Awesome new FET package :) :) :) Not a geek? you won't care

liveforphysics said:
So, yeah, you get a 24 fet equalivlant controller in a package the size off a 6 fet.
Click to expand...
Still the same amount of heat though, isn't it? That can bump the size up a bit.
But, that's manageable with a good design. :)
 
$250 collosus controller!!!! that might actually fit on the bike!!, or in it!!

This might be able to find a use. mabye on a budget build. :p

lyen you paying attention?
 
hydro-one said:
$250 collosus controller!!!! that might actually fit on the bike!!, or in it!!

This might be able to find a use. mabye on a budget build. :p

lyen you paying attention?
Click to expand...

Hi Hydro-one,

Yes, I am watching things closely. But the challenge is the initial investment and implementation. I will see what I can do. :)

Regards,
Lyen
 
The Ixys looks great hope this will work out. The isolated tab is already worth it ! It even looks very good for producing them factory-wise in many numbers , fantastic future product to be.

My knowledge isn't that great concerning mosfets but I tried to compare the Ixys to TO-247 , it's about the equivalent as 2 x IRFP4368 or 3 x IRFP3077 current-wise and only slightly more expensive. Rds also practically around same area.
Interesting though is the drain current vs. case temperature , max current is gradually getting less from 25 degree c while the TO-247 packages can deliver the same current until about 120 degree c and only then taper off. So there is a certain cross-over point where the TO-247 packages are better ! The IRFP4368 is certainly very good. Size-wise the Ixys is the same as 2 TO-247 IRFP4368 next to each other , it should be possible to make the same size controller with both.

What I mean to say is : you can still build a compareble controller with the older stuff , but only if some one would start making them :wink:

But the coming time will be very interesting , I think a lot of people are going diy controller crazy at the moment (atleast in their minds) , so it's only a matter of time !
 
Fligh High said:
Interesting though is the drain current vs. case temperature , max current is gradually getting less from 25 degree c while the TO-247 packages can deliver the same current until about 120 degree c and only then taper off. So there is a certain cross-over point where the TO-247 packages are better !!
Click to expand...


You're looking at it a bit incorrectly. :)

The TO-247 package irfp4468pbf has an RdsOn increase from 25degC to 175degC of ~240%
The TO-220 package irfb4110pbf has an RdsOn increase from 25degC to 175degC of ~255%

This slick package has an RdsOn increase from 25degC to 175degC of ~205% (about 15-20% less effected by temp increase than the leading IR TO-220 and TO-247 parts).

The limit you're seeing on the other FETs is the package limit (how much current the leg/bond-wires can hold) showing a plateau. This FET has a lead structure designed around handling many hundreds of amps, so you don't see this package limit plateau, and you see only the silicon limits reflected in it's temp vs current graph.
 
liveforphysics said:
The limit you're seeing on the other FETs is the package limit (how much current the leg/bond-wires can hold) showing a plateau. This FET has a lead structure designed around handling many hundreds of amps, so you don't see this package limit plateau, and you see only the silicon limits reflected in it's temp vs current graph.
Click to expand...

Sure , I think I understand the silicon/bond-wires limits , but in real life use wouldn't 2 x Irfp4368 be the same current-handling wise as 1 Ixys if they reach "working" temperature in practical use ?

How would you program the max phase current-limit in a controller btw , because it can't be a fixed 500 amps in case of the Ixys because it is less capable when temperature rises but the 2 TO-247 Irfp4368 limit can be set allmost to a fixed 390 amps because of the plateau (all veeerry loosely compared of course)

You know a lot about this stuff and I am learning all the time , please tell me how you think Irfp4368 would hold up in a diy controller untill the Ixys becomes available because I don't want to wait for the Ixys as you might understand :wink: Shouldn't we be working on the TO-247 stuff in the mean time ? It's all we have....
 
Fligh High said:
Sure , I think I understand the silicon/bond-wires limits , but in real life use wouldn't 2 x Irfp4368 be the same current-handling wise as 1 Ixys if they reach "working" temperature in practical use ?

How would you program the max phase current-limit in a controller btw , because it can't be a fixed 500 amps in case of the Ixys because it is less capable when temperature rises but the 2 TO-247 Irfp4368 limit can be set allmost to a fixed 390 amps because of the plateau (all veeerry loosely compared of course)
Click to expand...


Yep, the IRFP4368 is a very very good FET in a nice package, I based my controller design around it's 100v brother, the IRFP4468, which is also outstanding. The 4468, as well as the IXYS gigamos TO-264 series FETs are what I based my previous controller FET stages on, because they kick ass. Unfortunately, when switching these current levels, the inductive path becomes a huge factor in controller design and the ability to share current effectively at high PWM frequency (required for the low L+R motors) as well as complex cooling and cap layout/packaging needs. The only option to go beyond a 12-fet design and keep it inductively balanced is to use radial FET arrangements, which is the path I took on my controller layouts, because my linear layout design experiences (like my cell tester, which uses 8x IRFP4468s in linear parallel) taught me it's a fool's game to play with these current levels.

With this FET, the inductive path from having the entrance and exit on opposite sides, along with the sinking opposite to the connection side just makes these things a damn dream to layout! :) You can do a 6fet layout neat as a bugs ear with an inductive path that would make an RF engineer happy, and 12FET designs are equally fantastic, not to mention so much easier to construct than the radial designs. These things have seriously made my last few months of work on FET arrays seem like huge waste of time, because it's effortless to just bolt these things down to the backside of a CPU sink and have a beautiful and balanced compact simple and superior inductive layout.

TO-247 and TO-264 both are very good, worlds better than TO-220. However, this thing just dominates from the angles those packages miss on. :)
fet1.jpg
 
I had a dream last night that Luke and I went to a factory and were trying to get some boards built around these fets :lol:
 
If you could get your rc motor guys to use these fets, you would have a serious monster product. :)



johnrobholmes said:
I had a dream last night that Luke and I went to a factory and were trying to get some boards built around these fets :lol:
Click to expand...
 
If they were for a 48V device, what's the voltage limit of this things? If I can safely go to mid 82V, then up my guarantee from 24 units to 50.

John
 
liveforphysics said:
I got this skinny on this part.

There will be no 100v part. :( :( :(
Click to expand...



Do you see this changing in the near future 3 months 6months?
Have you got your samples that you ordered yet?
 
vehiculeselectriques said:
Hi,

>4 years have passed ... is there some usage news about this FET ?
Click to expand...


Yes, I have all I need of them now. They are extremely impressive devices. 75v version, 100v versions, and 150v versions. You can just buy them on digikey.

20130828_100213.jpg



They are perhaps the best suited towards motor-control power stages built around individual mosfets around AFAIK.

I've morphed my interests towards inherently lower inductance combined 3-phase leg modules now though... lol
75v version:
http://ixapps.ixys.com/DataSheet/MTI200WX75GD.pdf
100v version:
http://ixapps.ixys.com/DataSheet/MTI145WX100GC.pdf


I think those modules make the most sense for ebike power-level controllers now.
 
liveforphysics said:
I've morphed my interests towards inherently lower inductance combined 3-phase leg modules now though... lol
75v version:
http://ixapps.ixys.com/DataSheet/MTI200WX75GD.pdf
100v version:
http://ixapps.ixys.com/DataSheet/MTI145WX100GC.pdf
I think those modules make the most sense for ebike power-level controllers now.
Click to expand...

where does one get these from? Sample-only so far? Any pricing data?


boostjuice said:
Indeed, the single FET units are finally available in retail small quantities!

75V http://www.digikey.com/product-detail/en/MMIX1F520N075T2/MMIX1F520N075T2-ND
150V http://www.digikey.com.au/product-detail/en/MMIX1F360N15T2/MMIX1F360N15T2-ND
Click to expand...
Now that is a bit pricey... I wonder what segment are they targeting it at. 6*45 = $270 of mosfets alone.
 
circuit said:
Now that is a bit pricey... I wonder what segment are they targeting it at. 6*45 = $270 of mosfets alone.
Click to expand...

I was quoted at low as $14 each in forum-friendly group buy qtys.
 
Looks like those integrated 3 phase bridges have a similar heatsink path? Dissipation numbers look good, would indeed make a great ebike power stage. Almost idiot proof to lay that out.
 
johnrobholmes said:
Looks like those integrated 3 phase bridges have a similar heatsink path? Dissipation numbers look good, would indeed make a great ebike power stage. Almost idiot proof to lay that out.
Click to expand...

It would make the highest continuous power RC controller stage in the world as well if someone wanted to put it on a little heatsink with a high performance fan on it.
 
Its one helluva large footprint to work with though, in RC terms. Would be fine for 1/8th scale stuff, but tough to work into a 10th scale controller where the market could move enough volume to make it worthwhile.


But on an ebike, a 50x50 package could really output serious power. Wouldn't even need forced cooling at "bicycle" power levels.


Thanks for sharing Luke 8)
 
johnrobholmes said:
Its one helluva large footprint to work with though, in RC terms. Would be fine for 1/8th scale stuff, but tough to work into a 10th scale controller where the market could move enough volume to make it worthwhile.


But on an ebike, a 50x50 package could really output serious power. Wouldn't even need forced cooling at "bicycle" power levels.


Thanks for sharing Luke 8)
Click to expand...


My pleasure friend. :)

Fortunately, most 1/10th scale, or even 1/8th scale RC doesn't need a 10kW controller. :) The stuff it would fit fine in is the stuff that could use it's power. Hydroplane. 1/5-1/4 scale cars, but 1/4 - 1/5 scale RC cars, big helis and planes etc.
 
Hey, I can dream about shoehorning it into a 1/10th scale and watching the driveline explode :lol: You are right though, the craft that need so much power have the room for it.


I've been working on an off road mobility "chair" for a friend, and our biggest hangup is the controller. He wants to really bind the sucker up and push through tough terrain, which means a lot of low speed high duty throttle. We plan to limit power at 4000 watts per motor and use the stator pattern like a 9c, stack height TBD. im kinda hoping that we kill off the shelf controllers so we have a reason to build our own.
 
You must log in or register to reply here.
Back
Top