Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Serious IGBT overshoot

Swend

Full Member level 4
Joined
May 14, 2019
Messages
237
Helped
10
Reputation
20
Reaction score
9
Trophy points
18
Location
Roskilde, Denmark
Activity points
2,304
Hi Everybody!

I'm working on a power-supply but have run into some serious overshoot problems. As you can see I tried with a RCD snubber, but that didn't change anything, do you perhaps have any suggestions how to deal with this?

schemeit-project(1).png

SDS00035.PNG

Regards Svend.
 

barry

Advanced Member level 5
Joined
Mar 31, 2005
Messages
4,815
Helped
1060
Reputation
2,130
Reaction score
1,043
Trophy points
1,393
Location
California, USA
Activity points
26,232
Where are your decoupling caps?

Where is your waveform from?

Where is your PCB layout?

Where is your overshoot?

What are your driving signals?
 

kripacharya

Banned
Joined
Dec 28, 2012
Messages
1,209
Helped
182
Reputation
360
Reaction score
175
Trophy points
1,343
Location
New Delhi
Activity points
0
Your circuit and scope trace is not understand able as presented. Please add details to better explain the problem, the investigations done, and the results obtained so far - with proper cross reference to your diagrams.
 

Swend

Full Member level 4
Joined
May 14, 2019
Messages
237
Helped
10
Reputation
20
Reaction score
9
Trophy points
18
Location
Roskilde, Denmark
Activity points
2,304
Hi barry

Where are your decoupling caps?
I don't have any yet

Where is your waveform from?
It's generated by a microcontroller, then entering a driver that is connected to the on/off resistors R3-6

Where is your PCB layout?
I don't have any, it's a bird nest, no breadboard, only soldering.

Where is your overshoot?
I don't know if I used the correct technical term, so I have marked it in red.

What are your driving signals?

I'm using 1EDC20I12AHXUMA1 as driver with single supply +15V, and the outputs are connected to the on/off resistors R3-6

Since my last post I removed the RCD snubber and changed values of R3-6, then I got rid of the turn-on "overshoot", but I still have a problem with turn-off.

schemeit-project(2).png

The yellow trace is "OUT" and I'm using a 1000x probe, so you have to multiply by 1000

SDS00036.PNG
 

Swend

Full Member level 4
Joined
May 14, 2019
Messages
237
Helped
10
Reputation
20
Reaction score
9
Trophy points
18
Location
Roskilde, Denmark
Activity points
2,304
Your circuit and scope trace is not understand able as presented. Please add details to better explain the problem, the investigations done, and the results obtained so far - with proper cross reference to your diagrams.
Hi kripacharya, please see my reply to barry
 

barry

Advanced Member level 5
Joined
Mar 31, 2005
Messages
4,815
Helped
1060
Reputation
2,130
Reaction score
1,043
Trophy points
1,393
Location
California, USA
Activity points
26,232
No decoupling caps and no PCB, and you're wondering why your circuit works like crap???

You probably have a LOT of stray inductance, not to mention inductance from the ground probe of your oscilloscope.
 
  • Like
Reactions: Swend

    Swend

    points: 2
    Helpful Answer Positive Rating

Easy peasy

Advanced Member level 5
Joined
Aug 15, 2015
Messages
2,318
Helped
848
Reputation
1,696
Reaction score
838
Trophy points
113
Location
Melbourne
Activity points
13,254
I see you are switching on & off quite slowly with those high-ish gate resistors - but you still need 1uF 630V say across the pair of IGBT's from top collector to bottom emitter to soak up the turn off wiring inductance energy - without seeing the entire schematic and a picture of the hardware - it is hard to make positive suggestions ...

- - - Updated - - -

zoom right in on those over shoots for a good look ( 500nS/div & 100nS/div ) it may be just CM pickup ... how many amps are you trying to switch ?
 
  • Like
Reactions: Swend

    Swend

    points: 2
    Helpful Answer Positive Rating

Swend

Full Member level 4
Joined
May 14, 2019
Messages
237
Helped
10
Reputation
20
Reaction score
9
Trophy points
18
Location
Roskilde, Denmark
Activity points
2,304
Hi Easy peasy!

I see you are switching on & off quite slowly with those high-ish gate resistors
Yes, this is the gate signal

SDS00040.PNG

The plateaus disappear if I turn off V1 and V2, what could cause that?

I - but you still need 1uF 630V say across the pair of IGBT's from top collector to bottom emitter to soak up the turn off wiring inductance energy
I tried with a 0.1uF (didn't have 1uF) and the only thing it seems to do is to distort the positive cycle.

SDS00038.PNG
Note: 1000x probe


I zoom right in on those over shoots for a good look ( 500nS/div & 100nS/div ) it may be just CM pickup ... how many amps are you trying to switch ?
Like this?
SDS00037.PNG
Note: 1000x probe

how many amps are you trying to switch ?
At the moment around 3-400mA

I - without seeing the entire schematic and a picture of the hardware - it is hard to make positive suggestions ...
I know, but it's too ugly to show to other people. And I'm not asking for miracles, just a few hints and pointers from experienced people would be appreciated.

- - - Updated - - -

No decoupling caps and no PCB, and you're wondering why your circuit works like crap???
I'm sorry about that, but it's my first attempt at an IGBT application, and I didn't have anyone else to ask before posting.

You probably have a LOT of stray inductance, not to mention inductance from the ground probe of your oscilloscope.
You are probably right about that.
 

KlausST

Super Moderator
Joined
Apr 17, 2014
Messages
17,111
Helped
3862
Reputation
7,722
Reaction score
3,755
Trophy points
113
Activity points
113,770
Hi,

The plateaus disappear if I turn off V1 and V2, what could cause that?
I recommend to read some MOSFET or IGBT driving tutorials.

The plateau should be well explained in every good tutorial.

Klaus
 

KlausST

Super Moderator
Joined
Apr 17, 2014
Messages
17,111
Helped
3862
Reputation
7,722
Reaction score
3,755
Trophy points
113
Activity points
113,770
Hi,

Your document is about snubbers, whuch ar connected at the output of a Mosfet / IGBT.
But I recommended a "driving" tutorial, which is at the input, the gate.

The plateau is called "miller plateau". Do an internet search, you will find a lot of informations.

Klaus
 

Easy peasy

Advanced Member level 5
Joined
Aug 15, 2015
Messages
2,318
Helped
848
Reputation
1,696
Reaction score
838
Trophy points
113
Location
Melbourne
Activity points
13,254
Your gate drive is possibly too soft, the overshoot is well reduced with the 100nF 630V cap across the supply right by the IGBT's the distortion tells me your 200V supplies are very soft ( the top one any way ) - at least you won't cook your devices at 400mA ( unless they have no heatsinking )

try 1uF 630 V and all your overshoot may disappear ... ( 2 x 2u2 400V in series or similar ... )
 

Easy peasy

Advanced Member level 5
Joined
Aug 15, 2015
Messages
2,318
Helped
848
Reputation
1,696
Reaction score
838
Trophy points
113
Location
Melbourne
Activity points
13,254
you haven't zoomed in on the overshoot at turn off - what you have captured is a blip ... please try again ...
 

Antor

Member level 2
Joined
Mar 1, 2019
Messages
51
Helped
Reputation
0
Reaction score
0
Trophy points
6
Activity points
501
Where are your decoupling caps?
I've been searching for decoupling caps for IGBT like 1 hour a day like 2-3 days. I checked many tutorials, datasheets and application notes. I haven't seen any decoupling capacitors.

What is the "decoupling capacitor" in IGBT commutating / driving ? Do you mean to putting a capacitor between collector and base ? Collector and Emitter ? Base and Emitter ? For what purpose ?

I'm on designing switching layer and I need urgent reply.
 

Antor

Member level 2
Joined
Mar 1, 2019
Messages
51
Helped
Reputation
0
Reaction score
0
Trophy points
6
Activity points
501
but you still need 1uF 630V say across the pair of IGBT's from top collector to bottom emitter to soak up the turn off wiring inductance energy
That's basically putting a capacitor at source.
soak up the turn off wiring inductance energy
Do you mean there will be an energy stored in PCB because of an stray inductance and bypass it via that capacitor ? What if the input is AC not DC ?

- - - Updated - - -

Hi,

Your document is about snubbers, whuch ar connected at the output of a Mosfet / IGBT.
But I recommended a "driving" tutorial, which is at the input, the gate.

The plateau is called "miller plateau". Do an internet search, you will find a lot of informations.

Klaus
I searched many manufacturer IGBT tutorials, application notes, datasheets, datasheet tutorials, miller plateau voltage papers and now I'm confused like a cat watching pinball game. Could you please point out a spesific tutorial so I can print out and read line by line.

Thanks..
 

Toggle Sidebar

Welcome to EDABoard.com

Sponsor

Sponsor

Design Fast


X
Top