Greentime controllers

If people want to run 24s LiPo with a 72v rated controller then simply charge the LiPo's to 4.0v. You will give yourself some leeway with the margin of error AND you will probably extend the life of your LiPo's. Personally I think trying to get 100v out of a 72v-rated controller is just asking for trouble in any case. Granted, its a PITA to have to switch to the 4115's but really, 88v is plenty. How fast do people want to go? And for how long? I agree with Jeremy on the reliability thing. 2000w in my X5305 in a 20" wheel at 72v {24s of LiFe} is plenty. Plenty to get up any hill, plenty to stay with traffic. Let us count our lucky stars that the Germans used to over-engineer their stuff too.
 
methods said:
I did high reliability government work for 10 years and we used IR fets almost exclusively. They made their big money supplying the government and military so they already have the margin built in - to the point where we could just use their parts off the shelf and not even raise an eyebrow at a design review. This was in an environmental where we would commonly spend $10k - $20k qualifying a part (every new part).

Obviously your advice is sounds (as it always is) but I would have to say that street testing has shown that running 24S on a 4110 controller can be reasonably reliable - especially if extra attention is paid to the capacitance. I think the key here is that the 100.8V surface charge burns off almost instantaneously. As soon as a load is applied we are talking about more like <95V with the standard 20C cells and maybe 98V with the super high C cells. On my bike it seems that the only time it really sees 100V is while it is sitting on the charger.

Of course all that depends on the pack size, wiring length, internal resistance of the cells, current limit, etc.

Of greater concern for me are the terrible Chinese capacitors that are used in almost all of the kits from China - My money says those electrolitics will dry up and pop long before the Internal Rectifier mosfets will let loose.

As a sidebar - Keywin is now offering high quality 3000 hour and 6000 hour caps with his kits. I was just expressing to him the other day the importance of ripple current and ESR over rated hours so he was going to talk with his Asia cap rep.

-methods

Hi Methods, supposing I was going to open up my controller and replace the capacitors in a 72v controller, but want to keep it as a 72v controller:

1. What procedural standards or processes would I need to follow??

2. What brand and/or rating of capacitor should I buy?

3. Where should I buy them??

Thanks. :D
 
I will release my capacitor BOM a little later - but basically all you have to do is open it up and replace the 4 or 5 big 100V 1000uF or 470uF with new caps. I just got done with a painful review of all the latest offerings from all the big companies - their availability, cost, performance, lifespan, ESR, ripple, sizing, etc. Then I made up a huge order with Digikey to evaluate them all and ran out of R&D money. Spent the remaining money I had buying 20pcs 18fet kit for a new run of those... I have to make some money and then I can invest more - (but anyway - that is the part of being a developer folks dont get... the needing to make money or go broke part lol)

Yea -
I am going to do a full up thread on how we build these controllers so that the DIY guys can DIY. Look for the cap information there.

It is going to be a "normal guy friendly" build - no crazy stuff. No matching mosfets or machining down heat sinks or anything crazy - just procuring quality parts and installing them in a way that will give us big burst power with reasonable continuous handling. Anyone should be able to duplicate the build with standard shop tools and no EE knowledge. BOM will be available from 2 or 3 sources like Digikey, Mouser, and a few bits from Ebay and Amazon.

We will call it the: "Everyday mans 420 controller" (for reasonable ebikes that need 400A burst capability)

-methods
 
I popped a cap on an infineon 12fet at a bit under 90v ( 100v caps), I just doubled them all including the little red caps, had no probs since
 
Awesome methods ! Ill be on board for these 420 controllers!! :mrgreen: I havent done a controller build , sick of caps exploding, ruining boards, etc. !!
 
The Mighty Volt said:
How do you double them? Double the number, double the quality, double the capacitance??

Cheers.

What's needed are capacitors with a high enough voltage rating for the voltage you want to run at (with a margin for reliability) and a low equivalent series resistance (ESR) so that they can cope with the high current pulses on the supply rails without getting too hot. The ESR is probably more important than the capacitance, as is making sure the capacitors are fitted with the leads as close to the power rails as possible (to minimise lead inductance). I suspect some failures we see on tweaked controllers are really down to the capacitors not having been uprated when the current/voltage was increased. Good capacitors with a low ESR mounted close to the FETs make life easier for them, meaning better reliability.

Not all capacitors are equal, even those with the same voltage rating and value. It pays to dig around in the specs to find ones with the lowest ESR that you can physically fit inside the case.
 
Well since were all talking caps, ive just scored a 18fet Xiechang/Infineon/Ecrazy controller, an will be doing the rails/caps ect to run 18s lipo
the standard caps are 63v 470 uf , but i just happen to have scored some rubycon 220v250uf caps on a pcb out of a stripped tv, should i use them and as the uf is low what is the worst that could happen.
 
Emoto said:
Well since were all talking caps, ive just scored a 18fet Xiechang/Infineon/Ecrazy controller, an will be doing the rails/caps ect to run 18s lipo
the standard caps are 63v 470 uf , but i just happen to have scored some rubycon 220v250uf caps on a pcb out of a stripped tv, should i use them and as the uf is low what is the worst that could happen.

Depends on what series they are, not all Rubycon caps have a good ESR. Check the part number against this: http://www.rubycon.co.jp/en/catalog/aluminum-Products_Eng.html and it should tell you whether or not the caps you have are low ESR ones or not (low impedance is what you're after, together with a high ripple current rating).

For example, a 470 µF, 100 V, YXH series Rubycon has an ESR of 0.04 ohms, whereas a 470 µF, 100 V ZL series Rubycon has an ESR of 0.033 ohms, so will have around 17% lower losses and be more effective at suppressing ripple.
 
A 200V and a 63V cap in the same can size will have dramatically different values as well... the price you pay as you go up in voltage is that you have to go down in capacitance. A perfect example is that you can make a 200V cap from two 100V caps by stringing them in series. If they were originally 1000uF each the new total capacitance will be 500uF. Otherwise - obviously we would all stuff the biggest 500V caps in these controllers we could find. Engineering tradeoff - capacitor head room for capacitance.

For me - I would rather have 1000uF 100V caps on my board than 470uF 200V caps all other things equal on a 4110 controller. I have not had caps blow up in one of my controllers since my very first 12fet 100V 100A controller back in the day.

I totally respect what Jeremy is doing - what I fear though is that in the ebike world people are not even close to starting to accept what the actual cost of building margin into a controller is. They will literally cost double and perform half as well... as in literally I will charge you double and it will have a smaller voltage range and much lower current limit.

There is a place for high reliability for sure.... but I think what most people what is high value dollar for dollar with reasonable reliability.

Chinese Lipo is a perfect example of this. Hobby King is a total craps shoot but everybody loves it... they very well could go for the expensive stuff but they roll the dice. The rich guys can afford the 15Ah EiG pouches and we can deal with the returns desk at Hobby King.

-methods
 
http://www.ebay.co.uk/itm/2x-NEW-1000UF-100V-105C-BC-058-LOW-ESR-CAPACITORS-FOR-AUDIO-/170743839288?pt=Vintage_Electronics_R2&hash=item27c1203a38

Opinions? Is this what we are looking for?
 
The Mighty Volt said:
http://www.ebay.co.uk/itm/2x-NEW-1000UF-100V-105C-BC-058-LOW-ESR-CAPACITORS-FOR-AUDIO-/170743839288?pt=Vintage_Electronics_R2&hash=item27c1203a38

Opinions? Is this what we are looking for?


They are 3cm wide and 3 cm tall .. that's a lot of volume needed.
 
My secret weapon is going to be drilling holes in the pcb to accommodate caps with spacing other than 7.5mm and 10.0mm
Second leg does not even have to land on a rail - just needs a clear run for a piece of 10AWG.

There are a ton of really nice caps out there in many flavors.

-methods
 
20s is a better RC Lipo ebike pack voltage anyway, because then the pack can be nicely paralleled with 23s LiFePo4 to the benefit of both battery types in the form of mutual support, same bulk charging (83V cutoff), etc. Keep your options open. Oh, and it's good for the controller too.
 
Those caps look good mightyvolt

Jeremy looked at that link, my caps are bxc-p good ripple what do you think [first time ive moded a controler.] i will have to lay these caps on there side and extend the legs, apart from the colored band down the cap body is there a right way to mount these.

Methods im looking forward to the everymans controler hack fest thread.
 
They're not ideal (but they are free!). The problem with fitting bigger caps is that you need to try and avoid having long leads to them, which can be difficult if they have to lay down to fit the board (as is the case with a lot of ebike controllers).

In the overall scheme of things caps aren't that expensive - the array of 220µF, 100 V, Rubycon ZLs that I've used in my home made controllers were around the equivalent of $1 each.
 
Jeremy Harris said:
The Mighty Volt said:
How do you double them? Double the number, double the quality, double the capacitance??

Cheers.

What's needed are capacitors with a high enough voltage rating for the voltage you want to run at (with a margin for reliability) and a low equivalent series resistance (ESR) so that they can cope with the high current pulses on the supply rails without getting too hot. The ESR is probably more important than the capacitance, as is making sure the capacitors are fitted with the leads as close to the power rails as possible (to minimise lead inductance). I suspect some failures we see on tweaked controllers are really down to the capacitors not having been uprated when the current/voltage was increased. Good capacitors with a low ESR mounted close to the FETs make life easier for them, meaning better reliability.

Not all capacitors are equal, even those with the same voltage rating and value. It pays to dig around in the specs to find ones with the lowest ESR that you can physically fit inside the case.

Cheers Jeremy. Low ESR seems to be a recurring theme in the Ebike world!!!
 
Jeremy Harris said:
They're not ideal (but they are free!). The problem with fitting bigger caps is that you need to try and avoid having long leads to them, which can be difficult if they have to lay down to fit the board (as is the case with a lot of ebike controllers).

In the overall scheme of things caps aren't that expensive - the array of 220µF, 100 V, Rubycon ZLs that I've used in my home made controllers were around the equivalent of $1 each.
Well after some thought i better do this the right way, and wait for some ebay caps 100v- 500/1000 uf do you have any good/cheap ebay suppliers.
do i only need 5?
Thanks again jeremy.
 
A good number of electronic components I've bought from ebay have turned out to be fake/counterfeit, so I can't honestly say that I'd recommend that source.

You have Digikey and Mouser down in your part of the world, I think, as well as some more expensive suppliers, like Element 14 (Farnell) and RS. Here's some links to probably the best, Rubycon ZL series capacitors, you will have to check to see how much room you have in the case to see which ones, or combinations, will fit in there:

http://www.digikey.com.au/product-detail/en/100ZLH680MEFC18X35.5/1189-1035-ND/3133962

http://www.digikey.com.au/product-detail/en/100ZLJ560M16X35.5/1189-1046-ND/3133974

http://www.digikey.com.au/product-detail/en/100ZLH820M18X40/1189-1037-ND/3133964

http://www.digikey.com.au/product-detail/en/100ZLH470M16X31.5/1189-1033-ND/3133960

http://www.digikey.com.au/product-detail/en/100ZLJ820M18X40/1189-1048-ND/3133977

http://www.digikey.com.au/product-detail/en/100ZLJ470M16X31.5/1189-1045-ND/3133973

Be wary of just any old capacitor you're offered, for example a Nippon Chemicon KZE series 470µF, 100 V will have an ESR around three times greater than a Rubycon ZL series of the same capacitance and voltage rating. The Nichicon PM series aren't too bad, if you can't get Rubycon ZLs, they are maybe 20% or so worse.

Also, watch out for obsolete types being sold off on ebay, like this ad: http://www.ebay.com/itm/Rubycon-capacitors-16mm-25mm-100V-470uF-470-105-1-PCS-/220881518852?pt=LH_DefaultDomain_0&hash=item336d908104 which is for a type Rubycon stopped making a long time ago. It's also a pretty crap capacitor for a controller!
 
Ebay is fail for electronic components. The only time I have been lucky is buying large lots and that is usually for power resistors or other things that dont go bad. Blow off the cobwebs and apply power. Caps dry out and are very easy to counterfeit... friend of mine sent me pictures of entire shopping malls where you can just buy "genuine" labels, stickers, heat shrink tube, etc. Then pictures of guys sitting there putting Intel stickers on chips....

-methods
 
methods said:
friend of mine sent me pictures of entire shopping malls where you can just buy "genuine" labels, stickers, heat shrink tube, etc. Then pictures of guys sitting there putting Intel stickers on chips....

-methods

That's the massive electronics market in Huaqiang North Road, Shenzen. Probably the place a great deal of the stuff we see on ebikes comes from..............
 
Back
Top