Compact Field Oriented Controller, ASI + Grin, limited run

Hi Bill,

Thanks for your reply.

I did have the poll button pressed (it turned blue). The read button turns blue only until you select something else, so I think it just reads once. Neither selected does not do anything. Do you remember if it is normal for the 1 or 2 in the motor discovery box to instantly switch back to 0 when write is pressed?

I did not realize that the status button was not suppose to be red. I thought it was telling me it was connected. Opps. I held my cursor over it and it reports 'bit 5 Motor Halls sensors fault 15 LED flashes'. Try a quick end run and go to sensorless mode. This clears the status light but when I go to Basic Motor (sensorless) and put 1 in motor discover mode, I get 'bit 7 POST static gating test (3 LED flashes). The one does not instantly disappear from the motor discover mode box but there is no autotune results.

I guess I will try the method outlined in the BAC controller manual ver 1 to manually check halls.
 
So I research the 'bit 7 POST static gating test (3 LED flashes)' error and found that either the controller has an insulation fault or one of the motor phases is shorted to ground or power. The approved test is disconnect motor phase leads and repeat test. http://www.accelerated-systems.com/KB/?p=62

This clears the error which means the controller is fine and my motor is to blame. I do not have a megger to do a motor test and basic Ohm test does not help.

I disconnect both my temperature sensors and tape off their wires so they should not be a ground path. I move my phase wires so they are not touching the chassis. The error clears and allows me to perform the auto tune. I get 232 for Ls and 25 for Rs. I repeat a number of times and the results are very close. I fill these to the write/save window and write to flash.

I press the throttle and can read that the throttle signal is in the 'GO' range but the motor does not spin. Since I am running sensorless, I try a hand spin but no joy.
 
A cheap LC meter for 20$ is a good tool for seeing if the motor is shorting internally. Check the inductance of all phases. Multimeters' resistance readings aren't accurate enough
 
ProEV said:
I disconnect both my temperature sensors and tape off their wires so they should not be a ground path. I move my phase wires so they are not touching the chassis. The error clears and allows me to perform the auto tune. I get 232 for Ls and 25 for Rs. I repeat a number of times and the results are very close. I fill these to the write/save window and write to flash.

I press the throttle and can read that the throttle signal is in the 'GO' range but the motor does not spin. Since I am running sensorless, I try a hand spin but no joy.

Did you do both auto-tuning steps? The first is the static (write "1" in the entry box), where the motor buzzes and whines for a few seconds and returns Ls and Rs values. The second is the dynamic (write "2" in the entry box), after which the motor will spin at approximately half maximum speed for several seconds (and return a Hall offset and mapping, if you're tuning as sensored). Only after both tuning steps are executed will the throttle function as expected.
 
Hummina Shadeeba said:
A cheap LC meter for 20$ is a good tool for seeing if the motor is shorting internally. Check the inductance of all phases. Multimeters' resistance readings aren't accurate enough

Thanks Hummina. I had no idea what an L/C meter is. Sounds like it will help me establish if I have a problem with the motor windings. I found a video that walks through the testing here. http://forum.flitetest.com/showthread.php?20346-LC-Meter-to-check-motors . It sounds like I should dissamble the motor to do the test.
 
Bill,

Since I am getting an error with the Halls, I set the controller to sensorless and then did the autotune on the sensor motor page. It did both the 1 and 2 parts. This time I heard the high pitched sound when I did autotune 1. The motor spun as described. The Halls autotune showed -1 for all Halls sensors.

The motor still did not respond to throttle, so I did some simplifing. In command inputs, features, I change bit 4 to 0 so analogue braking is disabled. Next I decided to change Cutoff brake sensor type from 0 (brake1) to 1 (brake2).

Now the motor responds to throttle. Brake1 is where my regen pedal is wired. Brake2 is unwired and floating at 4.XX volts. I do not know what cutoff brake sensor is but...

I run the motor a couple of times. Switch off and restart and it works fine. Then start packing up the wiring to do a track test. I get the'bit 7 POST static gating test (3 LED flashes)' error again.

I'm going to try moving the wires one at a time and see if I can localize the problem.
 
Well, the saga continues...

The motor would spin nicely or it would kick up one of two errors, either 'bit 7 POST static gating test (3 LED flashes)' or, for variety, 'bit 3: Current sensor over current (flash code 1,4)'. I called it a day, disconnected everything and came back at it a couple of days later.

Now, I can't establish connection between the computer and controller. Same computer. Same cables. Same pack, motor,controller.

I reboot etc a couple of times. I check USB to serial cable by plugging into cycle analysis. Works fine. I continuity check from serial connecter to female plugs on connector to controller. Controller has power and is putting out 5 volts to throttle.

Is the controller damaged and no longer willing to talk?

My L/C meter arrives. I pull the motor, remove the rotor and check readings. I get:

Measure L red to black .168 mH,
red to green .134 mH
black to green .135 mH
Black to ground .085 mH,
green .041,
red .084

Does this mean the motor needs to be rewound?
 
I don't understand you with so many measurements with the LC meter. You should get three readings for the three phases. If you look on YouTube you should find it. From what you did get they do seem imbalanced and internally shorted, but I'd watch the video and I've limited experience. Cool tool though.
 
cavallo pazzo said:
With last software allowing variable regen and "good" tuning on Phaserunner, decelarating is nice and smooth. Total regen didn't really changed, as I more often use mild current and vary on a wide range from 0.4 V to 0.0 V (at 20 A max) instead of having 7-8 A all the time. Sometimes too strong, sometimes too weak and I had to anticipate to make good use of regen.

I really feel high current regen as an alternative to classical "heat" brakes, for everything but emergencies. Now I can slow down in a 10 % downslope, that's impressive !

Only drawback ?
Everybody wants one, and I struggle to explain I'm THE lucky guy using it so far.
I'll be more than happy when production will raise to semi industrial volumes to allow some units to arrive here.

You will soon be not the only lucky guy Cavallo. Things are setup again at our new facility and we've got the capacity now to make about 30-40 pieces a week and are back in the business of taking orders!
Phaserunner_2.0.jpg
The 2nd run of Phaserunners makes a lot of enhancement tweaks which were painful to implement but I think will be appreciated and hopefully worth the wait. I've mentioned these pending changes earlier but will re-iterate what's been implemented.

Firstly we brought the total wire count down to just 4 cables exiting the Phaserunner.
  • A 90cm long Battery Cable (with extra pair of wires for charging/accessories)
  • A 100cm long Motor Cable
  • A short Throttle Cable (with hidden wires for remote on/off button and hardware fwd/rev)
  • A short CA Cable

Phaserunner_Connectors.jpg

The battery cable uses our 4 conductor cable with 2 heavy 13g wires, and two thinner 18g wires. That let us hook things up so that current flowing through the thinner gauge plug still goes through the controller shunt and shows up on the CA display. This provides an easy means to monitor your charging current on the CA, or the current draw of any lights or other accessories that you run off the main battery pack, and is a direct response to some of the feedback from early adopters of the pilot run last summer.

We got rid of the TRS communications wire and instead molded a jack directly into the casing, and right beside it we added an on/off switch to the back of the controller too. But instead of a normal latching on/off switch that cuts the V+ power to the logic circuitry as is normally done, we've implemented a mosfet soft switch circuit so that the on/off control is done with a momentary style button which tend to be much more robust.
Phaserunner_Jack_Button.jpg

And finally, the speedometer signal is now no longer coming directly from the hall signal but is instead being generated by the controller firmware, so that you will still get speed readings on a CA-DP device even when you run a sensorless direct drive motor without any halls connected. Overall I'm pretty damn happy about the final result and will really look forward to seeing these get used by the ebike community. It really is the dream controller that I always wished I had for my personal EV projects over the years and hope that others feel the same!
 
The other thing we've been working hard on is the phaserunner setup software. I think the last post I had here was for the V0.6 release, we're now at V0.91 and have both the Linux and Windows downloads available:

http://www.ebikes.ca/downloads/?file=PhaserunnerSoftware_Windows
http://www.ebikes.ca/downloads/?file=PhaserunnerSoftware_Linux
(MacOS version just going through some adjustments)

The basic setup tab now has visual sliders for the battery voltage rollback setpoints, and more importantly a graph showing the sensorless self-start map as an RPM vs Time plot. The hope is that with this visual portrayal of parameters their functions and meaning should be easier to grasp. And if not, we also have added pop-up tooltips on all the parameter fields, so if you hover the mouse over a parameter name it will provide a detailed description.


There is a new "Advanced Setup" tab where we've put both the control loop parameters and also the exact throttle and regen voltage maps, also shown graphically as well as numerically:
View attachment 1

And finally there is a new dashboard tab for viewing all the motor signals in realtime, and also graphing and recording data while the motor is running. You can choose whether any parameter is plotted or not and whether it scales to the left or right axis, and export the data to a .csv file if you want to do logging or analysis.
PhaseWarmup_Graph.jpg

We're hoping that this helps makes the advanced controller more accessible and less intimidating for newbies to setup. In principle, it should be that all you need to do is connect the phaserunner controller to your motor, battery, and computer. Hit the "Read Parameters" button on the bottom to see the current settings on the Phaserunner, then click the "Autotune" button to have the controller automatically determine the motor parameters, hit "next" a few times to step through the process, and then be done! You would then set the battery and motor current limits and voltage cutoffs to suite your pack and motor. The default parameters should generally work with most setups, although tweaking the sensorless starting map is likely required if you are running sensorless.

Any feedback on this will be most welcome.
 
That is really nice. Both the recessed programming jack and the on off switch. It might be time to sell my original and get yet another one!
 
Hi Justin:

Thanks for working to improve the Phaserunner and its programming software. You and the rest of the Grin staff have done an excellent job of providing a quality near-turn-key BLDC controller for us DIY'ers.

One question remains for me as I consider whether to use this controller on my crank-drive ebike.

Have you or ASI addressed the problem of "springy" drivetrains causing a resonance with the torque throttle feedback timing when shifting the rear gear cluster? Or, is a simple voltage (PWM) throttle option available?

Thanks.
 
Most ebikers wouldn't ride purposefully through rain, but...sometimes it looks like you can go for a ride, and you get hit with a sudden and unexpected downpour. This Phaserunner is the ULTIMATE water-proof controller.

The design of the heat-management also means that it could be located inside a battery triangle back for an uncluttered build, and in 99% of cases it would not get hot, because it does not depend on a fan with air-circulation to work...
 
Though you would still need to create a heat-path out of there for it, like a heatsink bolted thru the battery compartment/bag/etc to the heatsink side of the controller. ;)
 
With aluminium frame, as on my Catrike, cooling is sufficient with simple contact on tube.

I run a 14s setup with 40 A max and regen at 20 A, running a 2808 9C motor. No heat troubles at all, bottom of controller is barely hot to the touch after a 250 m elevation trip under full load. Didn't checked correctly with thermal camera on a defined trip, but will do soon.
Efficiency seems very high in real conditions with load change and speed variation, and power ranging wildly between 350 and 1850 W with most of the time 1100-1250 W.

Yes, it's a dream controller. Thanks !
 
With aluminium frame, as on my Catrike, cooling is sufficient with simple contact on tube

My main ebike right now has an aluminum seat-tube that is perfect for this. However, I will stick my neck out and say that since Justin spec'ed efficient components instead of the cheapest components, I suspect many builders will be very surprised at how cool this controller runs. Heat sinking to an aluminum frame is the ultimate combo, but even a steel frame would pull heat out (with good contact surface).

If I didn't already have six controllers under my roof, I'd buy one of these right now. My wife is now getting jealous of my ebikes, so she has been hinting that she might be willing to pick out a bike for me to electrify. If she goes through with that, she will be getting my controller, and I will be getting a Phaserunner.
 
Good for you Ron, recycle your stuff to the wife and get your self the new parts :)

I am also smitten with the phaserunner. I really like the small footprint and the fact it is all weatherproof.
Too bad the phaserunner does not have a angry bigger brother. 12-14 kw peak power and I would sign up for one.
Otherwise I need to find a 6 phase motor first and run dual phaserunners :)

IIRC phaserunner tops out at about 8 kw peak, maybe little less now with the new version as they lowered the max voltage from 100v to 90v?

Has there been any rumors about any larger phaserunners being in the pipeline?
 
I run it maxxed out on amps at 66v (16s lipo)

It does need a heatsink for this... I can go about 30sec of WOT before thermal kickback cuts in. With a heatsink, it goes forever.
 
Some questions:

When to expect the MacOS version of the Phaserunner setup software?
(has to work with OS-X 10.8.5 since I will NOT update to a newer OS-X version!)

Is the Phaserunner still available in custom colors? Like blue?

A blue Phaserunner would look rather pretty on the blue frame: :D
blue-ice-24-26-header.jpg


How much power is the Phaserunner able to handle without heat sink?
I'm using a Bafang BPM (code 10) with 15s battery at 35A at just shy of 2kW in the trike, but steel planetary gears are on the way and 20s is an option... ;)

How much power is the Phaserunner able to handle with heat sink?
 
Marc S. said:
How much power is the Phaserunner able to handle with heat sink?

How much POWER is going to depend on your voltage.
With my setup of 16s lipo (66.4v off the charger) and a small heatsink, it will hold 3800watts with some 4kw peaks until you run out of battery. never gets warm.
with the same setup, but it zip tied loosely to the frame without any heatsink, it last for about 30seconds at 3800. it does see some wind of course. It will hold at 2500watts while it cools off, but its an abrupt chop.
 
FluxZoom said:
justin_le said:
Here's a better scaled closeup. It seems to average a steady 3% improvement in efficiency right across the board once both controllers are in their current limited regime. And 3% is still no small matter when it comes to overall system efficiency. Trying to eek that much improvement via copper fill, thinner laminations, higher grade steel etc. would be a lot of work compared to just changing the controller's drive scheme.

There are plenty of controllers out there that do sinusoidal commutation without field oriented control (FOC). Any chance you might or have done a similar comparison with this non-FOC type of controller to trapezoidal and FOC?

Hey FluxZoom, I haven't yet but as we have a large number of random china sinewave controllers kicking around there is no reason not to do this and then have the complete picture. My expectation is that the sinewave controllers will still have some efficiency and torque gains over the trapezoidal controllers at low speeds just like the FOC controllers do, but what you won't get is the much stiffer torque increase as your speed slows down at high speeds. To illustrate, if you look back at my post here comparing a trapezoidal drive to the FOC on a MXUS hub, you can see that the slope of a the torque vs RPM curve is quite a bit different:
https://endless-sphere.com/forums/viewtopic.php?p=1087084#p1087084
file.php


Theoretically what you'd expect is that in this section of the graph, to a first order the motor current would increase in proportion to the difference in battery voltage and back emf voltage divided by the winding resistance.
Imotor = (Vbatt - Vemf)/Rmotor = (Vbatt - RPM*Kv)/Rmotor,

and the expected relationship between I and RPM is -Kv/Rmotor. A low resistance motor would have a steep slope. But in a trapezoidal controller the measured current doesn't increase as steeply as you'd expect because the inductance of the windings limits it further. With a field oriented controller this inductance doesn't have any effect and the slope is as you'd expect based on the motor resistance. In this MXUS hub example running at a fast 575 unloaded RPM, you can see that at about 530 rpm, the FOC controller is already able to put in maximum power to the hub while the trapezoidal controller is only at 50% power. To get the trapezoidal controller to reach full power the motor needs to be slowed down even further to just over 500 rpm.

Trap vs FOC Slope.jpg

I'm pretty sure that if we do this same test with a regular sinusoidal (non-FOC) controller, the results will be just like the trapezoidal controller in terms of this current vs. RPM slope.

I would assume that the non-FOC type would be the least efficient, but I've wondered more than once by how much. I think this sort of idea might help others see a tangible distinction between these types of ideas.

I'll try to get measured empirical results to clarify this. So far it's only been FOC vs. Trapezoid, but to be complete we should have FOC vs Sinusoidal vs Trapezoidal
 
Marc S. said:
Some questions:

When to expect the MacOS version of the Phaserunner setup software?

We got it! Though we only have a really old mac hardware to test it on so if any people have newer machines it would be great if you can confirm that it works OK.
http://www.ebikes.ca/downloads/?file=PhaserunnerSoftware_Mac
Also, the windows and Linux version have been updated to V0.92 as well, which just has some minor tweaks over the 0.91
http://www.ebikes.ca/downloads/?file=PhaserunnerSoftware_Windows
http://www.ebikes.ca/downloads/?file=PhaserunnerSoftware_Linux

Is the Phaserunner still available in custom colors? Like blue?

For this run we are keeping things simple and efficient and only doing batches in clear and in black. We might do select runs of other colours, but it will be more limited than last time since we want to produce inventory for stock rather than building them all on demand.

How much power is the Phaserunner able to handle without heat sink?
I'm using a Bafang BPM (code 10) with 15s battery at 35A at just shy of 2kW in the trike, but steel planetary gears are on the way and 20s is an option... ;)
How much power is the Phaserunner able to handle with heat sink?


Talking about "power able to handle" is the wrong language, it's really just the phase amps that causes heating and not the power flowing to the motor. So if you are running a slower winding motor at say 72V with 30 phase amps, that's 2100 watts that it would handle effortlessly. But if you are running a muh faster wind motor at 48V that pulls 70-80 phase amps it will get hot pretty quick, even though you might be drawing well under 1000 watts from the battery pack.

I promised some time ago and still promise to follow through with a wind tunnel test that shows the phaserrunner heatsink temp vs. time at increasing phase amps, and do that test in still air, with airflow, and with an external heatsink in airflow, so that people will be able to extrapolate numbers for a given scenario.
 
justin_le said:
We got it! Though we only have a really old mac hardware to test it on so if any people have newer machines it would be great if you can confirm that it works OK.

It seems to work great on a current model Macbook Pro :)

Cheers,

Jason
 
Back
Top