Controlling Rate of H-Bridge Switching.

Thread Starter

james31207

Joined Sep 19, 2018
72
I have an H-Bridge required to allow reversal of polarity in an electromagnet at 4 Hz. In particular, the specifications state 600 V, 80 A using IGBTs; in a situation in which only 50 A is delivered by a 12 V DC source. I have two questions: first whether such an H-Bridge is appropriate; and second, whether a simple 555 IC timer can be wired to it to cause the desired rate of switching.

I have a general idea of how a 555-timer works to deliver an output pulse to the IGBTs; but am unsure whether on its own it is capable of triggering the necessary effect in such a case.
 

danadak

Joined Mar 10, 2018
4,057
The electromagnet is highly inductive, so using a HV part appropriate for
margin, even though H Bridge may have clamp diodes to handle L load
disconnect transients. Although I must admit 800V seems overkill in a
12V system.

No on just a 555 timer. You need a gate driver as well to handle the high capacitance
load the gates present to the driver, and prevent thermal problems switching the IGBT
too slowly.

Another issue are you reversing the polarity in the electromagnet ? At a high switching
rate ? If so you need to implement dead band in the H Bridge to insure the top and bottom
transistor in a vertical leg of the H Bridge are not on at same time, eg. present a short to power
supply. Most modern processors have PWMs and dead band settings.



http://www.ti.com/lit/pdf/slua618

https://www.fujielectric.com/produc...pplication/box/doc/pdf/REH984e/REH984e_07.pdf

https://www.ixys.com/Documents/AppNotes/IXAN0010.pdf


Regards, Dana.
 

Thread Starter

james31207

Joined Sep 19, 2018
72
Thank you Dana. The documents are very useful, especially the last one on IGBT gate drivers and so on. Actually, I'm connecting 6 EMs in parallel, each with an inductance of about 0.1 H, so total inductance 1/6 x that. I've calculated that for di/dt of 500 A/second, the back EMF through the H-Bridge controlling the switching for the whole 6 will be less than 10 V. I can't find H-Bridge modules rated for 50 A (drain current) with less than about 600 V drain-source voltage.

Also, I'm aiming to switch polarity in the EM at only 4 Hz; but it may be that the IGBTs in the H-Bridge are probably capable of handling the heat at that low rate of switching provided the correct gate-driver is used.

I had also thought to try a reversing contactor assembly as an alternative method, but am unsure whether switching polarity in the internal coils of such devices can be controlled by a 555 IC timer. Thanks again.
 
Last edited:

MisterBill2

Joined Jan 23, 2018
18,167
Can you explain what would be driven by such powerful reversing magnetic devices? This sounds like an interesting application. Also, it sounds like one from a couple of months ago.
 

Thread Starter

james31207

Joined Sep 19, 2018
72
Same one Bill. I'm up to the part where I need to acquire the correct H-Bridge module, since I can't possibly build one myself. They're quite expensive [Vincotech], so I need to understand how the transistors function better than I do--but I'm making progress. If you have any specific information about a particular H-Bridge capable of handling 50 A (since there are 2 parallel circuits, and I now know that my 12 V car battery delivers only 90 A), it would be gratefully accepted.

One of the problems might be the switching time, which is around 200 ns for the H-Bridge I'm considering; since at 50 A, even though the total inductance of my coils is very small (~0.02 H), implies a back EMF of thousands of volts. The switching time would have to be slower--or slowed. Any suggestions?
 

Attachments

MisterBill2

Joined Jan 23, 2018
18,167
The first way to reduce the inductive spike voltage is indeed to slow the switching speed a bit, BUT that puts the switches in the linear mode for more time so that they dissipate a lot more power. So snubber resistor in series with a Varistor could also absorb a bunch of that energy.
Then another thought comes to mind is to consider driving the coils with a sine wave instead of a square wave. No big spike, and probably a simpler control scheme. I don't know if that would work in this application.
And one more option would be to use transistors with a rating high enough to withstand the spikes.
 
Top