Roboteq HBL-2360 + Invacare Brushless motors

[Burgerman]

Torque mode = no speed control...
Closed loop in, any mode, feels horrible on any powerchair and is only really useful if you are driving with strange controls or non proportional control. So it can track straight without correction. Works like a RC servo. It gives no feedback or feel.

You don't want to use it. Use open loop.
Motor compensation algo in the script that Lenny included does much the same thing but leaves the feedback we need for control "feel" during control or driving.

[LROBBINS]

Closed loop speed mode will maintain constant speed and allow torque to vary. Torque mode keeps torque constant, allowing speed to vary - you definitely DON'T want this on a wheelchair, except perhaps for (some day I might try this) a truly active anti-tip that responded to CG shifts by applying only enough force to keep you from tipping. Ciao, Lenny

[Williamclark77]

The new Roboteq FBL2360 advertises compatibility with trapezoidal and sinusoidal switching as well as hall, encoder, and sensorless rotor sensing. It's a max 2x 60a output though. Will dual 60 amp output be enough for a powerchair?

I will read Irving's link as soon as time allows. The hbl2360 Roboteq I use is "only" dual 75 amp, but when paired with the low resistance motors I'm using is more powerful than any 24v powerchair I've used. And, that's geared for 14ish mph. I'm sure they're weaker than BM's BM3 43v brushed setup, but is definitely powerful enough for my Willchair with me and my 50 lb younggun in it. As is, my 14mph gearing would be too high for a fat user or heavy chair. Or feels like it would be too high rather for smooth movement at very low input.

[Burgerman]

Remember that your running 48v too though. Equal to having motors that are half the resistance that your already low impedance motors are.

[Williamclark77]

And also, how does it compare torque wise if the brushless motors are ran in delta or wye configuration? I know the way each configuration is energized, but not how that affects the current carrying or torque ability at low input values. For a comparison of exact motors anyway. I know technically that wye pulls more current and produces more torque, whereas delta produces more RPMs, but have never seen an apples to apples comparison of one motor on the dyno in each configuration.

My motors can "easily" be ran in either configuration with some wire swapping and rotating the pcb inside the housing. "Easily" done for testing on a bench that is. Not so easy when on the chair. I have mine in wye configuration and geared for it. If it wasn't a major task (and would be nearly $1000 for different planetary gearboxes) I would try mine in delta and geared to give the same wheel speed as wye to compare. I wonder if delta would give equal amounts of torque with the gearing compensated lower for the increase in motor RPMs? I've never seen any testing of this.

I hooked mine up in wye because that's what they recommended to maximize torque and bought gearing accordingly. It just seems to make sense that the exact same motors that spin about 3500 rpm in wye would produce equal or more torque (and more efficient???) in delta at 7000 rpm when geared 2x lower, even if they produced 1/3 less torque at the shaft (and if they actually turned double the RPM in delta. I've read conflicting data).

Anyway, this is all academic for me at this point. I would like to learn it though for when I build another.

[Irving]

Thank you DMC-man & Irving for your replies.

Another 2 questions.

Q1, the log above is captured when motor was spinning at around 170 rpm (6.5mph @ 13.5" tyre)

If chair goes 15mph , motor will be spinning at 400 rpm. The data rate from the hall sensor will also increase.

Can you tell from the graph that HBL-2360 has enough bandwidth / processing power to do it ?

No, there's no way of telling from 'recorded' data what the internal processing capability of the controller it. However I would imagine that the motor 3-phase bridge controller is independent of the MCU used for overall control and housekeeping so data rate from the hall effect sensors shouldn't be an issue, but I can't guarantee it.

Q2, In CLOSED LOOP SPEED MODE with brushless motors, the controller will keep motor speed to the command with variance in load. That seems to have the same effect of the CLOSED LOOP TORQUE MODE , itsn't it ?

Then, which mode should be best to drive a wheelchair ?

Thanks.

As has been said, neither. These modes are generally for servo (torque) or spindle (speed) control... e.g. a CNC coil winder needs torque control to maintain a fixed tension as the bobbin fills and revs decrease... a drilling/milling spindle wants to maintain speed as the bit hits the workpiece to maintain chip load on the cutter...

@williamclark77 - Y configuration is usually preferable, has 1.7x the torque of the same motor in delta, delta has 1.7 the kV of Y. Since we are (mostly) dealing with direct drive we need torque not kV. Also perceived wisdom is Y has better efficiency and more stable open-loop speed characteristics...

[Burgerman]

You need OPEN LOOP if you want to retain normal driving characteristics and feel.

I tried a few chairs with closed loop control and its terrible. There's no feel or feedback. EG try to push a door open gently with your toes. and if its locked you will literally break your toes. Same on leaves or slippery surfaces, or angles on ramps. You cannot tell how much torque is being used until its too late and you have suddenly tipped out... IT decides, you don't.

[shirley_hkg]

WOW ! It's a bunch of information . Need some time to digest it.

Still a long way to put on chair to test it. Just ordered 5 pieces of chips for signal conversion.

[stevelawiw]

I love Invacare GB motors

To be able to increase the volts on GB motors is a very exciting prospect, but I'm afraid my input is going to be limited to asking questions, so

Williamclark77 wrote:
The new Roboteq FBL2360 advertises compatibility with trapezoidal and sinusoidal switching as well as hall, encoder, and sensorless rotor sensing. It's a max 2x 60a output though. Will dual 60 amp output be enough for a powerchair?

Does this mean that the FBL2360 would be compatible with the Invacare GB motors without any need to decode the output of the existing hall sensors?

Link to Roboteq page
http://www.roboteq.com/index.php/roboteq-products-and-services/brushless-dc-motor-controllers/fbl2360-detail

[shirley_hkg]

Does this mean that the FBL2360 would be compatible with the Invacare GB motors without any need to decode the output of the existing hall sensors?

He said so when we exchanged email with their engineer.

[Irving]

Sensorless is useless for a wheelchair...

To develop the commutation signals you need to know rotor position. Sensorless does so by measuring back-EMF when the PWM current to the winding is in the off-state, great for closed-loop speed or torque control on a motor that's already spinning, or can be manually spun (e.g. electric plane, or electric pedal bike). But a wheelchair that's stationary generates no back-EMF so therefore no torque... and that's when you need max torque.

Sorry, you need sensors or encoders....

[stevelawiw]

Hi Irving,

The idea is to use the existing hall sensors already in the GB motors and use the feature of the new controller that will accept the output directly without a need to decode, are you saying this is not possible?

[Burgerman]

Yes its possible.

Sensored, sensoreless, brushed can all be used in torque mode, speed mode, open loop or closed loop, these are all different things.

[Irving]

I was referring specifically to sensorless operation as that's what I read into your question, though I see now I'd read it wrong! The new controller suggests it can ONLY do sinusoidal with encoders and Field (Vector) oriented control in that mode is a Q3/14 firmware update. However, you need to be sure that the Invacare motors are suitable for sinusoidal operation - I don't know, and not sure how you'd find out. Though at a guess, since they have encoders fitted it is likely.

Another resource worth reading... http://cladlab.com/electronics/circuit- ... or-control and I quote from there:

"The motor windings can be wound to either give trapezoidal or sinusoidal feedback, which relates to the control methods mentioned further down on this page. As quoted by Shiyoung Lee, Ph.D., A COMPARISON STUDY OF THE COMMUTATION METHODS FOR THE THREE-PHASE PERMANENT MAGNET BRUSHLESS DC MOTOR),

The simulation results verify that mismatch of the back-EMF waveform and commutation method produces ripple rich torque. Therefore, the BLDC motor and trapezoidal(six-step) commutation and the PMSM and sinusoidal commutation are the most desirable combination to produce minimum torque ripple.

So it pays to get the right motor for the right job!"

[stevelawiw]

Thanks Irving and BM for the info. I need to get my finger out now and finish my second chair which is more of an indoor/outdoor rather than the outdoor/indoor of my first chair, all using the Invacare GB package for motion. Main differences are the indoor chair has shorter wheelbase, narrower at the front and lead powered.

Getting back to the Roboteq FBL2360, if I rework my LiFePO4 105aH 24v 7p 8s pack into 4p 14s giving 44.8v 60aH that should give me more speed, torque and greater range, right? Can't wait!

Shame there's no way of charging 17s or I could have gone for 3p 17s giving me 54.4v 45aH but only using 51 of my cells.

[Williamclark77]

Getting back to the Roboteq FBL2360, if I rework my LiFePO4 105aH 24v 7p 8s pack into 4p 14s giving 44.8v 60aH that should give me more speed, torque and greater range, right? Can't wait!

Shame there's no way of charging 17s or I could have gone for 3p 17s giving me 54.4v 45aH but only using 51 of my cells.

Major improvement in range! I'm using those same 15ah cells in 3p and 14s configuration with the Roboteq hbl2360 and custom wound brushless motors that turn 3500ish rpm with 70 mOhm of resistance. Combined with a 4:1 planetary gearbox and a 2.92:1 secondary reduction with a 14.5" tire (a little less than 14" tall in use at low pressure) for roughly 14 mph. My rough testing so far that I'm doing now is showing about 3.5 to 4x more of USABLE range than my Permobil c350 with 60ah mk gels. You would be thoroughly pleased with 4p range! Combined with Lenny's amazing script it is as controllable and smooth as any manufactured powerchair. See the videos I just posted in my "Willchair Build" thread.

I would love to go 15 or 16s, but like you said, no simple way of charging currently. In use my pack sits on 46.5 volts. I was afraid 3p would not be enough, work the cells too hard, and get too much voltage drop when I designed my system. It's not. Under HARD use I'm only getting a 1/2 to 3/4 of a volt drop. And it recovers within 5 seconds of taking the load off.

I would love to see pics and specs of your chair build!

[Burgerman]

That's because the cylindrical headways have a higher C rate than the prismatic cells. And because as you increase the voltage you decrease the amps at the battery.

[woodygb]

Here's one of Steve's chair.

[Burgerman]

Getting back to the Roboteq FBL2360, if I rework my LiFePO4 105aH 24v 7p 8s pack into 4p 14s giving 44.8v 60aH that should give me more speed, torque and greater range, right? Can't wait!

Almost.


More speed = yes. Double.
More torque = same PEAK torque, but presuming the motors are low enough impedance to bul full amps at 24v then you will STILL have full torque at half the new max speed.
More range = same range. At double the speed. OR shorter gearing allows more range (and more torque) at the same speed.

[stevelawiw]

Does anybody have any insight into why the FBL2360 controller is only 60A and doesn't have any box to protect it?

As I understand it this is the only RoboteQ controller that can interpret the output of the Invacare GB motors, so should I go for it and buy one, or maybe wait for the HBL controllers are upgraded to be able to control the GB motors

60A seems a bit low, would this be enough to make enough of a difference to make the effort worthwhile?

[Burgerman]

I don't know how they are measuring 60.

On a BRUSHED controller, then the controller is limited to a max of 120 or 150 amps by software. That is motor amps. When one motor is at say 50A and at a 50% pulse-width then the battery amps will be 25Amps. So battery amps are not used here. They vary like a gearbox.

On a BRUSHLESS controller it may also be motor Amps. But how would you measure this? On one wire? Or 3? And how does this add up or relate or compare? Is 3x 60A equal to 120? Or 60? Or 180A? These are switched signals too. Is it averages across all 3? Or what. Or some other figure based on a 3 phase AC type supply?

What I am saying is that its not so simple. And I don't know. Maybe there's a page somewhere... Remembering that brushless are more efficient you wouldn't need as high output anyway. Again its not so easy to compare.

[stevelawiw]

Ok thanks BM.

Even if it really only is 60A then I'm still going to get the improvements in range because it will be getting 44.8v to power it.

Why do you think they are selling these without a protective box? Seems a bit weird to me

[Burgerman]

Many controllers, esp small ones are used for hobby stuff, embedded in machinery, or cost and weight is an issue. Most of my hobby speed controllers are just a board with a bit of heat shrink.

See http://www.hobbyking.co.uk/hobbyking/st ... llers.html

And there are big ones, small ones, reversible ones, programmable ones, current limiting ones, low and high voltage ones, water cooled ones etc...

[stevelawiw]

After another read of the FBL23xx Datasheet
http://www.roboteq.com/index.php/docman/motor-controllers-documents-and-files/documentation/datasheets/fbl23xx-datasheet/203-fbl23xx-datasheet/file
(which is still marked as 'Preliminary - Subject to Change') on page 6 is this:-

The controller requires the Hall sensors inside the motor to be 120 degrees apart.

Given that this is not the case with the Invacare GB motors, assuming the specs are correct and do not change, am I right in thinking that this controller will not be compatible with the GB motors?

[Burgerman]

I don't know how they work on Invacare motors. But 120 degrees refers to the distance between poles. If there are 3 poles.

If there are 12, 24 or something they can be at all kinds of physical angles and still see 3 120 degree electrical spaces between poles. So you will need to examine the motor. Or stick 3 wires into a sillyscope and take a look... If you see even spaced pulses then you are good to go... You could also use headphones or 3 LEDS and turn by hand and see if they switch at 120 degree intervals.

[shirley_hkg]

Signals are tapped and I'm glad that they can be converted . However, the output signals are not quite useable yet. Need to re-program the chip to do the job right. Will update here with any progress.

[shirley_hkg]

[shirley_hkg]

The student volunteer back off. No one follows, though he confirmed that the chips can do the conversions.

All inputs are welcome here.

Besides, did they mention that new controller will accept other forms of signals, including sin/cosine ?

[shirley_hkg]

[Williamclark77]

Details.....