Changing the LVC in a 72v generic controller

auraslip

10 MW
Joined
Mar 5, 2010
Messages
3,535
Edit:


Ok. It doesn't help that none of the parts are labeled in this section of the board.

Basically, from VCC there are SMD two resistors in series. After them, there is one SMD resistor that leads to the MCU and another SMD resistor that leads to ground. This last resistor is the one you need to remove and raise the value of. You can do this by replacing it with the appropriate SMD resistor, or by drilling a hole in the trace and attaching a regular resistor to it and to ground. In the pic above the hole is drilled on the left, and the resistor connects to a random pre-drilled hole that connects to ground on the right.

Drilling a hole in the trace is a bit of a pain though. Need a real small drill bit.

edit by request

I ordered one of these on a lark while browsing alibaba. Unfortunately for me I don't actually have a 72v battery pack and this controller, to my knowledge, can't be reprogrammed to suit my needs. Because I can't test I'll just sell it as is, but I don't think you'll have a problem!

Oddly enough another got a batch of the same controllers as me at the same time! He is selling his for 80 pounds, which might be fair. I really have no idea what to price this at! Bmsbattery and evassemble sell a similar priced controller for $30, but with shipping and shipping times it's not even worth it.


decent coat of thermal paste

100v caps

board and ic number. I can find absolutely no info about this online



not too shabby traces

Decent waterproofing
iFf8ll.jpg

Any one know what the numbers mean? 1.1-4.2v? Adjustable LVC? I might not sell it if I can get it to run on 48v :)
 
auraslip said:
Unfortunately for me I don't actually have a 72v battery pack and this controller, to my knowledge, can't be reprogrammed to suit my needs.
Does it not power up on your existing pack?

decent coat of thermal paste
Does the bar actually make contact with the entire surface/length of the case? My very similar one does not, over a great deal of it's surface area. Kinda makes it difficult to transfer the heat out that way. :lol: Even more important with your size of controller than with mine!
http://www.endless-sphere.com/forums/viewtopic.php?f=4&t=21830

Any one know what the numbers mean? 1.1-4.2v? Adjustable LVC? I might not sell it if I can get it to run on 48v :)
That's probably the throttle input voltage range.
 
auraslip said:
Does the bar actually make contact with the entire surface/length of the case?

Yup....poor design not to...
Yeah, but mine doesn't, either because of poor machining or a factory handling problem that resulted in bending the bar or the case. ;)
 
Hey amberwolf - do you want to trade controllers?

I made an attempt to hook it up to my 48v pack, and it put out 4.8v though the hall and throttle, but I couldn't get anything else to move!

If not, I'll just order from Lyen. I'm running low on funds though!
 
Well, first I'll have to see if this one works on 48V (probably), but sure, I'd go for that--you would not be getting the good end of the trade, though. ;)

I'm guessing yours doesn't do anything because it thinks your battery is below LVC. :( If you had the time, we could probably figure out how to get past that.
 
Well, first I'll have to see if this one works on 48V (probably), but sure, I'd go for that--you would not be getting the good end of the trade, though. ;)

I'm guessing yours doesn't do anything because it thinks your battery is below LVC. :( If you had the time, we could probably figure out how to get past that.

Yeah it's a shame because someone with the know how would be able to adjust the LVC in a matter of minutes, but I'm staring at the board for like half an hour and getting no where.
Actually I'm going to take it to my programming teacher tomorrow night. He did a lot of work with stepper motors in electron microscopes as well as ICs, so he might be able to reverse engineer the software or maybe play with the LVC settings.
 
I would guess that changing the resistive voltage divider made up of the big power resistors (like the one that gets hot on your other controller) so that it is calculated for yoru pack voltage instead of 72V, would probably also change the LVC on it, if it taps off of this divider to get the voltage for the MCU to read (since it can't do it directly anyway, tapping off this is simplest to do).

If you look in the "Infineon" threads that discuss changing the LVC (at least one each by Methods and by Knuckles, maybe even by Lyen and Steveo and others, too), there should be a ratio of resistance on those that you can use to calculate what it would be for 72V. If that same ratio applies to the resistors you have in yours, then you can use the same calculations to figure out what to change it to for your pack voltage. I haven't done that part yet, though I have wnated to for a while; just never time when I have mine out and am looking at it. :(
 
I would guess that changing the resistive voltage divider made up of the big power resistors (like the one that gets hot on your other controller) so that it is calculated for yoru pack voltage instead of 72V, would probably also change the LVC on it, if it taps off of this divider to get the voltage for the MCU to read (since it can't do it directly anyway, tapping off this is simplest to do).

So after the voltage dividers but before the 5v regulator I should look for a trace leading to the mcu?
 
Probably, but it may go thru buffer electronics like an op-amp or transistor, confusing things.

Changing the whole resistive divider values (proportionally) in those power resistors should do it, but you'd have to know what they should be.

I can tell you what mine are for 36V 12FET version:
file.php

Looks like 1Kohm for all three visible in that pic. I can open it back up and verify that.

Yours are all three 1.5Kohm, from the pic showing them above:
AIH9ql.jpg


What's yoru pack voltage again? 48V, right?

Hmm...72V's LVC would be what? Assuming 20s of 3.6V cells...3.0V per-cell LVC, probably 60V pack LVC.

36V of 3.6V cells would be 10s, so 30V pack LVC.

Some napkinmath that's probably screwed up ;) :
1.5K + 1.5K + 1.5K = 4.5K (60V)
1K + 1K + 1K = 3K (30V)
60 / 4.5 = 13.333333
30 / 3 = 10

13.3333 / 10 = 1.33333 v/kOhm? No, that doesnt' work.
10 / 13.3333 = 0.75 ? No, that also doesn't multiply or divide with the other numbers to give back a useful value. :?

I dunno, there has to be an easy way to derive some "ohms per volt" for the LVC, but danged if I know how. :oops: I have enough trouble getting 2+2 to equal five. Oh, wait. :( See? :lol:



So, anyway...48V of 3.6V cells would be 14s (50.4), so 42V pack LVC. If 13s (46.8V) then 39V LVC.
 
Auraslip, I have a brand new 48 volt Infineon controler from "Comcycle" for $125 about two years ago.

It has all the new connections like "CycleAnalyst" direct plug-in and, 100V caps.

Never used it even once, you can have it for $50, but you pay shipping, should be cheap though if you live close.

I'm always willing to help out a broke e-biker.
 
I used this calc to find parallel resistance http://diyaudioprojects.com/Technical/Electronics/parallel-resistor-calculator.htm

1.5k x3 is 500ohms --- 500ohms / 60v lvc = 8.3 omhs per volt
1k x 3 is 333ohms ---- 333ohms / 30v lvc = 11.1 ohms per volt

8.3 + 11.1 / 2 = 9.7 averaged ohms per volt for the lvc.

lvc for a 16s lifepo4 pack is 2.8v per cell 44.8 total?

44.8v * 9.7omhs per volt = 435 omhs

So I can just replace one of the 1500ohm resistors with a 1000ohm resistor..... I'll go to "the shack" tomorrow and see if I can find that resistor!
 
It seems much more reasonable the way you're doing the math. :) Me and numbers just don't get along. :lol:

I was thinking those resistors are in series rather than parallel, but now that I look at the pic, I can clearly see one end of them appears to be all on one connection. Can't see the other.

I guess you can measure yours with a meter first to verify the value you calculated. I just did that with mine (as now I can't get this out of my head, so I went and dug it out) and it comes out 331 ohms; I can clearly see they're all parallel on mine.

Hopefully even though they're in parallel the voltage divider I think they're part of is still what will fix the LVC. I had thought for some reason they were in series, and the divider would be coming off of them, so that means this might still be more complex than that. :(
 
auraslip said:
So I can just replace one of the 1500ohm resistors with a 1000ohm resistor..... I'll go to "the shack" tomorrow and see if I can find that resistor!
Hi folks, amazed and delighted to see someone dealing with exactly the same problem that was on my mind. I'd just figured out that these are 1.5K resistors needing a change to 1K, but my math was a bit simpler:

48/72 * 1.5K = 1K (You'll need only 2/3 the resistance value for 48V as you would for 72V to bring it down to the value you want)

However, you're talking about changing just ONE of the resistors. - Are you sure you have this right? As I understand it, each of these resistors serves one of the FET banks, of which there are THREE, therefore they are SERIES resistors serving a different section of the FET array. The layout of the tracks on the back would seem to support this: http://zenid10.wordpress.com/?attachment_id=1999
- Feel free to chime in, anybody, if I have this wrong...

I see from your later post that this hasn't worked. Try changing ALL THREE. Would love to hear if this works, because I was about to do the same thing myself :)

P.S: A piccy of my own 72V controller board with those same resistors is here:
http://zenid10.files.wordpress.com/2011/03/pcb1-1600x1200.jpg
 
Zenid - It'll blow your mind to find out that I also got some of the same dc-dc converters for sale that you were looking on selling :D
Looks like we've both been on alibaba lately 8)


In any case (I think) we were talking about swapping out these large resistors.


I'm not sure if it worked as I currently have no throttle to try it with and I've run out of enthusiasm to change out all the connectors. The light was blinking, but it was also blinking when I added a 12v sla in series with the 48v charger I've been using. So at 66v the light was blinking (wtvr that means)

I'm not sure if this is even worth anything though because the controller is supposed to be limited to 20 amps. If we can find a way to run it at 48v and push that to 1500w or 2000w (30-40 amps), than these controller might be a worth quite a lot to the budget ebike community! I mean they are 15 fets, so they should be good for quite a bit more than 1500w. People are pushing like 100 amps through Lyens 12 fets with the proper mods.
 
auraslip said:
I'm not sure if it worked as I currently have no throttle to try it with and I've run out of enthusiasm to change out all the connectors. The light was blinking, but it was also blinking when I added a 12v sla in series with the 48v charger I've been using. So at 66v the light was blinking (wtvr that means)
What else was not connected? It might have enough smarts to know if a motor/etc. is connected or not, and if so then blink an error. I suspect that if there aren't any halls connected, it will detect that and error out from that. Probably the same if there's no throttle, but it might not care about that if it has a pedelec option.



I'm not sure if this is even worth anything though because the controller is supposed to be limited to 20 amps. If we can find a way to run it at 48v and push that to 1500w or 2000w (30-40 amps), than these controller might be a worth quite a lot to the budget ebike community! I mean they are 15 fets, so they should be good for quite a bit more than 1500w. People are pushing like 100 amps through Lyens 12 fets with the proper mods.
Assuming these FETs are of similar quality to the ones in Lyen's (which I doubt ;)), then it would be possible to do it by modding the shunt, even if programming it is not an option. That would take some experimentation to figure out, but if yours is like mine then there are extra spots for shunts, and you could just solder some more in (off a dead controller, for example). I think mine has two shunts, with a spot for one more, meaning that adding one would increase current capability by half again what it was before. (if 20A, it'd now be 30). Assuming, of course, that the FETs/etc can take it. :)

Zenid said:
However, you're talking about changing just ONE of the resistors. - Are you sure you have this right? As I understand it, each of these resistors serves one of the FET banks, of which there are THREE, therefore they are SERIES resistors serving a different section of the FET array. The layout of the tracks on the back would seem to support this: http://zenid10.wordpress.com/?attachment_id=1999
- Feel free to chime in, anybody, if I have this wrong...
On my controller, and AFAICT on Auraslip's, these resistors are definitely in parallel with each other, and they are not connected to the FETs, but rather are the dropping resistors from Battery + to the low-voltage regulator system that serves the MPU/etc.
 
auraslip said:
Zenid - It'll blow your mind to find out that I also got some of the same dc-dc converters for sale that you were looking on selling :D
Looks like we've both been on alibaba lately 8)
Ha ha nearly. I used another similar Chinese-eBay type dealer, but have already sold three so am quite happy (no one else in the UK stocks these). Yes! I noticed they're the same "Hua Tong" brand, but I'm not surprised as these are absolutely everywhere at the moment (the first one I bought for my upgrade was this exact brand) :)

However, mine have a current limit of 45A like the other 'vanilla' controllers I've bought. Surprised that your's would be rated so low for a 72V controller...

[Edit] I just saw your photos again... Why do you say they're limited to 20A? According the specs on the panel they have a current limit of 45A(±2A)just like mine and the others I've bought...
 
However, mine have a current limit of 45A like the other 'vanilla' controllers I've bought. Surprised that your's would be rated so low for a 72V controller...

72v*45amps = 3240 watts. That's twice the 1500w rating on the case. I haven't tried mine out, so I don't know what the actual current is limited to. Yours goes up to 42amps? That is cool
 
auraslip said:
72v*45amps = 3240 watts. That's twice the 1500w rating on the case. I haven't tried mine out, so I don't know what the actual current is limited to. Yours goes up to 42amps? That is cool
I was going purely by what both of ours say on the panel (45A). I think that's the peak it can draw, but only for short periods of time... It's confusing as - like you say - it contradicts the 1500W spec on the controller... :?

[Edit] Either way, these things work fine, as I can push 40mph, as opposed to the flimsy 28mph I got on 48V with my old controller... So I can attest to the fact that they're perfectly good, basic controllers, just not as fully featured as more expensive eCrazyman or Lyen ones.
 
amberwolf said:
Zenid said:
However, you're talking about changing just ONE of the resistors. - Are you sure you have this right? As I understand it, each of these resistors serves one of the FET banks, of which there are THREE, therefore they are SERIES resistors serving a different section of the FET array. The layout of the tracks on the back would seem to support this: http://zenid10.wordpress.com/?attachment_id=1999
- Feel free to chime in, anybody, if I have this wrong...
On my controller, and AFAICT on Auraslip's, these resistors are definitely in parallel with each other, and they are not connected to the FETs, but rather are the dropping resistors from Battery + to the low-voltage regulator system that serves the MPU/etc.
What I mean is they must be running off into three separate, parts of the circuit, and therefore any change to one should surely be done to all three, right?

If they were literally stuck together, end-to-end in parallel, that would make no sense at all, as you they would just act as ONE 500Ω resistor (if my parallel resistance math is correct). :?
 
Zenid said:
What I mean is they must be running off into three separate, parts of the circuit, and therefore any change to one should surely be done to all three, right?
Not necessarily. Since they are in parallel, a change done to any affects all three. So changing one to a lower value will change the circuit by about 1/3 of the value it would change it if you swapped all three to teh same lower value.

If you want a large change, changing two or all three is easier, but a little change is easy enough with just one.

If the were literally stuck together, end-to-end in parallel, that would make no sense at all, as you they would just act as ONE 500Ω resistor (if my parallel resistance math is correct). :?
That is correct, assuming they are three 1.5K resistors. (though they are side-by-side connected for parallel, not end-to-end, as that would be series ;)).

They do that so they can get better power handling, by dividing up the power (heat) across all three of them, instead of having to buy larger and more expensive heatsinked ceramic resistors. It's all about money. ;)
 
If they were literally stuck together, end-to-end in parallel, that would make no sense at all, as you they would just act as ONE 500Ω resistor (if my parallel resistance math is correct). :?

If they were end to end they'd be in series.... As far as why 3 1.5k rather than one 500ohm.... I'd imagine it's because it would require a larggggggge resistor high wattage resistor rather than 3 smaller ones.

Err.. what he said...

Do you run these with a CA or watt meter of some type? I'm curious to know what type of power you pull!
 
amberwolf said:
If the were literally stuck together, end-to-end in parallel, that would make no sense at all, as you they would just act as ONE 500Ω resistor (if my parallel resistance math is correct). :?
That is correct, assuming they are three 1.5K resistors. (though they are side-by-side connected for parallel, not end-to-end, as that would be series ;)).
Yes, I meant BOTH ends of one to BOTH end of the others side-by-side, not series end-to-end if that's any clearer (probably not hehe) but yes, I get it, - in parallel:)
amberwolf said:
They do that so they can get better power handling, by dividing up the power (heat) across all three of them, instead of having to buy larger and more expensive heatsinked ceramic resistors. It's all about money. ;)
Ahh! I see. That's what was confusing me. - I just assumed that the three different resistors must represent branches of the circuit - it never occurred to me that they'd stick three of the same one together in parallel...

In that case, yes, the math looks fine. 333.3Ω as opposed to 500Ω. Which is what you would expect to get a proportionate drop from 48V as opposed to 72V. Wonder why it didn't work :(

I wonder what went wrong? Could there be an innocuous explanation like not having the throttle cable connected? I can easily hook this up to my bike complete with all connectors for testing, but is there any chance I could damage the controller or anything else if our logic is all wrong?
 
auraslip said:
If they were end to end they'd be in series.... As far as why 3 1.5k rather than one 500ohm.... I'd imagine it's because it would require a larggggggge resistor high wattage resistor rather than 3 smaller ones.

Err.. what he said...

Do you run these with a CA or watt meter of some type? I'm curious to know what type of power you pull!

Yeah, I get it now... Three little resistors rather than one great big one :)

As for how much power I pull, I have no clue. All I know is I have an almost identical one on my bike at the moment, and used this exact type of controller for a few weeks prior to swapping these over. I originally put this one on just to test it so I would have a working one as a spare, but it was absolutely identical in performance so I kept it on and used the Hua Tong as a spare instead. My speed tops out at about 40mph (speedo) or 36-37mph satnav.

From what you're saying, I gather you've not stuck it on you bike and actually connected everything, in which case I'm not surprised you've got a warning light. Perhaps it means "you've not stuck it on you bike and actually connected everything" ;)

I'm tempted to make the same change and stick it on my bike, rewiring the bank to 48V and running it (if it works) until the batteries are discharged to what should be there new lvc (about 40V or so). My only concern is actually damaging something doing this (specifically the hall sensors or motor). Is that even remotely possible?

What power rating resistor did you use for your resistor, BTW - it looks lots smaller than the others... Is a ceramic (white, long square type) power resistor OK instead? These don't seem that expensive at all... How do they differ from the metal type?
 
Back
Top