Yuan Lang controller shunt idea

Svard75

10 mW
Joined
Apr 17, 2020
Messages
26
Hello. Newb here.

I have a controller which indicates 50amp max limited on the side. I opened it up and found 100v caps and 503 fets which can handle 400 amp burst 100amp continuous. There are 4 shunts and appear to be aluminum wire. What are your thoughts if I replace these aluminum shunts with copper? Copper has 15-16 mohm vs aluminum 20-30mohm.

Thanks in advance
 
I think I may have found something related. Shunts are typically copper so this one may be tinned copper. I’ll use a knife and see.
 
First, the battery current of 50A doesn't have anything directly to do with the FET current ratings, as those dont' see battery current, they see phase current, which is already much much higher (in pulses) than the battery current.

If you increase the battery current, you multiply the phase current, and it can quickly get beyond what the FETs can handle.

Additionally, higher phase current means higher voltage spikes on the phases from the inductive kickback from the motor coils when the controller ends any particular current pulse to the motor. So the FETs can be damaged from the voltage spikes even if the current isn't too high, if the spikes exceed what the FETs are rated for.



Next, there's a lot of threads about "shunt mods" if you look around.


Just remember that if you replace the shunts with copper wire, the controller now has no way to sense current flow, and it cannot protect itself. The results of this tend to be predictably dramatic. :/

Even if you just modify the shunt a little bit, the controller no longer knows what the actual current is, and if it allows more current even momentarily than things can handle, you can end up with failures as well.


(Shunts aren't copper; they're typically Manganin. The shunts have a specific resistance, very low but there, and the copper wire of that length has virtually no resistance, and so wont' have the necessary voltage drop for the controller to sense).
 
amberwolf said:
First, the battery current of 50A doesn't have anything directly to do with the FET current ratings, as those dont' see battery current, they see phase current, which is already much much higher (in pulses) than the battery current.

If you increase the battery current, you multiply the phase current, and it can quickly get beyond what the FETs can handle.

Additionally, higher phase current means higher voltage spikes on the phases from the inductive kickback from the motor coils when the controller ends any particular current pulse to the motor. So the FETs can be damaged from the voltage spikes even if the current isn't too high, if the spikes exceed what the FETs are rated for.



Next, there's a lot of threads about "shunt mods" if you look around.


Just remember that if you replace the shunts with copper wire, the controller now has no way to sense current flow, and it cannot protect itself. The results of this tend to be predictably dramatic. :/

Even if you just modify the shunt a little bit, the controller no longer knows what the actual current is, and if it allows more current even momentarily than things can handle, you can end up with failures as well.


(Shunts aren't copper; they're typically Manganin. The shunts have a specific resistance, very low but there, and the copper wire of that length has virtually no resistance, and so wont' have the necessary voltage drop for the controller to sense).

Thanks amberwolf for your input. What do you think I should do to gain a small amount of current. Here’s my setup and why I’m looking for a touch more umph.

It’s a Chinese emotorbike rebranded to a local company known as the emmo dx. It appears to have a QS rear hub motor. The controller is as indicated and the battery is 72v 24ah. Not much about 1700whr. I believe the motor is 2000w. During my commute there are a number of hills and one is quite steep. It dies drastically. About 15km/hr slower and a voltage sag of 8v. I’ve opened the battery pack and the bms is a 50a unit. I’ve upgraded the wires right from the cells to 2x16awg per cell with 4 banks all joined to 8awg right to the controller. This seemed to provide a little more juice (about 1km/hr more on the uphills).
What do you think would be the best approach to get just that little bit more umph uphill. The rest of the ride is perfectly good. I have no issues reaching speeds of over 60km/hr.

Thanks in advance.
 
Svard75 said:
It’s a Chinese emotorbike rebranded to a local company known as the emmo dx. It appears to have a QS rear hub motor. The controller is as indicated and the battery is 72v 24ah. Not much about 1700whr. I believe the motor is 2000w. During my commute there are a number of hills and one is quite steep. It dies drastically. About 15km/hr slower and a voltage sag of 8v. I’ve opened the battery pack and the bms is a 50a unit. I’ve upgraded the wires right from the cells to 2x16awg per cell with 4 banks all joined to 8awg right to the controller. This seemed to provide a little more juice (about 1km/hr more on the uphills).
What do you think would be the best approach to get just that little bit more umph uphill. The rest of the ride is perfectly good. I have no issues reaching speeds of over 60km/hr.

Given these specifics, we should do a little analyzing and troubleshooting before modifying anything. ;)


How steep is the hill? If you have a smartphone you can get free apps that measure inclines, and even map them with GPS to a path on a mapping site. Knowing what the actual slope is will help you determine the power required to go up it at a specific speed. If you can't measure the slope, at least see if it is listed anywhere, like on a warning sign for the road (sometimes they do this on mountains, for instance). It's better to overestimate it than underestimate it, for these purposes.

You'll also need an approximate weight of you plus the bike (because it takes a certain amount of power to go a certain speed up a certain slope). A good guess is ok, just overestimate rather than under.

By "dies drastically" do you mean that it actually powers off?

If so, then that means the BMS has shut off the output to protect the BMS and cells from damage.

If it doesn't actually power off but the motor ceases to run, then that means the battery is sagging so much that the controller is shutting down to protect the battery from damage.

In either of those cases, you need either a second battery in parallel with the first, or a complete replacement battery that is capable of the power level you're after.

If it is just slowing down radically, but the motor doesn't stop moving, then there are two possibilities, and it could be a combination of both, or either by itself.

The first, most likely, is the battery simply can't output the current needed to provide the power, so the voltage drops so much that the maximum available motor speed drops.

The second is that the controller can't output enough power to the motor to maintain that speed under that load. We'll need to know how much power that load is before we can determine if this is part of the problem or not, so we'll need the hill slope and the weight of you and the bike.


BTW, the BMS has a shunt in it too, so it can protect the battery and prevent damage (and fires). Did any of your wire upgrades bypass this? It may look like the ones in the controller, or it may look like a little black square with metal tabs on the end, surface mounted to the board. There may be several in parallel. I ask because if they are bypassed, then the battery isn't protected, and forcing the controller to draw more may then allow the pack to be damaged. I'd recommend not bypassing or modifying the BMS shunt even if you do the one in the controller.

Most of the cheap batteries are cheap because they use cheap cells, cheap construction methods, and little or no QC. This leads to problems with them even in normal usage, and can be worse when used outside their original design limits (which typically already push the limits of what they coudl do in the first place). It isn't always a problem...but ti can be, and the harder they're pushed, the more likely it is.



Does the systme you have have a display screen with buttons on it to change controller settings or modes? If so, there may be a setting for current limit or power limit in there, that you can change safely rather than the relatively uncontrolled shunt mod method. ;)



If a shunt mod is your only option, then this is the way I would do it. It will take some tracing out of circuits, but Fechter's method works and is controllable and undoable, where actual shunt mods are neither. :/

This thread is one that has posts describing the process. https://endless-sphere.com/forums/viewtopic.php?f=2&t=29097#p1428919 If you're willing to trace circuits in the controller and draw them up, we can help you figure out where to put the potentiometer.


If you prefer one of the uncontrolled shunt mods, we can try that too. While you're fairly certain to get enough more power out of the controller to go up the hill, it's less certain if the battery and controller will survive that undamaged over time. Depends on the quality of the battery itself, as well as the design of the controller and quality of it's parts (and whether they are counterfeit or real), whether there will even be a problem or not (many people do shunt mods without failures, but there are no guarantees; I've blown up controllers the very first time I used them after modding. :(
 
888BB5C9-F41C-4894-A547-A5CB09464E1D.jpegThe hill is about 300m long and slopes to about 7 degrees.

I weigh 240lbs with gear and backpack full of water, food lunch etc. the bike weighs 220lbs dry. The battery is about 40lbs. 500lbs total.

The cells are lifepo4. Not sure about the quality. I’m not able to pull the pack out of the alloy box it was installed into. The cost of the pack is $2200 so one would assume it’s good quality for 24ah. 🤷🏻‍♂️

The bike doesn’t completely stop it just drops from 60-65 km/hr to 54km/hr. At that speed the voltage on the display reads 71.5v from 79v. I would be ok if it went down to 69v but not more.

I haven’t dissected the bms. Don’t think I need to so I won’t be messing with the shunts.

The controller is closed. No visible way to program anything. Actually I’ll see if there is a connector not connected to anything and see if it’s usb (4 pin with two being + and -). In any case I would need software. Where do I get that?

Here’s a closeup of the shunt area. I can’t take the board out to see the rear so not able to fully map it out. The large connector seen on the other side of the shunt is the positive lead to the battery. If this isn’t enough to roughly estimate location of pot then maybe I should just buy a new controller 🤷🏻‍♂️

E1E38604-28C6-46CB-A882-B2CDDAE2C922.jpegAA8BB9B6-83C2-4B0C-9BB3-DF0B64F21DD6.jpeg

Much appreciated for your advice. I will be purchasing a second, same battery and run them in parallel.
 
You can try running a wattmeter in series with the battery and measure what current it’s actually drawing; sometimes it won’t get near the ratings. A 12FET or 18FET controller is usually fine with a mod of 10% or so as they share load over multiple devices.

The quick and dirty way to shunt mod is by wrapping some bare stranded wire around all the shunts a few times, then soldering it with a hot iron until you’re covering 20-50% of the shunt with solder. Then re-measure your current and add more solder until you’re satisfied.

Caveat is that it may smoke itself.

Another thing to look at is the fat traces on the bottom of the PCB. sometimes they are supposed to have a thick layer of solder but this gets missed on the production line. Adding more solder to the traces from the battery and FETs is a good idea if it’s looking a little thin
 
electric_nz said:
You can try running a wattmeter in series with the battery and measure what current it’s actually drawing; sometimes it won’t get near the ratings. A 12FET or 18FET controller is usually fine with a mod of 10% or so as they share load over multiple devices.

The quick and dirty way to shunt mod is by wrapping some bare stranded wire around all the shunts a few times, then soldering it with a hot iron until you’re covering 20-50% of the shunt with solder. Then re-measure your current and add more solder until you’re satisfied.

Caveat is that it may smoke itself.

Another thing to look at is the fat traces on the bottom of the PCB. sometimes they are supposed to have a thick layer of solder but this gets missed on the production line. Adding more solder to the traces from the battery and FETs is a good idea if it’s looking a little thin

I think I might want to try the pot method. It’s super tedious to the controller, lots of plastic panels need to be taken off and battery box removed as well. I would prefer to get a pot, set it at maybe 10-20% max and leave it. If it smokes then I’ll get a new controller.

The lines are above as seen here not below. I think. What do you think?

6848B88F-B6E1-4A8F-ADDA-8930774CADF3.jpeg6883BE76-7667-4028-9AD7-E82435534740.jpeg6CAE050E-C119-4433-959C-5F6550E85596.jpeg
 
I reached out to yuan lang inquiring if I could buy the schematics of this controller. If I get it I’ll post it here. If not I guess I could take the board out. The problem I’m concerned with is the clamp that secures the fets to the heat sink are extremely tight. It’s easy to remove but putting it back on creates scratches in the FET enclosure. I’m worried I may crack it.
 
I doubt they'll even respond, but it's pretty unlikely that they'd provide schematics regardless. Nobody else ever has. :/ AFAIK all the schematics on ES are stuff members drew up after they traced controller circuitry out.







That said, it is very likely that on the shunt end closer to the YL17G-C marking, the tiny little trace coming off the top and leading to waht is probably R55 or R28 is the detection lead from teh shunt. Tthe other end of the shunts look like ground (batteyr negative)--is that right?

I can't read the marking on the 8pin black chip near those resistors, but if it is an op-amp (741, etc) then that is the circuit for preamplifying the shunt measurement voltage, so the pot you would install would go in that circuit. I don't see a lot of vias there, so it is possible that most of that circuit is on this side of the board, and could be traced and drawn out on paper using a multimeter on continuity or 20ohms, and then just note down all the component nubmers next to where they are on the schematic (R28, R55, Q1, C24, etc, everything in the area is probably part of the circuit, probalby just what fits under your thumb-print area).

If you trace taht out and it turns out more info is needed to determine the circuit enough fo ryour purposes, then we can worry about having to open up the rest of the controller. :)
 
Gotcha. I’ll trace it myself. I think there’s pretty good schematic software to make it look official and to validate.

Might take me some time. Thanks for all your help, much appreciated.
 
I spent a bit of time looking closely at the circuits. It’s impossible to correctly document it without taking it out. There are plenty of jump points that connect through the board to the other side. I don’t think I feel comfortable taking the heat spreader clamps off either and risk damaging the FET package.
I’m going to risk using the copper wire and solder around the shunts as a first pass. If I fry the controller I’ll get a new one spec’s correctly from the go.

Thanks all. Truly appreciate your time and advice.

Best
 
Looks good- you’ll need to heat the shunts up a bit so the solder will flow, you could practice on some old thick wire first of all.

Did you measure the current so you know what you’re aiming for? Or just going to wing it
 
electric_nz said:
Looks good- you’ll need to heat the shunts up a bit so the solder will flow, you could practice on some old thick wire first of all.

Did you measure the current so you know what you’re aiming for? Or just going to wing it

Wingin it :lol:

I don’t have a proper amp meter. I have a digital multimeter but 10a is max before it probably throws a fuse.
 
Not really any way to know what will happen until you test it. Without a way to measure current before and after, also no way to know what change was made, other than by performance.

If it works, you're all set. If it doesn't, you can experiment further with the same method and see if it gets better without smoke. ;)
 
amberwolf said:
Not really any way to know what will happen until you test it. Without a way to measure current before and after, also no way to know what change was made, other than by performance.

If it works, you're all set. If it doesn't, you can experiment further with the same method and see if it gets better without smoke. ;)

If it’s too much wouldn’t I be limited by the batteries capabilities. My bms states 50a. I assume if it attempts to pull more the bms will cut.
 
I decided to pull the board. Some very interesting design ideas. I think this is an example of over engineering.
I found the sense line so I can solder a pot.
BAFA7FE1-C5C1-4D0E-82A5-BEDC7BD413DD.jpeg

Look at those phase segments. Beefed up traces with copper plates. 24 fets. I feel a little more confident in the trim pot mod now.
D3CD7D43-90E4-42D6-A521-AEB56EB4AA6D.jpeg
 
Trace is cut but soldering to the tiny remaining trace wasn’t fun. I had to apply epoxy to the wire to ensure it doesn’t rip the remaining trace off. My plan is to mount the potentiometer so it’s accessible from outside the controller.

5A432610-11A4-4DA3-B958-7B390F6C2B42.jpeg
 
Question. I could only find a 25ohm 3watt potentiometer for this project. I’m not certain I understand this circuit mod completely. My question is should I be concerned that I’ve only got 3watt when 72v is passing through the two end points in the potentiometer? I think I read somewhere that voltage is like water it will take the path or least resistance which might not necessarily be the shunts. I don’t really know their resistance at this time and there’s 4 of them. How can I put this all together and make sure this mod will actually work?
 
I guess I’ll just keep talking to myself :lol:

I’ve been reading up on the power triangle to help me understand the maximum voltage the potentiometer will take. So far the calculation is as follows

484F2560-A647-4D5E-886B-2F08C9422D69.jpeg

V = P/I

Power is not really known yet because I need to measure the resistance of the four shunts. I know the pot is 25 ohms.

Let’s assume that the shunts are 0.5 ohms. I plugged it into a power calculator and get this
802EFD04-7E0B-4D4D-A597-810981702204.jpeg

9E6B714D-2F86-4FAC-BF32-8E9276E4E2BE.jpeg

So I guess my next steps are to calculate actual shunt resistance.

Going back to the power triangle. I now can assume that I=1.4257 amps and V=72v therefore P=102.465watts???

My pots gonna smoke :confused:
 
sorry, i'm still busy with a city-ordered cleanup project so will be at least several days before i can look into it and respond. :(
 
amberwolf said:
sorry, i'm still busy with a city-ordered cleanup project so will be at least several days before i can look into it and respond. :(

No sweat amberwolf. Real life is more important. I live in Canada, riding season is over anyway and the bike is apart due to servicing the front forks. These things came with hydraulic fluid I think. And overfilled. Planning to look at the rear shock as well, however I heard the rear are typically filled with nitrogen 🤷🏻‍♂️
 
Here is a photo of the assembled controller. The three wires are 16awg and very short. About 1” each. I was able to measure resistance at the lowest point at 0.5 ohms. I suspect this is still quite high. So at this point I have no idea what will occur. In the meantime I am shopping for a new controller. I noticed quite a few on alibaba.
591878A4-62A4-4A5A-94B5-6E11B8E52B61.jpeg
 
Back
Top