How To, 50A BBSHD Controller Mod

[Matador]

Ebikes4Real, I just want to verify some info about the SMD type Shunt resistor before I pull the trigger on buying these components :

1 - The shunt resistors are of the 2512 size (6.35 x 3.05 mm ; 0.25" x 0.12") <--- Verified and edited this 2017-06-23

2 - The MOSFETs, I found these IRFB3077 mosfets for 5.12$ (CAD) for https://www.digikey.ca/products/en?keywords=IRFB3077PBF
Is that the good component ??

I might try the shunt mod without changing the mosfets a first, buy I will get the new FETs on the same order though. Unless you can suggest some even better ones

View attachment 5


Fechter, I cant believe 30Amps are going through such tiny SMD components !!! This ???

 

[fechter]

Yes, 30A going through those tiny things. And you can bet they get hot. They have pretty high temp ratings, but eventually you can melt the solder off (bad). Most BMS boards use the same little resistors now. I haven't seen anybody melt them off yet.

If the PC board has heavy copper and good thermal connection to those resistors, the heat gets spread out over a much larger area and they seem to do fine. Stacking them might not be ideal. Much better if they are all stuck to the board, but there's not much room. If you remove the two and scrape the solder mask off to the sides, you might be able to squeeze in 3 in a row. I bought some 3mohm ones the same size I can replace them with.

In the closeup pic of the shunt resistors, you can see the 8-pin LM358. Voltage from the shunt goes to that and gets amplified to a level the MCU can use. Someplace I reverse engineered that part of the circuit. It's pretty simple. If you wanted a remote switch to lower the current, you could tap into the little resistors going to it. It's easy to lower the limit in software too.

The 3077s are good. Way better than the 75NF75s. These will generate much less heat than the old ones. There may be some better ones out there, but too expensive for me. These seem to be the most economical.

 

[Matador]

The old low-end STP75NF75 FETs are 11 mOhm @ 40A, 10V : https://www.digikey.ca/product-detail/en/stmicroelectronics/STP75NF75/497-2788-5-ND/603813

The high-end replacement ones are 3.3 mOhm @ 75A, 10V : https://www.digikey.ca/products/en?keywords=IRFB3077PBF

Seems much better.

I'm also wondering.... heavier gauge wire to the battery ?

 

[Matador]

Yes, 30A going through those tiny things. And you can bet they get hot. They have pretty high temp ratings, but eventually you can melt the solder off (bad). Most BMS boards use the same little resistors now. I haven't seen anybody melt them off yet.

If the PC board has heavy copper and good thermal connection to those resistors, the heat gets spread out over a much larger area and they seem to do fine. Stacking them might not be ideal. Much better if they are all stuck to the board, but there's not much room. If you remove the two and scrape the solder mask off to the sides, you might be able to squeeze in 3 in a row.

Thanks Fetcher, as usual great knowledge here !

Tiny indeed! In original OEM configuration, these 2 parallel 5 mOhms get 15 amps each (P=UI^2) that's 1.125 W each... The 5 mOhm "2010" SMDs I've sourced on Digikey are rated for 1.5W.... Flirting with the resistors limits here.... Though I'm not exactly sure of what the exact OEM part rally is in terms of sizing, my bet is "2010" package, though I hesitate with "1812" or "2510" package...

I bought some 3mohm ones the same size I can replace them with .

Eh eh clever ! So with two 3mOhm in parallel, you get 1.5 mOhm instead of 2.5 mOhm... Exactly 50Amps instead of 30 amps (or 45A with 3 x 5 mOhms).. I'm a bit concerned about power ratings though .... Each resistors would have P = 0.003 x (25)^2 = 1.875 Watts.... I'd check the 3 mOhms resitor power rating just to be sure... if they are 1.5W, pushing 1.9W might overheating them, possibly making their resistance value drift ? Unless there rated for more than 1.5W ?

In the closeup pic of the shunt resistors, you can see the 8-pin LM358. Voltage from the shunt goes to that and gets amplified to a level the MCU can use. Someplace I reverse engineered that part of the circuit. It's pretty simple. If you wanted a remote switch to lower the current, you could tap into the little resistors going to it. It's easy to lower the limit in software too.
The 3077s are good. Way better than the 75NF75s. These will generate much less heat than the old ones. There may be some better ones out there, but too expensive for me. These seem to be the most economical.

I kinda get it.... So there's maybe like 50-75 mV @30A OEM (or in this order of magnitude ) that go to the LM318. Tapping that little resistor going to it would'nt be a problem, as the voltage and current going throught it is so small, that even a long wire going to the handlebar will not significantly affect it (should be okay with 14awg)...

Fechter you're way ahead on this BBSHD controller game. I'd like to know more

 

[fechter]

The shunt resistors would be over the rating, but unless you keep the motor at max current for like 5 minutes straight, they probably won't overheat. I haven't tried this on a BBSHD yet, but lots of experience with the ones on BMS boards which are running the same kind of current. Also the watt rating of those resistors is highly dependent on how much heat sink they have. With enough cooling they can probably do 10W. On another project I put Kapton tape over the resistors as an insulator, then used silicone glue to attach a piece of aluminum over the tops. Probably overkill here, but something to consider if one actually melts off.

When I get back to work, I think I have some of the current amplifier information. I remember its a dual amp but only one side is used. The circuit also has an offset so the MCU sees over a volt at zero current. You could tweak one of the two resistors in the amplifier circuit to change the gain.

 

[Matador]

The shunt resistors would be over the rating, but unless you keep the motor at max current for like 5 minutes straight, they probably won't overheat. I haven't tried this on a BBSHD yet, but lots of experience with the ones on BMS boards which are running the same kind of current. Also the watt rating of those resistors is highly dependent on how much heat sink they have. With enough cooling they can probably do 10W. On another project I put Kapton tape over the resistors as an insulator, then used silicone glue to attach a piece of aluminum over the tops. Probably overkill here, but something to consider if one actually melts off.

When I get back to work, I think I have some of the current amplifier information. I remember its a dual amp but only one side is used. The circuit also has an offset so the MCU sees over a volt at zero current. You could tweak one of the two resistors in the amplifier circuit to change the gain.

Concerning the rating of the resistor, I made some measurments on the picture (using documented size of LM358 and the rule of three). The OEM resistor really look like they have the "2010" size factor (5.0 x 2.5 mm ; 0.20" x 0.10") (Please correct me if I'm wrong). With that size, I'm pretty shure each 5 mOhm resistor has a 1.5W max rating (from datasheet : http://www.ttelectronicsresistors.com/datasheets/LRMA.pdf)

I also measured the size available on PCB : The 2512 packet form (6.35 x 3.2 mm ; 0.25" x 0.12") of equivalent resistance value should be rated for 2 or even 3W !
I think it could be possible to cram in two 3W - 2512 current sensing resistor in there. intead of the 1.5W 2010... So we might have another alternative to do this mod if one goes wrong.

Anyways, I like your idea better Fechter, tweeking the resistor value in the LM358 op-amp circuitery.... I'd really like to have a multiposition switch (maybe even a trim pot !) on the handlebar (or hidden under my seat, or on the controller itself) so that at 48-52 volts we could have :
Setting A - 10A Max (500W legal limit in Canada)
Setting B - 15A Max (750W legal limit in USA)
Setting C - OEM 30A
Setting D - Hotrodd 40A
Setting E - Hottttttroooooodd 50A

I think 60A might be a bit too much (tradeoff in reliability)....

BTW, on cut section the BBSHD phase wire seem to be CCA (copper clad aluminium) rather than real copper wire... Did bafang cheap out on the cables ?

LM358 datasheet : http://www.ti.com/lit/ds/symlink/lm158-n.pdf

 

[Matador]

I've just bought a blowned 30A BBSHD controller on eBay for a best offer of 30 USD (+8$ shipping).

Seller said : "Not working! Accidentally shorted communication cable to LCD. On start error 30 shows up on the LCD. Potting compound carefully removed, but nothing looks burned out. The damage is most likely some communiaction IC. Got a new one".

It's exactly the pictures I posted in my previous posts here. I figured It would be a good way to train myself in modding BBSHD controllers before touching a good one and risk destroying it (especially for the unsoldering the mofset part). Also will have spare crappy mosfets for the 30A BBSHD controller I plan to keep in addition to the new working one I plan to buy and modify.... Since each crappy mosfet retails for 2.68$ and there is 12 FETS, i figured it wasn't such a bad deal...

I wonder if the blown controller I bought could be fixable...

 

[fechter]

I wouldn't have much hope of fixing it, but you never know.

I practiced on a dead one first myself. Removing the FETs without damaging anything is not easy. The gate traces don't have much copper on them and pull off pretty easy. The source and drain have big copper so you need more heat. I used a solder sucker and eventually got them all.

 

[Matador]

I wouldn't have much hope of fixing it, but you never know.

I practiced on a dead one first myself. Removing the FETs without damaging anything is not easy. The gate traces don't have much copper on them and pull off pretty easy. The source and drain have big copper so you need more heat. I used a solder sucker and eventually got them all.

Yeah I figured I should'nt have much hope either. I'm already in the process of ordering 1 or 2 new BBSHD controller. Will probably get a quote on monday if the seller stocks them.
How about solder wich (fluxed copper braid) to such the solder.... I have some at home.

Also Have 3 soldering irons : A weller W100P (100watts, fat tip), probably too hot. A weller soldering gun (140W), probably to much, and a cheap chinese 25 or 40W IIRC...

 

[fechter]

I used a small iron on the gate legs and a big, temperature controlled iron (140W) for the source and drain legs. Be sure to remove the heat sink before starting as this will suck up a lot of heat on the drain leg. I've never been real successful with solder wick, but it's worth a try. Probably good for the gate legs.

Once you get all 3 holes looking clear, I wiggled the part back and forth while watching where the leg stick through the hole. If you got them all, all 3 legs will wiggle. If one is not moving, it needs more sucking. Once you get the part to wiggle, keep wiggling it until it gets loose, then try pulling it out.

I've also tried heating all 3 legs at the same time and pulling. Damaged gate traces.

There are lots of different approaches to try.

 

[fechter]

I measured the space on the board and there is NOT enough room to fit 3 shunt resistors. It was close, but no cigar.

So, I finished reverse engineering the current amplifier. The LM358 takes the voltage across the shunt and amplifies it x10 and gives it a slight zero offset to make sure the zero reading is within the input range of the MCU. The bottom half of the 358 is not used and both inputs are tied to ground.

The schematic:



Parts location:



As far as I can tell, the offset resistor is 30k. The offset resistor is on the other side of the board. At zero current, the input to the MCU is about 1.70V.

So, for example, if you placed a 10k resistor across the existing 10K gain resistor, the amplifier gain would be 1/2, which would double your current limit. But its not quite that easy. The gain will also affect the offset. If the MCU sees a zero current signal that's out of a specified range, I'm sure it will trigger an error. If the offset is in the required range, then changing the gain resistor will change the current limit.

Well, it wouldn't be too hard to try paralleling a 10k resistor across the gain resistor and see if the MCU boots up with an error or not. If so, then the offset resistor would also need to be tweaked. With some math, we could figure out exactly how much.

 

[Matador]

Excellent !
Waiting to get my busted controller to experiment whit it a first...

It is true the BBSHD is way overbuilt for 48V@30A (1440W input)...

I mean, the BBS02 runs 48V @25A.... That's 1200W input.

But when you look at this video, you can see that the BBSHD, in comparison, has a rotor and stator 5/3 the size of that of the BBS02 (5 cm deep instead of 3 cm deep). Check the vid at 28:45, you'll see what I mean.
[youtube]E46OW4ubK84[/youtube]
To me, that means the BBSHD should be handeling 2000W quite easily. At 48V, 2000W is 42 Amps...
Hence 42 Amp on the controller should be as reliable as 25Amps on the BBS02 with that ballpark math reasoning.

I asked a seller for a controller clearly stating my intend of modding the BBSHD controller. He's been experimenting on his side too. He tells me that he burnt the BBSHD windings pushing it 90V@40A (that's 3600W).
I dont know which external controller he used.
I've heard the BBSHD uses a trapezoïdal wave OEM controller.
Anyways, 3600W at 48V is 75 A...... At 52V, it's 69A....

So I'd be hesitant pushing the OEM controller to 60A. I'd rather blow a new controller than fry the windings of a 1000$ BBSHD.
40A seems like a safe bet.
50A is a little adventurous IMHO.
I'd avoid 60A.

Come to think about it, there's just 7% power increase by using a 52V rather than a 48V battery. That 7% increase is already enjoyable. Then 40A at 52V is 2080W... That's a 44% increase in power compared to running it at 48V 30A.... Bet I'd be really fun !

 

[fechter]

It's hard to say how much is too much. I'm sure it can take a lot for short bursts. If you monitor the temp and keep it under control, it might do OK with a higher current limit. If you set the shunt for 50A, you could then lower it in software to, say, 40A and test. This way you can change it over a good range without needing to solder every time.

The stock controller is trapezoidal output, but appears to also have synchronous rectification.

 

[Epyon]

Hey guys, new here, and very interested in performing this mod.
@fechter Have you made any progress with the current amplifier?

@Matador Any links to recommended shunt resistors other than the 5 mΩ ± 1% 1.5W on DIgikey? Or do those work just fine?

Thanks guys.

 

[Matador]

Hey guys, new here, and very interested in performing this mod.
@fechter Have you made any progress with the current amplifier?

@Matador Any links to recommended shunt resistors other than the 5 mΩ ± 1% 1.5W on DIgikey? Or do those work just fine?

Thanks guys.

Actually, I didn't try them. Although I do have two more original resistor I could unsolder and use from a dead BBSHD controller I bought on eBay.... The thing is , i didn't buy a spare controller to mod yet. and still working on my batteries to be able to actually deliver 30A reliably over the course of a full discharge cycle, without getting to warm.... You'll see... nut many commercial pack are designed for medium to high amps... unless you're willing to carry a heavy battery !

 

[fechter]

Tweaking the little resistors in the current sensing amp circuit will be tough simply because the parts are so small and hard to access. I think it is much easier and more foolproof to simply replace the shunt resistors.

From my measurements, the case size seems to be 2512.

3 milliohm (would give you 50A if you replace both)
http://www.mouser.com/ProductDetail/Vishay-Dale/WSLF25123L000FEA/?qs=sGAEpiMZZMtlleCFQhR%2fzRv21WPTCkfbxjfArIWQsbn42115l%252bxy7w%3d%3d
$0.52ea.

If you just replace one of the pair with 3mOhm, you would get 40A.

Another approach would be to leave the stock shunt resistors and place another one stacked on top. Less work, but I'm not sure about the heat dissipation.
If you stacked another 5mOhm on top, it would give you 45A.

You can always turn down the max current in the programming.

 

[Matador]

I tried using the BBSHD with a 61 volt battery. It refused to come on until I dropped it to 60 volts. Then it worked fine. Also, the BBSHDs capacitors are rated 60 volts.

I have the BBSHD controller opened in fron of me (label : CR R10E.1000.SN 852-7-501 1603310055 48V30A).
There big capacitor (1.65 cm diameter x 2.6 cm height) is a Rubycon 63V 1000 uF XYJ with numbers 3M1548 105°C PET
There smaller capacitor (1.04 cm diameter x 2.05 cm height) is a Rubycon 63V 220uF YXF with numbers 3M1514 105°C PET


Just to confirm, I did measure the shunt resistor size directly on the board using a caliper ruler
I get 0.61 cm or 0.62 cm lenght by 0.31 cm wide.

So this means the current sensing resistors are SMD packed sized 2512 (0.25" x 0.12" = 6.35 mm x 3.05 mm) rather than the 2010 I thought in my previous posts.

Matador

 

[Epyon]

Tweaking the little resistors in the current sensing amp circuit will be tough simply because the parts are so small and hard to access. I think it is much easier and more foolproof to simply replace the shunt resistors.

From my measurements, the case size seems to be 2512.

3 milliohm (would give you 50A if you replace both)
http://www.mouser.com/ProductDetail/Vishay-Dale/WSLF25123L000FEA/?qs=sGAEpiMZZMtlleCFQhR%2fzRv21WPTCkfbxjfArIWQsbn42115l%252bxy7w%3d%3d
$0.52ea.

If you just replace one of the pair with 3mOhm, you would get 40A.

Another approach would be to leave the stock shunt resistors and place another one stacked on top. Less work, but I'm not sure about the heat dissipation.
If you stacked another 5mOhm on top, it would give you 45A.

You can always turn down the max current in the programming.

Thank you. Just ordered a few.
Now, if I replace both, I'd set current limit in programming to half of actual?

 

[fechter]

Thank you. Just ordered a few.
Now, if I replace both, I'd set current limit in programming to half of actual?

Stock limit is 30A, so 50A is not quite double. Actual should be about 5/3 x programmed setting. To get 30A, program limit to 18A.

 

[tomjasz]

Thanks Matador!

 

[Epyon]

A big thanks to you guys. I swapped out my shunt resistors and capacitors.

 

[Matador]

A big thanks to you guys. I swapped out my shunt resistors and capacitors.

I'm curious.... Did you try running a voltage higher than 60 volts to the controller ?
What shunt values did you use ?

Matador

 

[Epyon]

A big thanks to you guys. I swapped out my shunt resistors and capacitors.

I'm curious.... Did you try running a voltage higher than 60 volts to the controller ?
What shunt values did you use ?

Matador

The HVC is at 61.5v I believe. I haven't played with it much to see exactly. Any help on bypassing the HVC would be awesome. Still waiting on a bare controller housing from Luna before I can put it on the bike.
I used the resistors from Mouser that Fechter linked above.

 

[Matador]

A big thanks to you guys. I swapped out my shunt resistors and capacitors.

I'm curious.... Did you try running a voltage higher than 60 volts to the controller ?
What shunt values did you use ?

Matador

The HVC is at 61.5v I believe. I haven't played with it much to see exactly. Any help on bypassing the HVC would be awesome. Still waiting on a bare controller housing from Luna before I can put it on the bike.
I used the resistors from Mouser that Fechter linked above.

Nice. At 61.5V HVC, you could use a 15S Li-Ion battery, charging to 4.10V max. At 63V, you could charge that 15S to 4.20V. But the HVC won't allow the controller to work.

I'm curious, why did you change the capacitors to higher if the bottleneck is the 61.5V HVC ? Is there a way to change the HVC value ??? I't be more interested to up the controller voltage than the amps it can take !

I have an empty controller case. Is there a place where we can buy the controller without the metal case (for cheaper than the whole thing) ??? If so, I'd buy a few and start up experimenting !

Matador

 

[Epyon]

Nice. At 61.5V HVC, you could use a 15S Li-Ion battery, charging to 4.10V max. At 63V, you could charge that 15S to 4.20V. But the HVC won't allow the controller to work.

I'm curious, why did you change the capacitors to higher if the bottleneck is the 61.5V HVC ? Is there a way to change the HVC value ??? I't be more interested to up the controller voltage than the amps it can take !

I have an empty controller case. Is there a place where we can buy the controller without the metal case (for cheaper than the whole thing) ??? If so, I'd buy a few and start up experimenting !

Matador

I originally had 15s LiPo running, but killed a few cells in testing (cheap Zippys).

I'm hoping that I (or someone much smarter than myself) can figure a way around the HVC. I already had the caps, so figured might as well.

I got the controller from Highpowercycles on E-Bay. Well, actually I got 2, the first one they sent looked like it was pryed out of the case by a gorilla. It was severely twisted. I removed the potting and heat spreader and flattened out the PCB. Surprisingly the soldering looks intact but I haven't actually tried to power it up yet. The replacement controller (the one shown) arrived in good condition. I paid ~$60 for it and $25 for a case. Doing it again, I'd just get a $99 unit from Luna with the case.