[ disclaimer: I'm not doubting of what you say, I'm just trying to learn
something from you ]

At 11.16 20/03/2004 +1300, you wrote:
>> > I am in urgent need of some help. I need to switch a 160V DC @150mA
>> > signal (hopefully with PWM around 100Hz) using a PIC I/O. The solution
>> > I had, which I have used for up to 50V DC @ 5A was like a 4th of July
>> > celebration when hooked up to the 160V DC.
>
>> > http://www.geocities.com/researchtriangle/lab/6584/high_v_switch.bmp
>
>> Come on guys, take a look at the circuit again. The high side switching
>> device Q2 is P-channel MOSFET. What gate voltage is being applied? When
>> Q1 turns on it will apply -160V to the gate of Q1 relative to the Source
>> terminal. You can't do that. With a Vgs absolute max spec of +/-20V no
>> wonder it complained!
>
>
>Brent is right. I'm amazed that this would have worked on even 50V as is.
>
>> you could add a 10 to 15V zener to Q2, cathode to Source, anode to gate.
>> Leave R11 there as the turn off resistor, you will then need a series
>> resistor to Q1 drain to limit the zener current.
>
>As he's in a hurry let's add some values.
>You MUST have a resistor from Q1 drain to Q2 gate.

Why? I thought that gate MOSFETs act like capacitors, and do not draw
current (besides the one necessary to charge the capacitor).

Assuming that the max gate voltage rating is not exceeded, what's wrong if
one quickly grounds (or, on the opposite, quickly puts the highest acceptable
voltage) on the gate?

The MOSFET will switch faster, which should be a good thing, not?

What am I missing? Why the gate resistor on Q2?


>100k should be fine but higher may be OK depending on PWM speed.
>A 100k will dissipate about 0.3 watt when 100% on so make it at least 1/2
>watt rated. Preferably more.

But how can the gate resistor dissipate power when the MOSFET is 100%
turned on? Doesn't it draw current only when charging or discharging
the gates' capacitance?

Are we still talking about this schematic?
http://www.geocities.com/researchtriangle/lab/6584/high_v_switch.bmp


>A 1 M may be OK.
>Note that at 160v you are approaching the voltage rating of SOME resistors.
>make sure you use a part that's OK at 160 V.
>
>> For a measure of peace
>> of mind add a series gate resistor to the gate of Q1, so that if it fails
>and gets
>> 160V on the gate the current back into the PIC would be limited. Say 22K
>to
>> keep fault current below 10mA for the PIC. R10 would have to increase.
>
>R10 would have to be MUCH higher than the 22k so the gate voltage is not
>divided down by more than a small amount.
>Say 220k. Q1 (type unspecified on diagram) MUST be a "logic level" part that
>switches with 5v gate drive. As it worked on 50V you should be OK.
>
>R11 47k is unnecessarily small. You will dissipate 160^2/47k = about 1/2 a
>watt. Size accordingly or increase resistance for 160v operation. Also has
>160V across it when Q1 is on.
>
>> An easier circuit would be a single N-ch MOSFET doing direct low side
>> switching if the load permits.
>
>IF you can stand the LED string NOT being grounded at the bottom end this is
>a MUCH better choice.
>Only 1 FET, N Channel are easier to get in high voltage ratings, no high
>voltage gate switching, cheaper, simpler.
>I would only go to a P channel high side FET if it was essential
>
>
>
>        RM
>
>--
>http://www.piclist.com hint: The PICList is archived three different
>ways.  See http://www.piclist.com/#archives for details.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body