Paul & Sabrina's cheap DIY 144v motor controller - Page 453 - Fuel Economy, Hypermiling, EcoModding News and Forum

harlequin2 · 03-22-2011, 04:41 PM

Horses for courses, as they say! I'm not fixated on one or the other and the reasons for using either IGBTs or FETs have been covered in these posts. I note that the LogiSystems 1000A controller that I started out with used IGBTs rated at 600V and had enough of them to carry about 2500A. But it still blew up, so poor design will be a much more important factor in the longevity of these things than device selection alone.
I think you will generally find that most designs up to 200V use FETs and above that IGBTs.
By the way, if you study the manufacturer's spec sheets, you find that not all IGBTs have a negative temp coefficient. Many of them are almost zero and some are slightly positive and you can parallel those ones very readily. IR have some good app notes on this. I read a lot of them while repairing the LogiSystms thing!

sawickm · 03-22-2011, 06:53 PM

Quote:

Originally Posted by sawickm

I just opened Paul's email, I have his files and will be uploading them to my file server, and modifying the Revolt wiki with the new links in the next few days.
-Mark

The 1000amp Cougar files that Paul gave me have been added to the wiki.

Open ReVolt - EcoModder

- Mark

isaac_alaska · 03-22-2011, 08:56 PM

Quote:

Originally Posted by harlequin2

By the way, if you study the manufacturer's spec sheets, you find that not all IGBTs have a negative temp coefficient. Many of them are almost zero and some are slightly positive and you can parallel those ones very readily. IR have some good app notes on this. I read a lot of them while repairing the LogiSystms thing!

I read this somewhere, i think i might have even posted the findings on here somewhere. one of them was a really good write up on igbts, and i think it included some test examples as well. i do remember that for the most part, NPT type igbt's have a somewhat positive temperature coefficient, and IIRC it actually increases with temperature and current. I'm not too worried about it, as i'm pretty sure that two of them in parallel would be more than enough for my application, i was going to add a third to increase efficiency somewhat by reducing current through each device. Paul is setting me up with the bg2a igbt driver circuit, which takes two vla500 modules. i think that each of them is good for controlling at least two igbt's.

would it be bad practice to use one of the vla500 modules to control 2 igbt's and the other vla500 to control the remaining igbt? or am i trying to make it way bigger than it needs to be, and i should just stick with running two 400 amp igbts?

MPaulHolmes · 03-22-2011, 09:36 PM

I think it would not be best to drive 2 with one driver,and 1 with another driver. I think the 1 would turn on faster maybe because of less gate charge? I don't know about that. One driver would be putting out more current than the other. Hmm... I don't know. I just think that symmetry is really important at very high currents. How about 4 igbts! ya! You want the other competitors to just quit and go home, you don't want it close.

isaac_alaska · 03-22-2011, 09:51 PM

Quote:

Originally Posted by MPaulHolmes

How about 4 igbts! ya! You want the other competitors to just quit and go home, you don't want it close.

i like your style! 4 igbt's it will be. i'll be able to drive a tank with it

isaac_alaska · 03-23-2011, 05:47 AM

NEW! Fuji IGBT Module 1MBI300F-060 600V 300A - eBay (item 190318302428 end time Apr-21-11 09:37:55 PDT)

found these. i bought four of them cause they seem dirt cheap, especially for NEW devices. also, the switching times on them is pretty nice. few micro seconds on and a few more to turn off. should be able to run at 20 khz pretty easily. (would like to stay out of the audible range.)

eight left on ebay if anyone is looking for igbt's. i think they'll work nicely

adamj12b · 03-23-2011, 08:07 AM

Some things to remember about igbt modules. You need a 2 pack module minimum to build a controller. You need a switch and a diode. So with those modules being 1 pack modules, you will need 2 of them to build a controller. That means with the 4 modules you got, you will be able to build a 600A controller before derating. it would be wise to cut that 300A limit each back to probably 225A to 250A max.

-Adam

Nevyn · 03-23-2011, 10:43 AM

Quote:

Originally Posted by adamj12b

Some things to remember about igbt modules. You need a 2 pack module minimum to build a controller. You need a switch and a diode. So with those modules being 1 pack modules, you will need 2 of them to build a controller. That means with the 4 modules you got, you will be able to build a 600A controller before derating. it would be wise to cut that 300A limit each back to probably 225A to 250A max.

-Adam

I'm going the IGBT route myself. I sort of understand what you mean (and what Jack talked of earlier, but I want to make sure I know EXACTLY what I'm doing before I turn money into blown IGBTs.

If you've got 'dual' IGBT mondues, how do you connect them? I know one side is the power side and one side is the freewheel side; or at least I think I know that.

Is it something easy to explain or diagram for people who feel like they're losing to a box of rocks when it comes to understanding all of this?

isaac_alaska · 03-23-2011, 12:17 PM

Quote:

Originally Posted by Nevyn

I'm going the IGBT route myself. I sort of understand what you mean (and what Jack talked of earlier, but I want to make sure I know EXACTLY what I'm doing before I turn money into blown IGBTs.

If you've got 'dual' IGBT mondues, how do you connect them? I know one side is the power side and one side is the freewheel side; or at least I think I know that.

Is it something easy to explain or diagram for people who feel like they're losing to a box of rocks when it comes to understanding all of this?

I'm in the same boat. i was under the impression that the diode simply went in parallel with the igbt and pointed back the other way. but with a two pack module...they're in series. but the other part of the module couldbe configured in parallel i suppse, if the emitter of the first was tied to the same point as the collector of the second....i think there's something important here that i'm missing. can you maybe give us a simple picture or drawing of how they have to be connected? also, a physical second piece could be used as a diode, correct? (as you mentioned that with the ones i bought, one would have to be set up as a diode)

if so, i have two of these which i was kind of planning to use as diodes, once i know how to connect them..
2 - Hockey Puck SCR 800V 550A T720 Welder Power Supply - eBay (item 300539348666 end time Apr-20-11 11:53:48 PDT)

as a side note, how well do diodes parallel? i'm guessing it's not a huge issue, since the mosfet power board must use a dozen diodes in parallel...i think.

ElectricZX2 · 03-23-2011, 01:13 PM

Quote:

Originally Posted by jackbauer

Don't use scrs. They are and were used in forklift controllers with a capacitor discharge commutation circuit. real stone age stuff. Have a look at the videos on my website in my sig to see the controller build. I honestly never heard of vce rising at higher currents. If it did i don't think desat detection would work correctly. In any event i can get the hardware current limit to kick in without causing the devices any distress.

I second not using scrs. I ran a 500amp model for a while. They fail in funny ways if you run voltages too high above their rating. Mine failed on as I was coming up on a red light with a long line of cars...like forgetting you had the cruise control set but your brakes don't do anything. Luckily there was a dirt shoulder and the ignition contractor still worked. Getting home was interesting using controller bypass...