BIPOLAR ALTERNATIVE TO H BRIDGE

[green]

An easy way to measure the current slope is with an RX coil. Current slope will induce a +/- offset in the RX coil.

Thanks, I can control ramp rate(20A/sec to 12500A/sec) or hold current constant at .5A with my target tester. Connected it with separate Tx and Rx coils(Rx amplifier gain=500) to see how low a ramp rate I could detect. Current is in a control loop, looks good when scoping current feedback resistor, but Rx signal to noisy when Tx is on. Current not in a control loop with moodz circuit so it should work. Reply back when I get the circuit working.

Maybe moodz could try with his circuit?

[green]

Hi moodz,
Thanks for sharing your circuit! Can you please explain how it is performing better than the standard H-bridge circuit with diodes instead of your isolation switches? Is there any benefit in using the MOSFET body diodes instead of normal diodes? In terms of efficiency your additional switches add a bit more resistance in the current path when turned on, though that would be negligible if low on-resistance MOSFET is used.
Is a single gate driver capable of driving all 3 the transistors at once?
[ATTACH]55539[/ATTACH]

I have a similar bipolar Tx circuit on the bench. Trying moodz circuit to compare. [With the same change in Tx current at switching] target signal should be the same. Moodz circuit should subtract less so Rx signal should be higher for some targets. Moodz circuit would have higher average current but maybe less power supply watts. Moodz circuit looks better to me, getting constant current slope low enough is one thing that concerns me.

[Carl-NC]

An easy way to measure the current slope is with an RX coil. Current slope will induce a +/- offset in the RX coil.

[green]

Ordered some parts to make the circuit, including some polypropylene and NPO ceramics to compare across the coil. Reply #25 suggests constant current slope should be around .2A/ sec or less. Haven't been able to with spice, wondering how to measure slope in real circuit. Maybe a test to verify slope isn't causing a problem?

[moodz]

Good job, Moodz! I've been playing with your circuit in LTSpice and I'm eager to try it in real. I understand that only 2 high voltage mosfets are required since the rest don't see the flyback voltage. It's amazing how well it works with such a low power supply voltage.

Thanks Lucifer ... I have not tested the lowest effective voltage operation but 1.24 volts was very good IMHO.

[eclipse]

CBB such as these is just a regular polypropylene film capacitor

[Altra]

Hi Green,

I've been using this style of metal film.

This seller has various values rated at 2KV. You pay shipping on first and the rest are free

https://www.ebay.com/itm/163632399053?hash=item26194042cd:g:mVQAAOSw-9NcpSIR

[green]

I would prefer polypropylene film. Obtain few caps so you adjust the capacitance to obtain desired voltage.

Searched polypropylene capacitors. They have an AC and DC rating. If peak volts=500V, p-p=1000V. What rating should I purchase? I'm thinking (I squared L=E squared C). With .5A, 300uH and 500p, peak volts calculates about 387V. Why is simulation volts(355V) lower than calculated?

[eclipse]

I would prefer polypropylene film. Obtain few caps so you adjust the capacitance to obtain desired voltage.

[green]

Wanting to try the circuit. Need to order some parts, capacitors to add across coil to prevent avalanche if needed being one of them. Thinking 100p 1000V NPO ceramics. Good choice or not?

[lucifer]

Thanks Lucifer ... I found that the circuit I posted had twice the current magnitude and twice the speed and ran off lower voltages than an equivalent H bridge.

I used 500 volt mosfets for the 2 isolating switches and 4 milliohm low voltage mosfets for the rest.

I used a garden variety dual o/p inverting mosfet driver to drive the common gate drives nothing special .... still achieved 500 ns current transitions. no point using external diodes to increase this speed as the flyback voltage gets very high ... too high and the mosfets avalanche.

The driver was a TCC4428 1.5 amp mosfet driver actually.

Good job, Moodz! I've been playing with your circuit in LTSpice and I'm eager to try it in real. I understand that only 2 high voltage mosfets are required since the rest don't see the flyback voltage. It's amazing how well it works with such a low power supply voltage.

[moodz]

Hi moodz,
Thanks for sharing your circuit! Can you please explain how it is performing better than the standard H-bridge circuit with diodes instead of your isolation switches? Is there any benefit in using the MOSFET body diodes instead of normal diodes? In terms of efficiency your additional switches add a bit more resistance in the current path when turned on, though that would be negligible if low on-resistance MOSFET is used.
Is a single gate driver capable of driving all 3 the transistors at once?
[ATTACH]55539[/ATTACH]

Thanks Lucifer ... I found that the circuit I posted had twice the current magnitude and twice the speed and ran off lower voltages than an equivalent H bridge.

I used 500 volt mosfets for the 2 isolating switches and 4 milliohm low voltage mosfets for the rest.

I used a garden variety dual o/p inverting mosfet driver to drive the common gate drives nothing special .... still achieved 500 ns current transitions. no point using external diodes to increase this speed as the flyback voltage gets very high ... too high and the mosfets avalanche.

The driver was a TCC4428 1.5 amp mosfet driver actually.

[Carl-NC]

Hi moodz,
Thanks for sharing your circuit! Can you please explain how it is performing better than the standard H-bridge circuit with diodes instead of your isolation switches? Is there any benefit in using the MOSFET body diodes instead of normal diodes? In terms of efficiency your additional switches add a bit more resistance in the current path when turned on, though that would be negligible if low on-resistance MOSFET is used.
Is a single gate driver capable of driving all 3 the transistors at once?
[ATTACH]55539[/ATTACH]

The standard H-bridge drive like you show is not capable of square wave current drive. That is the purpose of Moodz' circuit, it is a completely different class of operation.

[lucifer]

Hi moodz,
Thanks for sharing your circuit! Can you please explain how it is performing better than the standard H-bridge circuit with diodes instead of your isolation switches? Is there any benefit in using the MOSFET body diodes instead of normal diodes? In terms of efficiency your additional switches add a bit more resistance in the current path when turned on, though that would be negligible if low on-resistance MOSFET is used.
Is a single gate driver capable of driving all 3 the transistors at once?

[moodz]

Thanks, could you describe your coil to get 300uH, .2 ohms and 100p. Diameter, number of turns, wire size, winding method(bundle, spiral, ?)and cable length.

The coil I use for bench testing is "old faithfull" .. which admittedly has seen better days. The resistance is actually 0.29 ohms. It has 17 turns of very expensive .. at the time I bought it several years ago ... litz wire ... might be 2 mm x section area .. and the diameter is 18 inches.

Its unsheilded ... and the winds are flat spiral basket weave for minimal capacitance. 296 uH last time I measured it.

But that is not the point ... a coil is a current operated device .. you dont want series resistances if you can avoid ... just adds to charging times if reactances are involved and power loss.





Probably worth considering also that the I would probably add capacitance ( eg 1000pf ) across the coil to limit peak flyback as once the current gets up to +/- 1 amp level the peak flyback is exceeding 1 KV. ( using SIC 1.4 KV mosfets for the output devices ).