DIY Electric Car Forums banner

KELLY controller KDH14501 Post mortem 2

9331 Views 16 Replies 10 Participants Last post by  mamodbasher
Hi guys, my first posting after a Kelly experience. (I had hoped on a good outcome, but..):(

I carefully choose the KDH14501 for my 96V system. 144V (180V max), 250A continuous, 500A max.
I had 2 of them. Each one diseased smoking after about 30 miles.

After the first one blew, Steven Li was so kind to sell me an other one for about 500$.
He asked me to dispose of the bad controller, not to tell about it, and promised to send me a new one, 4 inches longer and also stronger. (why would some one ask for such a thing?)
When I looked through the manual, and saw that not only the dimensions had stayed the same but also they had reduced the ampere rating down from 250 to 200A, Steven Li confirmed, that “the case is longer”, and the 200A was due to a “different Testing procedure” but really the new ones are stronger.
After having read Tesseract`s analyses of a KDH12600, I know that Kelly had realized people actually expected to see anything close to what’s published on the label, and readjusted the performance data labels.
“Different testing procedure”? ?
I told Steven my doubts, and he finally sent me another KDH14501 for a reduced price, my 2nd controller. I was a little bit astonished, since I had left the choice open for Steve to pick any controller he found suitable, I had told him that I had no need for regen. Actually no one knew why the KDH`s fets had blown, but the new controller was supposed to be a completely new designed series, much stronger and even had an optocoupller!
Well, it neither had different dimensions, nor was it any different to the old one: Same place, old procedure, after 30miles engine out, controller smoke, I was going downhill this time- couldn`t have overloaded.
I`ll continue this post, so that’s why I`ll go into details later.
Now it was time for me to get to know Kelly`s praised customer service.
If you pay, you can buy something-
Steven found the reason for the failure right away trough a distant diagnosis. I had told him, that a dcdc converter, supplying the fieldwindings with about 3A at 70V, was connected in parallel to the controller. Here`s what he wrote: “The dcdc loaded the prechargresistor. So the prechargeresistor won`t be able to charge the controllers internal capacitors….”
Up to this point I wasn`t sure where to place Steven Li. A mix of used car salesman and electronic engineer? I read through some of his statements on the internet, also the statement where he outed himself as an engineer. Now I know, those technical statements didn`come from him. It took him some time to respond too. Well it didn`t take him much time to come up with this beauty- “The dcdc loaded the prechargresistor”

After having gone through installing and deinstalling and a thorough analyses, I can say: Kelly controllers will die for 2 reasons if properly installed, of course with a prechargeresistor.
1. Transient voltage, which is the case when using them on 96V or higher. Transients build up to hundreds if not thousands of volts in the form of spikes, that last only may be for 10^-9 s. Kelly uses completely unprotected Fets rated at 200V. Again, I`ll write details later.
There is no way to adjust for a shutdown if the voltage reaches 136V transients!!!
And here goes the whole VB Kelly program, its for playing.
2. Overcurrent is the logic consequent failure cause for a Mosfet assembly with a gate circuit layout that`s built under the premises that “electricity travels fast”. It sure does, and the first FET gets all the current- every time its turned on- and that’s pretty often.
The Kelly labels represent the components max values added up, not real tests, those the paying customer does - or they change the testprocedure- until the desired outcome takes place- see above.
Others have found out, unfortunately I read teseracts analyses just too lat.
In my case it sure wasn’t over current, since Kelly had obviously down rated the current to under 50A. One can easily imagine that my car didn’t drive very fast, with the second controller with only a max of 60V on the engine. But I didn’t mind, hoping that it would at least last longer than the first one. Wrong- 30miles- smoke- out.

Anyway, I lost money. I went through a lot of install- and reinstall. I even got to drive, and I gained wisdom: Sometimes saving can be expensive. And I found truth: Don’t trust Chinese labels. 144V 400A!!! They smile at you, and it`ll cost you-
After over 2000$ spent, I not only bought 2 controllers, but also other gear- some didn’t break yet, I have no controller.
I had sent the 2nd KDH145xx to Kelly for repair. Steven Li had earlier pointed out, that in addition to being stronger and longer, the new KDH would have optocoupling, meaning the units would be repairable in the future.
Kelly stuff is not repairable: When a Fet, that lacks preventive circuitry is operated beyond its limits repeatedly, it explodes and flames can be the result. (That’s why in some applications protective circuitry is necessary). Fets fail open in all directions! Meaning a single failure would lead HV to the gate and other Fets. A optocoupler can prevent HV to go to the uC section, but that’s the only thing left- maybe.
Now the brown sauce, that Tesseract described, starts to make sense:

First it helps to hold everything in place, especially the fets, that are fixed to the board with tiny plastic screws.
Second, who wants to freely exhibit such simple unprotected circuitry? (Is that why Steven asked me to not show, and to dispose the first KDHB?)
And third, Kelly`s electronic is completely irreparable! One single failure- and everything goes! So why not put brown sauce on such a masterpiece?
Fourth: Why repare? When Kelly sell for a reduced price- looking inside the kelly- I saw what tesseract meant- The parts for the 14501 are worth maybe 100$ that is in the US, or EU! Excluding the metal case, which they`ll use over again. So every unit sold for a reduced price means profit. Why should they build repairable units?

Of course Steven Li allowed me to buy an other controller- for a reduced price. And he pointed out, that the KDHx14xxx is again stronger. And he put emphasis on checking the prechargeresistor, and he wrote, that other products also breake if not handled just right.
Bottom line, I`m to stupied to do it right, everybody has to learn… he wrote.
I have a degree from a European university in electronic engineering, with scope on power electronics. When I read Kellys manuals, I thought “looks good”
Now I remember: There was a company that had bought old Intel processors, overpainted the components markings and sold them world wide as new chips.
Got to be careful trusting labels - especially when something is sold cheap! :confused:
See less See more
1 - 1 of 17 Posts
[quote SCG

I do have a Kelly (KDH09401), and I am not being a fanboy or anything, but I just want to understand why controllers have been failing. So far I have found all the failures to be "misuse" of product. Also, the KDH14xxxx have had the most problems I think.[/quote]

May I ask how to avoid the"magic smoke" I have read about with the Kelly controller. I have the KDH 9500 to use with adc 9" and the PB6 potbox with the third wire added. The 300k precharge resistor was included,along with small diodes for the contactor. Any help would be greatly apreciated. Take care Watt.
1 - 1 of 17 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.