high side IGBT simple resistor biasing.

[raymundor]

using a mosfet driven by a microcontroller, this circuit seems to work. I have been looking for similar examples around the internet and haven't found one quite like it. that leads me to suspect I am overlooking something. I have an IGBT equivalent in this circuit. I can see that it might be slow, but at 1khz it seems to work fine. any thoughts? the IGBT is an 40n60 with an input capacitance of 1075 pf

 

[FvM]

Not quite clear what you want to achieve, with 72V connected to the resistor foot point, the IGBT will never turn on.

Be careful when operating an IGBT in linear mode.

 

[KlausST]

Hi,

this circuit seems to work.

What do you mean with "work"?

What do you expect it to do? I see an input, but where is the output?
Does 72V and 92V make sense? These are quite unusal voltages.


Klaus

 

[dick_freebird]

Such a weak drive may allow misbehaviors like
local oscillation, or such a long switching plateau
that self-heating damages the device. IGBTs
and MOSFETs are generally meant to spend most
of their time in low-dissipation (on, or off) and
not at high I*V. You should look at the "Joule slug"
that attends a switching event to see whether you
violate any peak power or time averaged continuous
power limits.

 

[raymundor]

FVM, the output is the 72v source. so there is a differential of 20v. the zener diode keeps the voltage to 12V through the 10K resistor. the other resistor pulls that to ground weekly so the voltage remains around the 72v level when the mosfet is switched on.

dick_freebird, the time constant for the gate is 1.07E-5 seconds being that the switching frequency is 1khz, that's about 1% of the switch time. does that seem to be spending too much time in the switch event? the package is the TO3P, actively switch about 20-30V differential and about 20A max

KlausST, these are the voltages expected to switch against on the high side.

thanks for your comments.

 

[Easy peasy]

with 72V connected to the resistor foot point, the IGBT will never turn on.

aaah, yes it will, when the driver mosfet turns off - but it will be biased only just above linear mode - so may well get hot.

 

[FvM]

Yes, you are right. Need to check if the IGBT can stand power dissipation.

 

[raymundor]

FvM, the resistors don't behave like a divider. the Zener diode wont let the voltage go below that the IGBT is at. and so it wont go below a little less then 72v. when the mosfet is off the top 10k biases the voltage to supply, and the zener holds it to 12v , or 14v depending on the zener selected. (with respect to the 72v of the emitter.) so about 84v with respect to ground.

--- Updated Aug 19, 2021 ---

this is the actual circuit for clarity, the circuit above had a IGBT equivalent for Falstad simulation.

 

[Easy peasy]

For the above image, when Q is on, the gate will attempt to sit at 92V/ 2 = 46V

thus D1 will be forward biased and carry current from the 72V batt to R2, about 7.1mA, thus the IGBT will be off

when Q is off, R1 tries to lift the gate to 12V, but if the 92 volts is solid there will be a huge current flow to the 72V batt,

if the 72V batt has some series R that limits the current, the the emitter if the IGBT will rise up to about 5V below the Collector - which is enough to hold the IGBT just on, as Vge will be the same - i.e. the threshold voltage of the device...

This is not a good circuit ...

 

[raymundor]

Easy peasy, I don't really doubt its a bad circuit (which is why I posted it here) however in your explanation, how can the resistor network settle at 46V when current is flowing forward through the zener towards R2 (conventional + to - visualization) from 72V? there is a voltage across the zener of 26V, and so the emmiter would be lifted to 72V as the resistance path is much less from 72v then 10K from the 72 to ground. The current flow when Q is off is 0.002A its basically a standard Zener voltage regulator or reference circuit at that point.

here is a simulation of what I am saying. for clarity I have omitted the mosfet and IGBT, and just addressed the behavior of the voltage divider. I show it in both states.

I also don't understand what you mean by the huge flow to 72V when R1 lifts the voltage to the Zener voltage of the diode.

 

[Easy peasy]

please improve your English, note the word attempt ....

--- Updated Aug 20, 2021 ---

if the 92V and the 72V are true volt sources, large currents will flow

 

[FvM]

There are no doubts that the IGBT is turned on and off in the circuit, the details your are focussing on in post #10 are not really important, e.g. if the gate node goes to 46 or 71 V. Problem is that the IGBT will be most likely operated outside its safe operation area, with or without a 1 ohm current limiting transistor.

Vge at Ic of 10-20 A can be expected around 8 to 9V according to datasheet.

 

[raymundor]

Easy peasy, sorry I missed that (English is my second language, my first being grunts and giggles).

I see what you mean, the voltage drop a crossed the IGBT will be essentially zero and since the drive voltage relies on that difference it will immediately settle at the threshold voltage of the gate.

thanks.

 

[Easy peasy]

The volt drop across the device will be about the same as the Vge...

If you try linking any two voltage sources by a low impedance - what current do you get ?

--- Updated Aug 20, 2021 ---

.. many concepts of power electronics are hard to grasp - esp linear operation of devices when you may be expecting something else ....

 

[raymundor]

A high current, in this case the impedance of the voltage sources are not very low. the battery bank wont behave like a low impedance source. and the solar panels will drop off quickly on their voltage at high currents. short circuit current being about 15A per panel in parallel. so in the practice of this circuit and its application the maximum current will be around 20 or so amps. in its current configuration. I am just going to do what I usually do and use a high side chip and IGBT. It will handle the current just fine.

 

[Easy peasy]

a far better idea than a made up linear ckt that will only cook the igbt.