Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Spike on H-bridge output

Status
Not open for further replies.

munnaka

New Member
Hi,
can anyone help with this one? I am running an H-bridge supplied at 80V with four mosfets (2xZVP3310F on the high side, 2x ZVN3310F on the low side) triggered by a PIC running at 5V - No intentional inductances, just tracks, the load is resistive.
The triggers are pulses about 5V 60 - 100 usecs on alternate gates to reverse the output polarity. (they are not simultaneous!!!!)
Two more ZVN3310F provide the inverted pulses for the ZVPs.
On the scope I can see a fast largish spike (about 1 - 2 usec, around about 10-15V amplitude) at the very beginning of the output pulse across the load.
(rough outline bottom right of the attached drawing)
This voltage is more than the drop due to the current through the load (rectangular outline).
I have asked a few other people but they are puzzled as well as to what exactly is going on. That pulse is mains sensitive, although the circuit is battery powered. Running the pulses at 100 Hz it goes up and down slowly, if I run them at say 30 Hz they go up and down faster, so it looks as if a mains pickup is involved. In fact, it also grows enormously if I touch the ground, making me thing it could be related to the scope ground . However, the pulse is rather solid, does not go away with any decoupling I have tried and it is still there, although reduced, if I use a battery powered scope.
I have substituted a bipolar for the ZVP (MMBTA92) and the effect is still there.
I am attaching a schematic of the output stage for clarity.

Thanks to any kind bright brain out there
 

Attachments

  • test4.pdf
    14.1 KB · Views: 190
The resistor values at the gates of the Mosfets are so high that there will be considerable time for the Mosfets to turn on and turn off due to the high capacitance of the gate of a Mosfet. Maybe the delays cause the spikes.
 
I assume you a looking at one of the outputs single-ended with the scope grounded. Since this is a bridge circuit you need to look at the voltage across the load, which is floating, not from one side of the load to ground. That's why touching the ground changes the spike value.

To measure across the load, use a dual-trace scope with the two inputs in a differential (subtracting) mode. Then connect one channel to one side of the load and the other channel to the other side. This will give you a picture of the true load voltage so you can better understand what the load spikes really are.

Make sure both inputs are set to the same sensitivity setting, but one with a negative polarity. (To adjust this for best operation, apply the same voltage to both channels and fine adjust the sensitivity of one channel until applying and removing the voltage to both channels causes no trace movement).
 
Resistors and scope

Hi guys,
thanks very much for your ideas; I have reduced the resistors on the divider on the high side (I thought those were the ones in question) going down to 500 ohms in place of the 22k - the spike was reduced a bit but still of the same order of magnitude. In reply to the scope query, yes, I did try with a single probe across the load, but then I did the difference between two probes referred to ground - and it was still there.
I have thought of possible effects of delays, and I have tried to use the inverter transistor to obtain both pulses, so they would be there virtually at the same time (one resistor on the source to ground to follow the incoming pulse and another one on the drain to invert).
This did seem to reduce it further, but it was only on the pulses of positive polarity, where it nearly disappeared.
I thought this may be due to asymmetries in the transistor parameters, but where I got really baffled is when I inverted the scope connections across the load and the spike was still on the positive pulses!
I am lost! The only thing I can think of is an interaction with the scope itself,
pickup or something like that. The problem is, if I put a cap across the load
(2200 pF) the spike comes back with a vengeance, it looks as if it's got real energy.

Maybe it's getting too much like 'one of those things', but I would really like to understand what is going on. I understand the difficulty of tackling something like this at a distance, and I am grateful to you for trying.

Riccardo
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top