Amberwolf's Lebowski Controllers project (For SB Cruiser)

[amberwolf]

EDIT: This thread's purpose has changed; it's now going to be about building a Lebowski controler that does what I described in the original title. (or rather, a pair of them).

They are to run the rear left and right hubmotors of the SB Cruiser trike:
https://endless-sphere.com/forums/viewtopic.php?f=2&t=67833&p=1250812#p1250808
(present configuration as of 12-25-16, with Yogi in foreground)

Presently these motors are an old Crystalyte X5304
https://endless-sphere.com/forums/viewtopic.php?f=4&t=64279
and an old much-abused 9C 2807, but the 9C will be replaced with an MXUS 3K 45H
https://endless-sphere.com/forums/viewtopic.php?f=9&t=84337
as soon as I can get the wheel swapped from one to the other.


The goal is a pair of controllers that can do:

-- reliable hauling of me and dogs and cargo; this *is* my transportation.

-- proportional braking (more lever pull = harder braking like mechanicals)

-- very hard braking (at least as hard as mechanicals can do), ideally so that it actively fights the wheels' rotation, not just slowing it.

-- braking down to as close to zero speed as possible, from as high a speed as possible, because I need to be able to stop the very heavy SB Cruiser trike (several hundred pounds with rider and dogs and/or load and/or trailer of cargo) in as short a space as possible. Legally the brakes have to be able to skid the wheels, and since I'm using Shinko SR241 "20" wheels on the pair of motor wheels, it's not that easy to skid them, especially with a load or dog back there. My front mechanical rim brake can be set to skid the front wheel, which is sufficient to meet the legal requirements, but I'd love to have the ability to electrically brake that hard, too, with both rear wheels. As heavy as the trike is, it could be critical someday.

--instant start, for as quick an acceleration as is possible from a stop, to get it up to speed (just under 20MPH, the legal limit here) as quickly as possible, regardless of load. If instant start impractical, then as fast a ramp-up as possible under whatever limitations I end up with.

--high power; exact level unknown. Present 12FET generics provide about 30-35A battery current each, using a 14s1p 20Ah EIG NMC pack, something around 1500-1600w peak on each wheel, IIRC. This presently gives several seconds from 0-20MPH, when unloaded. Is too long in traffic with some cars growing impatient as I accelerate. Not so bad once past 8-10MPH but not quick enough before that. Ideally less than 3 seconds 0-20MPH.

--programmable, to experiment with various settings, to eventaully optimize it for my rides.

(other requirements to be listed as I remember them or figure them out)



Whereswally offered PCBs and possible help sourcing components; and Lebowski the brain chips.

I'll accept any other components needed for it that anyone else has to offer, or if they're cheap enough for me to afford, but I should have many, perhaps all, of the passive parts already around in my stuff already. I also have FETs to test with, and probably make the full operational pair of them, as long as I don't need more than 100V types. Beyond that I'd have to save up to buy them.

I may also have FET driver chips in DIP form but they may not be the right pinout or functions for these PCBs
https://endless-sphere.com/forums/viewtopic.php?f=31&t=81990
Have to find them in my stuff and see which ones they are.




Links to other Lebowski-controller-related threads, will add as I find them (am also reading and rereading them all to learn what I can before diving in):
https://endless-sphere.com/forums/search.php?keywords=lebowski&terms=all&author=&sc=1&sf=titleonly&sk=t&sd=d&sr=topics&st=0&ch=300&t=0&submit=Search

https://endless-sphere.com/forums/viewtopic.php?f=30&t=36602
https://endless-sphere.com/forums/viewtopic.php?f=30&t=82160
https://endless-sphere.com/forums/viewtopic.php?f=31&t=81990
https://endless-sphere.com/forums/viewtopic.php?f=30&t=83788
https://endless-sphere.com/forums/viewtopic.php?f=10&t=80468
https://endless-sphere.com/forums/viewtopic.php?f=10&t=81681
https://endless-sphere.com/forums/viewtopic.php?f=30&t=81889
https://endless-sphere.com/forums/viewtopic.php?f=30&t=55641
https://endless-sphere.com/forums/viewtopic.php?f=30&t=63572
https://endless-sphere.com/forums/viewtopic.php?f=30&t=57877



Links to various sections of this build thread as I get to each stage will be added here:





Original post:
******************************************
Sorry for the long post, but I'm not totally sure what everyone calls each of the things I'm after, so thought it might be easier to describe what I need, and let readers decide if they know of products that meet my needs.


In the relatively near future, I may have a battery/etc upgrade to the SB Cruiser trike, so am now looking at new controllers for the dual rear hubmotors on it (presently one X5304 and one 9C 2807, in 20" wheels. Will be relacing the 2807's wheel onto an MXUS3k "45H 3T" after I find washers for my spokes to fit it, or redrill the flange, or get new spokes), for more equal torque on both sides.

The battery may be a 28s, so I'd need a controller capable of reliably handling 120V hot-off-charger voltages, meaning it would need components (caps, fets) actually rated significantly higher than that, so I don't have to worry about anything failing under whatever wierd circumstances that might come up.

The trike needs to be reliable while hauling heavy loads (potentially at least half a ton, to plan for the future, presently at least 600-700lbs, including myself and the trike and the dogs). So the controllers have to be reliable under those loads.

I don't know for sure what battery current it needs to be able to handle, but the 12FETs I am using now can take 60-80A between the pair of them, during the several seconds it takes to get from a stop to the max of 20MPH, and it takes too long right now, would prefer to get that down to 3 seconds, which might be some massive torque, and I don't know that the motors I have are capable of that, but I'm willing to test them out. I expect it'd probably take 80-100A battery current, at least, to do what I'm after, for those few seconds.


The controller also needs to have the "EABS" type of regen braking, that actively powers the wheel against the rotation direction, to force it to stop faster. Simple regen braking isn't really sufficient; I'd like to be able to stop quickly and completely with just the EABS (even though I also have mechanical brakes), and regen on any controller I've used so far is totally insufficient for this--but the EABS on the X5304's controller *would* do this if I had it on both rear wheels. I'm totally ok that this *uses* power rather than recovering any.

Because that is some massive braking torque, if what I get using that now with just one little controller and the X5304 is any indication, I'd also prefer to get proportional / variable-with-the-brake-lever braking force. I don't want one that changes the braking force with the throttle; it needs to be a separate control (I can easily build a brake lever that has an analog hall sensor to change levels). This is partly because I use independent throttles for the motors, but I will probably use a single brake lever for both to avoid brake steer for various reasons.


I'll probably need to tweak settings for various things, so it'd help if the controller is programmable, too.


I think the Grin Phaserunner would probably do what I want if it were only high enough voltage. (and not so expensive, and so complicated to setup).




As a bit of additional info, the trike itself is linked in my signature. The areas I ride are relatively flat, so the only times I typically need high current are during acceleration, and during short uphills into driveways/parkinglots.

Controllers get mounted next to the wheels under the cargo deck, so they get plenty of airflow for cooling and the phase wires stay pretty short.


If there are questions to clarify things; I'll copy the answers up here in the OP as well for later readers.

 

[eee291]

Good luck :wink:

 

[amberwolf]

Which one is that? I found this one:
https://endless-sphere.com/forums/viewtopic.php?f=31&t=75888
that says it "can potentially be upgraded to 150v" for one version and "up to 150v possible with upgraded caps", but not what that cost would be, or it's exact specifications, or how reliable it has been tested to be at these voltages.

But it does not have proportional braking--you have to program the braking strength. It also does not appear to have the "active braking" (EABS?) that actively fights the wheel rotation, just regen, which is not enough. It lists in one of the updates that it has "variable regen" but since no other details were added in that first post, I assume that this is only available via programming, which isn't useful to me; it would also have to be thru a separate variable input and NOT the throttle.

(I did look thru the rest of the 26-page thread, but must've missed it if it is stated that it has these things somewhere else--I found posts asking for some of it, just not stating that it is in there).

FWIW, I cant' use the BT stuff so its' a waste of technology; I can only do it thru serial/usb.

 

[whereswally606]

Cause its you Amberwolf I will send you a Lebowski board for free (if you want it), Hopefully others can send you other components you might need and the rest you can build yourself.

There are some 247 fets that will do 150v. PM me if you want to go ahead an I will send it gratis. Hope that others will chip in too. Would be great to have you on team Lebowski.

 

[amberwolf]

I definitely appreciate the offer; I've PM'd you with a more detailed response to find out if this would work out for this trike.

 

[whereswally606]

I pm'd Lebowski about the technical aspects and he said he was happy with me reposting the reply here:

So the bit about EABS is that something that the Lebowski controller can do?
I offer a free pcb (or two) in the hope others could help crowd source the rest since he has skills enough to build such a controller. Hope you can help answer the technical bits.

It used to have that (EABS) but Arlo1 had bad experiences with it with the car. When you are standing still and applying
EABS it has a tendency to shake (between two hall trigger points), so I changed it to where the regen ramps down as erpm drops...

I would have no problem giving him 2 chips for the controllers, but only if they really get build, not when they just end
up being unused... But it does not have the EABS het wants.

 

[amberwolf]

I wouldn't be using the EABS to hold the trike still, just to get it down to almost stopped, then the mechanicals do the rest. So perhaps I'll ask in the main Lebowski thread if the EABS could be enabled to do that; it wouldnt' have such problems that way.

(and figure out if the regen ramp down can be setup so it is almost not a ramp down at all; so it doens't ramp down until it's almost at a complete stop already)


If the EABS isn't in there, it may not have braking strong enough for the trike. I wish I had some way to measure the braking force I already have, and then I could post a number to say how much I need at a minimum, to know if a controller is known to be capable of that amount of force with motors similar to mine.

But I suppose I could just say that it needs to be able to cause braking force equal to mechanical brakes that aren't fully locking up the wheel (with Shinko SR241 tires), on a trike with it's load over those wheels, weight from 400-500lbs up to 600-700lbs+ to be stopped from 20MPH in as short a distance as possible (trike's own length would be good). I have front mechanicals too, but those can skid because most of the weight is on the rear (more is on front now after my rebuild but still most is on rear).



Still reading threads; I havent' yet determined even if the Lebowski controller can control the variable regen via an analog sensor on the brake lever/etc itself, rather than using the throttle to do it like all the other controllers with variable regen that I have found so far.


I did think of a way to "switch" throttle sensor over to one on a brake lever durign braking and back over ot throttle when the brake lever is released, but I don't know how well it will actually work or what side-effects/problems it might cause. (and it makes it more complicated, and thus one more failure point possibility).

 

[Lebowski]

With my controller you can set the erpm limit for braking at almost 0 and for it to ramp down almost instant.

The braking force you can expect is the same as the propulsion force, they both depend on the phase current.

There are 2 indepent analog voltage input throttles, they can independently be used for power and regen. So you can have regen strength be dependent on an analog voltage.

Two controllers for two motors can share throttle information digitally, only need to connect the analog throttle/brake to one controller and let it share info with the other. This can even be done with two different motors where you match motor torque by adjusting each controllers max phase current...

 

[amberwolf]

With my controller you can set the erpm limit for braking at almost 0 and for it to ramp down almost instant.

The braking force you can expect is the same as the propulsion force, they both depend on the phase current.

Ah; so if it's capable of breaking tires loose from pavement then it'd also be capable of locking up the wheels in braking.

I guess that's good enough, if I can build one capable of doing that without frying FETs. :lol:

Of course, I don't intend to use the motor drive ability to do that, but it'd be good to be able to brake almost to that point. In fact, if I *could* apply that kind of braking force, then I would have motor braking good enough to be legal (it's required that the brakes be able to skid the wheel, but there are no parameters to that).

There are 2 indepent analog voltage input throttles, they can independently be used for power and regen. So you can have regen strength be dependent on an analog voltage.

OH! That's what the other one is for (I was having trouble figuring that out cuz I haven't read far enough).

Two controllers for two motors can share throttle information digitally, only need to connect the analog throttle/brake to one controller and let it share info with the other. This can even be done with two different motors where you match motor torque by adjusting each controllers max phase current...

That's really interesting...and I'd love to do that, since I have two different motors, but...I kinda like being able to steer with the motors, using the left rear one to turn right, and the right rear one to turn left. (or at least, to help me make much sharper turns than I could using just the steering, at speed).

I don't suppose there's any way to set it up so that I could cruise using both motors in this "shared info" mode, but then pushbutton override that so each throttle is independent?

I'd guess that simply interrupting the communication bus's wire between the two wouldn't work, so unless it has the feature built in I'd guess it doesn't work that way. :/ And I can't imagine many people wanting this feature so I don't know that it'd be worth the effort of adding it if it isn't there already.

I'm thinking based on what I've read so far I might want to build and use these anyway; but I'll have to be sure I can build them up powerful enough to handle the braking needed, first.

Anyway....back to reading.

 

[whereswally606]

As soon as you know you want to go ahead, pm your address and we can make the Amberwolf-Lebowski trike a reality. I'm pretty excited about seeing this happen.

 

[Lebowski]

That's really interesting...and I'd love to do that, since I have two different motors, but...I kinda like being able to steer with the motors, using the left rear one to turn right, and the right rear one to turn left. (or at least, to help me make much sharper turns than I could using just the steering, at speed).

I don't suppose there's any way to set it up so that I could cruise using both motors in this "shared info" mode, but then pushbutton override that so each throttle is independent?

I think for a setup like that, better not to use the digital throttle sharing as you can do this easy in the analog domain... Have two throttles, one which sets the common throttle voltage
going to both controllers, and a differential throttle which raises the throttle voltage to one controller and reduces it to the other, dependent on the direction you want to go. You can build
an aircraft style joystick throttle, forward/backward for general propulsion and left/right for sending more power to left or right motor with analog circuitry you can even have it regen-brake
when you pull back on the throttle stick....

 

[amberwolf]

That's a thought I once considered for one of the first trikes I was pondering, as I had a complete powerchair drivetrain/controller/etc that I would've just transplanted over, but running the output shafts thru a bike derailer system on each wheel (linked shifters) to gear them up and down for torque or speed, rather than direct wheel on output shaft like they originally came. I don't know how well it would really have worked, but the design ended up too complicated with too many failure points especially for my lack of skills to build it at the time. (not that they are much better now).



For the present SB Cruiser trike, I don't think it would work with a joystick, as the roads are bumpy enough in the areas with the worst traffic that I think it would affect control too much, and potentially be unsafe with cars passing sometimes inches from me (usually they go around in the next lane, but not always, and it's much worse during the holiday season (late nov thru early january) when many don't pay any attention to anything other than the next sale they are heading towards, and are angry at anything or anyone that might possibly have the potential to impede them by a millisecond). I'd have to operate it with a thumb, because I still have to keep both hands on the tiller handlebars, and that limits the type/size/range of joystick I could use on it quite a bit, which amplifies the problematic inputs.


The differential throttle might work, though it'd take getting used to. I suppose it would also make it more difficult for a thief to drive it away, should that ever happen. :lol:

But if I use the two throttle inputs for the forward motion, then there's no input for the analog braking sensors?

Or would it use a throttle input on each controller for throttle, and still use the second input on each for braking?


For reference, the present operational setup is:

Left throttle controls left motor forward speed.

Left lower brake handle initiates both rear motor braking ("EABS", active braking that applies power to wheel to counteract forward motion very quickly).

Left upper brake handle initiates brake lights, and would also control rear cable operated rim brakes if I had them built/finished. (but with the EABS I don't really need them; the plan is to still build them anyway eventually).

Right throttle controls rigth motor forward speed.

Rigth brake handle controls cable-operated front rim brake (and eventually add a brake light switch to that too).


The same setup is used on CrazyBike2; where I learned to use them this way and it works well, so I kept the same system on SB Cruiser.

At first I'd intended to use a common throttle so both motors would respond the same, but since I didn't have a matched pair of controllers and motors, and didn't want to try figuring out how to build some analog thing to compensate for one of the motors to match the other thru the throttle range (to prevent torque steer since one was much more powerful than the other), I haven't ever tried it out on SBC.


I did try out a common throttle on CB2 at one point, but found that independent control of each wheel was important for traction control in various situations, from wet roads to sand/debris in corners, so I could power just the wheel that wasnt' slipping (or steering).

It's possible that true torque control of the motors, with a smart enough controller to detect whether the wheel actually has any traction and remove power if it doesn't, would work as well or better, but I haven't ever had the stuff to test it out that way.

 

[Lebowski]

I mean, you can always have two independent power throttles going to each independent controller, and have a common braking throttle.

With a current controlled throttle / controller it is much easier to match the torque of two different motors, just take one as a reference and increase/decrease the max phase current of the other to make the torque match the reference...

 

[Highfreak]

Hi all. I am currently collecting parts for a similar design bent delta trike. In the process I have considered that same left/right power application on turns and one of my ideas was using a potentiometer or variable resistor mechanically linked to the steering and wired in series with left and right throttle signals to controllers. It could also be adjustable to help slightly with variations in straight ahead torque steer. I have already collected 2 x qs205v3 50mm stator motors 2 x 19" mc rims, 2 x 20s 10p Pan.PF (still un-assembled new cells)to be wired in parallel. and ordered 2 x powervelocity controllers. BTW the Bluetooth programming module is an OPTION that you pay for. Good luck with solutions! I will be watching this thanks.

 

[amberwolf]

I mean, you can always have two independent power throttles going to each independent controller, and have a common braking throttle.

With a current controlled throttle / controller it is much easier to match the torque of two different motors, just take one as a reference and increase/decrease the max phase current of the other to make the torque match the reference...

I don't yet know how to do that, but I guess I"ll figure it out.

I am going to go ahead with this project; just have to gather up all the parts now. I've PM'd Whereswally606 to continue the PM chain about it too.

 

[whereswally606]

Managed to get the boards in the post today. Christmas has really put a spanner in the works for shipping. Good luck AW.

Sent from my ALE-L02 using Tapatalk

 

[amberwolf]

Thanks! No worries on the wait; I have to gather up all the parts before they can be built, anyway.

 

[amberwolf]

PCBs arrived today just before I left for work.


.

 

[whereswally606]

Excellent I was thinking they should arrive soon. Merry Christmas AW and a happy New year.

Sent from my ALE-L02 using Tapatalk

 

[amberwolf]

I'm still going thru my parts as I find them for these, but I should have enough stuff to start putting them together "soon". Maybe by this time next year (at the rate I'm going) I'll be at the point of talking to Lebowski about the chips themselves.

I'm hoping I'll get a weekend to actually start stuffing and soldering passive parts at least.

 

[whereswally606]

Hey AW, good news you are back on this. I had BobC visit this weekend and help me construct another of my side projects. Ryan Zellar's (allted's) MPCNC mostly printed cnc http://www.viciousone.com. I am nearly done on that and should be able to route pcb, wood, acrylic and even aluminium soon so that's quite exciting.

The good news from Bob is that he has the Lebowski controller working with hall sensors on the emax 110s 2008 year hub. This is a major milestone in our project as now I should be able to get out emax scooters travelling at 50mph (maybe faster). I hope that we might make more iterations of it combining work from animalector (Andy in Oz) and his mini smd logic board and Bobc's MTI 6pack power board. Hopefully I can contribute milling some fancy scmancy boxes for the project and the custom PCBs in heavy copper.

Anyway I stop littering your build thread now. Good luck AmberWolf