On 01/12/2017 12:00 AM, Lee Hart wrote:
Jay Summet via EV wrote:
I have been researching replacement parts for my Curtis 1231c power
board a lot tonight. All of the parts and numbers are listed at this
blog post:
https://www.summet.com/blog/2017/01/11/curtis-1231c-replacement-power-board-components/
I believe I will be able to upgrade the diodes to 600V 75A with the
DIOTEC 7506R.
Don't go by the ratings advertised on the front of a data sheet. They
are usually absolute maximums. The part would soon fail if used at these
limits.
Look *inside* the data sheet for the test conditions. For example, that
"75 amp" diode says it has a 1.35v typical voltage drop at 75 amps.
That's 75 x 1.35 = 101.25 watts of heat.
Now look at the thermal resistance from junction to case. It is 0.8
deg.C/watt. Therefore, the junction temperature rises 0.8 deg.C for
every watt of power dissipation. 0.8 x 101.25 = 81 deg.C.
Although I haven't found a datasheet for the existing diodes in my unit,
if they are similar to others Curtis used I think they are around 24A
maximum rating. So I doubt I'll ever need to use all 75A of the absolute
maximum ratings of the new diodes.
I cruse down the road at 100-200 motor amps, and with 18 diodes, that's
under 11.1 amps per diode. I believe that would be an average thermal
dissipation needed of 15 watts, (0.8 * 15 = 12 degrees C junction to
case difference).
The real question is when I punch it up to 500-800 amps (44 amps per
diode) for a few seconds from a dead stop, how long it takes to get back
down to average temp after dissipating 60 watts, or a 48 degree C rise.
Capacitance does not matter. What matters is the capacitor's ESR (as low
as possible) and its voltage rating (at least 25% more than the maximum
voltage you ever expect it to see).
Electrolytic capacitors have the shortest life of any part in the
controller. As capacitors age, their ESR goes up, and their voltage
rating goes down. At some point, one or the other of these gets bad
enough, and the capacitor fails.
Sounds like a good argument for paying the $140 to replace them with new
ones. The new ones I'm looking at have the same tangent angle of loss
(dissipation factor) which is derived from ESR and capacitance as the
ones in there now, but a 250 volt rating instead of 200.
I'm also considering a 250 volt 110A MOSFET (IXTH110N25T) to get the
MOSFET and capacitor max voltages both up to 250.
There is also a marketing "ratings game" with MOSFET specs. You have to
read the data sheet pretty closely to see what they can really do. It's
hard to be sure what part is better just from the data sheet.
Yes, I'm generally just over-engineering the heck out of the parts and
never hoping to reach their absolute maximum ratings. Cost is also
something I'll need to look at eventually.
Given the Curtis design, the on-resistance of the MOSFET is important.
My biggest possible gain in heat savings is by using a MOSFET with a low
Rds(on).
If I can adjust the current limit set (CL_MAX) and current limit adjust
pots (CL_ADJ), and/or modify their resistance range with replacement
components I might be able to have a lower Rds(on) AND still have
current sensing cutback work correctly. (although I may lose accuracy as
the voltage to sense would be smaller). I don't know if the current
adjustment pots have enough range built in to deal with 1/5 of the
resistance on the "shunt" or not.
Seat of the pants electrical engineering: Turn them both all the way
down and then see if it current limits below 500 amps or not while
accelerating from a stop ;>
Jay
_______________________________________________
UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
http://lists.evdl.org/listinfo.cgi/ev-evdl.org
Read EVAngel's EV News at http://evdl.org/evln/
Please discuss EV drag racing at NEDRA (http://groups.yahoo.com/group/NEDRA)
