Questions on Electric Vehicle high power electronics

[Russell McMahon]

> No H-bridge. It's a single IGBT module.

Byron,

Do you have the IGBT module already?
And what is the part number.
If you haven't got it yet I may be able to supply something.
What specs (current and voltage etc) or part number?


        Russell

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Byron Jeff]

On Wed, May 28, 2008 at 08:07:10PM -0400, Apptech wrote:

> >> Do you have the IGBT module already?
> > PowerEx CM600HA-24H
> > http://tinyurl.com/648p67
> >
> > These are 600A 1200V modules. They are way overengineered
> > for the
> > application which is going to be 144V with a max amperage
> > of 350-400A in
> > short pulses.
>
> Ah. Better than mine.
> I just checked what I have and they are "Trilingtons" - 3
> transistors in a Darlington cascade. 1000V, 400A.
> IGBT will be much nicer to drive and the cost is low enough
> to not matter wrt the rest of the project.
>
> The 400A is close to their 400A max rating (for very very
> small values of close). Most things get grumpy near the
> edges. Using two may be a very good idea [tm] if you can
> current share OK. Also, as they are BJT output second
> breakdown SOA becomes an issue. I'd check data sheets VERY
> carefully. Best guide is what others have reliably achieved
> with them already.

Actually in my searches today I found a couple of extremely useful
resources. The first is that the Electric Vehicle Discussion List has a
wiki with everything from Donor car selection to DC motor controller design
tips. You can find it here:

http://wiki.saymoo.org/EvdlGems/CategoryEvdlGems

One of the pages echoed a post in this thread that stated that IGBTs are a
poor choice because of the awful voltage drop. So poking around Digikey I
happened upon some 200V 120A MOSFETS from Ixys that are running about $42
each. You can find the datasheet here:

http://ixdev.ixys.com/DataSheet/96538.pdf

I figure that parallelling 4 of them should get me in the right ballpark.
I plan to current limit in software anyway.

I'm going to finish reading through the build your own controller page
here:

http://wiki.saymoo.org/EvdlGems/CategoryMotorControllers?highlight=%28EvdlGems%5C/CategoryEvdlGems%29

Lots of useful tips.

BAJ
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Russell McMahon]

>> Do you have the IGBT module already?
> PowerEx CM600HA-24H
> http://tinyurl.com/648p67
>
> These are 600A 1200V modules. They are way overengineered
> for the
> application which is going to be 144V with a max amperage
> of 350-400A in
> short pulses.

Ah. Better than mine.
I just checked what I have and they are "Trilingtons" - 3
transistors in a Darlington cascade. 1000V, 400A.
IGBT will be much nicer to drive and the cost is low enough
to not matter wrt the rest of the project.

The 400A is close to their 400A max rating (for very very
small values of close). Most things get grumpy near the
edges. Using two may be a very good idea [tm] if you can
current share OK. Also, as they are BJT output second
breakdown SOA becomes an issue. I'd check data sheets VERY
carefully. Best guide is what others have reliably achieved
with them already.



        Russell


--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Russell McMahon]

> One of the pages echoed a post in this thread that stated
> that IGBTs are a
> poor choice because of the awful voltage drop. So poking
> around Digikey I
> happened upon some 200V 120A MOSFETS from Ixys that are
> running about $42
> each. You can find the datasheet here:
>
> http://ixdev.ixys.com/DataSheet/96538.pdf
>
> I figure that parallelling 4 of them should get me in the
> right ballpark.
> I plan to current limit in software anyway.

My Trilingtons have a Vcesat max of 3V at 400A with 8A base
drive :-)
Hmmm - that would be 1.2 kW dissipation in the device.

Abs max power dissipation is 3.12 kW.
Rjc = 0.04 C/W
So junction-case rise is 50C.
To limit case rise to say another 50C above ambient would
need a 0.04 C/W heatsink.
That explains the heatsinks they are on :-).
Not enough though - maybe ?0.1 C/W? with fan cooling?
Heat pipe open water boiling would do it. BUT that puts the
junction at 150C+ in extreme cases - a "bit naughty" [tm].
BUT sealed heat pipe at < 100C action would do well enough.

    "Why is the car body warm?"
    'I'm using full throttle.'


        Russell McMahon


--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Tobias Gogolin]

Looks like you gain only about a volt of reduced loss,
However you get mighty close to the voltage rating!
Maybe rather consider to further increase the system voltage rating instead;
to 360 V for example like some of the commercial hybrids with AC drive....

On Wed, May 28, 2008 at 2:38 PM, Byron Jeff <byronjeff@...> wrote:

> On Wed, May 28, 2008 at 08:07:10PM -0400, Apptech wrote:
> > >> Do you have the IGBT module already?
> > > PowerEx CM600HA-24H
> > > http://tinyurl.com/648p67
> > >
> > > These are 600A 1200V modules. They are way overengineered
> > > for the
> > > application which is going to be 144V with a max amperage
> > > of 350-400A in
> > > short pulses.
> >
> > Ah. Better than mine.
> > I just checked what I have and they are "Trilingtons" - 3
> > transistors in a Darlington cascade. 1000V, 400A.
> > IGBT will be much nicer to drive and the cost is low enough
> > to not matter wrt the rest of the project.
> >
> > The 400A is close to their 400A max rating (for very very
> > small values of close). Most things get grumpy near the
> > edges. Using two may be a very good idea [tm] if you can
> > current share OK. Also, as they are BJT output second
> > breakdown SOA becomes an issue. I'd check data sheets VERY
> > carefully. Best guide is what others have reliably achieved
> > with them already.
>
> Actually in my searches today I found a couple of extremely useful
> resources. The first is that the Electric Vehicle Discussion List has a
> wiki with everything from Donor car selection to DC motor controller design
> tips. You can find it here:
>
> http://wiki.saymoo.org/EvdlGems/CategoryEvdlGems
>
> One of the pages echoed a post in this thread that stated that IGBTs are a
> poor choice because of the awful voltage drop. So poking around Digikey I
> happened upon some 200V 120A MOSFETS from Ixys that are running about $42
> each. You can find the datasheet here:
>
> http://ixdev.ixys.com/DataSheet/96538.pdf
>
> I figure that parallelling 4 of them should get me in the right ballpark.
> I plan to current limit in software anyway.
>
> I'm going to finish reading through the build your own controller page
> here:
>
>
> http://wiki.saymoo.org/EvdlGems/CategoryMotorControllers?highlight=%28EvdlGems%5C/CategoryEvdlGems%29
>
> Lots of useful tips.
>
> BAJ
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
Tobias Gogolin
Tel. Movistar (646) 124 32 82
Tel. Telcel (646) 160 58 99
skype: moontogo
messenger: usertogo@...

You develop an open source motor controller at
http://groups.yahoo.com/group/GoBox
and an open source electric motor/alternator at
http://groups.yahoo.com/group/Performance_Axial_Flux
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Sean Breheny]

Hi Byron,

I think you are going to have to do some testing and/or simulation to
get a very good handle on the current involved (if you haven't already
done so)

In order of preference for "gotchas" and ease of use I'd say:
1) Single FET
2) Single IGBT
3) Parallel FETs

Parallel FETs have their own weird issues. For example, they might not
both turn on at exactly the same time so you have some fraction of the
time when (at worst) only one is carrying full current. Also, even
though they have a positive temp coeff on Rds_on (so that they
automatically share current well when fully on), that ONLY applies to
the fully on state. At least some FETs have a positive temp coeff on
CURRENT in the partially-on state. This means that the hottest one
will snap on first and take all the current (AND experience ALL of the
turn-on transient power dissipation). These are not insurmountable but
not at all trivial.

I would only advocate a single FET (or even parallel ones) if you
could be well within (say at least 20% below) their max Vds spec.

Also, bear in mind that motor current will not in general equal
battery current. This is because your motor inductance is keeping the
motor current fairly steady while the battery only sees that current
during the "on" portion of the duty cycle. This is the buck converter
action going on. So, as a rough example, if you are running at 50%
duty and the motor winding current is 200A, your battery will only see
about 100A.

As for the diode across the motor - I looked at that page and it looks
to me as if it is the same as what I said - the diode is in a
different direction (as seen by the motor inductance) than the
freewheeling diode inside the IGBT. Think of the inductor current
during the PWM "on" time. Now think of the path it has to take to stay
flowing in the same direction when the IGBT switches off. You will see
that the internal diode in the IGBT (or FET) tries to block the
current, but one right across the coil does not block the current. The
diodes are in the same direction as far as the supply rails are
concerned, but from the point of view of the motor winding current,
one is a stop sign and the other is a GO.

You will ultimately need quite a bit of protection circuitry in order
to prevent something like a stalled motor or shorted wires from frying
your expensive controller.

I would HIGHLY recommend that you trying lashing up a quick drive for
say a 100W motor or even a 1000W motor as a learning exercise. You
could do it with a PIC or a function generator, gate drive IC, and
either a FET or an IGBT (or try both). You would be able to see the
kinds of things we are talking about.

Sean


On Wed, May 28, 2008 at 5:38 PM, Byron Jeff <byronjeff@...> wrote:

> On Wed, May 28, 2008 at 08:07:10PM -0400, Apptech wrote:
>> >> Do you have the IGBT module already?
>> > PowerEx CM600HA-24H
>> > http://tinyurl.com/648p67
>> >
>> > These are 600A 1200V modules. They are way overengineered
>> > for the
>> > application which is going to be 144V with a max amperage
>> > of 350-400A in
>> > short pulses.
>>
>> Ah. Better than mine.
>> I just checked what I have and they are "Trilingtons" - 3
>> transistors in a Darlington cascade. 1000V, 400A.
>> IGBT will be much nicer to drive and the cost is low enough
>> to not matter wrt the rest of the project.
>>
>> The 400A is close to their 400A max rating (for very very
>> small values of close). Most things get grumpy near the
>> edges. Using two may be a very good idea [tm] if you can
>> current share OK. Also, as they are BJT output second
>> breakdown SOA becomes an issue. I'd check data sheets VERY
>> carefully. Best guide is what others have reliably achieved
>> with them already.
>
> Actually in my searches today I found a couple of extremely useful
> resources. The first is that the Electric Vehicle Discussion List has a
> wiki with everything from Donor car selection to DC motor controller design
> tips. You can find it here:
>
> http://wiki.saymoo.org/EvdlGems/CategoryEvdlGems
>
> One of the pages echoed a post in this thread that stated that IGBTs are a
> poor choice because of the awful voltage drop. So poking around Digikey I
> happened upon some 200V 120A MOSFETS from Ixys that are running about $42
> each. You can find the datasheet here:
>
> http://ixdev.ixys.com/DataSheet/96538.pdf
>
> I figure that parallelling 4 of them should get me in the right ballpark.
> I plan to current limit in software anyway.
>
> I'm going to finish reading through the build your own controller page
> here:
>
> http://wiki.saymoo.org/EvdlGems/CategoryMotorControllers?highlight=%28EvdlGems%5C/CategoryEvdlGems%29
>
> Lots of useful tips.
>
> BAJ
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Alan B. Pearce]

>    "Why is the car body warm?"
>    'I'm using full throttle.'

Self heating in winter, though ...
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Byron Jeff]

On Thu, May 29, 2008 at 02:21:35AM -0400, Sean Breheny wrote:
> Hi Byron,
>
> I think you are going to have to do some testing and/or simulation to
> get a very good handle on the current involved (if you haven't already
> done so)

I haven't. I'm still in the design phase.

> In order of preference for "gotchas" and ease of use I'd say:
> 1) Single FET
> 2) Single IGBT
> 3) Parallel FETs

OK. Will at least at first glance single FETs are out. So it looks like
back to IGBTs.

> Parallel FETs have their own weird issues. For example, they might not
> both turn on at exactly the same time so you have some fraction of the
> time when (at worst) only one is carrying full current. Also, even
> though they have a positive temp coeff on Rds_on (so that they
> automatically share current well when fully on), that ONLY applies to
> the fully on state. At least some FETs have a positive temp coeff on
> CURRENT in the partially-on state. This means that the hottest one
> will snap on first and take all the current (AND experience ALL of the
> turn-on transient power dissipation). These are not insurmountable but
> not at all trivial.

Understood. It was one of the reasons I was looking to these single IGBT
modules. I understand they come with a basketfull of issues too.

> I would only advocate a single FET (or even parallel ones) if you
> could be well within (say at least 20% below) their max Vds spec.

That's the game plan. I was looking at 200V parts. I don't plan to run them
over 144V.

> Also, bear in mind that motor current will not in general equal
> battery current. This is because your motor inductance is keeping the
> motor current fairly steady while the battery only sees that current
> during the "on" portion of the duty cycle. This is the buck converter
> action going on. So, as a rough example, if you are running at 50%
> duty and the motor winding current is 200A, your battery will only see
> about 100A.

Got it.

> As for the diode across the motor - I looked at that page and it looks
> to me as if it is the same as what I said - the diode is in a
> different direction (as seen by the motor inductance) than the
> freewheeling diode inside the IGBT. Think of the inductor current
> during the PWM "on" time. Now think of the path it has to take to stay
> flowing in the same direction when the IGBT switches off. You will see
> that the internal diode in the IGBT (or FET) tries to block the
> current, but one right across the coil does not block the current. The
> diodes are in the same direction as far as the supply rails are
> concerned, but from the point of view of the motor winding current,
> one is a stop sign and the other is a GO.

Hmm. So that means that a separate diode is required.

> You will ultimately need quite a bit of protection circuitry in order
> to prevent something like a stalled motor or shorted wires from frying
> your expensive controller.

I plan to throw the entire array of safety circuitry at it. I'm still
trying to wrap my head around the essential components necessary to get
even a basic controller going.

> I would HIGHLY recommend that you trying lashing up a quick drive for
> say a 100W motor or even a 1000W motor as a learning exercise. You
> could do it with a PIC or a function generator, gate drive IC, and
> either a FET or an IGBT (or try both). You would be able to see the
> kinds of things we are talking about.

That was my plan. Does the voltage significantly impact the issue? By this
would testing at 12V give a good sense of the issues that are going on?

Thanks for the advice.

BAJ
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Sean Breheny]

Hi Byron,

Voltage does impact things (not the absolute voltage but how close you
are to the rating of the devices), but in order to keep the problem as
simple as possible, I would start out with a voltage which is several
times lower than the max of the part (i.e., like 12V on 36V or 50V
parts).

Sean


On Thu, May 29, 2008 at 4:15 PM, Byron Jeff <byronjeff@...> wrote:

>> I would HIGHLY recommend that you trying lashing up a quick drive for
>> say a 100W motor or even a 1000W motor as a learning exercise. You
>> could do it with a PIC or a function generator, gate drive IC, and
>> either a FET or an IGBT (or try both). You would be able to see the
>> kinds of things we are talking about.
>
> That was my plan. Does the voltage significantly impact the issue? By this
> would testing at 12V give a good sense of the issues that are going on?
>
> Thanks for the advice.
>
> BAJ
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Alan B. Pearce]

The OP on this thread may like to dig out this 'magazine' from Freescale, as
it appears to have a fair bit of (general) stuff on motor driving, although
it does appear to be mains powered stuff. I have put in the URL as it comes
off the Freescale email (but without my identifier in it). I don't know how
well this will work, and it expects you to log in to download.

If you want to try searching other ways it is Beyondbits3 that you are
looking for. May or may not be useful for your project.

http://www.freescale.com/webapp/site_cons.newsletter_metrics.framework?newsletter_id=NL242008&type=CLICKTHROUGH&url=https%3A%2F%2Fwww.freescale.com%2Fwebapp%2FDownload%3FcolCode%3DBEYONDBITS3%26amp%3BnodeId%3D0127260061788213CF333D%26amp%3Blocation%3Dnull 

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist