I've tried to make drivers for my H-bridge circuit using complimentary emmiters but the square wave outputs runat a constant 25v instead of +-10v square wave

what have I done wrong?

 

The image is a bit hard to read but it looks like your mosfets' gates and sources are wire together (or gate connection is missing). also you have 8 voltage sources - you might want to reduce that number.

edit: you might want to make your schematic more compact, save at a higher resolution and crop out the unnecessary stuff (like anything but the schematic) the image is bigger and more readable.

 

New image of circuit

 

Whils tinkering with the simulation I found that the reference voltage for the comparators was too high

setting it to 5v solved the problem but now LTspice crashes when I run the simulation

have I constructed the gate drivers wrong?

 

The high side drivers and their gate drivers (entire top half of the bridge) won't work.
What are you trying to drive? I doubt that it really has a resistance 0f 0.0001 ohms.

If you post the .ASC file for your circuit, I might play with it and see what I can come up with. You will have to either ZIP it, or change the extension to .TXT so that the forum will accept it.

Moderators - Can you get .ASC files added to the list of acceptable extensions? SwitcherCAD III is fairly popular here, and this issue has come up before.

__________________
Ron

 

know a good site for hosting zip files?

 

You can post it here. Just "Go Advanced" and then click on "Manage Attachments".

__________________
Ron

 

user1453, are you really sure the upper half of the H-bridge does what you think it does? In order to turn the mosfet on, you'll need to supply 50V + 12V to the emitter followers. At present, Q3,Q4,Q7,Q8 are only ever going to see at most 12V. If this were a real circuit, those mosfets would've probably blown their gates - followed by some more smoke.

Look up voltage translation circuits. Open-collector/resistor pull-up types are probably the kind you want, although AC coupling via capacitors/transformers would also do.

One side note - you're probably also better off using as many "ideal" spice components as possible - AC voltage sources for the (I think it's an oscillator, but I haven't built any in a long time), and get rid of the voltage divider op-amp too. Also get of unnecessary voltage sources - especially floating ones- in the circuit. It makes it much harder to read and isn't very realistic.

James

 

James, I started to explain all that, but concluded that he was probably never gonna get where he wants to be without some hands-on help. Even if he got the drive levels correct, he would most likely have problems with shoot-through due to not paying attention to drive overlap (lack of dead time).
I just decided (as you have seen) to ask for the file. Maybe he would learn more by letting him smoke some transistors, and ask a lot more questions, but I wonder if we or he has the patience to see it through...
I agree with you about cleaning up the schematic. At the very least, he could replace the oscillator and the inverting amp with voltage sources, at least in the beginning, thereby saving complexity and simulation time.

__________________
Ron

 

I think an easier design would be to use pmos fets for the high side drivers. they aren't that expensive these days.

 

The reason the op amps and comparators on the left exist is to stop shoot through

The components generate non overlaping two square waves with variable duty time 0-50% roughly

The final frquency of the pwm signal will be about 20khz I slowed it don abit to make the simulations quicker

What I was trying to do with the gate drive was +12v on -12v for off

heres the file I've removed two of the voltage sources and raised the voltage of the remaining one to 230v that is what the voltage will eventually be set to

Attached Files

pwm circuit.zip (1.8 KB, 14 views)

 

You posted the (.RAW) waveform file. As I said, you need to post the .ASC file.

__________________
Ron

 

thos the one?

Attached Files

pwm circuit.zip (3.4 KB, 21 views)

 

That's it.

__________________
Ron

 

Okay, I'll just play peanut gallery on this one. It's just that trying to figure out why Spice nukes itself is pretty annoying work. I've never had that many floating voltage sources before, and even though I didn't see any obvious conflicts with them, I can't say it inspires a lot of confidence.

James