Commuter E-Bike: Grin All-Axle Hub Motor with Detachable VESC & 27s3p P45B Battery Pack in One Box

Hi,

I have been planning my first e-bike build for months, and I think I am ready to start purchasing all the components I need. Below, I will briefly summarize my plans and would be happy to hear your thoughts and critiques.

The goal is to build a hardtail city/commuter bike with an average maximum speed of 30-40 km/h, which can also be powerful and fun to ride. For my daily needs, I expect around 400-600W of consumption before the next charging.


1. **Motor**: Grin All-Axle Standard Wind

I chose the standard winding because it is 1-2 percent more efficient due to its slightly higher copper fill.

The disadvantages are the need for higher voltage, which requires a controller with high-voltage MOSFETs that have higher power losses. Also, this comes with potentially dangerous voltage levels and fewer compatible components like BMSs and chargers.


2. **Controller**: Tronic X12 30s (VESC)

I like the VESC a lot. It is highly customizable and appears to be a well-polished project. My concern is that the settings for a hub motor may not be plug-and-play. Some people have reported problems, especially with cogging at higher speeds. I'm hoping for the best!

Choosing Tronic wasn't difficult; there are very few high-voltage VESCs. It has a very small form factor, seems to use high-quality electronic components, and while there aren't many reviews, the ones I have found are positive.


3. **Battery**: DIY Pack Made from Molicel P45B in 27s3p configuration.

I'll use a copper-nickel sandwich construction for the battery. I chose 27s3p because I need high voltage for that motor winding and have space for around 80 cells.

For a 27s configuration, my peak battery current will be 60-90A, with a maximum duration of up to a minute before risking temperature damage to the motor.

For everyday use, the battery current will be only ~10A (3.3A per cell).

Theoretically, for such currents, the Samsung 50S might be a better choice. However, I’m not entirely convinced that the slightly better capacity at low currents is worth it. The P45B has a better internal resistance, seems to run much cooler and is better known for quality then 50S.


4. **BMS**: Ennoid XLITE-V4 32S

Possibly the only BMS that can truly communicate with the VESC via CAN bus. I didn't find any alternatives.


5. **Charger**: Still searching

My requirements:
- Variable voltage and current with fast switching between different profiles.
- Manual selection is also acceptable (e.g., using knobs), but it should be convenient, not using a potentiometer requiring a screwdriver.
- It is important that the charger is quiet. Ideally, it should be silent during slow charging (1-2A) and have a fan, that to turn on only when needed for fast charging.

---

My idea is to make a removable case that will accommodate both the battery and the controller and will slide into the bike’s triangle. It will be made from acrylic sheets and 3D-printed parts. The top will feature a 300x80mm heatsink, to which the controller and the battery BMS will be attached.

The phase wires will be equipped with bullet connectors, and all other communication will be managed through a D-sub 25 pin connector. The bullet connectors and the D-sub will be combined in a custom 3D-printed connector for quick connection and disconnection.

The charging port will use an XT30 connector.

Sounds like you're on the right track in a lot of ways, you'll appreciate the all axle's motor to frame interface a lot. Consider a smaller than normal rear wheel to improve the torque ( it's not a big motor )
Vesc will = good times on anything
Look forward to seeing it!

I'm curious why you're shooting for such a high voltage. That's going to make, basically, every part of your build more difficult (you've already hit controller sourcing, bms', dc/dc's will be hard, etc).

The Tronics are...from what I've read, a bit of a mixed bag. Some people are very happy with them, but I've seen enough failures to not really want to invest in them. 3shul has a much better reputation, you may want to look there; also vesc.

The ennoid is great; I've had one and it had some errors (one guy making PCB's!) but he got me sorted out right quick and with no big hassles, so even if it wasn't a perfect out-of-the-box experience, I still think it's a great way to go. Especially with the bms-to-vesc setup so you can have a charge-only tiny bms but not miss out of over discharge and low voltage cut out protections, etc.

As for connectors, I've used something like this a few times and it's a nice way to handle bunches of pins.

You say 30-40 kph but at 96v in a 26" wheel the std winding motor has an unloaded speed of 120kph which is massively overkill. Did you mix up mph and kph? Even then 96v is still very high.

chuyskywalker said:

I'm curious why you're shooting for such a high voltage. That's going to make, basically, every part of your build more difficult (you've already hit controller sourcing, bms', dc/dc's will be hard, etc).

Click to expand...

Epithemeus said:

You say 30-40 kph but at 96v in a 26" wheel the std winding motor has an unloaded speed of 120kph which is massively overkill. Did you mix up mph and kph? Even then 96v is still very high.

Click to expand...

Two reasons - the kV of the standard wind All-Axle motor and optimally using the space I have in the frame triangle.

I can accommodate around 80 cells, which means I can build either a 20s4p or a 27s3p battery.

Comparing both variants in the simulator at around 3.2V per cell shows a significant difference.

I'm concerned that the top speed and the performance above 30 km/h, especially at half charge or less, will noticeably decline, even for daily riding where I don't push the bike to its limits.

Of course, with a 20s configuration, I will still reach my desired speeds, but the e-bike will be operating closer to its limits.

I'm trying to apply the same thinking as car manufacturers: a good car has much more top speed and power than theoretically needed.

And lastly, this will be my first e-bike. I'm not completely sure what I need or want, so it's better to have more power than not enough.

chuyskywalker said:

The Tronics are...from what I've read, a bit of a mixed bag. Some people are very happy with them, but I've seen enough failures to not really want to invest in them. 3shul has a much better reputation, you may want to look there; also vesc.

Click to expand...

You are not the first to suggest 3shul. I'm currently looking for more information about their controllers, specifically the CL350 V4.0; maybe I'm missing something.

The reasons I didn't choose it as a favorite last month when I was deciding on a controller are:
- I haven't found anyone who has used it.
- There were some negative comments about 3shul; to be more precise, their products were described as not bad but also not top of the line.
- I couldn't find any information about which MOSFETs are built in or any reviews or pictures detailing the design and components of the controller.
- Additionally, the fact that they are based in India doesn't give me much confidence.

Why are you comparing at 3.2v/cell? If you adjust your simulator for a near fully charged battery, that motor heats up a lot faster than I'd want to deal with.

bengxe said:

Why are you comparing at 3.2v/cell? If you adjust your simulator for a near fully charged battery, that motor heats up a lot faster than I'd want to deal with.

Click to expand...

It is an approximation for a battery under high discharge current with less than 50% charge remaining.

The time before overheating depends on the phase current; the voltage doesn't matter. I have limited the phase current to 120A in the simulator, which results in 1 minute before overheating. The motor will be monitored by the VESC, and if needed, the power will be automatically throttled.

vladrov said:

- There were some negative comments about 3shul; to be more precise, their products were described as not bad but also not top of the line.
- I couldn't find any information about which MOSFETs are built in or any reviews or pictures detailing the design and components of the controller.

Click to expand...

Give this guy on FB a holler; he does a lot of 3shul stuff out here.

vladrov said:

It is an approximation for a battery under high discharge current with less than 50% charge remaining.

The time before overheating depends on the phase current; the voltage doesn't matter. I have limited the phase current to 120A in the simulator, which results in 1 minute before overheating. The motor will be monitored by the VESC, and if needed, the power will be automatically throttled.

Click to expand...

I get what you're saying, but that's not how the simulator works. You're double compensating for the voltage drop, if your battery IR is realistic. More volts = more speed = more phase current, since its not hitting 120 phase amps anywhere. That may end up being a small difference, but on your setup it made the simulated temp go red. And regardless of what voltage you put in the simulator, it's a spicier motor than I would want to deal with on 27s. But that's only because there are comparable/ cheaper motor options that wouldn't rely so much on thermal throttling.

bengxe said:

But that's only because there are comparable/ cheaper motor options that wouldn't rely so much on thermal throttling.

Click to expand...

Which motor would you recommend?

If light weight is important and you can limit your power and speed goal slightly, the slower wind all-axle would be fine. An rh212 or leaf would make more power safely, but would add weight. A smaller diameter wheel will help with any of these, if your frame allows for it.

neptronix said:

115nm can happen, but at a great efficiency loss under high load.
You will make quite a bit of heat too.

Also we have all the torque on one of your chainstays instead of on both.. seems dangerous to me.

if you want that kind of torque i'd consider a leafbike 30mm or RH212 with dual torque arms. Both should be able to take short blasts of 3kw here and there

Click to expand...

Neptronix pointed out that the theoretical maximum torque the motor could generate might make the frame the weak point in my build. Based on the information I have, I’ve tried to make the most accurate educated guess.

The short answer: theoretically, it shouldn't be a problem.




I modeled one side of the bike’s rear triangle in 3D. The chainstay and other tubes are hydroformed, meaning they vary in shape and wall thickness, making accurate modeling difficult. To simplify, I used the weakest cross-section of the tubes. I measured the wall thickness through small holes in the frame made during the manufacturing process and found it to be around 4 mm, but I used 3 mm in the simulation to account for manufacturing tolerances and paint thickness.



The force from the torque arm is 907 N, which corresponds to 115 Nm of torque based on the length of the torque arm. Additionally, I also considered the normal gravitational force (from me and the bike), multiplied by four as a safety margin and to simulate road imperfections, such as hitting potholes.

The simulation showed a maximum of 83 MPa Von Mises stress in the model, which represents the maximum stress the frame would experience at a given point. For comparison, the yield stress of 6061 aluminum, the material used for the frame, is 241 MPa. Yield stress is the point where the material can no longer return to its original form after deformation.

To be conservative, I assume that the welds are 0.7 times as strong as the base material, so I reduce the material strength to 241 * 0.7 = 169 MPa.

This results in a safety factor of 169/83 = 2.

These calculations include several safety precautions and conservative assumptions. My more optimistic estimations suggest that the frame could withstand not just 2, but 5 times more force before deformation begins, with complete failure possibly occurring at around 8-9 times the applied force.

Same calculations for a more realistic 4 mm wall thickness:

Hey there, I think that an unproven controller and 96 V is not a great idea for your first build, and not really necessary if your desired average speed is 40 km/h. Here is my reasoning:

1. Dry skin conduction –According to electrical safety standards, the safe limit for direct current (DC) voltage that humans can safely touch is generally considered to be 60 volts. 72 volts is risky, but pretty common (a fully charged 72 volt battery is 84V, 96 volts is getting pretty scary. Of course, we try not to touch wires, but it does happen, especially on first builds.
2. if you're running a Grin All-axle in a 26 inch wheel, 40km/h will be nowhere near the peak efficiency of the setup. at 40km/h you'll be at 73% efficiency. The peak would be 88% at 99km/h, but you're not likely to want to run that fast very often (but you wouldn't actually attain that speed as the power drops off significantly.
   

72 volts would be plenty for what you want, and if you run 72 Volts you can use a proven controller like the Grin Phaserunner. You can also then use the high power Grin Cycle Satiator charger which is awesome. I run a 72-volt system with 20-inch wheels (OD 21.7 inches, they are big tires) and my cruising speed is 55km/h, top speed is 70km/h

I get the desire to build something exotic, but I recommend that you start with proven stuff and voltages. A first time build can be fun, but full of frustrations. Sticking with a proven standard can at least alleviate some of the frustration.

Would you be more comfortable working with 48v live wires? I don't think "it does happen" is a healthy way to think about this.

On the simulator, if you use the throttle to reduce the speed you'll find the efficiency isn't as bad as you say.

Higher voltage has advantages, that doesn't stop at 72v. The only real downside is the cost of components, but OP seems to have found reasonable options for most of that.

Would you be more comfortable working with 48v live wires? I don't think "it does happen" is a healthy way to think about this.

On the simulator, if you use the throttle to reduce the speed you'll find the efficiency isn't as bad as you say.

Higher voltage has advantages, that doesn't stop at 72v. The only real downside is the cost of components, but OP seems to have found reasonable options for most of that.

Click to expand...

Of course, one should try to avoid mistakes, and high amperage 48V is just as likely to weld a misplaced tool but there is no getting around the fact that higher voltage is more likely to overcome the resistance of dry skin. However, I think that the availability of more proven controllers at 72V is a more compelling reason. A Phaserunner controller at 72 V would allow the OP to attain a very decent top speed and fits their criteria of having some extra top speed when necessary.

That's a good point about the throttle, it does make a difference in the efficiency graph.

5. **Charger**: Still searching

Click to expand...

As mentioned, going to 72 V would allow the OP to use a high-voltage Satiator charger, which fits their programmability criteria.

tigcross said:

Hey there, I think that an unproven controller and 96 V is not a great idea for your first build, and not really necessary if your desired average speed is 40 km/h. Here is my reasoning:

1. Dry skin conduction –According to electrical safety standards, the safe limit for direct current (DC) voltage that humans can safely touch is generally considered to be 60 volts. 72 volts is risky, but pretty common (a fully charged 72 volt battery is 84V, 96 volts is getting pretty scary. Of course, we try not to touch wires, but it does happen, especially on first builds.
2. if you're running a Grin All-axle in a 26 inch wheel, 40km/h will be nowhere near the peak efficiency of the setup. at 40km/h you'll be at 73% efficiency. The peak would be 88% at 99km/h, but you're not likely to want to run that fast very often (but you wouldn't actually attain that speed as the power drops off significantly. View attachment 359762
   View attachment 359763

72 volts would be plenty for what you want, and if you run 72 Volts you can use a proven controller like the Grin Phaserunner. You can also then use the high power Grin Cycle Satiator charger which is awesome. I run a 72-volt system with 20-inch wheels (OD 21.7 inches, they are big tires) and my cruising speed is 55km/h, top speed is 70km/h

I get the desire to build something exotic, but I recommend that you start with proven stuff and voltages. A first time build can be fun, but full of frustrations. Sticking with a proven standard can at least alleviate some of the frustration.

Click to expand...

The efficiency will peak at around 40 km/h (25 mph) - 83.8%.




I completely agree with your other arguments. I would also advise against this build for 99% of people.

The reason I’m still justifying my decision as not completely idiotic is that I study electrical engineering and I hope that all the difficulties and challenges associated with it will benefit me in the future.

But as you mentioned, I also can’t resist the urge to build something exotic.

With a clear head, this project wouldn’t be worth it. At this point, I predict around 4,000 euros for all the hardware and tools – and, if we count my time as time I could have spent working instead, maybe another 8,000 euros or more.

I'd strongly recommend less voltage or a slower wind considering your speed target of 40km/h.

83.8% efficiency at that speed on partial throttle is poor efficiency. Some motors are getting 87% efficiency in that condition. I suggest selecting the motor type/size and voltage it's run on based on optimal efficiency at your target speed.

8000 euro?

I built a bike that has this motor..
52v 20ah pack from em3ev = $600 USD shipped
all axle motor = ~$700 USD shipped
VESC w/brake handle and controller = $225 USD shipped

That's 1371 euro

top speed on 52v in a 24" is 30mph/48kph.. which is very close to peak efficiency in a 24" wheel.

vladrov said:

The efficiency will peak at around 40 km/h (25 mph) - 83.8%.

View attachment 359799


I completely agree with your other arguments. I would also advise against this build for 99% of people.

The reason I’m still justifying my decision as not completely idiotic is that I study electrical engineering and I hope that all the difficulties and challenges associated with it will benefit me in the future.

But as you mentioned, I also can’t resist the urge to build something exotic.

With a clear head, this project wouldn’t be worth it. At this point, I predict around 4,000 euros for all the hardware and tools – and, if we count my time as time I could have spent working instead, maybe another 8,000 euros or more.

Click to expand...

Hey man, I get it.

I would still recommend starting with a simpler, cheaper build. A big part of my e-bike journey has been learning what it is that I actually want in terms of performance, cost, and complexity. I can tell you that my current bike is quite different than what I thought I wanted at the beginning.