H-bridge driver for a 90VDC motor

I want to build an H-bridge driver for a 90 VDC motor, just like this one.

​

I've designed a (possibly quite clumsy) circuit in LTSpice for this effect, based on a previous design that @ronv was kind enough to help me with. This circuit will not be driven from mains, but rather from a 1:1 isolation transformer working at 120VAC/60Hz, so it should comply with the TOS.

​

The circuit is running at 1.25 kHz, and only M1 and M4 are being driven in the sim. M4 is doing the switching, since I'm under the impression that nFets are better suited than pFets for this purpose.

Questions:

• Is it normal that the rectified voltage accross the motor's coil never reaches zero?
• Is my design more or less passable, or are there important changes/improvements that should be made?
• @MaxHeadRoom, you may be able to help me determine the correct coil resistance for the motor (since I don't have it with me at this moment) and perhaps even its inductance. Maybe you know ballpark figures for these values?

I've attached the fet's datasheets, as well as the one for MPSA42 (a high voltage transistor) for reference.

 

Can you post the .asc file?
Driven by a micro?
Leisure time between reversals?

 

The asc is there already in my first post. And yes, it'll be driven by an mcu.
Lag time between reversals is quite long. In the order of several seconds, actually.
Thanks in advance for all your help!

 

I have used a Pic ECCP module, and the fig7 in http://tahmidmc.blogspot.ca/2013/01/using-high-low-side-driver-ir2110-with.html
With VCC 15vdc.
The armature resistance would typically be around 1Ω to 2Ω for that motor.
There will be a generated BEMF from the motor that opposes the applied DC, current will be a mean level dependent on load.
Max.

 

Hmm.
I'm guessing the inductance is quite a bit higher - maybe .5 H?
The zeners need to be changed to N=2 so they don't conduct under normal conditions.
I'm not crazy about the PFETs on top. There is no gate drive when the 120 volts is low. Might fix this with a couple of caps on the 120 volt line, but it might be better to use NFETs all around with a gate driver.

 

I agree with Ron, I am not comfortable with Pfets on the top. The reason is, there is a big difference in the Ron resistance between the P and N type Mosfets. I have been using the bootstrap type of high side Mosfet gate drivers with a lot of success , LM5104, 100 volt, but the UCC27714 might interest you.

 

The sim has M4 switching at 12.5kHz.

 

You're right. M4's period is 80us.
1.25kHz was a typo.

 

What μP are you thinking of using? I used 5khz switching for a 2hp 90vdc motor.
Max.

 

Atmel's AT89LP4052... I'm in love with that little sucker. It's extremely easy to use, has a single-cycle core with configurable I/Os, and of course, it also has an 8-bit PWM, among other things.

Thanks for the suggestion Les... I've tried to use mosfet gate drivers in the past, and I've failed miserably... but back then I didn't have you guys to go to if I ran into trouble. So I'm going to give your suggestion serious consideration. The UCC27714 looks a lot more viable for my application than the LM5104, since it can handle up to 300V, and my motor's being fed rectified 120VAC rms, which peaks at almost 170V (that's 120/sin(45°), is my calculation correct?)
Anyway, I was hoping to use the components shown in my first post, since I have them readily available. But both you and ron are right, pFets have a much higher equivalent resistance than nFets, and are probably not the best option for this application. How hard would it be to design a high-side nFet driver for this purpose?

So, if I were to use the UCC27714, would things be as simple as the diagram shown in its datasheet?:

​

I'm guessing I'd have to add a few TVS diodes here and there, to protect the circuit and make it more reliable, right?

It's too bad that this chip only comes in SMT... I try to avoid SMT chips at almost all costs, since I find them hard to install. I have no problem with SMT caps and resistors, and even tiny diodes. But chips definitely get on my nerves. But hey, I gotta face the fear and move forward, right?

 

Here is one I like.
http://www.irf.com/product-info/datasheets/data/ir2184.pdf
It has built in dead time (leisure time) so you don't have to worry about the top and bottom FET being on at the same time.
I think I may have given you models of this family. If not I have them.
The one thing you have to watch out for is that the duty cycle can't be 100%. It needs some time to recharge the bootstrap cap.

 

Here is a bootstrap cap calculator http://www.silabs.com/support/Pages/bootstrap-calculator.aspx
Just use a DC+ve above the motor rated by ~10% and you don't need to go to 100% duty cycle to obtain motor rated rpm.
Max.

 

I have a few IR2301 laying around... would they work?

 

That seems to be an invaluable resource that I hadn't heard of before... thanks!

 

Should work.

 

I guess they should.... but I tried using them before and I failed, as I said earlier...
Gonna try again, this time using Max's bootstrap cap calculator link... but before I do, I'm gonna draw a diagram, to make sure I got things right

 

Just in case.

 

Excellent! I'll draw the schematic tomorrow, and see how far I can get. Then I'll come back here with the results.
Thanks!

 

Ron, LTspice is notifying me that it can't find the IR21xx.sub file. I've checked, and couldn't find it in the zips you posted. And I haven't been able to locate it in Google either.

 

Found it!... you attached it last year in this post.