Repetitive peak collector current of an IGBT

[kappa_am]

Hi All,
I am dealing with designing a converter circuit. Before I chose IGBT with continuous current even more than nominal output current, but now pricing is critical and I need to keep it low. My question is about Repetitive peak collector current? what this current means? 1ms! Does not it depend on cooling system and thermal impedance? how long the switch must keep off after this period? In other word, how much can maximum duty cycle be?
Can a system be reliable if we design it using this current? regarding deteriorating
thermal filler , etc...

Thank you

 

[dick_freebird]

Different failure mechanisms may apply to peak
current, than time-averaged. If you look at things
like measured power-to-blow vs pulse width there
are two asymptotes - steady state (X asymptote
for infinite time) and instantaneous (Y asymptote,
for near-zero duration). For a -realistic- FWHM
pulse width you'd pick your place on that curve
and sandbag it a bit. Presuming you took the data.

A max repetitive peak current may also be where
the manufacturer ran out of knowledge, to say
that average current / power covers it all, or
believes this is where this "cusp" lies.

 

[d123]

Hi,

Doesn't the datasheet specify the actual duty cycle at the bottom of that page/that parameter section with those 'see note 1' points in brackets in parameter fields.

For various MOSFETs and BJTs I sometimes read 1% duty cycle in the datasheets notes section. 1%...! Surely they can't be serious...?

 

[KlausST]

Hi,

I agree with d123.
Why no link to the complete datasheet?
Why do you hide the part number and manufacturer?

Klaus

 

[kappa_am]

Thank you all for the information. I am not hiding; actually, I am looking for couple of components from ABB, Fuji and Infineon. I just shared a picture because it is easy for reader to look at rather than going through datasheet. The module I shared part of datasheet is Infineon FZ3600R12HP4. I cannot find any useful info in the datasheet! just a vague number. The datasheet can be found on below address.

Thank you for your time.

Datasheet

 

[FvM]

Read the specification as a quick selection parameter.

Does your application involve periodic peak currents of exactly 1 ms duration?
The complete information for arbitrary power dissipation profiles and the answer to all your questions above is found by using the transient thermal impedance diagram in the datasheet.

 

[kappa_am]

The problem is that the modulation is space vector modulation and the duty ratio varies through one cycle. Seems this curve is drawn for 50% duty cycle. Am i right? what's the cool off time (OFF time of the IGBT after one pulse)

 

[dick_freebird]

Winding capacitance is the repetitive peak
current's origin. Duty cycle is not relevant. It's
switching edge energy. If you have PFM then
you probably have some more figuring to do.

 

[FvM]

Single pulse specifications are for overcurrents events rather than continuous operation.

The datasheet give a specification of dynamic thermal impedance as chained time constants. It can be used to calculate the maximaL chip temperature for arbitrary current waveforms. You find ready made IGBT SPICE models with thermal modelling or can adapt the datasheet parameters in your own thermal model.

Your questions sound as you are trying to operate the IGBT near its absolute maximum ratings. Consider that it probably won't achieve long lifetime due to thermal wear.

 

[kappa_am]

Actually, it is not normal operation mode. For normal operation I usually I choose conservatively according to constant DC current= max current+20%. For this project Customer asks converter to have capability to supply twice nominal current for 2 seconds, in case there is a fault in the grid. I Don't want to have semiconductors paralleled ( stray inductance problem and hard DC bus manufacturing process) and there is no IGBT available in range of 8kA. I am dealing to come with a cost efficient solution, because cost also is a major factor for this project.

Any comment is appreciated.

 

[FvM]

There's no significant overload capability for 2 s time range, except for the margin given by heatsink heat capacity. What's the case temperature before the 2s transient? All necessary information is in the datasheet.

 

[kappa_am]

It is water-cooled. We also have limitation in flow and allowable outlet water temperature (42C). Total loss in the switch is 3960W I am tring to design to design for junction temperature 120-130C, in that case, case temp will be 52C. I am delving into datasheet and trying to come with solution. Unfortunately, Infineon online software doesn't simulate transients or with initial temperature values.

 

[FvM]

I'm not familiar with Infineon software. But steady-state case temperature + dynamic thermal impedance specification should allow you to calculate the maximal junction temperature for arbitrary transient power.