I was in fact not thinking too far ........... 😱
But how fast do you think it switches off in practice ?
Regards
Charles
He saw "Toff=1100ns" in ZTX851 datasheet, but this is actually the storage time for saturated operation at Ic=1A, Ib1=0.1A, Ib2=-0.1A. This is common emitter switching and has nothing to do with emitter followers. Switching performance of bipolar transistors depends a lot on the way the base is being driven. There is not a fixed 1100ns turn-off delay as he mistakenly understood.
For common emitter operation, the Ft figure is more useful, being 130Mhz for ZTX851 (F=50Mhz, Ic=0.1A, Vce=10V), which means that at 130Mhz base current will contribute as much as collector current to emitter current, in other words, this transistor still provides a 2x emitter follower current gain at 130Mhz (and it tends to rise at 6dB/oct below that).
For common emitter operation, the Ft figure is more useful, being 130Mhz for ZTX851 (F=50Mhz, Ic=0.1A, Vce=10V), which means that at 130Mhz base current will contribute as much as collector current to emitter current, in other words, this transistor still provides a 2x emitter follower current gain at 130Mhz (and it tends to rise at 6dB/oct below that).
My favorite:
ZXTP25020DFL: 20V,1.5A/6A, ft=290MHz, hfe=450
Also nice:
ZXTP25020CFH
and for higher voltages and higher dissipations
FZT953
But in general take care when using PNPs for improved pull down. If you run the turn ON current through the resistor which is usually connected in parallel to B-E it may happen that you bring the base-emitter junction to break through. Consider the output of IR2110 jumping to high within 20ns, while the MosFet Ugs is slowly moving up within 200ns ...
This effect will usually not cause a defect, but it might change the turn ON sloping speed, because effectively your are paralleling a zener to the turn ON-gate-drive-resistor. Most small signal transistors break through at inverse Ube around 5V...7V.
...with quite some tolerances, so some less mature designs might show some hard to predict variations in the range of several 10ns regarding the turn ON behavior.
ZXTP25020DFL: 20V,1.5A/6A, ft=290MHz, hfe=450
Also nice:
ZXTP25020CFH
and for higher voltages and higher dissipations
FZT953
But in general take care when using PNPs for improved pull down. If you run the turn ON current through the resistor which is usually connected in parallel to B-E it may happen that you bring the base-emitter junction to break through. Consider the output of IR2110 jumping to high within 20ns, while the MosFet Ugs is slowly moving up within 200ns ...
This effect will usually not cause a defect, but it might change the turn ON sloping speed, because effectively your are paralleling a zener to the turn ON-gate-drive-resistor. Most small signal transistors break through at inverse Ube around 5V...7V.
...with quite some tolerances, so some less mature designs might show some hard to predict variations in the range of several 10ns regarding the turn ON behavior.
Well, this was my idea of the implementation of active PNP-turn off, is anything wrong with it?
The resistor allows me to control the turn-on current and hence the turn-on speed, directly driven from the IR 211x chip.
ChocoHolic, you have confused me a bit with the B-E parallel resistor you referred to...¿¿??
The resistor allows me to control the turn-on current and hence the turn-on speed, directly driven from the IR 211x chip.
ChocoHolic, you have confused me a bit with the B-E parallel resistor you referred to...¿¿??
Attachments
Hi Pierre,
some designs do not use this diode. Some people simply would use that 10 Ohms resistor in place of your diode. In that case the PNP transistor would see reverse Ube during MosFet turn ON.... Usually even that is not troublesome, I also played around with that simplified circuit and did not get trouble. As long as the base emitter diode does not break through during MosFet turn ON it would act like your circuit. If the Ube diode would break through, then the turn ON would be slightly faster.
The circuit, which you are using is perfectly correct from my point of view. The diode is limiting the reverse Ube to an uncritical level and the 10 Ohms are defining the turn ON speed ...but of course you spend one diode more...
some designs do not use this diode. Some people simply would use that 10 Ohms resistor in place of your diode. In that case the PNP transistor would see reverse Ube during MosFet turn ON.... Usually even that is not troublesome, I also played around with that simplified circuit and did not get trouble. As long as the base emitter diode does not break through during MosFet turn ON it would act like your circuit. If the Ube diode would break through, then the turn ON would be slightly faster.
The circuit, which you are using is perfectly correct from my point of view. The diode is limiting the reverse Ube to an uncritical level and the 10 Ohms are defining the turn ON speed ...but of course you spend one diode more...
Hi Pierre,
So what type of results are you getting with this PNP turn-Off scheme....Have you considered using schottky like IN5819 in place of diode...Is the gate driver is cool now and what is the gate charge of mosfet you are using....what are the rise & fall times you are getting with this.
regards,
Kanwar
So what type of results are you getting with this PNP turn-Off scheme....Have you considered using schottky like IN5819 in place of diode...Is the gate driver is cool now and what is the gate charge of mosfet you are using....what are the rise & fall times you are getting with this.
regards,
Kanwar
Sorry,
Didn't get too much improvement. Have to measure the temperature of the driver chip with and without the PNP active turn off, but I think it is not too different.
Didn't get too much improvement. Have to measure the temperature of the driver chip with and without the PNP active turn off, but I think it is not too different.
Pierre said:
I have got an improvement in both driverchip temperature as well as less shoot-through for a given deadtime with active PNP.....😉
- Status
- Not open for further replies.