[dtbaker]

I am wondering what the general consensus is regarding buying used controllers. I see some come up in classifieds due to upgrades or 'retired' builds or whatever, and I am wondering if controllers have any kind of life expectancy based on regular use... or if it more like they'll keep going as long as they don't get fried somehow?

 

[dexion]

Thats really hard to quantify. The weaklinks vary from controller to controller. But a good rule of thumb is heat kills (among other things.) But you cant tell if a controller was abused by looking at it. I suppose you could crack it open and see is things arent in spec any more. That seems like more trouble than its worth. There are controllers with better and not so good reputations but there are always exceptions to those rules as well. I guess its a crap shoot. But assuming it works, keeping it cooler raises your chances of having it last.

 

[dtbaker]

The weaklinks vary from controller to controller. But a good rule of thumb is heat kills (among other things.) But you cant tell if a controller was abused by looking at it.

yeah, this is why of all things I am hesitating considering a used controller. Electronics are hard to spot former abuse until they pop.

 

[grayballs]

5 or 6 yrs ago I bought a dozen Curtis 1209B's (72v). Of the 5 that I have personally put into use all but 1 (operator error, no precharge) are still functioning. 1 of the units, I sold, died afer a period of time (I think,,,,the buyer never mentioned it to me directly). Whether that was a problem with initial hookup, precharge, use or ????? I don't know. Considering what I paid for the units, I think it was a good deal,,,, still have some left if (when) one I'm using fails.

 

[Tesseract]

...I am wondering if controllers have any kind of life expectancy based on regular use... or if it more like they'll keep going as long as they don't get fried somehow?

Perhaps unsurprisingly, the answer is, "it depends".

One inescapable problem for all controllers is thermal fatigue of the power semiconductor dice. Thermal fatigue is caused by repeatedly cycling the semiconductors between hot and cold (relatively speaking) temperatures and it eventually results in the actual semiconductor chip ("die") detaching from the heat spreader. The higher the ultimate temperature reached each cycle the fewer number of cycles that can be tolerated.

This number can be as few as 10 cycles if the ultimate temperature is 175C or as many as 10^6 cycles at 100C peak. Just depends on the type of solder used for the die attach and the difference in expansion between the dice and the spreader, etc...

The short conclusion: stop-and-go driving is somewhat harder on controllers than highway driving is (just like with gassers - whodathunkit?) and keeping the controller as cool as possible is always a good idea (e.g. - use liquid cooling where possible).

The other potential problem for controllers that use electrolytic capacitors is that they will eventually dry out from having to support excessive ripple current (which leads to internal heating). All capacitors are rated to sustain a certain amount of ripple at a certain external temperature (frequently either 85C or 105C). Unfortunately, withstanding high ripple currents is not the forte of electrolytics (a high capacitance per unit volume is), so what happens in most controllers that use them (including the Zilla) is a bit of statistical arbitrage - the designer uses capacitors with a total ripple current capability of as little as 1/20th of the calculated maximum and banks on the assumption that maximum ripple will be experienced rarely. Not the sort of bet I like to make, but, e.g., the Zilla only has about 150A of continuous ripple current capacity and it seems to do alright.

All that said, I think that if you are looking at decent controllers and the seller guarantees it to not be DOA then I would go for it.

 

[Woodsmith]

I suppose it also depends on how much it is going to cost for the used one.

My Curtis 1214 for my tractor was bought on Ebay for £60. The seller had two and I got one of them. He said that the other buyer got the one that had been installed, in a gokart, for testing as there was proof of it working. I got the one that was still in its box and never conected. The seller had no more proof of it being good then I did but for the price I wasn't arguing.

My next controller will most likely be an Open Revolt kit at $600 plus shipping but I think I would rather pay that for a new kit then the same for an unknown used controller (unless it was a gently tested Soliton 1 or Jr for about the same price ).

For me it is the fear of a costly sealed 'black box' of bits that I can't check for wear and tear in the same way I can with a transmission or a motor. A near dead power transistor or ic chip looks the same as a fresh new one to me.

 

[dtbaker]

For me it is the fear of a costly sealed 'black box' of bits that I can't check for wear and tear in the same way I can with a transmission or a motor. A near dead power transistor or ic chip looks the same as a fresh new one to me.

thats my fear too....

so, a question to the EE guys... is there any way to test a controller to see if it is in good shape versus been abused and may not last much longer?

 

[gottdi]

My first controller was a Golftech 72 volt 550 amp and it was never connected but OLD as a hatter and worked perfect. I put a good heat sink on it and it performed flawless. Sold it and it's still being used. I don't think you could open up a controller and tell if anything was wrong. It's not like you can see wear and tear like you can on worn metal parts.

Was the controller properly heat sinked is the biggest issue in my book for used. It's the only used controller I ever purchased.

 

[Woodsmith]

I suppose that's the other side to it.
If you get a working used controller then treat it really well, good cooling and heatsink, and not thrash the remaining life out of it too quickly.