High-speed, long-distance (heavy "30^3") Q128 commuter build

One point to consider is which of these motors is the easier to maintain or are they the same? Given the high mileage I suspect you'll put on these and that they use nylon gears, planned maintenance may be something to consider. For instance, simply changing the nylon gears out at set intervals before you have problems may be a good idea. That, or at least opening up the motor at such a time and evaluating the wear. Parts availability and cost may factor in as well.

There are 2 things that will effect the service life of a geared hub motor used on the street.
1)Over Volting, or more correctly, over-powering. A BPM will take more than a mini and both the MAC and BMC will take more than the BPM. Most experienced geared hub motor users have a good idea what "is safe" for a particular motor and there is plenty of anecdotal evidence documented here for those who care to search.
2) the other thing that can effect the motor is water, or more specificly, water w/ chemicals, like salt, in it. Winter riding can be very detrimental. The more expensive motors, like the Ezee and the BMC have dedicated seals on the axle, but even so, there will be some water ingress when riding in extreme wet conditions. Pure water, in itself, probably is not that harmful, for it will find it's way out, either by draining or evaporation. But dirty water will leave residue which can cause corrosion. Winter riders are advised to dis-assemble their motors and clean the insides and lightly oil in the Spring.

Geared hub motors that are run on reasonable power and dry conditions will last a long time. My first motor was a MXUS geared mini and although it was abused somewhat power-wise, it lasted many years and 1,000's of miles. What eventually did happen to it, was it wore out the axle bearings. They could have been replaced, but overall, it was tired and I gave it away. Last I heard, the kid I gave it to is still using it, noise and all.

Most geared motors have gears/cutch ASM's available. All the large(5 Kg.) do and most Bafangs.
The Q100 has gears available, but they are the high-speed (328) gears only.
I don't think parts are available for the Q128, but like the minis, it is inexpensive to replace the whole motor.

wturber said:

One point to consider is which of these motors is the easier to maintain or are they the same? Given the high mileage I suspect you'll put on these and that they use nylon gears, planned maintenance may be something to consider. For instance, simply changing the nylon gears out at set intervals before you have problems may be a good idea. That, or at least opening up the motor at such a time and evaluating the wear. Parts availability and cost may factor in as well.

Click to expand...

That's an excellent point, wturber. And thanks, motomech, for providing some insight into it.

My plan to, at least partially, address the all-weather Virginia riding is to enamel coat the stator (right name?) when the hub first comes, and then to see if I can run it in a small bath of ATF, as some folks have discussed. As I mentioned earlier, I do plan to buy two of whichever motor so that I have a hot spare.

I’ve noticed your plots have you riding at speeds to the left of the inflection point. For most people, this is a BAD thing. All out of box solutions will be to the right of this. However, it works for me and probably will work for you.

Some background.

Inflection point. On the red line (power output) the inflection point is when the controller hits the battery current limit, and current (and power) rapidly falls off after this. All out of the box motors cruise at speeds to the right of this point. This means that the controller is not maxed out (at maximum amperage), gives the highest speed for the motor, and generally prevents things from overheating. Your design and mine don’t operate this way without pedaling. Thus, they are not great designs for riding around in motor only mode (less torque, less efficiency, more potential for overheating), but it has worked great for me. Caveat: the controller does have to have good ventilation – a heat sink helps too (e.g. don’t put it in a bag and expect it to survive).

Speed with pedaling. My bike doesn’t have enough current to get to its happy point (right of inflection point). It does about 20mph on its own, maybe 22 down wind in a tuck. I do about 25mph pedaling it. So, by adding 200+watt with my legs makes a big difference.
My pedaling alone: ~20mph
Motor alone: ~20mph
Motor and my pedaling 25mph
Noload speed of motor 29mph.
Inflection point of motor 23mph (although it is does not have the power to get here on its own)
Pedaling makes a difference because the inflection point on my motor is 23mph – making it relatively easy to get from 20 to 23mph and a bit harder to go faster. And it keeps me in the safe zone to the right of the inflection point when I add 200watts. Ideally this is where you want to be.

So, for all off the shelf motors (except for very small ones) its near impossible to pedal hard enough to make the bike go faster. They cruise at speeds to the right of the inflection point. Here power falls off so rapidly, while the aero effects raise so rapidly that no one can make much of a difference pedaling (on a higher powered motor at full power). But with a moderate powered motor running solo to the left of the inflection point, pedaling can make a big difference (if you can live with the drawbacks).

Good point. Controllers are generally rated at half the maximum current, so a 20 amp controller will run happily at 10A, but if you run it at 20A continuously, it'll over-heat. The current that it delivers depends on the motor speed, where the back emf prevents the full current when above the inflection point.

The KT controllers use current control, where the 5 PAS levels range from 10% to 100% of maxpower, so you can also prevent over-heating by using level 4, but of course you only get 12 amps instead of 20. If you're going to run a high kV motor, you need to consider how to help the controller stay cool.

Boy, you guys are really helping me to unwrap the onion.

First, I'd like to explore what chas58 said that his design doesn't do:

chas58 said:

All out of box solutions will be to the right of [the inflection point].

Inflection point. On the red line (power output) the inflection point is when the controller hits the battery current limit, and current (and power) rapidly falls off after this. All out of the box motors cruise at speeds to the right of this point. This means that the controller is not maxed out (at maximum amperage), gives the highest speed for the motor, and generally prevents things from overheating.

So, for all off the shelf motors (except for very small ones) its near impossible to pedal hard enough to make the bike go faster. They cruise at speeds to the right of the inflection
point. Here power falls off so rapidly, while the aero effects raise so rapidly that no one can make much of a difference pedaling (on a higher powered motor at full power).

Click to expand...

The A/B variable in the following graphs is whether or not 200W of pedaling is applied. The only variable change between the graphs is that the current is ramped up in whole numbers from 15A-19A as the operating range approaches and then goes past the inflection point. You can see that:

• peddling does not increase speed, but instead
• peddling does reduce power consumption

Now, here's a bit on the operating range where chas58 suggests that pedaling makes a difference in speed:

chas58 said:

My bike doesn’t have enough current to get to its happy point (right of inflection point). It does about 20mph on its own, maybe 22 down wind in a tuck. I do about 25mph pedaling it. So, by adding 200+watt with my legs makes a big difference.

Pedaling makes a difference because the inflection point on my motor is 23mph – making it relatively easy to get from 20 to 23mph and a bit harder to go faster. And it keeps me in the safe zone to the right of the inflection point when I add 200watts. Ideally this is where you want to be.

But with a moderate powered motor running solo to the left of the inflection point, pedaling can make a big difference (if you can live with the drawbacks).

Click to expand...

Again, the A/B variable in these graphs is whether not 200W of pedaling is applied, first with an example in the flats and the second on a 3% climb. You can see that:
peddling does increase speed by a few mph, and this time
[*]peddling does not reduce power consumption

Clearly, though, there are sometimes when I would like to use more current even while operating to the left of the inflection point. Hills for example:

Here, the A/B variable is 10.2A/20A of current (~500W vs 1000W), and then changing from a 5% to a 10% grade.

It would be really great if the motor could, for short spurts, help me "flatten" the hills. Also, since I'm fully aware that I need to maintain a minimum speed of half the unloaded speed, so more current could help me keep me in safer territory in the face of moderate climb combined with a blistering headwind.

I've decided to go with the Bafang CST.

Even though I'm, frankly, enamored with the lighter weight of the Q128C, the feedback here along the modeling have convinced me the ~4 lbs. lesser mass of the Q128C pales in comparison with 10% better efficiency. (~78% for Q128C vs 84% for the Bafang CST) When running 500W into both motors but with the Q128C at 95kg and the Bafang CST at 100kg, the Bafang CST is still appreciably faster, whether in the flats or climbing; the climbing part really surprised me.

d8veh said:

Good point. Controllers are generally rated at half the maximum current, so a 20 amp controller will run happily at 10A, but if you run it at 20A continuously, it'll over-heat. The current that it delivers depends on the motor speed, where the back emf prevents the full current when above the inflection point.

The KT controllers use current control, where the 5 PAS levels range from 10% to 100% of maxpower, so you can also prevent over-heating by using level 4, but of course you only get 12 amps instead of 20. If you're going to run a high kV motor, you need to consider how to help the controller stay cool.

Click to expand...

Now that I've settled on the Bafang CST and have a much better understanding of the operating regime (thanks!), I'd like to figure out which controller to use and how to use it.

I really don't understand how the interface or controllers work, and my questions will probably betray that.

I'm don't understand how the user selects what level of current for the controller to pass. Will I need to toggle between PAS4 in the flats and PAS5 for the climbs? Or can I change the controller programming so that the (5) PAS levels are in a more suitable range for my use (like I think I understand that folks have done with the BBS02 controller)?

Would the SO9P, S12P or S12S controllers that BMSbattery offers be a better fit for the Bafang CST than the KT?

Or is the sort of current control that I'm after even possible in that price range? Am I after capabilities only found in svelt controllers like the Phaserunner?

Thanks in advance. You all have been a world of help.

S09P and S12P are square wave controllers for sensorless motors. That only leaves the S12S . The PAS levels are current limits, so you select whatever power you want. Bear in mind that the S12S is quite large.

The poblem with the Bafang CST is to get one with the right speed. Make sure you know what speed you're buying because they don't normally tell you and there's a lot of different variants. Also the 250w, 350w and 500w versions are actually different, even though they look the same.

d8veh said:

The poblem with the Bafang CST is to get one with the right speed. Make sure you know what speed you're buying because they don't normally tell you and there's a lot of different variants. Also the 250w, 350w and 500w versions are actually different, even though they look the same.

Click to expand...

Oh, bother, that's confusing!

I tried to approximate no-load speed in the simulator with a 1kg bike/rider and 0.0001 CdA, and the Bafang CST ran at 333 RPM.

I'll contact bmsbattery to figure out which motors they have. Now that I look more closely, I see that the spec listed says 270rpm@48V, but then the selections for purchase list 201rpm and 328rpm.

Thanks for catching that.

d8veh said:

S09P and S12P are square wave controllers for sensorless motors. That only leaves the S12S . The PAS levels are current limits, so you select whatever power you want. Bear in mind that the S12S is quite large.

Click to expand...

Assuming that I can indeed source the correct shred, would the KT controller work ok for the Bafang CST, instead?

The 36v 500w Bafang CST that BMSB used to sell was around 260 rpm and had a speed of around 22 mph on the road in a 26" wheel, so 48v would get you around 28 mph and a 700c wheel would be 30 mph. 14S would give you a bit in reserve. The problem that you now have is a relatively heavy motor, a large controller and you'll need a sizeable battery to power them. Your 30 pound target is now just a dream. With the additional weight of those items, you'll need a strong frame and wider tyres. I think you might have crossed over the point when you might as well get a DD motor, strong MTB, big battery, etc like all the others.

I still following this and I think what are seeing are folks voicing the difficulties and compromises needed to take a lite weight road bike into the 30's mph range, something I believe I brought up in earlier posts.
And where we are at now, i think deserves the consideration of one more compromise.
Assuming the motor is going to be the BPM CST doesn't necessarily mean a big battery, I would say that is more a function of the controller.
I run an Ezee V1 w/ an Elifebikes QQ 20 Amp controller that has been shunt modd'ed to 25 A. That's a 25 A peak only at start-up and in normal usage is more like a 17 A controller. It's kind of soft and if my bike was not 2WD, I probably would want a more powerful unit. But it's smooth, easy to modulate the throttle and provides good batt. economy.
So my proposal is, why not use the KT sine wave shunt modd'ed to 25 A on the BPM?
Yes, it would be under-powering the Bafang, given the conventional wisdom is the BPM should be paired w/ a 30 to 40 Amp controller. But, the Ebike CA sim. shows it would still get the motor up to the desired top speed w/out becoming Current limited., But the compromise comes to the fore as lazy aceleration to get there.
Given the OP wants to use his legs, pedaling to get up to speed would give him the opportunity, as I'm not sure how pedaling @ 30 mph will work out.
One possible cavet w/. this plan is over-heating the controller, so it deffinately should be mounted somewhere where it gets good air-flow.
Just a thought, if it doesn't work out, the controller is the easiest thing to change out in an existing system.

d8veh said:

Your 30 pound target is now just a dream. With the additional weight of those items, you'll need a strong frame and wider tyres. I think you might have crossed over the point when you might as well get a DD motor, strong MTB, big battery, etc like all the others.

Click to expand...

I knew that I would never, ever weigh in at 30 lbs., wihch is why I called it a "heavy 30^3."

Big battery? Check! 2kWh's, with a planned consumption of 500kWh each leg (charging at work) to maximize overall life.

I'm taking your criticism of my frame selection seriously, and I'm looking into loaded touring folks do, since that's somewhat corollary.

The problem problem with selecting a mountain bike frame is the change in geometry. I'm don't want to do direct drive because I want to freewheel.

But I appreciate the points you've raised and I'm taking them seriously.

motomech said:

I run an Ezee V1 w/ an Elifebikes QQ 20 Amp controller that has been shunt modd'ed to 25 A. That's a 25 A peak only at start-up and in normal usage is more like a 17 A controller. It's kind of soft and if my bike was not 2WD, I probably would want a more powerful unit. But it's smooth, easy to modulate the throttle and provides good batt. economy.
So my proposal is, why not use the KT sine wave shunt modd'ed to 25 A on the BPM?
Yes, it would be under-powering the Bafang, given the conventional wisdom is the BPM should be paired w/ a 30 to 40 Amp controller. But, the Ebike CA sim. shows it would still get the motor up to the desired top speed w/out becoming Current limited.

Click to expand...

Are you saying to add solder to the shunt to decrease its resistance so that it'll handle more than 20A?

I'm just plain confused as to how to control current, other than selecting PAS 1-5, as I mentioned earlier.
--Occasionally, for going up a hill, I'd like to be able to dump 20A/1000W.
--Otherwise, I'd like to restrict the current to around 10A/500W

motomech said:

But the compromise comes to the fore as lazy aceleration to get there.

Click to expand...

I'm willing to accept this.

My QQ 20 A controller uses two shunts. What I do is push the two shunts together and coat about 2/3's the length with solder(lightly). The result is, the peak Amps go from 20A to 25A(measured). Adding more solder will not increase that.

W/ the sine wave controller, all aux. functions are based on Current control. Changing the PAS speeds for example, limits the Current more and more, as the limit speeds decrease.

A regular controller, like my QQ, limits the PAS speeds by cutting the Voltage. As system power goes up, there is a surge when the limit speed is increased, the system wants to "rush" up to that speed.
Other less than desirable traits include a delay to activate and deactivate, on my bike 2 to 3 seconds. It can be annoying.
And lastly, with only 3 speeds, there can be gaps on a system that has a broader speed range.
On my 1000 Watt Ezee, that does 23-24 mph, I am right at the limit where I find the simple 3-speed controller acceptable.
I tried several sive wave controllers(S6OP & S12S) when they first became available, but had some issuses w/ the auto-ranging "seeing" the odd Voltages of my LiPoly. The "fuel gauge" was off, but more importantly, the LVC was erratic(super important w/ LiPoly). The new units may be more forgiving, but as I have pointed out, you may be on your own trying to run much more than nominal 48 Volts on a sine wave.

Keep in mind that the ebike.ca simulator tested the Q100 a long long time ago. There have been changes to it since then, and the efficiency numbers may have significantly. We know the Q100H and Q128 are a lot different then they were back in those days.

For controllers: "torque simulating" controllers modify the current, the older design controllers are roughly like a throttle to limit power (i.e. speed limiting, but often use full current below max speed).

I don't think 30-40 amps is what you want. I'm happy with a 12 amp controller as it more simulates me riding a bike - or more specifically riding a tandem with my brother. at 12amps I'm clipped in, and accelerate out of the saddle. It definitely encourages me to use my legs and not get lazy!

Personally I think a gravel bike is the best solution 32-45mm tires, (long, low, slack geometry) with alternates being a cross bike (more agile) or a touring bike. Gravel/Cross are built rather like a cross country mountain bike, but with more road geometry.

Rather than soldering the shunts, I would rather get a controller with the current I need, and then dial it back with the PAS settings. But again, your graphs showed 10 amp usage. If you are willing to sweat, you don't need a lot of current. The current is mostly going to get you low end torque anyway. Do you need that? The high current does not help you at all with speed - if you are traditionally to the right of the inflection point.

chas58 said:

Keep in mind that the ebike.ca simulator tested the Q100 a long long time ago. There have been changes to it since then, and the efficiency numbers may have significantly. We know the Q100H and Q128 are a lot different then they were back in those days.

Click to expand...

I know that they have, and I know that the Q128C is an improvement over the Q100, but, at least according to the model, the Bafang CST is even more efficient than the eZee or MAC, and they're both more efficient than the Q100. The efficiency gains of the Bafang CST over the Q100 are worth more than even an extra ten pounds. I wish that I had current data now, but I don't.

I'm game to pitch in if folks would like to buy a Q128C (328rpm?) for ebikes.ca to characterize.

chas58 said:

For controllers: "torque simulating" controllers modify the current, the older design controllers are roughly like a throttle to limit power (i.e. speed limiting, but often use full current below max speed).

Click to expand...

So of pswpower's torque simulating controllers, they have a 17A/350W and a 20A/500W spec'd at 36V/48V, and they're listed with both the same dimensions and weight. Any reason not to get the 500W? (Efficiency penalty?) They would work with either the Q128C or the Bafang CST, right?

(Their 12A & 15A are listed as 24V/36V.)

chas58 said:

Rather than soldering the shunts, I would rather get a controller with the current I need, and then dial it back with the PAS settings. But again, your graphs showed 10 amp usage. If you are willing to sweat, you don't need a lot of current. The current is mostly going to get you low end torque anyway. Do you need that? The high current does not help you at all with speed - if you are traditionally to the right of the inflection point.

Click to expand...

Translating the batteries that I'm carrying into a "power budget," I can only afford to average 750W, tops, so I can't afford to 20A+ constantly. I plan on running 10A with up to 20A on steeper or longer ascents.

chas58 said:

Personally I think a gravel bike is the best solution 32-45mm tires, (long, low, slack geometry) with alternates being a cross bike (more agile) or a touring bike. Gravel/Cross are built rather like a cross country mountain bike, but with more road geometry.

Click to expand...

The huge FalconEV triangle bag came in, and even though all the batteries fit in the bag, they don't fit on the bike, so the CAAD4's out of the running.

I agree with your general assessment of frame types. I went to look at the only seemingly good candidate on Craigslist on Sunday afternoon, but it was no good. Wrong time of year to be shopping.

In the meantime, I think that I'm going to repurpose a ~1993 Giordana with a 3/2.5 titanium (Litespeed) frame and an aluminum fork. It has enough clearance to fit a 28mm tire both front and back. I don't want to read too much into this, but it does have a nipple on the headtube for a frame pump and mounting eyelets above the rear axle for a rack or mudguards; still, its very much a racing bike. I'll keep my eye out for a better candidate.

motomech said:

Would it be too bold if I inquired what your profession is?

Click to expand...

Bottom line;
I have read here something along the lines of, The BPM (Bafang) is good up to 30 mph and above that, it's best to go w/ the MAC and believe this is about right.
26 to 28 mph avg.-Q128
27 to 30 mph avg.-BPM
30 mph and over- High-speed MAC

Click to expand...

I agree with most of what you wrote above, and I think it's all good general advice. There's more considerations than the motor, though. A hybrid or road bike with thin wheels and forks is much easier to propel, once the speed goes over 20 mph. It's still an option to use the lighter motors with such bikes to get your high speed, but when you get to around 30 mph, pedal effort becomes less significant compared with the motor's power. Don't forget that some guys can average 30 mph without a motor at all, though for that, you need low resistance clothing as well as a low resistance bike.

Another point for motor choice: You can use 8-spd DNP freewheels, but that makes the motor wide on one side. Normally, you have to compensate for that by making a substantial dish in the wheel, which means that the spoke tension on the brake side is on the low side. That can cause many problems, especially if you're a heavy rider. I think that it's better to use the more expensive cassette motors if you want to go over 7-spd. MAC, Ezee, Bafang and many others all have cassette motors now.

chas58 said:

For controllers: "torque simulating" controllers modify the current, the older design controllers are roughly like a throttle to limit power (i.e. speed limiting, but often use full current below max speed).

Click to expand...

So of pswpower's torque simulating controllers, they have a 17A/350W and a 20A/500W spec'd at 36V/48V, and they're listed with both the same dimensions and weight. Any reason not to get the 500W? (Efficiency penalty?) They would work with either the Q128C or the Bafang CST, right?

Click to expand...

Help!?

I’m wanting to place the order for everything, but I just noticed that PSW’s torque simulator KT controllers are all square wave (linked above in the quote). Do I want that?? I thought that I wanted sine wave...

Thanks.

Another benefit of a gravel bike. They tend to have large triangles specifically to fit frame bags like that (for bikepacking). A bike from the '90s (before compact frame geometry) has lots of triangle room.

Yeah, I have the classic 1993 specialized Allez racing bike (the one in the movie of that year), and even it has the mounts for the frame pump!

I use an older square wave. At 36v it is quite silent. People that use higher power tend to like sine waves because they are quieter.

the advantage of a 350watt vs 500watt controller is generally that the 350 watt won't burn out a motor as easily. remember, most people are going to plug it in and forget about it. Purchase the one based off the max current you will need. And of course, the lower current will encourage you to pedal more - I just get lazy with more current.

That and of course the max current ratings only give non high speed torque. It does not matter riding to the left of the inflection point. If you are riding to the right of the inflection point (for a 15 amp controller) a 15 amp and a 30amp controller give the same power and current.

mystryda said:

chas58 said:

For controllers: "torque simulating" controllers modify the current, the older design controllers are roughly like a throttle to limit power (i.e. speed limiting, but often use full current below max speed).

Click to expand...

So of pswpower's torque simulating controllers, they have a 17A/350W and a 20A/500W spec'd at 36V/48V, and they're listed with both the same dimensions and weight. Any reason not to get the 500W? (Efficiency penalty?) They would work with either the Q128C or the Bafang CST, right?

Click to expand...

Help!?

I’m wanting to place the order for everything, but I just noticed that PSW’s torque simulator KT controllers are all square wave (linked above in the quote). Do I want that?? I thought that I wanted sine wave...

Thanks.

Click to expand...

These are the Sine wave ones. Torque simulation = current control, which they all have. There's a setting in the LCD if you want speed control, though I can't believe that many would want to use it.