Snubber for igbt module FS50R06W1E3 - Infineon product -

[jmx66]

Hi all,

I'm trying to add a snubber to this igbt and found useful formulas in a thesis.

For Cs = ( Io * Tfi) / (2 * Vd )


Io : average current ( rated current i guess ? )

Tfi : current fall time transistor ( in FS50 datasheet , same meaning as " fall time inductive load " ? )

Vd: bus voltage



Snubber resistor: Vd/Rs < Irr

Rs : snubber resistor
Irr : reverse recovery current of Free Wheeling Diode



So reading datasheet:

Tfi : ( fall time inductive load ) 0,10 µs to 0,14µs see characteristic values p 2
Irr : 70A see diode-inverter characteristic values peak reverse recovery current p 3

About my circuit:

Io bldc rated current about 4,7 A
Vd Vbus about 340 V


I found : Cs: 1nF and Rs > 4,85 Ohm

Rs choice: thin metal film on a high purity ceramic core, not wounded wire to avoid inductor
Power : 2 or 3 W

My choice:

- Capacitors ( thanks for your last advice, FvM, about Wima FKP )

WIMA|FKP2O111001I00JSSD|CAPACITOR, 1000PF, 1000V, 5% | Farnell United Kingdom

VISHAY BC COMPONENTS|BFC238580102|CAPACITOR, MKP, 1NF, 2000V, 5% | Farnell United Kingdom

EPCOS|B32652A2102K|CAPACITOR, 0.001UF, 2000V | Farnell United Kingdom

Resistors:

WELWYN|MFP2-10R JI|RESISTOR, 2W 5% 10R | Farnell United Kingdom

MULTICOMP|MOR03SJ0100A19|RESISTOR, 3W 5% 10R | Farnell United Kingdom


Are these maths, and components, right for the job?

Thanks a lot.

jm

 

[FvM]

In my opinion, a correctly designed IGBT-H-bridge should need no RC snubbers. I hope I didn't lead you on a wrong track when mentioning snubber capacitors. I was referring to the capacitors that bypass the DC-bus-voltage directly at the IGBT-modules. Interestingly, the thesis comes to the conclusion, that no clear positive effect could be shown in simulation. The author placed a snubber anyway, because he doubted the simulation results.

Snubbers are of course a last resort, if a circuit can't be tamed otherwise.

 

[jmx66]

Thanks again FvM.

So I must only add a capacitor to bypass DC bus ?

Like on a schematic i have seen with capacitors , between phase to phase? They were labelled X2 , but it was for 3 phases motor or inverter, it seems to me.

I'm going to "googling" again, to know right value for this capa.

You didn't lead me on a wrong way, but in many textbooks in internet - manufacturers also - snubbers are " mandatory ", for using igbt and power switches without frying them....

Thanks a lot.

jm

 

[FvM]

I don't think that RC snubbers are mandatory for recent IGBT modules. They are reducing the overall efficiency and increasing the current load to the IGBTs.

 

[jmx66]

Thanks a lot.

Without adding this snubber, board will be easier to build; and Infineon products are recent IGBT modules so no need ....

Really thanks for your clear advices.

jm

 

[Orson Cart]

Close de-coupling using quality caps is pretty essential on a power switching bus, where you are driving inductive (motor) loads on long leads at lighter loadings you may see ringing between the rails when a bridge leg is OFF. If this ringing is able to propagate on long leads it may cause non-EMC compliance and is the usual reason you see snubbers used on commercial designs when driving inductive or motor loads on long leads.
Regards, Orson Cart

 

[jmx66]

Thanks a lot for your clear answer.

To close this useful post , i must draw my board like the one joined ?

Choose this capacitor following advices from FvM , in FKP range.

WIMA|FKP2O111001I00JSSD|CAPACITOR, 1000PF, 1000V, 5% | Farnell United Kingdom

Thanks a lot,FvM and Orson, for your professional advices.


jm

 

[Orson Cart]

Just having capacitors across each device as you have drawn is not the usual way to go, capacitors from +ve to -ve are better so the current in the load has a local path when the devices turn off, don't forget to add some large electrolytics across the power rail also, snubbers are typically R&C across each device (say 1n0 and 470-ohm for starters) Regards, Orson Cart.

 

[FvM]

Just having capacitors across each device as you have drawn is not the usual way to go

Absolutely, I didn't look at the schematic before. Additional capacitive load across the IGBTs will increase the switching losses and also cause additional transients at the DC bus.

 

[jmx66]

Hi,

Thanks a lot, for your helpful and pro advices, as usual.

jm

 

[jmx66]

Hi,

I did an other sketch, is it right or wrong ?

"capacitors from +ve to -ve are better so the current in the load has a local path when the devices turn off, don't forget to add some large electrolytics across the power"

Capacitors data - WIMA FKP only-are false - higher than needed -, because i didn't get rid of previous text....



47 KOhm resistor is to discharge capacitances , when igbt is off.

Better choice to use resistor and capacitor, or only capacitor as told Orson Cart?

Thanks a lot.

jm

 

[Orson Cart]

C1, C2 across the gate drive are a bad idea, the big caps across the power rails can go anywhere, but close to the 3 half bridges, you may still need RC snubbers across the C-E of the igbt's depending on the load configuration, good luck...! Regards, Orson Cart.