Using ACPL-H342 to drive IGBTs to control charge/discharge of a capacitor at 1 KV

Hello,

I've been trying to get the ACPL-H342 schematic given in Figure 33, page 15 of the document given below(*) working.

I've managed to collect the needed models and Spice simulation works fine, as expected. However, for some reason, on breadboard the output capacitor voltage never goes above 15 V - meaning that the high side IGBT takes hundreds of volts...

Can someone, please, take a look and make a sanity check, am I missing something?

(*) ACPL-H342 component datasheet: https://www.mouser.com/ds/2/678/V02-2526EN_DS_ACPL-H342_13Mar20130-909326.pdf

Thanks in advance,
regards.

There is only one capacitor in that schematic so:

Designers will
need to adjust the VCC supply voltage, depending on the
MOSFET or IGBT gate threshold requirements (Recom-
mended V CC = 18V for IGBT and 12V for MOSFET).

Click to expand...

you have to supply the VCC yourself from an outside source.
Also note the text about why the capacitor is necessary, by implication it must be wired to minimize impedance across VCC and VEE pins.

Brian.

Hi,

most probably the issue is not in the datasheet but in your circuit.
--> You need to show us your circuit.

Generally:
A breadboard usually is not suitable for testing switching_power_circuits.

Klaus

Hello guys,

thanks a lot - not at my PC right now, so I will post the circuit once I get back.

Hello,

really sorry about the delay.

First of all, I am really sorry - I pasted wrongly the link to the first app note I was trying. That one I didn't get to work (and it was quite strange, as its output is grounded), so please refer to Figure 4 of this app note:

**broken link removed**

Sincerest apologies for the confusion.

Here is the circuit I compiled based on that schematic (please don't mind the ugly symbols of IGBTs and wrong pin numbering of ACPL):



Simulations went fine - the capacitor at the output goes up to 1 KV as it should, but on a breadboard it doesn't go higher than 15 V, no matter how high the High Voltage source is.

I hope you can help, thanks.

- - - Updated - - -

KlausST said:

most probably the issue is not in the datasheet but in your circuit.
--> You need to show us your circuit.

Click to expand...

Agreed - I hope the the last post can help you help me debug

KlausST said:

Generally:
A breadboard usually is not suitable for testing switching_power_circuits.

Click to expand...

Thanks for that, however, I do need to test it before PCB, so I hope you can find some time to check my schematic attached in the last post.

Hi,

* Missing informations makes it about impossible to verify your schematic.
* no dead time in control signal --> expect cross conduction.
* In my eyes it´s not possible to "test" this circuit on a breadboard. It´s like trying to bake a cake in a dish washer - maybe you can eat the cake but the result is not the same as with an oven.
--> Thus I recommend to use a meaningful test setup.

Klaus

* In my eyes it´s not possible to "test" this circuit on a breadboard. It´s like trying to bake a cake in a dish washer - maybe you can eat the cake but the result is not the same as with an oven.

Click to expand...

:grin: Brian.

Hello,

thanks for the response.

However, could you please be more specific?

What information do you actually miss?

Concerning the control signals, please note that they are reversed in the TOP driver V1 is 0V and V2 is 5V, whereas in the BOTTOM driver V1 is 5V and V2 is 0V. Having in mind pins AN(ode) and CA(thode), that seems quite ok. Can you, please, correct this train of thought if you find it wrong?

Also, a nudge or two towards a meaningful setup would be really helpful - this is the best I could came up with, please can you point to the right direction (as I really don't want the cake out of the dishwasher )?

Thanks.

Hi,

What information do you actually miss?

Click to expand...

* R2 value
* bootstrap circuit values calculation. (I don´t think it can work with this low frequency. What happens on undervoltage?)
* did you really try to use a 30V rated 1N5818 for your 1000V application?
* are you aware of the huge current when you charge/discharge a 60uF capacitor at 1000V
* If you want us to validate it: a meaningful optcoupler symbol
* If you want us to validate it: a meaningful IGBT symbol

***
Reversed driver signals don´t involve dead time.

***
Use copper claded PCB for a good GND plane. Use at least 10mm of clearance and creepage distance. Use short, low impedance, twisted wiring to the IGBTs.

Klaus

Hello Klaus,

thanks for the clarifications.

R2 value is actually provided; namely I used it as a parameter during parametric simulation, so if you look at the top left corner of the schematic - it's there, RES = 100 and R2 is RES.

I did expect a large current at the output capacitor, but let me see into the bootstrap circuit values again and I'll post back the actual findings.

As for the diode, yes, well, that's a bad one - thanks a lot!

I'll get around and include appropriate symbols, sorry about that.

Ok, so, thanks all together and I'll report back with more information.

Regards