Electronics > Projects, Designs, and Technical Stuff
Searching a high-side driver ic for a p-channel MOSFET, do you know any?
Ian.M:
@Sourcecharge,
The problem with Zener clamping is excessive junction capacitance. Its not a major problem if you are directly clamping the MOSFET gate as its almost invariably swamped by the gate capacitance, and its effect is limited by the highish drive currents required to charge/dicharge the gate, but if you need to clamp a higher impedance node earlier in the driver circuit to limit the gate voltage without excessive dissipation in the driver circuit, having what approximately looks like 75pF in parallel when you add a small 12V Zener like the BZX84C12L is *really* not helpful if you want reasonably fast transitions.
See attached LTspice sim of Zener clamping and of simple limiting by pulse current control through a known resisto, which also has (near) ideal clamping and the same with 75pF in parallel for comparison. It also shows the vital importance of having a low enough resistor in parallel with Zener to discharge it if the drive is unidirectional.
There are workarounds e.g keep the Zener continuously biassed to provide a clamping rail then use a fast low capacitance diode to clamp to it, but they all add significant complexity.
Jajaho:
For me here in Germany a new day of work has started and I'm just blown away by your dedication to this issue, thank you all very much.
Spezial thanks to Ian for taking the time to explain his cicuit.
The capacitance of the zener is also a thing to keep in mind.
I will take every post here into consideration and once my circuit is done I will upload it here.
Best regards.
Jajaho:
Hello everybody it's me again, reviving this thread.
I have finfished my schematic aswell as the layout (except for some minor things ofc.)
I would really appreciate it if some of you guys could review it and give me your two Cents about it.
Thank you all in advance.
Greetings from Germany
PS: As I mentioned earlier in this thread this sucker has to switch a maximum of 10A at a maximum PWM frequency of 10kHz. Altough be it for a short duration (say maybe 1 sec), I undoubtably will need some beefy caps to support this switching action. I'm asking if anybody has any experience values he/she can recommend or alternatively an actaul way of determining the required capacitance.
Thank you very much.
Jajaho:
Hello to anybody who still gives a,
I have now ordered the pcb's posted above and I am expecting them to arrive in 2-4 days.
On the other hand I am still intriged by Ian's circuit posted earlier and so I'm giving that one a shot too.
Due to a limited choice in distibuters I am forced to swap the matched 2N3904 pair for something else.
The choice I'm currently planning on going with is the BCM847BS.
I looked through the datasheet and found it to be a fitting replacement.
I wouldn'T mind if anyone could double check that however.
I'll be shure to post my post my results here.
Greetings from Germany.
Ian.M:
Matching is fairly non-critical, so you don't need a precision matched pair. All parts in my design were fairly generic so there shouldn't be any problems with substituting broadly similar transistors. The BCM847BS Vceo and Ic ratings look suitable. At 32V, in my circuit, it will dissipate 100mW during the PWM on time, which is well within its ratings. and its certainly preferable to hand selecting individual 2N3904 or similar to match them then having to improvise some sort of thermal coupling for the pair.
N.B. it absolutely needs *GOOD* decoupling, as close as possible to the complimentary driver stage collectors and the output stage emitters. Remember, its trying to throw sub 1us 800mA current pulses around (if you don't tame it by deleting the output stage).
Edit: *DON'T* attempt to decouple the driver collectors as that would FUBAR the output stage base drive!
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version