Replace Stock BMS with Remote Mounted Active Balancer (RS19)

[0000]

Thought I'd bounce this idea off the collective wisdom of the forums...

So I don't know about you, but I've been low-key persistently unsatisfied with the state of my stock 24S2P pack BMS-functionality in my wheel, and I've been thinking lately about how I'd go about making the situation better. To sum it up, I'm not a fan of the top-end resistance voltage bleed "balancing" because:

1. No ability to check real-time relative health of individual parallel cell groups and potentially catch pack issues before they become a problem. Stock BMS forces reliance upon total pack voltage by default to gauge overall pack health.

2. BMS allows individual cell groups to go over 4.20V before** slow resistance bleed is initiated, exacerbating both the time you are stuck waiting for your packs to finish charging and causing unnecessary degradation of the cells by increasing the time they are spent at or over 100% capacity.

3. No ability to reliably and safely partial charge to 80% or 90% for shorter rides while maintaining overall pack health/balance because of required top-end bleed balancing maintenance.

**I've actually gotten multiple different answers as to when bleed balancing initiates depending on who and when I've asked. Over 4.10V, over 4.18V, and over 4.20V from some posts on the forum I think. I've also never seen a BMS spec sheet (not for lack of trying) though I can't confirm anything either.

Anyway, knowing the stock BMS size is pretty flat and basically specifically sized to integrated with each 24S2P pack, I've not bothered looking for a direct replacement. However, I was poking around the other day and came across this 24S 4A active balancer on Ali-express. As far as I know, there is no communication between BMSs and the motor controller so getting rid of them altogether shouldn't be a problem.

So basically as long as the EUC could fit this 130x140x25mm Active Balancer case, theoretically I could cut open each battery, desolder each BMS and replace the connections with paralleled 24S pigtails, re-heatshrink batteries, mount the active balancer inside the EUC case, and connect it together for a much improved battery balancing solution with bluetooth monitoring capability. Anyone see anything wrong with this solution other than the Balancer case being potentially too big?

 

That said, I bet the Balancer case is probably too big by a half inch or so even if I were to cut out the lower left mounting post, but this seems like a good way to go if it were to fit. Thoughts?

 

[goatman]

if you cut out the bms from both packs you wind up with 4-12s2p blocks

its hard to explain without confusing but if you flipped one of the 12s2p end over end you would be able to connect the packs in parallel instead of series to make it 12s4p on one side of the wheel and 12s4p on the other side of the wheel (without having to rip off all the nickel stripping. if you have to rip off the nickel from perfectly good cells, its just not worth it) that get connected in series then you only need 1 active balancer instead of 2. 

im in the same situation of trying to find one that will fit

[alcatraz]

It's cheaper and safer to just swap to an external bms. That bms can have bluetooth etc. You can set it to balance at lower voltages etc. It takes care of charging.

You do not however use the supplied overcurrent protection unless it's a ridiculously high number, because you'll add another failure mode during acceleration/braking/pothole potentially causing cutoffs.

What you do need to do is attach balance wires to all your packs.

You need to route the charge port to the charge input on the controller. You can attach temp sensors. I'd look into getting a bms with heatpad functionality, but that's a luxury.

I've seen reputable brand smart bms 24s for only 50usd (local price). Brand is "jikong" 极空. There are others but this offers a good looking app and it's got decent balancing currents.

[Chriull]

4 hours ago, Vanturion said:

1. No ability to check real-time relative health of individual parallel cell groups and potentially catch pack issues before they become a problem. Stock BMS forces reliance upon total pack voltage by default to gauge overall pack health.

Imho the biggest shortfall of (almost) all euc bms. Good newer models start with reporting!

4 hours ago, Vanturion said:

2. BMS allows individual cell groups to go over 4.20V before** slow resistance bleed is initiated, exacerbating both the time you are stuck waiting for your packs to finish charging and causing unnecessary degradation of the cells by increasing the time they are spent at or over 100% capacity.

One does not need to let the pack on the charger endlessly - manufacturers recommendation is about 50-60mA per cell as cut off threshold.

If imbalance is to big, the bms cuts off at some ~4.25-4.3V anyhow and one can start another balancing run after the overvoltage cell got bleeded. Once this happens one should think about replacing this cell slowly anyhow.

4 hours ago, Vanturion said:

**I've actually gotten multiple different answers as to when bleed balancing initiates depending on who and when I've asked. Over 4.10V, over 4.18V, and over 4.20V from some posts on the forum I think. I've also never seen a BMS spec sheet (not for lack of trying) though I can't confirm anything either.

The bleeding resistors voltage threshold makes sense to conform with the final charging voltage, i.e. 4.2V.

But as many used bms components have 5mV tolerance about most of your numbers could be true in real specific cases. @RagingGrandpa measured afair 4.18V threshold for some gotway bms: https://forum.electricunicycle.org/topic/18374-how-bms-works-gotway-pack-teardown/

Afair we also looked up the datasheet for in this bms used components detailed specifications?

In this link also the bms synchronisation measures are shown.

4 hours ago, Vanturion said:

As far as I know, there is no communication between BMSs and the motor controller so getting rid of them altogether shouldn't be a problem.

There is none with the motherboard, but between the two packs. As both bms have to stop charging once one encounters cell overvoltage.

4 hours ago, Vanturion said:

 Thoughts?

My very very very personal gut feeling is that most/many battery probs come from bad weldings/nickel strips, badly matched cells or bms defects.

I'm also not convinced that 80% charging could increase anyhow cycle count - especially in case of a RS19 with 4p 21700 5Ah li ion cells... They should have enough stress during riding that other measures won't count.

But, yes - a bms with active balancing and app support is cool and great. If one likes the "hassle" of rebuilding the batteries and one has the abilities and skill to perform this in a way one does not have to fear sudden faceplants (or worse) because something went loose or did not work out...

[0000]

12 hours ago, goatman said:

if you cut out the bms from both packs you wind up with 4-12s2p blocks

its hard to explain without confusing but if you flipped one of the 12s2p end over end you would be able to connect the packs in parallel instead of series to make it 12s4p on one side of the wheel and 12s4p on the other side of the wheel (without having to rip off all the nickel stripping. if you have to rip off the nickel from perfectly good cells, its just not worth it) that get connected in series then you only need 1 active balancer instead of 2. 

im in the same situation of trying to find one that will fit

I think I get what you're saying, in each 24S2P pack, you have 2 modules of 12S2P assemblies that are (already) connected in series together probably with nickel strips. I'm not talking about disassembling the existing nickel connections though, what I was thinking was to run 25 wires from each pack from the same location the stock BMS used and use a parallel harness so each 24S2P would become one 24S4P through use of the wiring harness to the 1 24S Active Balancer. A downside being that the exposed wires wouldn't be individually fused, so the wiring harness to the Balancer would need to be well protected within the wheel case.

[0000]

12 hours ago, alcatraz said:

It's cheaper and safer to just swap to an external bms. That bms can have bluetooth etc. You can set it to balance at lower voltages etc. It takes care of charging.

You do not however use the supplied overcurrent protection unless it's a ridiculously high number, because you'll add another failure mode during acceleration/braking/pothole potentially causing cutoffs.

What you do need to do is attach balance wires to all your packs.

You need to route the charge port to the charge input on the controller. You can attach temp sensors. I'd look into getting a bms with heatpad functionality, but that's a luxury.

I've seen reputable brand smart bms 24s for only 50usd (local price). Brand is "jikong" 极空. There are others but this offers a good looking app and it's got decent balancing currents.

I think we're basically talking about the same thing.

I looked at the Jikong BMS and if you take a look at the component board in the advertised pictures, most of it looks identical to the Active Balancer I linked, in fact they are both sold on Ali-Express by the "IC GOGOGO" store. So they are probably using the exact same active balancing function and hardware. Like you said, it looks like you get more functionality with the BMS over the active balancer, but with an inline battery fuse and non-integrated temperature sensor (I have the LiTech batteries from eWheels), I don't think I need any of the other safety features. Current limiting and protection is set and controlled by the motor controller, and I'm not worried about overcharging as I'm using a reliable set of meanwell power supplies for charging.

Being longer, I'm thinking the Jikong BMS definitely wouldn't fit in that remote empty space in the RS case, but being slimmer, it may have a better chance being substituted directly into the same space as the stock BMS board on each battery pack.

[Chriull]

9 minutes ago, Vanturion said:

Current limiting and protection is set and controlled by the motor controller,

Yes current limiting is not a feature of a bms, that's the motor controllers job.

Just short circuit protection is a necessary feature, as you mentioned imo with the 

11 minutes ago, Vanturion said:

but with an inline battery fuse

9 minutes ago, Vanturion said:

and I'm not worried about overcharging as I'm using a reliable set of meanwell power supplies for charging.

Overcharge protection from our bms is normally single cell (group) overvoltage protection - in case of "grave" imbalances. Works as protection for (not) too high input voltages, too.

[0000]

Chriull,

I agree with most of everything you said, especially about where these battery problems are coming from. That was one of my biggest motivators for going with a RS19 from eWheels to get a shot at using a different battery assembler other than Begode as I think history proves they are not a reliable or safe manufacturer for pack assembly specifically.
 

11 hours ago, Chriull said:

One does not need to let the pack on the charger endlessly - manufacturers recommendation is about 50-60mA per cell as cut off threshold.

In a 24S4P pack, you're looking at .05A * 4.2V * 4 * 24 = 20.2W, which is still dragging on quite a bit longer than needed if I could confidently stop at 50W or higher. Although I've been stopping lately at about 2W probably more out of paranoia of not being able to see where my cell groups are at. With a 550W charger this usually equates to roughly a 3 to 3.5 hour charging time for me.

 

11 hours ago, Chriull said:

Afair we also looked up the datasheet for in this bms used components detailed specifications?

I asked LiTech for the spec sheet, but they claimed it was proprietary to eWheels and eWheels wouldn't get back to me unfortunately when I emailed for it. I'm not sure if assuming the balance voltage threshold is exactly the same as stock Begode is correct or not, it could be.

11 hours ago, Chriull said:

There is none with the motherboard, but between the two packs. As both bms have to stop charging once one encounters cell overvoltage.

Good to know. As for the Active Balancer BMS replacement solution, I personally wouldn't have a problem manually monitoring charging for overvoltage conditions and manually disconnected if I see a problem developing, but I know many are the set-it-and-forget-it types and this wouldn't work for them.

11 hours ago, Chriull said:

I'm also not convinced that 80% charging could increase anyhow cycle count - especially in case of a RS19 with 4p 21700 5Ah li ion cells... They should have enough stress during riding that other measures won't count.

This is the one thing I'm going to hard disagree with you on. I've read several research papers on truncating both top-end and bottom-end li-ion to slow capacity degredation, and the effect is pretty well known. Particularly on the lower end (avoiding voltages below 3.2, 3.3, or even better 3.4 for slowing capacity degredation), although this is already limited by the motor controller. I've even done my own testing on several different cells including Molicel P42As here to confirm the effect. It is definitely beneficial to limit your depth of charge and discharge if you care about long-term capacity degredation. That said, I think there's a happy medium to be struck between actually using all of your battery's capacity and limiting the capacity decay, but top-end balancing doesn't really allow for it.

*edit, got my math wrong

Edited April 14, 2022 by Vanturion

[Chriull]

20 minutes ago, Vanturion said:

This is the one thing I'm going to hard disagree with you on. I've read several research papers on truncating both top-end and bottom-end li-ion to slow capacity degredation, and the effect is pretty well known.

Yes, i fully agree on this - would be stupid to discuss about research papers as "outsiders".

My statement to this was, as these tests, afaik are performed well within the current limits of the cells. GW wheels with 4p config very often not. I'd guess that's substantial harm for the cells, too. So i'd see this relativating high voltage stress.

24 minutes ago, Vanturion said:

Particularly on the lower end (avoiding voltages below 3.2, 3.3, or even better 3.4 for slowing capacity degredation), although this is already limited by the motor controller.

+1. Especially as for the same torque at lower voltage higher battery currents are needed. So higher burden from current and low voltage. Together with the danger of increasing imbalance and hit some weak cells with even lower voltages.

 

[0000]

10 minutes ago, Chriull said:

GW wheels with 4p config very often not. I'd guess that's substantial harm for the cells, too. So i'd see this relativating high voltage stress.

This is something I don't think I can agree with, but I'm not trying to be disagreeable for argumentation sake here! So I don't really know what's going on within the motor controller as to what the battery current and phase (motor) current limits are, but assuming, for the RS19 anyway, that the 2600W is the true and actual limit programmed power limit in the motor controller, you get the following:

• 2600W/100.8V (24S@4.2V/cell) = 25.8A Maximum battery current at full charge
• 2600W/79.2V (24S@3.3V/cell) =32.8A Maximum battery current at what I've been told is the lower-bound voltage limit.

So basically, the maximum current an individual cell can see in the 4P arrangement is 32.8A/4  = 8.2A/cell. This is obviously being well within the ~10A continuous rating of the 50E2 cells and the 15A peak rating. Considering that the RS19 that I have has actual 30A in-line fuses on the battery mains that haven't burst under hard acceleration conditions and hills yet, I think it's fair to say that these calculations are correct, particularly as performance (current) is notoriously limited below ~50% capacity. So if there are problems with the batteries (that we've clearly seen), I don't think it is with the battery current limits and they're probably staying under 30 Battery Amps at all times (with the RS19).

That is, unless Begode is programming their controller battery (and phase) current limits inconsistently between batches of wheels. I don't think that's the case, but who knows - no one outside of Begode techs has yet gotten into messing with the motor controller programming yet.

[Chriull]

12 minutes ago, Vanturion said:

So I don't really know what's going on within the motor controller as to what the battery current and phase (motor) current limits are, but assuming, for the RS19 anyway, that the 2600W is the true and actual limit programmed power limit in the motor controller, you get the following:

There is no 2600W power limit programmed in the firmware. GW is and was always proud to deliver "pure performance". By physical limits of the wires and components they have to limit the current for low speed operations. And here afair once a 200A phase current limit was measured/reported?

But during "normal" riding the current limit is the back emf of the motor, the battery voltage and the curcuit resistance.

17 minutes ago, Vanturion said:

Considering that the RS19 that I have has actual 30A in-line fuses on the battery mains that haven't burst under hard acceleration conditions and hills yet,

These fuses are very slow and can take quite high currents for some seconds...

And there is one for each battery pack?

BTW - the antique KS16C with 76.2V already reached peaks above 2000W delivered from the battery...

 

[0000]

We're getting pretty far off-topic from the viability of replacing the stock BMS with an Active Balancer, but just to summarize, in lieu of contradictory evidence I'm going to tentatively say that:

Yes it's possible and would work as long as you can live with the Pros and Cons already discussed (and it fits in the case). I don't know if I'll do the mod or not, I'd have to take some measurements first anyway, but I'm leaning towards no at this time even though I hate top-end balancing with no real-time voltage feedback. That said...

 

1 hour ago, Chriull said:

There is no 2600W power limit programmed in the firmware.

Can you really say that definitively? Every motor controller I've personally worked with (Bafang, Nucular, Infineon) work by setting battery current and phase current limits separately. I know it's not unusual for phase current limits to be set sometimes at 2 to 3 times higher than battery current limits too. I have been assuming when Begode claims that the motor power is 2600W, that they actually mean that is their programmed power limit in the motor controller and experienced through the motor. I guess I could be wrong about that.

Anecdotally, I've run hub-motor ebikes from power limits of 1200W up to 8000W and a 2600W power limit seems pretty in-line from the HT winding performance I've experienced riding my RS19. I can't really say I'm right about my assumptions here unless I were to physically check the measurements myself with an ammeter on the battery main wire under hard acceleration. But I'm not really keen to run that test anytime soon.

1 hour ago, Chriull said:

These fuses are very slow and can take quite high currents for some seconds...

And there is one for each battery pack?

True, and you're right there are 2, one near each 24S2P pack, so what I said about blowing the fuses isn't correct. They'd only trip over 60A and under some sustained current at that depending upon the peak. On the other hand, I have a hard time believing the battery current ratings aren't being respected, particularly with the HT winding that requires less current to produce the same amount of torque at low speeds than the HS motor. I guess I'd have to see some evidence one way or the other to believe the motor controllers aren't programmed with reasonable battery current limits.

In your KS cutout example (nice graph btw), the cutout appears to be the product of back-emf limitations (exceeding physical limit of torque generation due to motor winding at that speed) which I think we agree on. But my question is, what is going on with the motor controller once the "cut-out" manifests? In this example, is it possible that the motor controller cut-out occurred after it already sensed an over-lean condition had already been initiated so it wouldn't keep applying power knowing the rider is already headed to the ground? So what I'm getting at is in this example anyway, motor controller current and phase limits wouldn't apply because the cut-out wasn't a consequence of those limits if they exist in the controller.

I don't know, I could be stuck in the ebike controller mentality - I guess an easy way to clear up the misunderstanding is to ask if you or anyone knows definitively if Begode or EUC controllers in general don't program battery and phase current limits? And if they don't, what are the conditions required to trigger a low-speed motor cut-out, just some kind of physical component failure then if too much load is demanded? That doesn't seem right to me...

Edited April 14, 2022 by Vanturion
words

[Chriull]

6 minutes ago, Vanturion said:

I don't know, I could be stuck in the ebike controller mentality

Ebikes can easily limit power or current anytime - for an EUC this means a faceplant.

6 minutes ago, Vanturion said:

 - I guess an easy way to clear up the misunderstanding is to ask if you or anyone knows definitively if Begode or EUC controllers in general don't program battery and phase current limits?

They have, newer begode wheels afair were reported/measured/estimated to limit the phase current to 200A.

Low Speed, standstill, pendulums, etc need special limits, too.

After the battery fires begode started to use fuses and maybe started with some more limits?

KS had very bad reviews because the limited low speed torque a bit more for the S20 than begode wheels do.

Maybe @Freestyler already has some insights? https://forum.electricunicycle.org/topic/27262-euc-dash/?do=findComment&comment=411819

 

[Freestyler]

I have seen some values that are associated with current.

Every wheel has different names for them in their extended packets, so I'm going to use the mcm5 names:

 

phse: At first I thought this is the phase current, but it's only a counter of sorts.

It increments when going forward and decrements when going backwards.

So it is either a count of wheel rotations or a form of energy used / restored

Protect_Current: This is the phase current that is also reported in the main packets (although it's multiplied by 10 there)

MaA: This is the max phase current reported

MiI: Still not sure. Its calculation is the same between all wheels and has a maximum value of 9000.

RoatCurrMax: This seems like the max battery current value reported

I've tried to reset this value on each packet in order to get a real-time value, but it did not go well, I crashed the firmware

 

I did not find a direct limit enforced against these values (doesn't mean there is not), but I found a number (in the same place as these values) that looks like a plausible phase current limit.

150 for mcm5, 160 for tesla3, 200 for Nikola, 220 for rs / ex.n, 235 for the monster pro.

The number is increased the strongest the wheel gets, so it kinda checks out.

 

My personal feeling is that begode does not place additional software limits in its firmware.

@Vanturion you keep mentioning the nominal watt value, but every PEV I know peaks to almost double that value, so enforcing a limit based on the nominal value, seems kinda low.

The limits in other brands would be closest to the peak value. For example Inmotion v11 with a nominal rating of 2200w, peaks at around 4400w.

Edited April 15, 2022 by Freestyler

[0000]

21 hours ago, Chriull said:

Ebikes can easily limit power or current anytime - for an EUC this means a faceplant.

Yeah, you're totally right about that.

Until now, I assumed that most of the time EUC riders are operating within both battery and phase current limits set by the motor controller (ignoring cutouts at speed due to physical limitations of motor winding to produce torque at higher RPM). If that is not the case and there are no defined current limits, then I guess the fundamental or default logic of the EUC system designers is something like:

"Keep rider upright no matter what even if wheel go boom"

Which means that without fuses, the safety and survive-ability of the hardware components is entirely dependent upon the robustness of the hardware and rider technique/behavior. Behavior like not holding the wheel against an immovable obstacle and inducing acceleration or caking the wheel with mud and riding at low speed until something in the controller pops from overheating.

So without "safe" hardware limits (fuses) or software limits (phase current threshold) to detect increasing/threshold phase amp generation without the wheel's RPM increasing, under a constant load you would get increasing stress/heat buildup in the motor, controller, batteries, and wiring until something catastrophically failed.

I just read the other day that someone had gotten a stick stuck in their wheel and had to immediately dismount and get it out before "their controller burned out" or something like that. So yeah, I guess no motor controller current limits then. This is more unsophisticated than I was expecting, crazy.
 

8 hours ago, Freestyler said:

My personal feeling is that begode does not place additional software limits in its firmware.

@Vanturion you keep mentioning the nominal watt value, but every PEV I know peaks to almost double that value, so enforcing a limit based on the nominal value, seems kinda low.

The limits in other brands would be closest to the peak value. For example Inmotion v11 with a nominal rating of 2200w, peaks at around 4400w.

Thanks for weighing in on this Freestyler. Yeah, I guess I've been biasing my perspective from a battery-first point of view if that makes sense. Basically I assumed that the continuous and peak ratings of the batteries chosen was respecting to reasonable limits programmed into the motor controller that already accounted for all possible loads generated by normal and aggressive EUC riding in all possible conditions with all types of rider weight (basically deriving a maximum design load). This was apparently wrong.

So without software current limits, a rider can generate a load requiring very high phase amps, thus requiring very high battery amps, particularly at 0 to low speeds, that can induce high peak current from the battery cells beyond their peak current rating. This especially in battery pack arrangements with low parallel group count. Even for a short amount of time, exceeding battery peak current rating can't be all that great for the long-term survive-ability of high-energy cells; cells like the LG M50LT used in Begode's 900mAh packs which has popped up on everyone's radar due to wheel fires. On top of that, normal operation of an EUC forces the battery to sustain constant cycling between discharge and charge with normal braking and acceleration. I think it's asking a lot from these batteries, high-energy type cells in particular.

Yeah, I think I'm starting to get a clearer picture of the situation now. Because of the way the EUC system is designed in general, it makes even more sense to favor the high-torque winding for any off-road application (more torque/amp = less overall system stress including batteries). Also to go with high-power cells like the Samsung 40T and Molicel P42A for larger wheels (which require more current/unit torque), for battery packs with low parallel group counts, and for most planned aggressive riding too.

Basically, use a large-diameter, HS wheel with a low-parallel high-energy battery group count in off-road conditions at your own risk I guess.
 

8 hours ago, Freestyler said:

RoatCurrMax: This seems like the max battery current value reported

I've tried to reset this value on each packet in order to get a real value, but it did not go well, I crashed the firmware

This is what I'd be most interested in seeing if you had this info as this would validate if the cells are being used beyond their current ratings without having to bust out an ammeter and try and fit it on test runs. What is the maximum battery current values being demanded from each of these wheels under "normal" or "aggressive" operation?

Anyway, I just read through the EUC Dash thread, awesome work there btw

Edited April 15, 2022 by Vanturion
writing a book here

[Freestyler]

Inmotion and kingsgong impose current limits from my understanding.

For example check this video at the 20 minute mark, where they jam 2 inmotion V12 against a tree and they tiltback. https://youtu.be/ziHTOHlNmx4

 

I think I saw some people mentioning that extreme commander limits the current at speeds below 3-5 km/h, but I can't find the video right now, (I think it was in wrong way channel) and I don't have a commander firmware to check.

 

Truth be told, I didn't do much testing with RoatCurrMax. I was hoping people would do it. 😁 

[Menace]

Isn't that a litech pack? If it is I'm not sure it would be of benefit to swap out the existing bms, as it is supposed to be better than the begode ones to begin with.

[0000]

5 hours ago, Freestyler said:

Inmotion and kingsgong impose current limits from my understanding.

For example check this video at the 20 minute mark, where they jam 2 inmotion V12 against a tree and they tiltback. https://youtu.be/ziHTOHlNmx4

I think I saw some people mentioning that extreme commander limits the current at speeds below 3-5 km/h, but I can't find the video right now, (I think it was in wrong way channel) and I don't have a commander firmware to check.

Hmm, tilt-back seem like a good solution as long as it doesn't activate too early/easily. Man it'd be great if all of the manufacturers published their programmed controller current limits when implemented and at what point features like tilt-back is initiated. Better yet, give us direct access to view and modify controller settings!

I mean in a perfect world, maybe this sort of technical information wouldn't need to be public knowledge, but with each of the manufacturers having issues in the past with certain models for sometimes different reasons, as well as motivated distributors like eWheels using 3rd party assemblers to improve pack assembly and now move toward high power battery cells in certain model(s), I think these are all reasons to want to verify what's going on electronically and make sure component selection was appropriate for long-term reliability of these EUCs.

5 hours ago, Freestyler said:

Truth be told, I didn't do much testing with RoatCurrMax. I was hoping people would do it. 😁 

Haha, probably just a matter of time before info is shared to forum then..

 

4 hours ago, Menace said:

Isn't that a litech pack? If it is I'm not sure it would be of benefit to swap out the existing bms, as it is supposed to be better than the begode ones to begin with.

It is. Well, to be honest I'm not sure how much better, if at all, the LiTech BMS is over the stock Begode BMS, or if it's even different. The LiTech BMS is still a top-balancer unfortunately. I watched a video where LiTech shows that the BMS they used also balances between the two 24S2P packs in the RS wheels if they are at different voltages, and the video demonstrates different charge current going to each pack. But, I don't know if this feature is something that wasn't included in stock Begode BMS or if it was something that was even needed. Since the 2 packs are already in parallel, I don't know why it would be...

IMO, the biggest reasons to go for a Begode wheel with LiTech battery packs was for stronger cell holders (better to handle massive abuse from a lifetime of vibration and shock impulses), use of pure nickel inter-connectors (instead of poor conductivity nickel-plated steel that can generate unnecessary heat), larger dimension inter-connects (IIRC), better protection against water-ingress, multiple temperature probes and alarm, and most importantly, not leaving the assembly up to Begode for this safety-critical component.

It's just a shame though that while they made these nice improvements, they did not also deliver an active balancing solution with bluetooth integration for real-time battery status monitoring. That's my only complaint.

Edited April 16, 2022 by Vanturion

[Menace]

20 minutes ago, Vanturion said:

to be honest I'm not sure how much better, if at all, the LiTech BMS is over the stock Begode BMS, or if it's even different. The LiTech BMS is still a top-balancer unfortunately. I watched a video where LiTech shows that the BMS they used also balances between the two 24S2P packs in the RS wheels if they are at different voltages, and the video demonstrates different charge current going to each pack. But, I don't know if this feature is something that wasn't included in stock Begode BMS or if it was something that was even needed. Since the 2 packs are already in parallel, I don't know why it would be...

IMO, the biggest reasons to go for a Begode wheel with LiTech battery packs was for stronger cell holders (better to handle massive abuse from a lifetime of vibration and shock impulses), use of pure nickel inter-connectors (instead of poor conductivity nickel-plated steel that can generate unnecessary heat), larger dimension inter-connects (IIRC), better protection against water-ingress, multiple temperature probes and alarm, and most importantly, not leaving the assembly up to Begode for this safety-critical component.

It's just a shame though that while they made these nice improvements, they did not also deliver an active balancing solution with bluetooth integration for real-time battery status monitoring. That's my only complaint

Hmm, interesting. I didn't know that.

[Tawpie]

8 hours ago, Vanturion said:

"Keep rider upright no matter what even if wheel go boom"

You would think, right?

This seems to be the GW approach, they allow the rider to make their own decisions about how far to push past the warning systems. This very well may contribute to the mysterious fires some of their packs have had, or not. But there are good reasons to point at "battery abuse" as a general underlying cause of the fires.

KS on the other hand, and this is an unsubstantiated opinion, takes the opposite view... I have a hunch that their tack has been "If I let this continue and the wheel does go boom, then the rider is on their face—might as well save the wheel since the rider is going to be eating it either way". Then they warn early (and often), and do not let you disable tilt-back. I get it. But it does mean that KS has given up "performance" because their limits are conservative.

I'm glad to see all manufacturers (forced there by eWheels) acknowledge that perhaps the push for capacity/range is at odds with performance/long term battery reliability. Having the option to choose my battery is a very welcome development. I really hope it becomes universal and not limited to "super large" distributors.

Edited April 16, 2022 by Tawpie

[mrelwood]

 

On 4/14/2022 at 9:41 PM, Vanturion said:

I've read several research papers on truncating both top-end and bottom-end li-ion to slow capacity degredation, and the effect is pretty well known.

It is. But what many people miss is the cycle count where the differences become actually worth enough to care.

500 charge cycles is a common cycle count (for the State of Health to reach 80%) for the batteries. Take a 1800Wh EUC with an average max riding range of roughly 75km. EUCs stop you at about 10% of power left in the battery, so one cycle accounts for roughly 82.5km. 500 cycles gets you past 40000km.

The furthest I know of anyone riding on a single EUC is a little past 30000km, but it was on a Monster which has a much bigger battery, increasing the 500 cycle distance even much further.

 Same goes for the discharge. The charts I’ve seen have a scale of 5000 cycles, because the difference at 500 cycles is minimal.

 If you limit your riding between 30% and 80%, you are using 50% of your expensive battery, and it will still decrease to 85-90% SoH after 500 cycles. If you don’t worry about your battery, you’ll use 100% of the capacity for a long time, which decreases to 80% somewhere around 30000-80000km depending on the wheel.

 Be realistic, how far do you think you’ll ride with your wheel?

 

On 4/14/2022 at 10:31 PM, Vanturion said:

assuming, for the RS19 anyway, that the 2600W is the true and actual limit programmed power limit in the motor controller

As you’ve now learned, it is not. The announced motor power is the nominal power rating of the motor. Ie, that’s the power it should be able to handle indefinitely. The peak power of the motor is rarely announced, just like the maximum power delivery of the controller, but they are both usually around double the motor nominal.

While I believe there probably is an actual current limiter for the phase current, reaching 220A on the MSP or the Sherman is not even difficult for a 200lbs rider. On Inmotion and KS though the power limiter driven tilt-back should engage well before that. But the threshold is not fixed, so even if you would somehow hack the firmware to find out the limit, it would be a complicated function instead of a simile fixed value. And it can change between firmware versions, like it has on the V11.

[0000]

13 hours ago, mrelwood said:

It is. But what many people miss is the cycle count where the differences become actually worth enough to care.

500 charge cycles is a common cycle count (for the State of Health to reach 80%) for the batteries. Take a 1800Wh EUC with an average max riding range of roughly 75km. EUCs stop you at about 10% of power left in the battery, so one cycle accounts for roughly 82.5km. 500 cycles gets you past 40000km.

The furthest I know of anyone riding on a single EUC is a little past 30000km, but it was on a Monster which has a much bigger battery, increasing the 500 cycle distance even much further.

Well, I won't speak for others, but I do care, but for reasons not just related to slowing (even marginally) capacity degradation, particularly if I know I don't need the maximum range in advance. On principle, I don't like using components like batteries at their margins. However, with EUCs, a good reason not to charge to 4.20V is that the top-balancing allows cells parallel cell groups to exist over 4.20V up to 4.25V, in addition to this, if you add excessive braking (regen) into the mix fresh off the charger, you can end up stressing the cells in the leading to rapid capacity fade or worse, thermal-runaway. I don't know if that S20 demo wheel fire in NYC was an example of this or not, but something, I think, worth some consideration.

So it makes sense to me to leave some margin of safety given the limitations of top-balancing and not being able to the real-time voltage status of each cell group. If it weren't for top-balancing, I'd normally charge to 4.15V or 4.10V for shorter trips. Begode's motor controller already sets the lower-bound limit at 3.3V/cell which is helpful in limiting capacity fade as well as voltage drift between parallel cell groups so that's pretty good at least.

 

13 hours ago, mrelwood said:

While I believe there probably is an actual current limiter for the phase current, reaching 220A on the MSP or the Sherman is not even difficult for a 200lbs rider. On Inmotion and KS though the power limiter driven tilt-back should engage well before that. But the threshold is not fixed, so even if you would somehow hack the firmware to find out the limit, it would be a complicated function instead of a simile fixed value. And it can change between firmware versions, like it has on the V11.

For sure, I get it now, it looks like peak phase current can get a lot higher than I was expecting. My main focus in all this is trying to understand what's going on with the battery current, particular during these performance peaks. I'm wondering how many of these wheels, if any at all, have chosen cells and a current sharing arrangement that allows peak current to exceed their rated peak current capacity. Not just for reasons of early or excessive capacity fade, but also for reasons of safety... And based on understanding from the conversation here, it seems like some Begode wheels, lacking tilt-back feature, would be the first place I'd want to look (at max measured battery current).

Edited April 17, 2022 by Vanturion

[goatman]

i normally go 4.15 on charge

id rather see begode change from 3.3v to 3.0. theres way more capacity from 3.3 to 3.0  than 4.1 to 4.2 then youre operating inbetween the upper and lower shoulders of the discharge curve

[0000]

14 hours ago, goatman said:

i normally go 4.15 on charge

id rather see begode change from 3.3v to 3.0. theres way more capacity from 3.3 to 3.0  than 4.1 to 4.2 then youre operating inbetween the upper and lower shoulders of the discharge curve

Unless you got a different method of balancing, I probably wouldn't undercharge that much too often as voltage drift between parallel cell groups can grow over time without reliably activating bleed voltage resistors each charging cycle as I understand it. I've been charging to 100.5V or 4.1875V/cell measured with a multimeter at the end of the charging cable since LiTech said 4.18V/cell was the balancer activation threshold and I wanted to give at least a little margin for living in a hilly neighborhood for regen braking.

As for the low-voltage cutoff threshold, I hear ya on the loss of capacity. It's pretty bad for high-energy cells like the 50E as they have quite a bit more packed capacity packed in the low-end, especially with lower current discharge rates. On the other hand, I think it's much easier to accumulate voltage drift with a threshold that lengthens time spent in the non-linear part of the discharge curve, especially when demanding higher performance/discharge current with low battery.

I know a lot of people probably don't care about this technical stuff, but I like to min/max things like this. I think my ideal cycle with active balancing would be 4.15V to 3.2V/cell with a high energy battery pack for a good compromise between pack longevity, safety, and usable range.

Edited April 17, 2022 by Vanturion

[goatman]

yes, but if youre riding under load and the lvc trips at 3.0v, the at rest voltage of the cells will bounce back up to 3.3v.

with older cells like what i use in my builds as the cell ages it wont hold at 4.2v after charging, the at rest voltage might drop to 4.165v

when i charge a pack like that it will go out of balance right away after a few charge cycles so i might lower the charging voltage to 4.15v or 4.1v to keep the pack balanced but to try and force balance an older pack at 4.2v is asking for trouble in my opinion.

with the cheap blinky blink active cell balancers, i would balance the pack down around 2.7v to 2.5v where there is less capacity per 0.1v which creates the pressure required to balance the pack through energy transfer than trying to burn off the energy up at 4.2v that your trying to force into cells that cant retain it,

i dont need to balance those salvaged cell battery packs every time i charge, maybe every 50 charges id do it when id notice the at rest pack voltage had changed 0.2v from 69.7v to 69.5v (4.1v/17s packs) and if they had a programmable controller lvc that you could set for doing pack maintenance it would be nice and easy to do, like on my bikes