How To, 50A BBSHD Controller Mod

[E-mil]

Climbing the ladder!
Yesterday I finally did the shunt mod. I reused a 5 mOhm shunt resistor from a dead smart-BMS, and soldered it on top of the two other on the backside of the controller. Everything works, so currently my BBSHD is running 15S 45A, and the boost is certainly noticeable!

I started depotting the controller, but realized I only needed to clear a cut-out on the backside of the board where the shunt resistors are located. Hence, there is more de-potting to do when I at a later time replace the cap for a 100 V version. (Still waiting for package from China).

As of now only the shunt have been altered. However, there are some possible caveats.
The FETs are the lesser NF75 type, but let's see how much they take. I have so far tested the system for 30 km, and so far so good. It actually surprised me it was the old FETs, because the controller has never been more than body temperature, and this has not changed now. If the FET's are to be changed it will be to 100 V FETs for future voltage upgrade compliance.


It took me a ½ hour to get the controller out of its housing, and I chose a radical step to drill two holes in the FETs' cooling plate in order to better lift it out without possible damage to the wires, which I didn't like to pull too hard. In these bored holes some screws were mounted, and then they could be used to pry the controller loose. It worked flawlessly!




An interesting thing: What is this daughter board? Never seen that depicted before:

 

[Epyon]

Never seen that daughterboard either. Judging from the markings, a 6v regulator? But for what? Does yours have any extra wiring? Don't some have a lead for a light?
Mind linking to the 100v caps you got?

 

[tomjasz]

Never seen that daughterboard either. Don't some have a lead for a light?

It being for the added light in some versions, makes sense. I personally wish they hadn't added the 6V JST for a light.

 

[E-mil]

Never seen that daughterboard either. Judging from the markings, a 6v regulator? But for what? Does yours have any extra wiring? Don't some have a lead for a light?
Mind linking to the 100v caps you got?

Yes, my BBSHD does have the 5-6 V light connector, but I use my own rather customized lighting system, disregarding the BBSHD lighting output jack.
You might be right, it does look like a power board for the lighting connector.

Regarding the caps: These were the ones I bought, various values. I HOPE they are OK quality!
https://www.ebay.com/itm/100V-100uF-220uF-330uF-680uF-1000uF-2200uF-3300uF-4700uF-Electrolytic-Capacitor/232504120059?ssPageName=STRK%3AMEBIDX%3AIT
https://www.ebay.com/itm/Electrolytic-Capacitor-100V-150uF-220uF-330uF-680uF-1000uF-2200uF-3300uF-4700uF/332393927388?ssPageName=STRK%3AMEBIDX%3AIT

 

[Epyon]

Regarding the caps: These were the ones I bought, various values. I HOPE they are OK quality!
https://www.ebay.com/itm/100V-100uF-220uF-330uF-680uF-1000uF-2200uF-3300uF-4700uF-Electrolytic-Capacitor/232504120059?ssPageName=STRK%3AMEBIDX%3AIT
https://www.ebay.com/itm/Electrolytic-Capacitor-100V-150uF-220uF-330uF-680uF-1000uF-2200uF-3300uF-4700uF/332393927388?ssPageName=STRK%3AMEBIDX%3AIT

Do you know the dimensions on them? You can only do about 27mm vertical off the board, or lay it on its side with long leads.


A little update on mine. Been running great! Zero electrical issues, just one broken chain.
I've got 2 6s LiPos inbound, gonna jump to 17s (71.4v full). I may add a trim pot or such to adjust the logic voltage, still cant find a suitable regulator.

 

[GeoffS]

Mounting the HD unit for greater power later:
I'm starting with the 4th line down on E-Mill's summary (62.3V) and have had the battery pack built with 30Q cells in 15s 7p format with a programmable BMS (via phone and port) to make my life easier. <-- in the post. I'll review it when I have time to log a controlled discharge within specs at work.
I'm going for a semi-conservative start, as I'm a noob, and need to get my head around things and at least experience SOME success with it before I kill the machinery pushing it to the edge.
I've been looking at mounting systems with greater power later (and stability now) in mind. Has anyone tried bolting through the back of the drive gear casing directly into the ISCG05 mounts that are on many bikes?
It looks feasible, but I have a couple of concerns:
1. the front lower bolt hole will be close to the front edge of the pedal assist sensor board. Do the wires go directly inboard ? or do they snake forward?
2. The casing strength? (issue is putting a bolt hole close to an existing gap where that circuit board sits). I'm less worried about this as the casing looks massively over-built.
3. is this a classic noob move to start mods before even riding it that will end in tears?
( sorry, I can't get the pic to attach )

 

[GeoffS]

The back of the casing. The motor had been rotated as much as possible to facilitate a gap between the pedal sensor board and the hole.

 

[fechter]

The hole locations look OK as long as the screw heads are low profile.

I'm not following why you want to put screws there. I guess I need a picture of the mating part on the bike. I don't think I have anything like that on my bike. I also haven't had any issues with the motor moving in the mount.

 

[malkie0831]

ISG05 is an ISO Standard mount fitted behind the chain wheel around the BB for mounting chain guides, tensioners or chain protectors:

Common on MTB’s in my experience, particularly high end one’s.
Normally to fit a BBS XX you have to remove the bottom right lug as it hits the motor.

 

[fechter]

I guess I've never seen one of those before. The BBSHD normally mounts so the motor is against the downtube, so can't rotate in that direction, but could slip down if you go over a big bump. My guess is if you mount the drive properly, you would really only need one of those holes to keep it locked in place.

 

[E-mil]

The questions about screws through the gear-plate have, strickly speaking, little to do with the purpose of the thread.
None the less it is an interesting take on a non-existing problem AFAIK, because as mentioned the motor turns TOWARDS (up against) the downtube. I know that Luna sells a long "triangle" bracket for better mounting, but IMO you could just mount a metal band like on the Cyclone motor, or as I have been pondering, drill a hole or two in the cooling fins on the motor house and strap the motor to the downtube this way. Same, but different..

Back to main BBSHD tuning topics:
I am still waiting for the caps to arrive, but 15S/45A is still sooo much more engaging than my original 14S/30A, tho the bastard battery setup dips a little too much for my liking (ca. 6V drop under full load) . After all its BMS is only certified for 40A, and a 20Ah EIG cell piggybacks it as the 15th cell series.
And by the way, I have ordered the Grin Phaserunner + CA3 setup. It is brand new second hand but NOT set up for BBSHD, because not bought thru ERT. I expect a lot of cable/connector trouble and setup headaches, but certified top performance at 90 V / 50 A. 22S at 4.09 V/cell should be possible. Since this new controller is coming I have only the monetary value of the BBSHD's controller to lose, so to speak.


A very sweet spot with the BBSHD controller voltage could be the 18S at 4.15 V/cell for a system voltage of 74.7 V.
Benefits are NOT having to replace the 75 V FETs, as this operation seems vastly more difficult than just replacing the 63 V cap for 80 or 100 V version, and available battery capacity is about 95%. This is all provided that the display V mod works at higher voltages..


About the display voltage mod, I have an idea I will give a try when upping the voltage above 15S / 63 V:
Simply use LED(s) as the display's voltage drop. As already mentioned a diodes forward voltage drop (Vf) is vastly less dependent on the display's power draw than a resistor is. The LED I have in mind could be one of these:

Whats good about them? They are easy to come by, and only one component is required for a drop of 6 to 9 V.

A Vf chart for the two LEDs is according to my testing:
mA: White Vf: Amber Vf:
20 8.48 8.09
30 8.58 8.18
40 8.67 8.27
50 8.75 8.33
60 8.84 8.39
100 9.00 8.70

A combination of one of these LEDs with a standard diode (Vf ca = 1.3 V) can get a close to 10 V drop, and this will make it easy to calculate a factual battery pack voltage. A string of white or blue 3 Vf LEDs could just as well be used to get to the desired voltage drop. You could even let the LED's light be used as forward or side lighting

The best would of course be to alter the BBSHD controller circuitry, as specified by the wizards here earlier on. Then the voltage dropper becomes redundant!

 

[E-mil]

Reentry into the controller again today, as I got most of the capacitors ordered. But I need help!

The biggest problem is now to fit the cap(s) into the confined space between the other components in the controller and the motor housing itself. It seems that, with thorough measurements, it is possible to fit a 100V/1000uF cap IF the lighting daughterboard could be removed. The problem is that the cap needs two more mm to be flush with the controller board, and without these 2 mm it is too long to fit, and it is simply 1 mm too wide to squeeze down.
Another way out of the problem is to fit a 680uF cap in place, and another one outside the controller. Could this pose some inductive problems?

Another rather important question: The other smaller cap which is still halfway under potting; does this one also need replacing for 100 V?

 

[fechter]

Another rather important question: The other smaller cap which is still halfway under potting; does this one also need replacing for 100 V?

Yes. It's in parallel with the larger one.

Adding one outside is about the only solution. I've seen this done on other controllers and many RC controllers are like this. You want it to be as close as possible to the FETs, but a few inches won't be a big deal.

 

[E-mil]

Thanks for quick answer!
I managed to cram a 680uF in place of the 1000uF 63 V, and a 470uF in place of the 220uF 63 V.

If the caps truly are connected in parallel their values will add up nicely already, but I will fit an external one as well, either at 470 or 680 uF.

It took forever to replace them and hopefully all the circuitry is still alright.

Edit: Pic of new caps added on controller. It sure was a tight fit when mounting the controller to the motor house, but system booted up nicely at 15S. Tomorrow I'll post about the venture into the realms of 18S. Cross your fingers!

 

[fechter]

I just double checked on a dead controller I had here. The caps are not exactly in parallel. They are on opposite sides of the shunt. For our purposes, I think this will work out the same.

It probably wouldn't hurt to add one outside too, but it would be ugly. Here's how it looks on the old BMC motor:

 

[E-mil]

18S / 45 A -- IT'S WORKING !

Theoretically a power-up of 100% from 14S / 30A to at least 3 kW
Had a short test earlier today. No misbehaving whatsoever, though batteries (14S + 4S) where at ca. 3.8 V/cell.

I also mounted an external 470 uF cap just outside the controller, intravenously into the wires (no soldering! :? ). This brings the combined capacitance to 680 + 470 + 470 = 1620 uF, well over the original 1000 + 220 = 1220 uF.

For the DPC-14 / C850 display to allow controller to turn on, I used two LED's in series; the amber 10W 8.3 Vf + a smaller white 3W 3.2 Vf (ca.). The LED's do not heat up due to their size, so atm they are just wrapped in tape. The display showed 54.6 V, which is approx what the original 15S battery would display at this state of charge. Just doing a rough calculation, I will need another 3 Vf LED to allow charge to 74 V ..

I think I will quit the voltage game at this sweet spot (18S / 4.15 V/cell) for the controller, at least for now.
As previously mentioned, this is just under the limit for the FETs' 75 V max. Going any higher requires new FETs, and at 20S also another power supply chip in the controller (according to our sources).

Was it worth all the hours fiddling, depotting, pondering and soldering? Definitely yes. I learned a lot and it was quite fun.

Next step will actually be true to the thread's subject; going to 50 A by virtually throttling max amps to 25 A in the settings and putting another 5 mOhm shunt in parallel. This - of course - first after verifying controller temps. Bear in mind my controller has the inferior nf75 FETs! (Has anyone ever blown their BBSHD's nf75 FETs yet?)

I will bring some feedback later. Hope you guys post your BBSHD adventures in detail as well. Cheers!

(Edit: Corrected incorrect cell Vmax of 4.09 V to 4.15)

 

[Matador]

Wow nice one E-mil!
What voltage did you run exactly ? (no load voltage)

Matador

 

[Epyon]

Good stuff happening here!
I made my maiden commute this morning at 71.4v. Had to up the resistor(s) to 200ohms. I lowered the current to ~45A (27 in config) to keep the wattage the same (~3200w) for now. Off the line acceleration feels the same. Top end increased slightly. The pull at the top end is noticeably stronger. I might drop my gearing another tooth or two. Gotta squeeze as much fun in as possible before the snow hits. Already looking at motorizing another of my bikes as a dedicated snow machine.

 

[E-mil]

Wow nice one E-mil!
What voltage did you run exactly ? (no load voltage)

Thanks guys - it's always nice to have an audience however few we are upping the game for BBSHD!
I have done two 5 km runs yesterday and today at resting voltage of approx 71 V, but the sag of 5-6 V at 45 A effectively removes a little of the kick-ass factor. Nonetheless my bike is now easily brought up to 60 km/h whereas before the same was true for 50 km/h. I guess a top speed of a good 70 km/h is generally possible.

@ Epyon -- A word of advice, although speculative: Maybe it would be a better solution to use LED's instead of resistors for the display mod. I mean if its PSU chip is designed for 60 V max you could burn it. The LED method secures that there always is the Vf drop regardless of power draw from the display. The display could be in standby mode and have its caps charged even when turned off, and resistors do not drop the voltage at minuscule loads. Thoughts on this are welcome!

Anyway I soldered an extra LED in series, so I can be sure the display is not overvolted at full charge.
The Vf is now: 2.8 + 2.8 + 8.3 = 13.9 Vf. (at 50 mA load).
Not much headroom at full charge of 74.7 V (4.15 V/cell), since 74.7 - 13.9 = 60.8 V. (but it will suffice, and I won't charge higher than 4.0 V generally).


Hey, if the snow hits, please DO take it out for a spin! I had a blast on a late night run in February where the city streets for some hours had so much snow on them, that the evil salting machines hadn't been everywhere yet. This was with stock 15S controller, but leaning forward, it could still drift. Imagine with double the power now

 

[Epyon]

@ Epyon -- A word of advice, although speculative: Maybe it would be a better solution to use LED's instead of resistors for the display mod. I mean if its PSU chip is designed for 60 V max you could burn it. The LED method secures that there always is the Vf drop regardless of power draw from the display. The display could be in standby mode and have its caps charged even when turned off, and resistors do not drop the voltage at minuscule loads. Thoughts on this are welcome!

I don't run a display, the current draw without it is pretty steady around 20mA. Anything over 61.4v and the controller simply won't power on, so no worries about frying anything. Though I imagine there is a limit on whatever component controls that.

Hey, if the snow hits, please DO take it out for a spin! I had a blast on a late night run in February where the city streets for some hours had so much snow on them, that the evil salting machines hadn't been everywhere yet. This was with stock 15S controller, but leaning forward, it could still drift. Imagine with double the power now

I considered doing the necessary weatherproofing, but just decided to build a dedicated winter fatbike. Either gonna do another modded BBSHD or a Tangent. I enjoy the build even more than the finished product :lol:

 

[fechter]

The internal voltage regulator is rated for 80v.

The load on the power line is not constant, so using a resistor will give a variable amount of drop. Also keep in mind when you first turn it on, there will be a surge as it charges some caps. I tried a zener diode and it just fried. A TVS diode might work well though.

 

[fechter]

On my test controller, I measured the current at about 25mA on standby, to about 40mA with motor running and display lit. From this, and knowing the amount of voltage drop, we can calculate the amount of heat that needs to be dissipated.

There are a number of ways to drop the voltage. The simplest would be to pick the right TVS diode. Large LEDs stacked in series work in a similar manner. Just a bunch of regular diodes in series will work also, but you need some 1A rated ones like 1N4001 or similar. At this current, you get about 0.74v per diode.

This circuit will give a voltage drop of about Vz + 0.7v, where Vz is the voltage of the zener diode. The transistor will dissipate most of the heat so needs to be sized large enough. The transistor only needs to be rated for more than the drop voltage, but a higher rating won't hurt.



Small LEDs stacked could be used in place of a zener diode.

Yet another version using a MOSFET:



This one is adjustable but the amount of drop is dependent on the gate threshold voltage of the FET, which will depend on temperature and load current. Over the typical temperature range, might vary by 1 volt or so.

Edited to remove incorrect circuit...

 

[E-mil]

Thanks for the circuits and insight, Fechter.

About the safety voltage drop, I was talking about the display, not the controller. I don't know if the display's caps and voltage controller can take damage, and that's what I was worried about

 

[fechter]

About the safety voltage drop, I was talking about the display, not the controller. I don't know if the display's caps and voltage controller can take damage, and that's what I was worried about

That's a good question. Since the display unit I have is potted, it's not easy to open up and check. Mine still worked after I blew a 12v zener dropper, so it seems to be good for at least 72v. I suspect it uses the same LM5007 voltage regulator as the controller. 80v rated. I suspect the display also uses 63v caps. As long as the voltage dropper is good, the display won't see more than the normal voltage.

 

[E-mil]

Well, here's to the rescue, Fechter:

I can detail how to safely break it apart if you wish.. I rubber-glued it together, and it's without problems since.

Gotta say I opened it because "ERROR 30 - Communication FxxxxP". The whole freakin display connector had bad solderings. Took me 4 hours to locate and fix the wire bugs. Had to test each lead! Unfortunately it was not in scope to detail the components of the display at the time