Help with full bridge driver

Thread Starter

Krej

Joined Aug 29, 2008
7
Hi
I am struggling on a MOSFET h bridge driver for my inverter. I stumbled upon this forum and a circuit that I have attached and was wondering if it would work. I just want to be sure, because I am tired of blowing up drivers and MOSFETs.
There would be two of the stages shown in the diagram.
Thanks in advance for any help I get.
Jeofrey.
 

Attachments

mik3

Joined Feb 4, 2008
4,843
Hi
I am struggling on a MOSFET h bridge driver for my inverter. I stumbled upon this forum and a circuit that I have attached and was wondering if it would work. I just want to be sure, because I am tired of blowing up drivers and MOSFETs.
There would be two of the stages shown in the diagram.
Thanks in advance for any help I get.
Jeofrey.
Where is the motor connected? As i can see form the diagram the MOS will burn out because they are shorted (wire from drain of Q33 to drain of Q4).This wire must be removed and you have to connect the load there. At Vcc you will connect your control signal?
 

Thread Starter

Krej

Joined Aug 29, 2008
7
Hi,
I have changed the circuit and made it easier for your understanding. Hope the circuit is clearer now. Will this work?
The 2 points marked G1 are the signals to the driver.
Jeofrey.
 

Attachments

Thread Starter

Krej

Joined Aug 29, 2008
7
Hi,
I have changed the circuit and made it easier for your understanding. Hope the circuit is clearer now. Will this work?
The 2 points marked G1 are the signals to be given to this driver.
Jeofrey.
 

SgtWookie

Joined Jul 17, 2007
22,210
Where is the motor connected?
To me, it looks like the load would be connected to the unmarked terminal on the far right, just above the drain of Q33. This is an inverter application, not a motor driving application - however, still highly likely that the load would be inductive.

As i can see from the diagram the MOS will burn out because they are shorted (wire from drain of Q33 to drain of Q4).This wire must be removed and you have to connect the load there. At Vcc you will connect your control signal?
I don't see what you are talking about - I don't see a Q4 anywhere? :confused:

Q6 and Q33 are in parallel on the low side; Q5 and Q35 are in parallel on the high side. Q9 and Q ? (no number) control the gates. No clue as to what an A42 is (Q9) but I suppse a 2N2222 might do.

The big objection I have about this circuit (prior to experimenting via SPICE) is that there is no positive control to ensure the low-side MOSFETS are turned OFF before the high-side MOSFETS are turned ON, or vice-versa. The shoot-through would likely fry the MOSFETS in short order.

Additionally, there are no reverse-EMF diodes to protect the MOSFETs. Since the load is very likely to be inductive, the first time current flow to the load was cut, the reverse EMF would destroy the MOSFETs. The diodes would also need small ceramic caps across them to buy time for the diodes to turn on.
 

SgtWookie

Joined Jul 17, 2007
22,210
Oh, OK - I didn't see your 2nd schematic.

It is very hard to read the text in your schematics. Perhaps you are using bold fonts?

If you could turn off the bolding and save the image as a type of .PNG, that might make it a good bit more clear.
 

Thread Starter

Krej

Joined Aug 29, 2008
7
Oh, OK - I didn't see your 2nd schematic.

It is very hard to read the text in your schematics. Perhaps you are using bold fonts?

If you could turn off the bolding and save the image as a type of .PNG, that might make it a good bit more clear.
Yes I will do that. But is the circuit going to work?
Jeofrey.
 

SgtWookie

Joined Jul 17, 2007
22,210
Yes I will do that. But is the circuit going to work?
Jeofrey.
I'm having a hard time reading what your components are. If I don't know what they are, I can't properly evaluate it.

Even without knowing what the values are, I don't see reverse-EMF protection diodes for the output, nor do I see any provisions for ensuring that there is "dead time" between the high and low side drivers.

Either problem will cause a quick death of your MOSFETs.
 

SgtWookie

Joined Jul 17, 2007
22,210
Better now?
MUCH better! I can actually read it now! :D
[eta] However, you've removed the reference designators! :confused:

(Reference designators are like R1, Q3, C1, etc. These are shorthand names to make parts easy to refer to)
I noticed that you have changed the values of quite a few components. I don't know what a UF4007 is; looks like they were 1N4004's beforehand (best as I can make out). [eta] OK, now I know about them; fast versions by Fairchild. ;)
 
Last edited:

Thread Starter

Krej

Joined Aug 29, 2008
7
Yes, I have changed the values a little. They were 1N4004s before, but I thought that they should be fast recovery diodes, so I changed them to UF4007. A UF4007 is an ultrafast recovery diode, 1000v 1A 75ns.
As for the reverse EMF protection - the MOSFETs contain anti-parallel diodes in them, right?
The deadtime - the signals are coming from an SG3526, where deadtime is already set, so I think they should not be cross-conducting. If I am wrong, please stop me.
Thanks.
Jeofrey.
 

SgtWookie

Joined Jul 17, 2007
22,210
The BC337 transistors will disappear with a loud BANG and a bright flash with a Vce of 280V. :eek:

Why not use an H-bridge driver IC? Look at something like an L6384 high-voltage H-bridge driver from ST Microelectronics. Many other manufacturers make H-bridge drivers, too. This one has built-in provisions for adjusting/setting the dead time, too.

They're about $3/each if you buy them singly. It will save you space, and lots of agonizing trying to work it out yourself.
 

SgtWookie

Joined Jul 17, 2007
22,210
Yes, I have changed the values a little. They were 1N4004s before, but I thought that they should be fast recovery diodes, so I changed them to UF4007. A UF4007 is an ultrafast recovery diode, 1000v 1A 75ns.
Good idea. :)
As for the reverse EMF protection - the MOSFETs contain anti-parallel diodes in them, right?
Yes, that's true. I simply forgot. Your symbols don't show the integral Schottky diodes.

You might also consider STB25NM50N/STF25NM50N/STP25NM50N/STW25NM50N; higher voltage and an Rds(on) around 0.11 Ohms, Id=22A.
The deadtime - the signals are coming from an SG3526, where deadtime is already set, so I think they should not be cross-conducting. If I am wrong, please stop me.
Ok, the SG3526 DOES have provisions for adjusting the deadtime via Rd. However, that doesn't solve your Vce=280V problem - and the way you've shown the input to your driver transistors in the schematic made absolutely no references to an SG3526, and the high-side and low-side driver transistors even have their bases tied together.
 

Thread Starter

Krej

Joined Aug 29, 2008
7
So, I see the circuit's not going to work. I tried the IR2110 and IR2113. They worked fine at 12v and 24v but when I connected 300v to the Full Bridge, the IR211x were shorted out. I used UF4007 here as well, bootstrap capacitors of 10uF non-polar caps paralleled with 0.1uF ceramic cap.
I simply don't understand where I'm going wrong.
Please help....
Jeofrey.
 

SgtWookie

Joined Jul 17, 2007
22,210
Post your complete schematic of when you were using the IR2110/IR2113.

Were you using small caps across the MOSFET's drain and source terminals? While they do have integral Schottky diodes, it still takes a little time for them to turn on. During that time, the voltage across the MOSFET could get quite high with an inductive load. Small ceramic caps, perhaps 330pF, buy the integral diode time to turn on. You don't want the caps to be very large.
 
Top