FET resource page for MEV/HEV applications.

Looking for a FET suitable for creating a MEV (medium EV) to HEV (heavy EV) application? Did you happen across an exceptional choice in a FET, and want to make sure others can also check if it suits their needs?
Post them up as you find them, and I will update the thread. Include a link to the datasheet, and if you happened to price-shop them, the cheapest price you found as well.

Lets group the exceptional FETs we find by Vds.

300v:

200v:

Under $200/fet. Single FET. 1,600amp
http://ixdev.ixys.com/DataSheet/VMO1600-02P.pdf




150v:

$~16/fet. Single FET. 360amps (225a realistic)
http://ixdev.ixys.com/DataSheet/DS100181(IXFK-FX360N15T2).pdf

100v:

<$200/fet. Dual FETs per package, configured to be a phase leg. 680amps.
http://ixdev.ixys.com/DataSheet/L401.pdf


75v:

Under $200/fet. Dual FETs per package, configured to be a phase leg. 1,540amp.
http://ixdev.ixys.com/DataSheet/VMM1500-0075T2.pdf

$11.75/fet. Single FET per package. 500amp. Low inductance package design.
http://ixdev.ixys.com/DataSheet/DS100269%28MMIX1F520N075T2%29.pdf


When ever you guys stumble across a FET worth remembering, throw up a post in here, and I will update the top part as I see them. I don't know about the rest of you guys, but this will save me a ton of time trying to search through pages and pages of threads and searches trying to find a datasheet or FET name somebody shared. Should be a very handy resource for folks looking to push towards higher powered BLDC options, which are sadly lacking on off-the-shelf affordable and compact controller options.

Best Wishes,
-Luke

Maybe a bit small to qualify for MEV/HEV application, but a good option for high current 60V nominal eBike controllers.

75V- IRFP4368, Single FET, $7.70 > $6.93
TO-247AC
RdsON = 1.46mΩ
RθJC = 0.29°C/W
ID = 195A (package limited)
http://www.irf.com/product-info/datasheets/data/irfp4368pbf.pdf

http://search.digikey.com/scripts/DkSearch/dksus.dll?vendor=0&keywords= IRFP4368PBF-ND

.

I just had a box of the new FETs arrive.

These things are monsters! 8) 8)

Thats some might chunks of solid-state switching.

Is that VMM650-01F I can rad on the side.
Are they all the same?
Did they come one per box, so you got 6.

IXSYS have some strange naming methods on their spec doc's
VMM650-01F spec is called "L401" I mod'd the url to L402 ond got some nice Diade pakaged devices.
http://ixdev.ixys.com/DataSheet/L401.pdf

VDSS = 100 V
ID25 = 680 A
RDS(on) = 1.8 mΩ

Digi keys list price for non-stock was $212.82 each min Qty 2
Your earlier post showed them as L401 fro under $200.



The VMM reference above from the IXYS site is different to the L401 spec. It shows and your image the Right FMM style device ????
I have noticed quiet a few typos in the IXSYS documents. It's strange that they don't have a higher QC standard on their info.

I can't work out is it the ON Die Intrinsic or Free-wheeling diode. "fast intrinsic reverse diode"
Intrinsic to me means its part of the MOSFET characteristic and is shown as the arrow into the center from tha SOURCE pin.
I guess if the Intrinsic diode is not fast enough they have to add an extra one in the package or its fitted as a seperate device.

What do you plan to do the the diode.

I should know this but is the revese current controllable with these VMM650-01F devices as they don't have the Diode across the Drain to Source?

You'ce got some beefy buss bars to make, and heatsink. Please share your choices.

7circle said:

Intrinsic to me means its part of the MOSFET characteristic

Click to expand...

That is what Intrinsic means, referring to the diode, but:

and is shown as the arrow into the center from tha SOURCE pin.

Click to expand...

That arrow just shows you gate, AFAIK, and it's direction shows you if it is N-channel or P-channel. But it's been a really long time since I actually had any explanations for what the symbols were designed around, and stuff has changed since then.

Okay I freshened up my stall MOFET and Trench FET physics.
http://www.irf.com/technical-info/appnotes/an-937.pdf

So they have an "Intrisic Diode" but some types have poor characteristics.
ie Too slow, high Voltage drop, etc ..

So to meet the application they "fit" in an extra reverse current capablity in Diode Area.
But is this on the same subsrtate or a seperate substrate in parallel ?

Any way the spec says on : L401.pdf for VMM 650-01F:
VF (diode) (IF = 650 A; VGS = 0 V) Typ: 1.2 Max 1.5 V
trr (diode) (IF = 650 A; -di/dt = 500 A/μs; VDS = ½ VDSS) 300 ns

So all reverse current is back through a diode.
So any thoughts on Regen current control is by BOOST topology using motor winding inductance.

I see on http://endless-sphere.com/forums/viewtopic.php?f=30&t=16728&start=855#p304242 You've started on the layout concept.

IMHO, the best FET I've ever seen for a MEV/HEV motor controller so far is this 200v 1600amp unit. The crazy good combined Rth of 0.04C/W is what makes it dominate. If any of you guys ever find these for sale on ebay or whatever for a reasonable price, PLEASE let the forum know (or be sure to snatch them all up yourself).

VMO1600-02P
http://ixdev.ixys.com/DataSheet/VMO1600-02P.pdf


Sadly though, they are $240 each from future-electronics or $235 from digikey, and there is only 1 in each package/module, so it requires 6 of them to make something.
http://www.futureelectronics.com/en/Technologies/Franchised/Product.aspx?ProductID=VMO1600-02PIXS
http://parts.digikey.com/1/parts/index24203.html

On the topic of FET package limitations (a huge factor for motor controllers), sometimes a picture tells the story better than numbers. This is a TO-220 package FET (the fet package used in >99.9% of all ebike controllers because it's dirt cheap), sitting with a Y3-LI (what my controller uses) FET package.

I shared with Luke in a call that the IXYS FETs we all bought on eBay were introduced in 1998 and are still stocked but not current production, so we are not likely to get thousands of them if we hit on a good design. :x Trying to get a "real" data sheet on them with all the parameters, particularly Forward Bias Safe Operating Area, but so far, no response from IXYS... :x :x

BTW Luke sure knows how to tell a story with pictures, doesn't he? !!!

The TO220 is so small it needs a "you are here" arrow. :lol:

I think that big FET package is about as big as my whole Fusin controller.

amberwolf said:

The TO220 is so small it needs a "you are here" arrow. :lol:

I think that big FET package is about as big as my whole Fusin controller.

Click to expand...

LOL. Just the single FET is the size and weight of 3x of the Turnigy 100a ESCs. lol.

It kinda takes the term "package limit", and tells it to go "f" itself. Which is exactly what we need for these crazy new generation of BLDC motors.

Thanks for this thread guys!
Your second link doesn't work luke!

Arlo1 said:

Thanks for this thread guys!
Your second link doesn't work luke!

Click to expand...

You're welcome bro. Fixed broken link.

Any opinions on this:

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=IRFP4668PBF-ND

Seems like it could work well in a 24 FET package for a 200v system?

-Michael

When you get above around 120 to 150V it doesn't really make sense to use FETs, as the losses start to rise dramatically due to the high Rdson (9.7mOhm in this case). If you ran this FET at a modest 60 to 70 amps, then the voltage drop across the FET is going to be around the same as an IGBT, and an IGBT will hold that sort of voltage drop for higher currents and work at higher voltages.

This is the reason that the Prius PMBLDC motor controllers use IGBTs and not FETs, they are better at high voltage, high current than FETs (at the moment).

For lower voltages, where you can get FETs with a nice low Rdson, then FETs make sense. For example, the Ixys Hiperfet modules that LFP, Bigmoose and I have just bought have an Rdson on around 1.8mOhm, work at 100V max and will handle up to 680 amps. That low Rdson means that the voltage drop across the FET (and hence the losses in it) won't reach IGBT levels until the current gets up to around 350 amps or more.

You can drive big IGBTs the same way as big FETs, as their gate drive requirements are similar. The main difference is that, because they have a bipolar transistor switching element they have a fairly fixed voltage drop when hard on that is pretty much the same as a diode, around 0.6V at low currents, rising slightly as the current increases. This makes them lossy for low power applications, but great for high voltage, high current applications.

Jeremy

Jeremy is spot on. I put the transition to IGBT's also at the 120 volt range. For the reason Jeremy said and also for product growth potential. If you are at 120 you are going to want 180 sooner or later, and that is definitely IGBT territory. Note the UQM motor controllers are IGBT based at 240 volt supply. Note also that IGBT's switch slower than FET's; but as you watch Luke, Jeremy and I play with the big FETs you will likely see us slowing down the switching speed to near IGBT territory anyway to control inductive spikes... If we only had a supply of those EMPCs (Extraordinary Macedonian Poly Caps) we could shed 2 kilo's of low ESR electrolytics!

Here is some updated driver listings from IR
http://www.irf.com/whats-new/nr100825.html

Semikron has many products that would be suitable (but probably expensive) and are available from multiple vendors

They have 6 switch packages with various voltages and currents in MOSFET format here:
http://www.semikron.com/skcompub/en/semitop-105.htm
They also have IGBT in the same sort of packages, http://www.semikron.com/skcompub/en/semitrans-83.htm

Here is a complete controller (minus the software) in a MOSFET package
http://www.semikron.com/skcompub/en/skai_mosfet_system-3636.htm

Comes in an IGBT package aswell
http://www.semikron.com/skcompub/en/skai_igbt_system-3635.htm

How about these? http://ixdev.ixys.com/DataSheet/99022.pdf

Arlo1 said:

How about these? http://ixdev.ixys.com/DataSheet/99022.pdf

Click to expand...

It is OK, if you could get a zillion of them for cheap, it would definately be worth using. It uses the older technology junction design (more heat). The hot ones in that package style are the trench-fet designs by IXYS.

And what about these? http://ixdev.ixys.com/DataSheet/98653.pdf

Arlo1 said:

And what about these? http://ixdev.ixys.com/DataSheet/98653.pdf

Click to expand...

The IXFN 150N15 beats the IRFP4668 in case to sink thermal conduction 0.05 vs 0.24 deg C/Watt; but is not superior in other characteristics. For example RDSon 25deg C 12.5 vs 8.0 mOhms, and gate capacitance 9100vs 10720 pf > implies a larger silicon die on the IRFP4668. I wish IXYS would put complete data sheets in the public domain. You can't make a fair comparison without parametric graphs that depict temperature effects on the listed parameters.

Thanks Bigmoose.
I Am just looking at what else might work well at the 100 and 150 volt areas to try to find something a little more current.
On of my friends mentioned a lot of these fet manufactures stoped producion when the econamy took a dump.
But the way I see it things like the 401 (the 100v 680 phase amp) fets should be continued for the production of the coming hybrids ect.

What about theses:

http://cgi.ebay.com/CM600DY-24A-IGBT-3-pieces-/260654973296?pt=UK_BOI_Electrical_Components_Supplies_ET&hash=item3cb03f4170
1200v 600A IGBT that would make a pretty good start. Dual like LFPs Hyperfets.
Three of these and you would have a pretty nice controller.

Regards
/Per