2WD (Two-Wheel Drive) FAQ

[Kingfish]

Sangesf, is the single-battery comment directed towards my way? If so I’ll try to answer: I use arrangements of 5S1P to create battery assemblies. They are all the same make & model, I don’t mix chemistries, K.I.S.S. Thus, I can evaluate the “whole” battery assembly as one, having 2C charge, 15C discharge. But then I run with a large capacity and experience a voltage sag of 0.4V at WOT which I think is fine and good for two motors. (I hope that was helpful). And I’m with you on the observation: Pull easy and the batteries will be of less trouble. Mine get physically beat-up over time so I doubt I’ll ever see 1,000 cycles. Not complaining, just observing; we make the bed we sleep in.

Miuan: The next bike design is targeted at twice the speed and twice the distance. Allow me to shed a bit of light on what I have learned between the two cross-country trips I’ve made:

• P0 eBike was FWD, 9C 2806, 37V 10S9P. I had a crappy charger experience (issues documented and resolved in time for next year). Went 550 miles in 3.5 riding-days, longest was 173 miles with some opportunity charging (twice). Speed was about 24-26 mph (need to double check), and I could cruise at 29 mph on a flat easy enough. Me and the bike weighed 350 lb.s; batteries were 23 lb.s.
• P1 ebike is 2WD, Ebike-Kit (might be 2807-equivalent), 63V 15S26P. Had a better charger that used about 1kW/hour, and took roughly 8-9 hours to recharge my whole assembly. Farthest on a single charge was 165 miles of California Hwy-1 between San Rafael and Fort Bragg (hilly, windy, curvy). Average speed that day was closer to 23 mph (not worthy of bragging rights, but I got there), though I could easily cruise at 32 mph and could climb hills without any issues (other than what has been documented). I took 4.3 times more batteries to go about the same distance as last year, except that I had the pleasure and power of climbing hills without resorting to downshifting in most cases, and I towed a trailer with 50 pounds of additional gear on board (not counting batteries). Me and the bike: 450 lb.s, with 100 lb.s of that in batteries.
• Next bike: Proposed – is a 2WD electric motorcycle (someone coined eMoto and I like that term). Mathematically, the motors are designed to handle 75 mph, though realistically 65 mph is more fitting: Need to be able to travel on the freeway without getting run over. Minimum distance is 300 miles with average weather. Hence, twice the speed & twice the distance means 4X the batteries… optimistically. 5X would be better. Aside from weight, which is now 400-500 pounds, we have to find a way to charge it! The bike can’t come into the motel room like before; I have to go to RV Parks or EV Charging Stations (hey dnmun, you got yer ears on?). I spec’d out the “road trip” and 300 miles is barely making time for where I want to go. 300 miles / 60 mph = 5 hours in the seat. Now let’s flip this around and look at it from the other end of the telescope.

Most RV parks offer 30/35A service, and about 40% have 50A service. I need to consider 50A hookup, and here’s why: Let’s say I’m running 4X the batteries I ran this last summer. That’s 4 X 8kWh = 32kWh. My present charger uses about 1kW and is 88% efficient. This is plenty safe for motels. However the next pack will need an industrial-strength upgrade/replacement to keep up. In other words, the available charging infrastructure limits the practical size of the battery assembly.

Now we understand the limits: 8 hours (sleeping), 125 VAC, 50A service commonly available at 40% of the RV parks around the country (which BTW, vary from $11-35/day; much cheaper than motels – and the power is free). 125 * 50 = 6250W/hour MAX (likely theoretical at best), and 8 * 6250 = 50kW. Frankly, I’d be happy with 32 to 40 kWh if for nothing else but to keep the weight and the cost down.

Anyways, this is how I think about batteries and weight. Imagine now given all of this, that weight is not our friend, and that it is a drag (literally) on our system. How can we reduce that weight and improve our performance (aside from going slower)

Am I drifting off-topic? 2WD systems have a lot of potential beyond just bicycles.
~KF

[Alan B]

Just a note on continuous power ratings - 80% of rated - so 50A * 80%.

Note also that a lot of parks have separate 30A and 15-20A breakers and outlets, so you might set up your chargers that way. Again 80% rule applies.

[sangesf]

My observation was directed at anyone thinking of 2WD..
Your original FAQ was basically saying that 2WD is inherently ineffecient, when in reality, it really depends on the setup...
As an example...
Using a single motor, of let's say, 36v 30a max controller on a 36v 15Ah battery and then comparing it to two of those same motor/controller combos on the one battery/one throttle, and then saying it's inefficient is true...
However, if you took the same motors/controllers and hooked up two 36v 15Ah batteries and used two throttles, so that you can vary the load (re:torque) between the two motors and therefore needing less load (re:amps) on each motor respectively, your using less power (re:c-rate) to accomplish the same "pull" as a single battery setup.
Best way to explain is my experience.....
On my original bike, I had a 36v 20a controller and a 36v 15Ah battery... Best I was able to pull out of it was 21 miles at a 20mph avg speed. I then bought a second 36v 15Ah battery to double my range.. So 42 miles was about it... The VERY FIRST trip I made once I added a second motor and used the second battery in tandem, I was able to achieve 60 miles.. There was NO way I was gonna eke out an extra 10 miles from each battery separately on a single motor setup. The factors involved were these...
1.) Having two motors helps with original take off from a stop, not needing as many amps from each battery. (As opposed to a single motor setup)
2.) Keeping 20mph speed is a LOT easier on the 36v motors when there's two to help share the load.
3.) In relation, the entire system as a whole is just easier on the batteries.
(There's no reason to beat up your batteries, you MEANIE!)

As I said before, YMMV, but that has been my experience...

[Alan B]

KingFish, you never responded to my hypothesis on why your slave motor cut out under hard acceleration:

The current required for the throttle is so small I suspect something else was going on.

Ground Differentials on 2WD Controllers

My concern is the voltage drop between the two controller grounds. How much resistance was there between the controller PC board and the junction where the current split.

Not sure you want to discuss this in your FAQ thread, though it is on topic.

Take an example, assume:
[list=]
* 50 amps battery current to the primary controller
* 20 milli ohms from controller pcb to current split point
* 50 x 20 = 1000 milli volts of drop would then occur from the split to the controller ground.
[/list]

This 1 volt of drop would RAISE the ground potential of the throttle and the main controller compared to the junction point. This would also raise the +5 and the throttle output by the same amount.

Now this would increase the apparent throttle voltage to the secondary controller (that is not yet drawing much current). Which would tend to cause it to accelerate and 'catch up'. So on the surface it seems somewhat opposite what happened, and self correcting.

However, what if the throttle input to the secondary controller exceeded the max allowed input and caused it to go into 'shutdown'. Most controllers will reject a too-high throttle input value as a 'fault'. At WOT there is very little room to 'overdrive', so it would not take much voltage differential to cause it.

Feathering the throttle would allow the secondary controller more time to catch up at a lower throttle voltage before it was driven into cutoff. As it draws current the voltage difference between the controllers will be reduced as both of them will have a similar drop from the junction to the controller ground.

It is very important that the controller grounds be referenced together and have very low impedance between them if a common non isolated throttle is to be used. Or use an isolation amplifier to re-reference the throttle signal to the local ground on the slave controller.

Could this have happened to KF's setup?

[Kingfish]

Hi Alan

It’s possible that your hypothesis is valid; on the face of it, there is merit. We’d have to set up a way to test that in situ to be sure – and I would be happy to oblige if I knew how. EE is not my strong point; I am an E-M engineer, and I know a tiny wee bit about computers too: Although I could not register an opinion about your circuit, I could make fish animate across your screen

Regardless, like you - I feel strongly that the two controllers should be as close together as possible. I agree that every time there is a connector in line, the quality of the signal is at increasing risk. FWIW, I measured the throttle voltage between the two controllers and it was the same (and I repeated this test several times on several days), but then these were static tests without load; I couldn’t replicate the condition at the workbench. Perhaps with a dyno… (it’s on my Wish List).

The Sandman comith...
Z z z, KF

[llile]

This is a very useful thread. I am determined to build a dual front motor tadpole trike, for various reasons but also because I can

Obviously with dual matched front motors they'd have to be matched in speed and throttle and so on. Meh? If they don't match, I will just point the wheels a different way and go. It'll work even if it wears out one tire. I'd probably set it up so the batteries are amidships, right next to each other on either side of the frame, and make all the connections super short between them. Controllers also mounted close. With 'Bents you have a lot more room to put things where you want them, and since it is a custom frame, the options are endless.

Here are the questions I come away with:

1. One motor per battery, or two batteries in parallel? It seems it has been done both ways, which is best?

2. I'll be using a Cycle Analyst. Will I need two CAs, or one? ( I guess that depends on the answer about number of batteries..?) Two is more expensive, but that is still in the budget. Maybe two CAs check the health of both batteries? two CAs monitor individual motor power more closely, allowing fine tuning of match? I am still a little concerned about how to hook up the CA to this hack.

3. One throttle. Definitely. The controllers and batteries will be physically close, grounds will be heavy and short between the systems, throttle wires will be short and symmetrical. I guess if the throttle is powered from the cycle analyst there isn't a problem with the 5V power supply? If there is any interface electronics needed, can do. 5V Dc-Dc converrter if necessary to get the throttle working right. Instrumentation amp? Hell it is all 5V stuff, I might use a PIC or an Arduino instead of an instrumentation amp. I do that shit for fun anyway. But that seems overcomplicated. Can the throttle just be hooked parallel to two controllers or CAs?

4. I will probably be running at 72V. With a faring. I hope to build something with a 40 MPH speed on the flats. Don't tell the cops.

5. I ride a long way in rolling hill country. 2WD will be a good thing for those hills. Some are killer grades, short sections of 20% are encountered on most rides. 50 Mile round trip is also not uncommon. Battery requirements will be serious. I am thinking about two 30 AH batteries. Ping. Is that too much battery? This is 4X the battery I use now, at 20MPH same distance.

6. I'll still pedal. I love the feeling of exertion, I love being in shape, and it takes a good 1-2 hour workout to get there anymore. I hate sitting in a car, wishing I was getting some excercise. I'll probably cobble together a transmission that can still allow pedaling at 40MPH - maybe a 7 speed rear and a 3 speed intermediate hub. I've built crazier transmissions than that. Geared high, the pedals will be useless at normal non-Ebike speeds but who cares? I have never run out of juice before, due to good planning.

7. The other innovation I am seriously considering is a leaning frame. Front wheels don't lean, but the rear wheel, me, and the heavy batteries do lean. My friend has had his trike up on two wheels, doesn't sound fun. Leaning should improve cornering. But that is a whole 'nuther thread.

How about the 'lectric details, am I on the right track?

[hjns]

Hi Llile,

I do not have experience with 2WD yet, but I am building one, so I just wanted to share some of my thoughts. More experienced ES-members can correct me at will.

This is a very useful thread. I am determined to build a dual front motor tadpole trike, for various reasons but also because I can

Obviously with dual matched front motors they'd have to be matched in speed and throttle and so on. Meh? If they don't match, I will just point the wheels a different way and go. It'll work even if it wears out one tire. I'd probably set it up so the batteries are amidships, right next to each other on either side of the frame, and make all the connections super short between them. Controllers also mounted close. With 'Bents you have a lot more room to put things where you want them, and since it is a custom frame, the options are endless.

Here are the questions I come away with:

1. One motor per battery, or two batteries in parallel? It seems it has been done both ways, which is best?

Probably depends on the expected amps per motor and the c-rating of your batteries. As I use Lipo, I don't have to worry about the discharge rate, therefore, I find 1 large pack more convenient. Especially as I have rear- and front driven motors that are expected to draw different currents at the same time, using one pack leaves me with one worry less that there may be a difference in discharge between two packs. I guess the same counts for 2FWD, but less so, as the expected current should be more or less the same for both motors (unless you are driving circles or use different motors...)

2. I'll be using a Cycle Analyst. Will I need two CAs, or one? ( I guess that depends on the answer about number of batteries..?) Two is more expensive, but that is still in the budget. Maybe two CAs check the health of both batteries? two CAs monitor individual motor power more closely, allowing fine tuning of match? I am still a little concerned about how to hook up the CA to this hack.

I agree this is depending on the number of batteries, but more importantly, what do you want with it. If you use the CA only to monitor voltage, then there are much cheaper ways to do that. If you want to check the used Wh, then having two battery packs and two CAs means that you will have to do a lot of counting. It is one more reason to go with one battery pack, one stand alone CA with a separate shunt to measure current from the battery pack BEFORE the split to both controllers.


However, if you expect currents higher than 45A continuous, you may need a bigger shunt. I went with a 200A shunt from Ebay (expecting 100A per motor), which is on the high side.... better safe than sorry, though. To use it with my CA, I need to modify the CA (I have the DP) and recalibrate it for the new shunt. I am still working on that. If you get the stand alone CA, connecting the new shunt should be piece of cake.

3. One throttle. Definitely. The controllers and batteries will be physically close, grounds will be heavy and short between the systems, throttle wires will be short and symmetrical. I guess if the throttle is powered from the cycle analyst there isn't a problem with the 5V power supply? If there is any interface electronics needed, can do. 5V Dc-Dc converrter if necessary to get the throttle working right. Instrumentation amp? Hell it is all 5V stuff, I might use a PIC or an Arduino instead of an instrumentation amp. I do that shit for fun anyway. But that seems overcomplicated. Can the throttle just be hooked parallel to two controllers or CAs?

I am no electronics guy. Therefore, I went with the easy way out, as shown by ~Methods, and inserted another hall sensor into my throttle (see here). Took me 45 minutes. No high-voltage peak will mesh up my throttle signal.

4. I will probably be running at 72V. With a faring. I hope to build something with a 40 MPH speed on the flats. Don't tell the cops.

I guess that is 20S? Looks good to me, cause that is what I use now! On the flat, my HT3525 gave me 36mph. A faster motor (HS version) should give you more than 40 mph, but if you go that route, you will use more amps when hill climbing, increasing the risk of melted wires inside your motor.

5. I ride a long way in rolling hill country. 2WD will be a good thing for those hills. Some are killer grades, short sections of 20% are encountered on most rides. 50 Mile round trip is also not uncommon. Battery requirements will be serious. I am thinking about two 30 AH batteries. Ping. Is that too much battery? This is 4X the battery I use now, at 20MPH same distance.

I have no idea yet. Is also very much depending on which motors, max amps used, etc. You need to give us a bit more details about those, in order to give a better estimate. My rear HT3525 eats 20% hills at 20mph at 84V (20S lipo hot of charger) and between 20 and 30Amp. I guess, at a certain speed, you just need to 2kW energy to manage those 20% hills. However, with a consumption of 100Wh/mile, your mileage may vary.. Two 30 Ah batteries (72V?) sounds very heavy to me and gives you >4kWh. But with >4kWh you should be able to climb a 40 mile-long 20% hill at 100Wh/mile (assuming your motors don't melt). So, a 50 mile round trip sounds very manageable without hitting LVC.

6. I'll still pedal. I love the feeling of exertion, I love being in shape, and it takes a good 1-2 hour workout to get there anymore. I hate sitting in a car, wishing I was getting some excercise. I'll probably cobble together a transmission that can still allow pedaling at 40MPH - maybe a 7 speed rear and a 3 speed intermediate hub. I've built crazier transmissions than that. Geared high, the pedals will be useless at normal non-Ebike speeds but who cares? I have never run out of juice before, due to good planning.

Looking forwards to see that.

7. The other innovation I am seriously considering is a leaning frame. Front wheels don't lean, but the rear wheel, me, and the heavy batteries do lean. My friend has had his trike up on two wheels, doesn't sound fun. Leaning should improve cornering. But that is a whole 'nuther thread.

How about the 'lectric details, am I on the right track?

[llile]

I am no electronics guy. Therefore, I went with the easy way out, as shown by ~Methods, and inserted another hall sensor into my throttle (see here). Took me 45 minutes. No high-voltage peak will mesh up my throttle signal.

Oh so THATs what all this gibberish about throttle hall sensors is about. I never realized the hall sensor is how the throttles work. This sounds easy - two throttles in one. THe only problem is getting them both to make the same signal. It looks like yours were a bit mismatched, and with my dual motors integrated into the steering, I want a setup that matches as closely as possible.

If the throttle puts out 0-5V, it would still be easy to take a microcontroller and make it follow one signal input with two totally matched identical outputs. and also make sure it didn't mind being shorted out by the CA's throttle override. As long as there wasn't a ground bounce somewhere in the system that upset the signals, the two would follow. But I digress. If I decide to go this way I'll post all the details.

I'm still a little unclear how the Cycle Analyst interfaces with a standard motor controller, and how that motor controller interfaces with the various flavors of motors. Apparently some motors contain hall sensors as well, but I haven't quite figured out how the controller interfaces with them, nor how a non-sensor motor is different. Surely there is a FAQ around here that explains all the wires and nuts and bolts and signals. I've been looking into this but havent run across anything of sufficient detail yet. Can anybody point me to a good technical reference about this stuff? Once I get out of FAQ territory and into a build I will peel off a separate thread.

[hjns]

Maybe the CA manual is a good start?

[hjns]

It looks like yours were a bit mismatched, and with my dual motors integrated into the steering, I want a setup that matches as closely as possible.

Makes sense, but good luck. Alternatively, add some resistance, or just get two identical programmable controllers (Lyen comes to mind), and program them for matched speed to correct for any other interference.

[Alan B]

One bigger battery is better.

Matching throttle response is problematic and more important for trike. Grounds will bounce at these currents. Identical controllers will not have identical adc calibration.

Separate throttles could be fun for steering options. But tiring for long term.

[llile]

I'm still a little unclear how the Cycle Analyst interfaces with a standard motor controller, and how that motor controller interfaces with the various flavors of motors.

Here are some basic FAQ-worthy resources about this question:

One decent diagram of the signals coming out of a typical controller http://www.ebikes.ca/store/photos/C7225-NC.jpg I find this quite instructive.

What the hell does a HALL sensor actually do?: Here is a hint: Datasheet for some random hall sensor: http://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=&productId=1915915& Reading the data sheet, I find that, if the power supply to the Hall sensor is interrupted briefly (as from a ground bounce) then the resulting signal may be at a different voltage than the previous voltage for the same magnetic input. This may explain at a basic level why people have trouble with jerky behavior on 2WD setups and one hall sensor!

Of course, the CA manual as was just suggested: http://www.ebikes.ca/drainbrain/CA_Large_V223_Web.pdf. Paragraph 7.1 shows how the CA interfaces with the throttle signal. 9.2 has an interesting detail about on-the-fly speed limit, which might be useful if you hand your overpowered bike to a friend. Hit the speed limit button you hard wire into the controller, and they poke along. Chapter 11 lists various CA signals.

Method's original post on modding a throttle with two hall sensors: http://endless-sphere.com/forums/viewtopic.php?f=2&t=12356&p=220262&hilit=dual+throttle+methods#p222045

Here is a good technical paper by infineon about how a sensored brushless DC motor actually works. See Chapter 2 http://www.infineon.com/dgdl/ap1611710_speed_control_bldc_hall_xc164_davedrive.pdf?folderId=db3a304313719f4f011372b30f58008a&fileId=db3a3043139a1bac0113aacbc6ab02e6

[Willow]

...to chime in here...

My trike has dual Kelly KBL controllers - linked via the main negative as per the wiring diagram supplied by Kelly, and I'm using a Magura throttle.

There is a glitch (as Luke and others have pointed out) which favours one side. Nearing the end of full throttle - the left motor 'grunts' as if it is being over-run, or going into regen. The right motor gets a bit warmer than the left.

Currently installing a data logger to map exactly what is going on.... and in order to solve the issue, going to separate the controllers (both negs straight to the main shunt).. and build a device to balance the throttle signal between the two controllers. Will post the results in time.

Should have trike road registered this month.

S.

[llile]

...to chime in here...

My trike has dual Kelly KBL controllers - linked via the main negative as per the wiring diagram supplied by Kelly, and I'm using a Magura throttle.

There is a glitch (as Luke and others have pointed out) which favours one side. Nearing the end of full throttle - the left motor 'grunts' as if it is being over-run, or going into regen. The right motor gets a bit warmer than the left.

Currently installing a data logger to map exactly what is going on.... and in order to solve the issue, going to separate the controllers (both negs straight to the main shunt).. and build a device to balance the throttle signal between the two controllers. Will post the results in time.

Should have trike road registered this month.

S.

Wow that is fantastic! Keep us posted as you progress!

[llile]

Here is the definitive answer on using a Cycle analyst with two motors from Justin, who is a God amongst E-Bike Gods:


Hi Lawrence and thanks for the inquiry and nice to hear from a fellow addict!

[This seems like a hack - Lile]
It's fairly trivial to make a dual motor drive system operate with a single throttle, you just split the same thottle signal to go to each controller. However, you are correct then that the CA integration can be a little bit trickier. If both controllers and motors are identical, then you can in fact get away with using a direct plug-in CA on just one of the controllers, but set the RShunt value to 1/2 what it should be so that the current readings are all doubled. Then you are just assuming that each controller is drawing the same amount of power at any given time. For the over-ride limiting to work, you would connect the throttle over-ride pin from each of the controller CA-DP ports to each other.

[This seems like the right way to do it: -Lile]
The other option which is a little more accurate is to get the Stand Alone model of CA which will properly measure the combined current going into each controller, and then for the limiting features you hook up the CA as a current throttle as described in section 9.3 of the user manual. So throttle goes directly to the CA, and then the CA outputs its own throttle signal that goes to the two controllers.
[Hooray! The CA does the work of dividing the throttle signal, so no throttle-surgery is required! - Lile]


In either case you don't really need any additional circuitry or hacks, it's just a question of doing the appropriate parallel wiring connections on the existing lines.

Justin

[hjns]

Well, that is contradicted by the other God among E-bike Gods: Methods. See this thread about his experiences.

[Alan B]

Higher power levels create more noise. It doesn't take much noise into the hall signals or into the throttle to cause problems. Methods was pushing the envelope with a pair of 10KW controllers. So it is not surprising there would be more problems.

If you need good balance on two motors (and a 2WD bike doesn't really need great balance, but a trike does) you should consider matching motor currents. One throttle drives two feedback loops that are both controlling either battery current or phase current. Battery current is easier. You might want to add anti-spin logic to that which would not be hard to do, as applying equal power to the motors in slippery conditions can cause a hard spin on one if it loses traction, but the other will still get good power since the feedback loops are separate. For anti spin logic look at a hall signal from both motors and don't allow one to get too much faster than the other. Override the throttle on the faster one and reduce its power so it is only a little faster, until it regains traction and slows and then ramp power back up on that wheel.

With these large currents flowing the controllers will at times have different and varying ground potentials. Since the throttle signal is referenced to this ground it is important to compensate for this otherwise the throttle signals will be unequal and varying.

[Kingfish]

The simplest solution is to craft a PCB with two CPUs with accompanying driver channels: We just have to agree on the features and set about making one. However I like Justin’s thought on using the CA to manage both throttles; it seems promising and I’ve yet to try that.

BTW – I just want to say that 2WD and AWD systems are advanced topics for experienced users. I would strongly suggest noobs go build a single-wheel drive system first before venturing into this realm, because most of us here already know well the aspects of common components. In other words – if you’ve never built an ebike before, then this is the wrong thread to read and post your questions.

Thank you, KF

[sangesf]

I have to disagree slightly (or more importantly with an addendum)...

Dual wheel (awd) on a regular bicycle can be done by noobs easily...
Just use two throttles and controllers..
Done.

[llile]

The simplest solution is to craft a PCB with two CPUs with accompanying driver channels: We just have to agree on the features and set about making one. However I like Justin’s thought on using the CA to manage both throttles; it seems promising and I’ve yet to try that.

BTW – I just want to say that 2WD and AWD systems are advanced topics for experienced users. I would strongly suggest noobs go build a single-wheel drive system first before venturing into this realm, because most of us here already know well the aspects of common components. In other words – if you’ve never built an ebike before, then this is the wrong thread to read and post your questions.

Thank you, KF

Good point. I'd have to echo the warning. I've got a $150 soldering iron and a a multichannel oscilliscope on my hobby bench, (my fourth scope, earlier ones were inadequate), and have a box fulla several generations of microcontrollers and programming hardware, plus a shop with several types of welders that can fabricate damn near anything I can dream up. Most people are trying to bolt a motor on a mountain bike, I'm dreaming up new chassis configurations out of raw 4130 tubing and trying to push bicycle technology into something more like a small ultralight car. You gotta start somewhere, and I'm not trying to diss anyone personally, but I've been at this game of fabricating stuff for quite a long time, so I am looking for more and more extreme challenges like a drug addict looking for a bigger high. We were building our first recumbent bike frames, powered bikes, and three wheel vehicle chassis in the early 1980's. Some of those early experiments even worked, although many failed miserably and humorously. What's worse, this ain't no brag, because there are plenty of folks on this board who have fabricated stuff that I couldn't even attempt. Stay away from multimotor stuff until you've built something straightforward K.I.S.S. simple and made it work.

Ain't nuthin' more fun than building stuff that makes people go "Holy Crap!!?!! What the hell was that?" and that actually works.

[Lebowski]

The simplest solution is to craft a PCB with two CPUs with accompanying driver channels: We just have to agree on the features and set about making one. However I like Justin’s thought on using the CA to manage both throttles; it seems promising and I’ve yet to try that.

BTW – I just want to say that 2WD and AWD systems are advanced topics for experienced users. I would strongly suggest noobs go build a single-wheel drive system first before venturing into this realm, because most of us here already know well the aspects of common components. In other words – if you’ve never built an ebike before, then this is the wrong thread to read and post your questions.

Thank you, KF

Are you guys familiar with my 30F controller build (viewtopic.php?f=30&t=34231) ?

I'm building a full sinewave sensored/sensorless modified Field Oriented Control controller IC with everything configurable
to the user over RS232 (or USB with a USB<->RS232 cable )

The options I built in for throttle control are ideal for multi-motor setup.

In my setup multiple controller IC's can be linked together via a CAN bus. One controller IC can be configured as throttle master, the
rest as throttle slaves. The throttle master has 1 or 2 analog throttle inputs (can be hall or potentiometer based). 1 can be used as
throttle, the 2nd can be used as regen (or a 2nd throttle). The analog throttle voltages are both converted to a 16 bit signal in
the 0 to 1 range. These two 16 bit words are transmitted over CAN by the master to the slaves.

Now each controller, independent of whether its a master or slave, uses the throttle inputs to regulate the motor phase currents (torque).
This is done according to

Code: Select all

x1 = throttle_1 [0..1]
x2 = throttle_2 [0..1]

y1 = a1 * x1 + b1 * x1^2 + c1 * x1^3       (a1, b1 and c1 are in [-8 .. 8] range)
y2 = a2 * x2 + b2 * x2^2 + c2 * x2^3       (a2, b2 and c2 are in [-8 .. 8] range)

motor_phase_current = max_motor_phase_current * (y1 + y2)

Coefficients a, b and c can be used to configure all kinds of linear / exponentional type throttle curves.
Negative values can be used to obtain negative motor_phase_current which is basically regen.

The beauty in a multi-controller setup is that all controllers share the same throttle_1&2 values
but that each has its own a,b,c and max_motor_phase_current. So the maximum torque and throttle
respons can differ between the motors. You can have a 350W Mac in the front and a 500W Mac in the
rear, both operating on the same twist throttle and thumb regen 'throttle'. Then during motoring the
coefficients can be such that the powers to the motors match their ratings. For regen you can enable
(by using a2,b2,c2) regen strongly in the rear motor and weakly (or not with a2,b2,c2 all 0) in the front.

On the physical side, the CAN interface can be build to work with a shared ground supply or, if
you use opto couplers, you can have completely independent batteries for your motors...

[Kingfish]

Are you guys familiar with my 30F controller ...

Lebowski, kind Sir –

In a word: Absolutely! I have been an avid lurker on both your awesome thread and with Ricky_nz’s. I am just a little bit out of my league to comment in a meaningful way other than to applaud your progress. But flag me as a subscribed lurker! I personally wanted to wait until either of you gentlemen was a bit farther along with your developments and in conjunction with my prototype before requesting direction/advice/solicitations.

Retreating somewhat, the idea of this thread is to identify the common off-the-shelf components that can be utilized to create 2WD (and AWD), and their related issues. Yet at the same time I pine to create new hardware specifically designed for 2WD/AWD, including motors, controllers, and throttling systems with ancillary features. If we proceed down the development path in our conversations then this thread no longer becomes a FAQ, but instead a spelunking Build-Thread.

We could always start a collaborative thread to separate the two goals. I would like nothing better to have a state-of-the-art controller and AF motor (x2) and go blow the doors off the Chinese competition! (I wouldn’t expect it to be cheap to build… just accomplished in the right way with extensible features.)

Open to ideas.
~KF

[llile]

Are you guys familiar with my 30F controller build (http://www.endless-sphere.com/forums/vi ... 30&t=34231) ?

I'm building a full sinewave sensored/sensorless modified Field Oriented Control controller IC with everything configurable
to the user over RS232 (or USB with a USB<->RS232 cable )

The options I built in for throttle control are ideal for multi-motor setup..

Nice! If my simpler cruder methods of balancing the two motors fail, I'll be getting out the soldering iron and following along. How fun!

[Ian]

Many thanks for an interesting thread everyone. I would appreciate any feedback for my particular circumstances, as sometime this year I should be able to sink some more money into the bike, up to $1,000, and I don't want to waste the money as once its gone, that's it for a while! I currently run a rear 9c2810 on 45v to 67v. evg bike, 26" wheels. I tend to use 45v to 48v for long, slower rides, and go to 63v just for fun, rides under 40 miles. Amp limit I peg at 26amp whatever the voltage so 1600 watts tops. It's a rugged, reliable motor at these levels and handles the hilly areas in my area pretty well. My setup is pings for 48v 20ah and I use lipo packs to add to the voltage.

Anyways, my problem is not speed or range but a 1.25 mile, steep gravel driveway to my home, that starts at 2,300 ft and ends at 2,750ft. For about 3/4 mile the grades are real steep, a few sharp turns and ruts, making the going a bit tricky. it will be paved one day, but maybe not for a couple years...

So far, the solution I have put into practice was to get the 2810, and then once I had experience of its performance, I got a small dedicated 67v nano pack just for getting up the drive. The 9c on 67v does the 1.25mile without stalling but it is definitely at its limit heat wise - especially as the motor can be quite warm already at the bottom of the drive, so I'm not starting with a cold motor. Also, I am sometimes a bit tired after a long ride and would like to somehow evolve the setup so I can get back up with less effort. So I got the following ideas each with their pros and cons methinks. (If I had the money, I'd try all of them just for the fun of it, but that's not gonna happen!) I don't want to over-think all this but just kind of move to the most sensible-feeling option if you catch my drift...all batteries are in good shape, motor is good so far, so would like to keep what I got whenever possible. So, what's the best next step?

1. buy a 9c2812 and buy more batteries so as to go from 60v minimum for long rides, to 75v for getting up the hill. HMm...gets a bit heavy adding another 12v to the rig, I am already maxed out. replace some of the ping with lipo would work. And would the 2812 really make much of a difference. not sure.
2a. Keep setup the same but add on a small geared hub to the front - the mxcus from cell-man perhaps - using separate controller and 2nd throttle. Just use this to help with getting up the drive. Seems like a small, no drag hub, should give me 500 watt on the front wheel for that 1 mile and not have a heat or gear problem. BUT... got to lug the darn thing around tho the rest of the time...
2b. As 2a option, but switch the 2812 for a 2808 on the rear, for faster cruising and overall lower voltage, maybe a tad more efficient for most of the ride, and just figure to use the mxxus front hub more often on the hills, keeping it within its 350 watt rating.
3. go to permanent 2WD setup with one throttle/2 controller using 2 x cellman 500w macs with voltage around 40v. I could refigure my packs and add a couple more 4s bricks to get a permanent 39v 40ah. 22mph top speed, and run a peak 750 watts on each motor for the drive . How much noise would 2 simultaneous motors make tho...what kind of efficiency would I get at 20mph cruising, around 1ah per 2 miles?
4. Replace the 2810 with one of the new crystalyte low speed (HT) motors, keep my voltages as-is for 20mph to 25mph top speeds. But I can't find any real world reviews on the HT. It seems like a nice motor, a few pounds heavier than the 9c and a bit more cogging from what I can gather. So maybe just a tougher motor generally?

well that's it for the options - I think!!!!!

[Alan B]

Or get a Cromotor. Wait, this is a 2wd thread. Never mind.