My tester device reports gate capacitance as 7nF. Very high, specs read 4nF (IRFP 260).
I tried gate stoppers up to 240 Ohms, mosfet oscillation went down from 2.5 MHz to 2 MHz.
Small Zobel (100 Ohm + 47pF from G to D) on every mosfet didn't make any difference.
Only upper mosfets were oscillating, bottom ones were ok.
Now I'm trying another board without output devices, and I see smaller/different oscillation from VAS.
Will try to get rid of this first, and then try single pair of mosfets...
Just wanted a confirmation from somebody who actually built mk2 version, that this amp is fixable.
Oscillations on upper mosfets (2.5 Mhz kind-of-sinus wave) gates were happening at all times,
even without input signal.
With signal, say with 1kHz sinus, I would see thicker (4 times thicker than normal) 1kHz sinus
on the output.
I built 4 amps nmos200 (with 1 pair, so I guess it should be called nmos100
without any issues, ever. I searched this forum, and couldn't find even one confirmed build of mk2 version.
One thing I would definitely say is that the quality of the compensation capacitors matter. Don't use ceramic disc capacitors here, as these will undoubtedly have poor stability. Look for multilayer ceramic capacitors of NP0/C0G type dielectric. These days they are easier to get compared to polypropylene/polyester.
Do people really go to the effort of faking MJE340/350 devices?!
I note that _ID indicated he used 2SC2240 in the input stage... genuine ones look like they should work, but I would bet they are fakes. 2N5551 would work but you would have to swap collector and base pins to make them fit the PCB. It should be possible to fit them, though.
Do people really go to the effort of faking MJE340/350 devices?!
I note that _ID indicated he used 2SC2240 in the input stage... genuine ones look like they should work, but I would bet they are fakes. 2N5551 would work but you would have to swap collector and base pins to make them fit the PCB. It should be possible to fit them, though.
All my parts are from mouser, no substitutes - all exactly like in quasi's schematic (mk2).
Output devices are irfp260, rails (for now) 60V.
Comp. caps are C0G.
In the other Quasi designs IRFP450 is used as output devices and now when I look at the Nmos350mk2 there is no recommended output device in the schematic.
Is it not possible that Nmos350mk2 is meant to use IRFP450 like the other designs? If you compare IRFP450 and IRFP260 the latter has twice the input capacitance and more than twice as high transconductance two things that will make stability harder to achieve.
Is it not possible that Nmos350mk2 is meant to use IRFP450 like the other designs? If you compare IRFP450 and IRFP260 the latter has twice the input capacitance and more than twice as high transconductance two things that will make stability harder to achieve.
I noticed this, but:
1) my nmos200s with irfp260s and the same gate stoppers and compensation worked fine
2) I have a bucket of irfp260s to use up
3) mk2 oscillates (VAS) even without output devices connected
I noticed this, but:
1) my nmos200s with irfp260s and the same gate stoppers and compensation worked fine
2) I have a bucket of irfp260s to use up
3) mk2 oscillates (VAS) even without output devices connected
So far it seems that no one actually successfully built mk2..
This is dead long time ago but since i am giving a go to a NMOS 350 here is one thing to start with
Having your drivers at 18 ma that will dissipate X watt times 2 in a common heatsink with another stage Vas at 12 ma times 2 that will also produce <X watt times 2
will actually mean that no matter how big is the heat sink ( which in our case is not that big due to pcb design ) bias will take some time to stabilize and in all cases the higher temp of the drivers will eventually stabilize to some temp and bring also VAS stages in the same temp ..... That is thermal run away any way you slice it
in my first attempt at 71+71 rails yes with a small heat sink and the amp biased at about 30 ma per transistor i had almost 65 degrees on the small heat sink at no signal conditions ,,,,Now no way an 340- or 350 can operate properly at this temp
I removed the drivers an locate them else were and temp dropped to 35 degrees down and bias setting become much easier to set and more stable regardless room temp or amp hot or cold ...
ok that is one issue solved .... but i think that there will be many more details to look at ....It seems to me that the circuit is not refined
Having your drivers at 18 ma that will dissipate X watt times 2 in a common heatsink with another stage Vas at 12 ma times 2 that will also produce <X watt times 2
will actually mean that no matter how big is the heat sink ( which in our case is not that big due to pcb design ) bias will take some time to stabilize and in all cases the higher temp of the drivers will eventually stabilize to some temp and bring also VAS stages in the same temp ..... That is thermal run away any way you slice it
in my first attempt at 71+71 rails yes with a small heat sink and the amp biased at about 30 ma per transistor i had almost 65 degrees on the small heat sink at no signal conditions ,,,,Now no way an 340- or 350 can operate properly at this temp
I removed the drivers an locate them else were and temp dropped to 35 degrees down and bias setting become much easier to set and more stable regardless room temp or amp hot or cold ...
ok that is one issue solved .... but i think that there will be many more details to look at ....It seems to me that the circuit is not refined
Halo Sakis,
Good to hear from you, hope you are doing OK. I have build the 350 amp about 8 years ago using 970/2240 input transistors which I suspect were not original. I have used IRFP460 output devices that were original.I have used IRFP150/250 as well, I'm not really sure.Drivers were MJE34/350. Their origin were also suspect if I think of it today. They were bought from a local shop and not from a reputable company.It was very difficult to set the bias, the pot were very sensitive. I think now that it was oscillation / instability. When everything worked the Quasi sounded very good on the ear. Cross conduction did happen sometimes and then it blew I think R20, some Fets and the drivers as well. If IRC the positive signal clipped a little before negative, That is why the circuit were modified to ver2 to get symmetrical clipping.
Good to hear from you, hope you are doing OK. I have build the 350 amp about 8 years ago using 970/2240 input transistors which I suspect were not original. I have used IRFP460 output devices that were original.I have used IRFP150/250 as well, I'm not really sure.Drivers were MJE34/350. Their origin were also suspect if I think of it today. They were bought from a local shop and not from a reputable company.It was very difficult to set the bias, the pot were very sensitive. I think now that it was oscillation / instability. When everything worked the Quasi sounded very good on the ear. Cross conduction did happen sometimes and then it blew I think R20, some Fets and the drivers as well. If IRC the positive signal clipped a little before negative, That is why the circuit were modified to ver2 to get symmetrical clipping.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.