Kelly naked - all washed up...with pictures

[ZapPat]

One thing you might want to change along with those FETs are the gate drive resistors, because those cheap IXYS FETs they use now have very low gate charge compared to IR4110's. If you don't do this you might get excessively long switching times, making more heat specially under low RPM + high current use conditions.

Those IXYS FETs might well have poor Rds-ON, but are actually good switching devices compared to IR FETs in general. This along with the fairly good looking electrical layout might suggest that they optimize their switching timings, and if so care should be taken when heavily moding the power stage.

 

[dnmun]

i feel like he should populate all those capacitor spots too. if he had 400+ amps switching, the di/dt could be what ate up his FETs.

from the statement about the amount of charge needed to switch the gate versus the higher resistance FETs, is this related to the drain source resistance?

the channel is more heavily doped to get the low on resistance? so the gate charge has to be larger to suck the carriers into the conduction band since there are so many more of them to suck on? so thicker gate oxide dielectric too? or just the opposite, thin oxide with low doped channel for the higher breakdown voltage?

 

[ZapPat]

i feel like he should populate all those capacitor spots too. if he had 400+ amps switching, the di/dt could be what ate up his FETs.

from the statement about the amount of charge needed to switch the gate versus the higher resistance FETs, is this related to the drain source resistance?

the channel is more heavily doped to get the low on resistance? so the gate charge has to be larger to suck the carriers into the conduction band since there are so many more of them to suck on? so thicker gate oxide dielectric too? or just the opposite, thin oxide with low doped channel for the higher breakdown voltage?

Good idea about the caps for sure.

As for the gate charge, I am no semiconductor pro or anything close, but the most important difference here would likely be the area of the die being larger on the 4110. Of course there are many other differences between the IXYS device and the IR one such as doping, gate channel shape/size/insulation thickness and others I'm sure. What I have noticed is that some companies have concentrated on fast switching technologies (like infinion, IXYS,?..) where others like IR have gone for brute low resistance techs. However what is really important here is the actual differences in device caracteristics, and that these differences might require some re-tuning of the gate drive circuit, specialy so if the circuit was optimized to start with. "Regular" ebike controllers use pretty basic FET switching setups and are slow switching, and thus have much room for change. The kelly might well have a better designed switching circuit for higher efficiency at partial duty, but then might also be more fineky about changes made to it. Only a scope would tell for sure...

 

[CroDriver]

Hall, you pulled that 470Amps out of a 1p setup?!

That's 47C!!

I suppose that these are the cells you gave me. It will be very interesting to test their capacity.

I found some good battery testers: http://www.alibaba.com/product-gs/267700013/Battery_tester_BTS5V3A.html

They have only a Chinese site. www.neware.com.cn Google translate does a great job translating it to English

They also have testers capable of 100+kW. That would be nice equipment for a lab

 

[HAL9000v2.0]

Hall, you pulled that 470Amps out of a 1p setup?!

That's 47C!!

I suppose that these are the cells you gave me. It will be very interesting to test their capacity.

I found some good battery testers: http://www.alibaba.com/product-gs/267700013/Battery_tester_BTS5V3A.html

They have only a Chinese site. http://www.neware.com.cn Google translate does a great job translating it to English

They also have testers capable of 100+kW. That would be nice equipment for a lab

No, those are the old cells yours are newer. And yes, 470A from 10Ah pack but that was not my intention after that they lost some 20% capacity.


The lab will wait...

 

[CroDriver]

The lab will wait...

Where did I wrote that I'm talking about your lab?

I have some room left in my garage

 

[dequinox]

Wow...

ed that one didn't ya?

That sucks but I am also always impressed with your builds, they are something out of this world!

 

[enoob]

so far im rather exited about this thread so bumpski back to the top.

dude my neighbor is from croatia . ill get him online hes right into what im doin scored me some batts from his work (and so his work shall remain nameless )

And now im gonna pretend im aussie jester and say wheres the vid man ? racing a motorcycle ? video man video. dont care who won . fact that it went on for more than 10 seconds tells me i missed something

any latest on the mod job ? ill be trying to follow suit in the not to distant future im guessing.

 

[orpheedulogis]

My experiment

http://motoelectrique.unblog.fr/2011/04/26/problem-with-kelly-controller/

Don't buy waterproof controller !

 

[Face86]

Hello everyone!

I opened a KBL72301 and unfortunately I found hard epoxy potting inside. I'm sure that is epoxy not wax, acetone didn't work for me.