets, both in the commercial and DIY sector.
Favorite because of their sound, and for their ruggedness.
The 55/175 i used once, for a design that needed higher ratings.
Both types used with sufficient cooling survive fuses with easy.
unless the heatpipe in the 6A collapsed, it is unlikely that the fets have been damaged first.
The 6A has a beautifull casing, an admirable build quality.
In high contrast with some other "Swiss made" that were filthy expensive, covered with "mumm is the word" , with an internal view that looked like a first attempt in DIY.
Unfortunately, like anything Swiss, too much for too little.
The 6A was very lovingly made, one of the few at the time that used a heatpipe for cooling the outputs.
It is 2SJ50/2SK135 and 2SJ49/2SK134, as reffered in another posting.
I built a number of amplifiers based on the 50/135's, in the 80s the favorite Mosfets, both in the commercial and DIY sector.
Favorite because of their sound, and for their ruggedness.
The 55/175 i used once, for a design that needed higher ratings.
Both types used with sufficient cooling survive fuses with easy.
unless the heatpipe in the 6A collapsed, it is unlikely that the fets have been damaged first.
The 6A has a beautifull casing, an admirable build quality.
In high contrast with some other "Swiss made" that were filthy expensive, covered with "mumm is the word" , with an internal view that looked like a first attempt in DIY.
Unfortunately, like anything Swiss, too much for too little.
Think about what happens in an amp with feedback.
When the amplifier output ceases to rise with rising input we say the amplifier is clipping.
Feedback causes the gate drive to go as high as it can possibly go so that the output can follow the input.
When the gate to source voltage exceeds about 15V it perforates and shorts to the source.
An example of this would be the Hafler DH200 which can drive 200W into 2 ohms. The peak supply to the front end is ±63V, 200W/2R is 28.3V peak. If the amp is now driven too hard the gates will never see more than about 35V or so, as it goes into clipping the front end will slam the gates with the total rail voltage to try and raise the speaker output. While there is a 10V 400mW zener to protect the gates it doesn't last long when being hammered by the 10W driver transistor. After the zener goes, the gate follows.
Some big Infinity speakers got down to 2R, and I also saw this in PA/DJ situations (mainly with the DH500).
When the amplifier output ceases to rise with rising input we say the amplifier is clipping.
Feedback causes the gate drive to go as high as it can possibly go so that the output can follow the input.
When the gate to source voltage exceeds about 15V it perforates and shorts to the source.
An example of this would be the Hafler DH200 which can drive 200W into 2 ohms. The peak supply to the front end is ±63V, 200W/2R is 28.3V peak. If the amp is now driven too hard the gates will never see more than about 35V or so, as it goes into clipping the front end will slam the gates with the total rail voltage to try and raise the speaker output. While there is a 10V 400mW zener to protect the gates it doesn't last long when being hammered by the 10W driver transistor. After the zener goes, the gate follows.
Some big Infinity speakers got down to 2R, and I also saw this in PA/DJ situations (mainly with the DH500).
Thanks djk,
I misunderstood your first post in the issue since I thought L-FET's might deal with some "strange natur" in combination with low impedance load not known to me since you didn't mention what was the real deal to protect the gate, eg. not to exceed the max Vgs which can destroy any FET necessary nothing to do really with the load.
But thanks again for explaning!
Cheers
I misunderstood your first post in the issue since I thought L-FET's might deal with some "strange natur" in combination with low impedance load not known to me since you didn't mention what was the real deal to protect the gate, eg. not to exceed the max Vgs which can destroy any FET necessary nothing to do really with the load.
But thanks again for explaning!
Cheers
djk,
what don't I understand?
You first said it depends on the load, in your next post you talked about clipping... I don't care, for me the mechanism is clear.
Clipping has nothing to do we the load either because it can happen with a lousy design with zero load so I don't understand what you are trying to argue here, exceeding rated Vgs WILL destroy the FET even if there is a very small load and that's what I said in the sentence you raped.
Cheers!
what don't I understand?
You first said it depends on the load, in your next post you talked about clipping... I don't care, for me the mechanism is clear.
Clipping has nothing to do we the load either because it can happen with a lousy design with zero load so I don't understand what you are trying to argue here, exceeding rated Vgs WILL destroy the FET even if there is a very small load and that's what I said in the sentence you raped.
Cheers!
You're both right....
but DJK is more right.
UT, if the load is very light, then the output will follow the voltage at the gate. The only way in that case to exceed the Vgs rating of the device is if the driver stage has a rail voltage that is much higher than the rail of the output device. While it is theoretically possible for this to happen, I have never seen a design where this was true (except for a hybrid design with a tube front end).
But as DJK points out, if the load is very heavy, then the output does not follow the input so well. In fact in the case of a dead short, the output does not follow the input at all. In this situation it is very easy for the driver stage to apply enough drive voltage to exceed the Vgs rating of the output MOSFET.
So I would concur with DJK's original statement, "Inadequate gate protection when driving low impedance loads is another frequent cause of failure."
Hope this helps,
Charles Hansen
but DJK is more right.
UT, if the load is very light, then the output will follow the voltage at the gate. The only way in that case to exceed the Vgs rating of the device is if the driver stage has a rail voltage that is much higher than the rail of the output device. While it is theoretically possible for this to happen, I have never seen a design where this was true (except for a hybrid design with a tube front end).
But as DJK points out, if the load is very heavy, then the output does not follow the input so well. In fact in the case of a dead short, the output does not follow the input at all. In this situation it is very easy for the driver stage to apply enough drive voltage to exceed the Vgs rating of the output MOSFET.
So I would concur with DJK's original statement, "Inadequate gate protection when driving low impedance loads is another frequent cause of failure."
Hope this helps,
Charles Hansen
Hi Charles, et al,
well I am not focusing so match on the load issue, the whole thing started beacause I thought there could be a "special issue" with L-FET's gates when djk said in post #5:
Inadequate gate protection when driving low impedance loads is another frequent cause of fai;ure.
I just started to wonder what that meant because there was nothing mentioned that the Vgs could be exceeded was the issue, the mechanism was NOT explained so I solely started to wonder are there OTHER issues with L-FET's I didn't know yet.
And of course it's pretty clear that Vgs grows with load, i have not opposed that.
Wheter the load is "pulling down" Source respective to Gate and increasing the Vgs or the Gate is "pulled up" respective to Source buy the driver is semantic discussion, I still say that exceeding somwhere above rated Vgs will destroy the FET and that's all I tried to say here, just look at it alone.
This thing could be watched from the componet (L-FET) viewpoint alone (which I did more) or the whole amplifier design criteria viewpoint.
So for me everything has been pretty clear from post #8, well, it's not easy to conversate over internet...
Cheers
well I am not focusing so match on the load issue, the whole thing started beacause I thought there could be a "special issue" with L-FET's gates when djk said in post #5:
Inadequate gate protection when driving low impedance loads is another frequent cause of fai;ure.
I just started to wonder what that meant because there was nothing mentioned that the Vgs could be exceeded was the issue, the mechanism was NOT explained so I solely started to wonder are there OTHER issues with L-FET's I didn't know yet.
And of course it's pretty clear that Vgs grows with load, i have not opposed that.
Wheter the load is "pulling down" Source respective to Gate and increasing the Vgs or the Gate is "pulled up" respective to Source buy the driver is semantic discussion, I still say that exceeding somwhere above rated Vgs will destroy the FET and that's all I tried to say here, just look at it alone.
This thing could be watched from the componet (L-FET) viewpoint alone (which I did more) or the whole amplifier design criteria viewpoint.
So for me everything has been pretty clear from post #8, well, it's not easy to conversate over internet...
Cheers
Ultima Thule said:Whether the load is "pulling down" Source respective to Gate and increasing the Vgs or the Gate is "pulled up" respective to Source by the driver is semantic discussion, I still say that exceeding somwhere above rated Vgs will destroy the FET and that's all I tried to say here, just look at it alone.
So for me everything has been pretty clear from post #8, well, it's not easy to converse over internet...
Cheers
Hello UT,
Your description of the circuit action is very clear and very well explained. And you are correct in saying that exceeding Vgs will destroy the device. In my experience, the main way that Vgs is exceeded is by driving low impedance loads with inadequate gate protection, which I think is what DJK was saying.
You are also right that it is sometimes not easy to converse over the internet!
Best regards,
Charles Hansen
Charles Hansen said:
Stupid question here... in the case of 1 step PSU, i.e. drivers sharing the same rails as OP devices, when the OP stage goes into clipping, it has probably done so because the rails have sagged (real life) and as such if the rails sag with any given point in clipping, the drivers have less of a rail potential to work with and to dump the voltage into the OP devices correct?
The only issue would then be if the driver's Vgs params are better than the OP MOSFET's VGS by such a magnitude that it kills the Zener in the OP, then yes I agree with you guys, if not wont this be an issue primarily with topologies using separate rails for the front end?
Thanks
K-
Hafler used a 10V 400mW zener in their gate protection circuit.
The driver transistor is rated for 1A current.
The amp will clip at about 20V into 2 ohms on a sine wave.
The power supply sags from about 60V to 40V under load.
On music the front end voltage will be higher than this because of the filter cap isolated with either a resistor (or diode, depending on vintage).
After the zener burns out there can be well inexcess of 20V on the gates with music.
The outputs gradually develope gate to source shorts and eventually have no gain.
The driver transistor is rated for 1A current.
The amp will clip at about 20V into 2 ohms on a sine wave.
The power supply sags from about 60V to 40V under load.
On music the front end voltage will be higher than this because of the filter cap isolated with either a resistor (or diode, depending on vintage).
After the zener burns out there can be well inexcess of 20V on the gates with music.
The outputs gradually develope gate to source shorts and eventually have no gain.
OK so the drivers did have a pseudo rail that was higher than the main rails under load. thanks DJK.
PS: I just got a DH-500 at a good price. I am planning to get a second Trafo from a surplus store and make it Dual mono.
Because the rails will not sag as much, I am tempted to replace the Hitcahi OP stage with 250w devices from Profusion etc, What devices do you recommend as drop down replacements?
thanks!
PS: I just got a DH-500 at a good price. I am planning to get a second Trafo from a surplus store and make it Dual mono.
Because the rails will not sag as much, I am tempted to replace the Hitcahi OP stage with 250w devices from Profusion etc, What devices do you recommend as drop down replacements?
thanks!
You want to look for the same size package.
(some higher output TO3 cans are thicker and have fatter pins)
You want to look for the same range of Vgs for the given bias & drive the circuit originally wanted.
You want to match each rail's devices if possible.
Obviously equal or higher Vds...
(take a peak at the gate capacitance, just to make sure it is similar)
Other than that, pretty much you can substitute at will...
_-_-bear
(some higher output TO3 cans are thicker and have fatter pins)
You want to look for the same range of Vgs for the given bias & drive the circuit originally wanted.
You want to match each rail's devices if possible.
Obviously equal or higher Vds...
(take a peak at the gate capacitance, just to make sure it is similar)
Other than that, pretty much you can substitute at will...
_-_-bear
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.