Yes, parasitic capacities, inductances etc. can make things worse when building it for real.
But a model from the manufacturer for a recent MOSFET will not be so bad that it is not possible to simulate a circuit like this regulator correctly. Spice is used widely in the industry. If it would not even allow simulating this rather simple circuit (no offense, I could never design myself), then it would be completely useless.
So, if a MOSFET does not work in the circuit in spice, I think this can be caused by the circuit reacting less robust / stable for some of them.
To not choose one which makes this trouble in the "real world", I would rather select one behaving in Spice at least.
But a model from the manufacturer for a recent MOSFET will not be so bad that it is not possible to simulate a circuit like this regulator correctly. Spice is used widely in the industry. If it would not even allow simulating this rather simple circuit (no offense, I could never design myself), then it would be completely useless.
So, if a MOSFET does not work in the circuit in spice, I think this can be caused by the circuit reacting less robust / stable for some of them.
To not choose one which makes this trouble in the "real world", I would rather select one behaving in Spice at least.
Ahh, you're an expert on circuits and parts modeling? That's good to know.
Anyway, it's not a matter of 'it doesn't work', it's a matter of correctly simulating the impact of unmodeled properties and parasitics.
Spice will do whatever you want, it's the models that are not fully complete, as well as the PCB and construction parasitics that you need to input!
But hey, it's your project!
Jan
I basically just said "the model of the circuit in my spice now is simpler because it is not having parasitics modelled and thus more likely should behave better than the real thing later". You think that is wrong?
You can't really say that. Parasitics can work any way.
Mostly cause instability but also just cure it.
It's a toss up.
Jan
Mostly cause instability but also just cure it.
It's a toss up.
Jan
I have now finished Jan's T-Teg5 with my own layout (just a little bigger than Jan's original, but I like to do my own if I can).
In my test setup, using a variac as input, followed by just a bridge rectifier and a 100u cap, I met the targeted 175V almost on the spot (I used spice to find the right voltage-defining resistor at 50% pot setting for 175V).
The regulator is intended for a 300B B+ supply at 425V, so the 175V was just a test, but I am confident that will work too, given the flawless operation at the test voltage.
Regarding MOSFETs, I have tested with IRFPG30PBF, as well as with IRFPE40PBF, both from Vishay. Both are working perfectly without any change, like the Zobel circuit I had used in my sims (see some posts above).
Thanks, @jan.didden for publishing the circuit.
In my test setup, using a variac as input, followed by just a bridge rectifier and a 100u cap, I met the targeted 175V almost on the spot (I used spice to find the right voltage-defining resistor at 50% pot setting for 175V).
The regulator is intended for a 300B B+ supply at 425V, so the 175V was just a test, but I am confident that will work too, given the flawless operation at the test voltage.
Regarding MOSFETs, I have tested with IRFPG30PBF, as well as with IRFPE40PBF, both from Vishay. Both are working perfectly without any change, like the Zobel circuit I had used in my sims (see some posts above).
Thanks, @jan.didden for publishing the circuit.
Last edited:
Correction, copy-paste error. I have successfully tested with IRFPG30PBF and IRFP460BPBF. For testing purposes I had also bought IRFPE40PBF, but did not test with that yet.
Also, be aware the above are TO-247, which will not fit on Jan's original board, but there might be identical with the footprint Jan has used.
Last edited:
Yes TO-247 does not fit on the PCB but the pinout is there on the edge so you can mount it on a heatsink and solder directly on the PCB.
I thought about that ;-)
Jan
I thought about that ;-)
Jan
I have built another T-Reg using the IRFPE40PBF this time and it also works fine.
Moving to the real application now (300B mono, one reg per side, 425V, 80mA idle), what is the recommended raw DC circuit? 1000u capacitance after the rectifier makes no real diff to omitting it in Spice.
I guess it might make still sense to lower the AC part in the wires going to the regulator? Any other thoughts?
I guess it might make still sense to lower the AC part in the wires going to the regulator? Any other thoughts?
The changes before the T-reg have very limited effect.
Probably not worth the effort/cost.
Jan
Probably not worth the effort/cost.
Jan
Hi @jan.didden , thanks for your response.
I did not mean if this would positively affect the noise or absence thereof on the output of the regulator. More thinking out loud if reducing the ripple on the wires between transformer and T-Reg from about 1V to 100mV (at least it is like that with a 1000u cap added in LTSpice) would be beneficial in not spreading hum in general.
If the transformer is very near to the regulator, then this is probably a non issue anyway. I am still doing the case layout though and do not exactly know the distance between transformer and regulator. Of course I will try to have them near together.
I did not mean if this would positively affect the noise or absence thereof on the output of the regulator. More thinking out loud if reducing the ripple on the wires between transformer and T-Reg from about 1V to 100mV (at least it is like that with a 1000u cap added in LTSpice) would be beneficial in not spreading hum in general.
If the transformer is very near to the regulator, then this is probably a non issue anyway. I am still doing the case layout though and do not exactly know the distance between transformer and regulator. Of course I will try to have them near together.