Measuring output from H-bridge destroys mosfet drivers

Thread Starter

Jorne

Joined Feb 28, 2020
6
Hello,

I am using two IR2110 mosfet drivers, one for each side of my H-bridge. I visualized the output of the H-bridge (a 100kHz square wave between +12V and -12V) on an oscilloscope, using a differential voltage probe. That was fine.
However, when I connect a load to the H-bridge, and I measure with a multimeter the AC voltage across the load, one of my drivers gets broken.
It also happened once while measuring across the load with the oscilloscope instead of the multimeter.
When measuring with both, I see that the square wave on the scope collapses completely when measuring the voltage with the multimeter. I can't understand why, as a voltmeter has a very high input impedance. It even happened once when connecting only one probe of the multimeter (one sided)!
My load is a LC resonant circuit for wireless power transfer.

I'm really looking for a clarification, as this frustrates me.
 

KeithWalker

Joined Jul 10, 2017
1,308
You must be using a grounded power supply for the load. When you make measurements with a grounded instrument, the low probe will be grounded and will short out part of the bridge.
To make measurements across the load, you need to isolate the supply from ground or use a differential measurement input. When using a scope, use two vertical input channels, both set to the same vertical deflection and both centered on the screen. Connect a probe at each end of the load. Switch the input mode to A-B. If your scope only has A+B then invert the polarity of the B channel input. The measurement displayed will be the difference between the two inputs.
Regards,
Keith
 

BobTPH

Joined Jun 5, 2013
2,605
Is the multimeter you are using battery operated? If so, it is not the grounding problem. You should be able to probe any two points with an isolated multimeter in voltage range without causing a problem.

Bob
 

Thread Starter

Jorne

Joined Feb 28, 2020
6
You must be using a grounded power supply for the load. When you make measurements with a grounded instrument, the low probe will be grounded and will short out part of the bridge.
To make measurements across the load, you need to isolate the supply from ground or use a differential measurement input. When using a scope, use two vertical input channels, both set to the same vertical deflection and both centered on the screen. Connect a probe at each end of the load. Switch the input mode to A-B. If your scope only has A+B then invert the polarity of the B channel input. The measurement displayed will be the difference between the two inputs.
Regards,
Keith
Thanks for your reply, but i'm sure it's not the problem. First of all, I'm fully aware of this, and therefore I used a differential (isolated) probe for the scope. Secondly, it also happened twice while using a Fluke multimeter (on AC volt). So I don't think I'm shorting anything. Sorry.
 

Thread Starter

Jorne

Joined Feb 28, 2020
6
Is the multimeter you are using battery operated? If so, it is not the grounding problem. You should be able to probe any two points with an isolated multimeter in voltage range without causing a problem.

Bob
Yes, just a Fluke multimeter on a battery. It cannot be a grounding problem indeed. And the probe I'm using for the scope is a differential (isolated) probe. Maybe it was just a coincidence that the driver broke down a couple of times when I was measuring.
 

Thread Starter

Jorne

Joined Feb 28, 2020
6
Please post the schematic. Label the supply voltages. What is the duty cycle?
/*Note from moderator: Please do not post huge files. The file size has been reduced to 5% of its previous size.*/

Here is a schematicSchematic.png
 
Last edited by a moderator:

ronsimpson

Joined Oct 7, 2019
1,026
I don't have an answer but;
I have seen IR2110s not work right when on a board like below. The IC does not like long wires.
1582983605088.png
Can you send a picture of your setup?
 

Thread Starter

Jorne

Joined Feb 28, 2020
6
I don't have an answer but;
I have seen IR2110s not work right when on a board like below. The IC does not like long wires.
View attachment 200241
Can you send a picture of your setup?
I don't have the setup at home, but it is a decent PCB with soldermask. The capacitors and bootstrap diodes are close to the drivers. The PCB functions very well actually. A STM32 Nucleo is connected to the headers on the PCB as a shield. The problem came when I measured with a multimeter (AC Volt) over the load. On the scope I saw my square wave across the load collapsing, but no increase in current from my supply. And then a driver fails. Maybe some high frequency coupling...
 

brighamj

Joined Mar 30, 2019
1
Hello,

I am using two IR2110 mosfet drivers, one for each side of my H-bridge. I visualized the output of the H-bridge (a 100kHz square wave between +12V and -12V) on an oscilloscope, using a differential voltage probe. That was fine.
However, when I connect a load to the H-bridge, and I measure with a multimeter the AC voltage across the load, one of my drivers gets broken.
It also happened once while measuring across the load with the oscilloscope instead of the multimeter.
When measuring with both, I see that the square wave on the scope collapses completely when measuring the voltage with the multimeter. I can't understand why, as a voltmeter has a very high input impedance. It even happened once when connecting only one probe of the multimeter (one sided)!
My load is a LC resonant circuit for wireless power transfer.

I'm really looking for a clarification, as this frustrates me.
At 100kHz, did you consider the capacitive and inductive load of the Fluke leads. If they were in close proximity to each other it could have exceeded the load on the driver. Separating the leads as far as possible should prevent an overload.
 
At 100kHz, did you consider the capacitive and inductive load of the Fluke leads. If they were in close proximity to each other it could have exceeded the load on the driver. Separating the leads as far as possible should prevent an overload.
At 100kHz, did you consider the capacitive and inductive load of the Fluke leads. If they were in close proximity to each other it could have exceeded the load on the driver. Separating the leads as far as possible should prevent an overload.
 

Hugh Riddle

Joined Jun 12, 2020
72
Hello,

I am using two IR2110 mosfet drivers, one for each side of my H-bridge. I visualized the output of the H-bridge (a 100kHz square wave between +12V and -12V) on an oscilloscope, using a differential voltage probe. That was fine.
However, when I connect a load to the H-bridge, and I measure with a multimeter the AC voltage across the load, one of my drivers gets broken.
It also happened once while measuring across the load with the oscilloscope instead of the multimeter.
When measuring with both, I see that the square wave on the scope collapses completely when measuring the voltage with the multimeter. I can't understand why, as a voltmeter has a very high input impedance. It even happened once when connecting only one probe of the multimeter (one sided)!
My load is a LC resonant circuit for wireless power transfer.

I'm really looking for a clarification, as this frustrates me.
Is the voltmeter mains-powered? Might incorporate an EMI filtering capacitor between common terminal and ground.

Could stray (capacitive) coupling between the voltmeter lead(s) and the inverter's drive circuitry be upsetting the generation of drive pulse(s)?
 
Top