In such a case there wouldn't be regulation, since the NMOS will conduct with constant current, regardless of the output voltage.
yes! I think that, because I use the Salas Simplistic Simplistic in put Max 350V- out put 250V, is very long times. It is very good! I like it because it has fewer components!
And the sound is exelent!
Hello Nick,
I've started tests with 10m45 CCS at 30mA 2sk170 8,2mA and MPSA94 instead of MPSA92 with R8=470+100 ohm.
Load=18Kohm
Vin =333V
Vout=330,9V
VbQ5=330,6V Before 47ohm
VbQ2=164,4V Before 47ohm
VgQ3=1,6V Before 220ohm
Vin =333V
Vout=310V
VbQ5=301,3V Before 47ohm
VbQ2=152V Before 47ohm
VgQ3=3,7V Before 220ohm
It seems work, the regulation is from about 50V to331V
I'v not tested noise and so on, I think in the next days.
The regulation in not linear, a multiturn trimmer may be better.
What do you think about.
Ciao
Guglielmo
I've started tests with 10m45 CCS at 30mA 2sk170 8,2mA and MPSA94 instead of MPSA92 with R8=470+100 ohm.
Load=18Kohm
Vin =333V
Vout=330,9V
VbQ5=330,6V Before 47ohm
VbQ2=164,4V Before 47ohm
VgQ3=1,6V Before 220ohm
Vin =333V
Vout=310V
VbQ5=301,3V Before 47ohm
VbQ2=152V Before 47ohm
VgQ3=3,7V Before 220ohm
It seems work, the regulation is from about 50V to331V
I'v not tested noise and so on, I think in the next days.
The regulation in not linear, a multiturn trimmer may be better.
What do you think about.
Ciao
Guglielmo
You use one MPSA94 or 2 in cascode? OK, Q5 means cascode I guess. Normal trimmers are better in long term reliability with DC 1mA and over on the wiper than multiturn. Just sub a steady resistor after measuring the trimmer value for chosen Vout in the final build. From your measurements I see that enough current runs on R5 when you employ some Vin-Vout margin as expected. So the single MPSA struggling before assumption, looks like holding water. Also we should be recommending 25V In-Out difference for 300V+ B+ regs. So cascode, we keep that for now. Lets see if it will work with IRF(P)9xxx 200V devices in the CCS too, presuming no output shorts or use of a Zener clamp.
Hello,
next I'll try the IRFP9240.
To make this tests I remuved the zener clamp on IRF840, I think that it was a problem.With it I burnt 2 IRFP9240.I don't know why, I check many times but there was a problem.
For the final application I've no problem, my Vin is about 375V for a Vout of 330V.Now I use my test Variable PSU that have a limit of about 340V.
Ciao
Guglielmo
next I'll try the IRFP9240.
To make this tests I remuved the zener clamp on IRF840, I think that it was a problem.With it I burnt 2 IRFP9240.I don't know why, I check many times but there was a problem.
For the final application I've no problem, my Vin is about 375V for a Vout of 330V.Now I use my test Variable PSU that have a limit of about 340V.
Ciao
Guglielmo
Good. It seems to be very steady. We sniffed its problem upon first try? Maybe. The cascode bjt driver stage that did not oscillate even, means more bandwidth also. Can have a subjective benefit even. That remains to be evaluated later. If good maybe we transfer it as standard for all voltages, one circuit up to 450V in. Lets see if the IRF 200V components for CCS can hold too. But leave the trimmer where it is, so it will not start beyond 200V dif, and be careful, don't make any output accidental shorts. Are Supertex DN TO-220 depletion easy to source somewhere in Europe? So if we get difficulties between 300-400V we may revert to that and have no Leds even. One more part in the driver, but less hassle in the CCS, remains simplistic. What Iset source resistor value it took you for the 10M45S to settle at 50mA on such Vin?
The Iset resistor for 50mA on 10m45 is about 17 ohm.
In my first tests with 9240 with the old circuit was stable at 50mA but when Vin rise (may I say more then 300V but I'm not sure because the voltmeter of my bench psu is broken and the max is about 340V) the current go rapidely to about 90mA.
But now I'll made a new test with the new shunt.
Ciao
Guglielmo
In my first tests with 9240 with the old circuit was stable at 50mA but when Vin rise (may I say more then 300V but I'm not sure because the voltmeter of my bench psu is broken and the max is about 340V) the current go rapidely to about 90mA.
But now I'll made a new test with the new shunt.
Ciao
Guglielmo
OK. Lets clear out the max in with IRF(P) devices so we know if we have a practical reliability limit. So the 300+ we are going to recommend will be one solid thing with certain CCS choice. We will wait so you will use it first a little on your 330V B+ application for subjective evaluation, reliability, or noise too.
If you see a bit behind, we opted for sharing the dissipation at such voltages between 2 BJT TO-92 PNP -300 to -400V devices than one, because I suspected that the weird problems guglielmope got when he was trying to set between 300-400V out, were due to the single MPSA 94 struggling or changing some parameter not manifesting in the simulator. So I suggested a cascode with 2 TO-92 driver devices. He did so with spare MPSA 94s he had, and it worked OK both with a depletion Mosfet 10M45S CCS and with enhancement mode as usual.
The circuit change is in that post.
The circuit change is in that post.
Hello Gugliemo,
Very nice, here is something for all working. To anyone having a depletion Ixys 10M45S or Supertex DN2540N5 (TO-220) I would suggest it on practical grounds for avoiding adding LEDS and having no fear of output or set trimmer mishandling because they can take 450V and 400V max respectively. They are not widespread available.
On the other hand the IRF(P)9xxx enhancement mode 200V PMOS devices are in everyone's fast reach in local shops, or in some parts box in the drawer, left over from some DIY. But they will not survive the whole max voltage on bad incidents. All the regs made by now use them, and they survive the time test when left alone with no accidental output shorts. A 150V 5W Zener across them (cathode on source pin, anode on drain pin) may help though.
I just hope that nothing changes in the subjective performance of the reg when substituting for depletion mode. This is for you to tell since you are the one who has the only 300V+ Simpler Simplistic build right now with both. I hope that the driver cascode has a positive impact too, beyond solving the practical problems in 300+ region.