there is no way to turn all the fets OFF.

The bottom FETs are turned OFF by transistors Q2 and Q3 turning ON.
And when one bottom FET turns ON, the opposite top FET turns ON, shutting OFF the adjacent top FET.

But you bring up a good point.
I did not properly consider the state of the top FET gates which are floating, and that could cause a feedthrough spike when one side is turned on.
I'll need to think some about that.
So, HelpMe2020, please wait to finalize the design.

 

The bottom FETs are turned OFF by transistors Q2 and Q3 turning ON.
And when one bottom FET turns ON, the opposite top FET turns ON, shutting OFF the adjacent top FET.

But you bring up a good point.
I did not properly consider the state of the top FET gates which are floating, and that could cause a feedthrough spike when one side is turned on.
I'll need to think some about that.
So, HelpMe2020, please wait to finalize the design.

Sure, I think kennybobby introduced very valid points. So, we do need to find a solution to discharge the gate charge and also to charge/discharge the big cap. Looking solutions about how to fix it.

 

Okay, I found that there is shoot-trough due to coupling from the top MOSFET gate-drain capacitance. when the bottom MOSFET turns on.
A brute-force solution is to add a capacitor in parallel with each of the top Zeners to absorb this feedthrough pulse, with an added 10kΩ resistor from V+ to each side of the load (to keep the floating voltage, and thus the top MOSFET's Vg, at V+ when no MOSFETs are on).
Simulations show that a 10nF capacitor works well to suppress the shoot-through.

This will, of course, slow down the turn-on of the top MOSFETs, but that shouldn't be a problem for this low duty-cycle application.
How often will you be reversing the motor?

 

How often will you be reversing the motor?

May be once in 5 seconds.

 

@crutschow

So, below is an updated schematic. I guess this is what you intend to implement. Yes, it does improve switching of P-FETs.
I also added R11 (10K) in parallel to big CAP (C1 = 440 uF) for bleeding purpose of the big CAP.

In addition, how may I implement precharge for the big CAP to avoid inrush? Adding resistor in series would drop the voltage. Should I use a resistor (1K, 10W) across the contacts of relay to let it precharge before triggering the relay?

 

Should I use a resistor (1K, 10W) across the contacts of relay to let it precharge before triggering the relay?

If you have it to both supplies, won't that charge to the average of the 12V and 24V supplies?
But it doesn't need to be a 10W resistor, a 1/4W will be more than adequate.

 

If you have it to both supplies, won't that charge to the average of the 12V and 24V supplies?
But it doesn't need to be a 10W resistor, a 1/4W will be more than adequate.

I didn't get that. Both +12V and +24V are independent power supplies and independent circuits on schematic PCB. They are not connected to each other in any way. So, how come there would be an average?

 

Actually, there are two separate independent H-Bridges. One runs on +12V and another on +24V.

 

Actually, there are two separate independent H-Bridges.

Okay, I didn't realize that.
Ignore what I said (except for the power rating of the resistor).

 

Hi,

I have added ICL in two different fashions around the relays along with in-line fuse to protect the circuit from inrush current and overloading respectively. Please recommend me which one is better. I found both of them useful other than the fact top one will deprive me of keeping control on the switching on and off the Vin but will give me ability to bypass the ICL after precharging. The second one will give me complete control on switching on and off the Vin but will engage ICL all the time and will not entertain burst of inrush current pulses in case that happens.
(Note: I don't know but I feel like bypassing the ICL after precharging all the CAPs is better than engaging ICL all the time but then I lose control on disconnecting the power supply and relay doesn't make sense here much in the switching context)

One more Q, is common mode choke a better solution to suppress Inrush current in addition to noise?

Anyone with past experience may provide inputs. I guess after this I might have confidence to conclude on this thread.

 

I have added ICL in two different fashions

So what's an ICL?

One more Q, is common mode choke a better solution to suppress Inrush current in addition to noise?

No.
A common-mode choke does what it says, suppresses common-mode noise.
It has no effect on normal-mode noise/current which is what your current is.

 

So what's an ICL?

NTC Thermistor aka Inrush Current Limiter

 

TS, why did you put such a large capacitor on your DC buss (440uF)?

What do you see as its function? How did you select the value?

 

I guess you may need active ICL with a high power P-MOSFET attached across the 1K, 2W resistor. Your load will connected to the drain and you can have control on switching from precharging to direct drive by switching mosfet off-on.

 

TS, why did you put such a large capacitor on your DC buss (440uF)?

What do you see as its function? How did you select the value?

Actually, it's 220uF on actual schematic, not 440uF.

Well, I have had an old motor driver board which uses 24V/12V and 220uF is put across the supply to avoid any voltage fluctuation. So, I though 220uF would be great.

Honestly, I don't know how to exactly calculate that Cap. All I understand is that bigger voltages need bigger caps to filter away unwanted stuff. :|

 

I guess you may need active ICL with a high power P-MOSFET attached across the 1K, 2W resistor. Your load will connected to the drain and you can have control on switching from precharging to direct drive by switching mosfet off-on.

Can you please provide more details around it? It'd be very helpful!

 

The problem I see with the circuit is the power-up behaviour. This problem exists in the original circuit and the rework. Switching the power on through relay will give a high rising time of the supply voltage. A Mosfet has capacitance from drain to gate and gate to source. Depending on the speed of the transient the critical situation may be when both M1 and M3 turns on through this capacitive coupling, because there is no passive discharge on the gate on M3. So focus also on that point of the problem and simulate power-up/down transient. Especially a relay may clutter, so try also to simulate that behaviour.
Nevertheless the shoot through is also a problem of the original circuit. But according to the first failure description it was stated that it happens at power up, so this is the main flaw of the circuit. A 10 Kohm from M3 (and M4) to ground may give some improvement, but I recommend to measure the transient voltage rising with a scope and try to simulate that. Then improve the circuit with respect to power up and shoot through.

 

use two BTN7960 - works
Important - by Development & Programming allways +++++ allways KILL the supply for the brigde

 

With respect to the original schematic posted:

What are you using to switch on the +24V power--how is that being done?
[edit] i see you are using a relay

Do you have a precharge resistor for filling the big capacitor? or a drain resistor to bleed it when it is OFF?
[edit] i see no precharge resistor.

There is no path to bleed off the charge on the gates of the FETS, once they are turned ON they will stay ON until the gate charge has been removed. i think your problem is due to stored charges holding FETS ON, and then the top P-FETs get turned ON when the +24V is applied.

The relay contacts are bouncing and arcing due to the high inrush to C42, and the contacts are likely frosted over with carbon causing intermittent contact, and this is causing the state of the FETs to be uncontrolled.

i've never seen an H-bridge using p fets and n fets such as this. i suppose it could be made to work, but as it is now there is no way to turn all the fets OFF.

Another consideration is that with the possibility of high dv/dt switching transients, your circuit could be going into high frequency oscillation with high current draw destroying the MOSFets. Parasitics on the PCB may be a contributing factor here. Take a look with a scope and see if there is any evidence of HF ringing or sustained oscillation with your case of no load. Bypassing the MOSFETs to subdue the oscillations could solve the problem, if you see any of the above behaviors, but I'm not sure where the placement of bypass caps would be most effective - probably gate to GND?

 

use two BTN7960 - works
Important - by Development & Programming allways +++++ allways KILL the supply for the brigde

That part is obsolete. And the OP has not spoken if this is high volume price sensitive or one off design (or I have not read that in this thread). But it's a good idea. BTN8982 is a replacement.