Adding burtie internal optical sensors to Astro 3210

Joined
May 26, 2008
Messages
951
Location
Chicago area
http://www.recumbents.com/wisil/e-bent/sensors

I had been using this Astroflight 3210 RC motor with an RC controller (ESC). RC controllers do not use sensors, because RC motors are designed to always run at high RPMs. E-bike motors on the other hand run at low RPMs when starting. It always annoying me that the motor would cough and hiccup until I got up to about 10 MPH (at which point it would of course launch my bike like a rocket).

Ian (burtie) has been making sensor conversion kits for sensorless RC motors. Sensors tell the controller when to send a pulse of power to a particular phase of the motor, which makes the motors run much better at low RPM, and can make them run more efficiently at high RPM.

Bertie's kit comes with a circuit board, a washer, a sheet of paper with the black/white patterns on them, and a piece of balsa wood.
The idea is to space the washer out about 5mm from the rotor on the axle. Instructions said to use the balsa and silcone, but I decided to make a washer out of Lexan (polycarbonate) for a spacer, and use JB Weld to glue it all together.

IMG_0586.JPG


The Astro motor has an aluminum spacer between the rotor and the bearing. The washer and Lexan spacer sit on this 1/2" OD spacer.
Because the washer has a bit of slop on the aluminum spacer, I used a punch to make a series of dents close to the ID of the washer, which closed the gap.

Here's the spacer and washer glued to the rotor. I roughed up all surfaces with fine sandpaper before glueing.

IMG_0587.JPG


I cut out and glued the sensor pattern to the washer.

IMG_0588.JPG


Added the wires to the sensor board and mounted the circuit board to the inside of the motor cover. Right now it is just a friction fit.

IMG_0589.JPG


Adjusted the circuit board until there was about 2mm between the sensors and the pattern.

IMG_0590.JPG


Sealed up the motor temporarily as the circuit board will probably need to be twisted a couple degrees to adjust the sensor timing.
Used a 5V power supply connected to the 5V and Gnd wire of the sensor board to ensure that the sensors switch between about (about) 0V and 5V. It works!

IMG_0591.JPG


-Warren.
 
THANKS for the tutorial. I see Burtie is out of stock, but when they come back I'll bet they sell like hotcakes!
 
Now, the next question is, how do I tell if it's adjusted properly or if I need to adjust the timing?

Burtie sells a little Blue Button that will adjust & advance the timing on the fly! (old cartoon referance)

but yes, adjust for the lowest amp draw at max throttle & see where the low speed responce is....my best performing motors are just slghtly advanced from that point. like 1/2 an amp increase in no load draw...you will hear a noticable/measurable increase in rpm's also. be carfule...its easy to get too much.

But I would love to hear some RC results using burties timing advance circute...at least before I get mine hooked up for some test runs. :mrgreen:
looking good Warren.
 
I need to change the direction of the motor, so I will swap two of the phase wires. I'm making a short video of the motor running. To adjust the timing I'll take all the screws out of the motor's cover and twist it back and forth at a set throttle speed to see how that effects the RPM.

-Warren.
 
Looks good Warren.

I am still too busy building drives to find the time to mess around with one of these.

I am glad others are doing it, however.

The only concern I have is getting full RPM out of these motors with the current E-bike controllers. Most Astro motors are higher KV than yours. But, experience will answer that question.

I look forward to your finalized setup on your bike. It will be good to see you up and running. :D

Matt
 
recumpence says:
The only concern I have is getting full RPM out of these motors with the current E-bike controllers.

I have tested the sensored XieChang controllers with 7pole pr motors beyond 13k rpms. They will have no problem spinning a 2 pole pr (or is it 4?) Astro even faster.

the sensoless versions of xieChange bike controllers fall on their face somewhere in the 8-9k range IIRC. BUT...the main reason to not use the Castle is for lowspeed sensored control.

These Kits are a great idea...& I really beleive they are a corner stone for progress.
 
recumpence said:
The only concern I have is getting full RPM out of these motors with the current E-bike controllers. Most Astro motors are higher KV than yours. But, experience will answer that question.
Matt

any more feedback from using those sensors please? :)
 
wojtek said:
recumpence said:
The only concern I have is getting full RPM out of these motors with the current E-bike controllers. Most Astro motors are higher KV than yours. But, experience will answer that question.
Matt

any more feedback from using those sensors please? :)

I have not run them yet. That is why I am excited that Warren is doing this. I am tired of only having one true option to run these motors. This sensor setup looks to be the solution we have been looking for.

Wojtek,

You are the PERFECT candidate to run this type of setup! :)

Matt
 
there is absolutely no reason why these sensors will not do the job very well, they suit inrunners very well as they are fitted inside of the motor and there is no possibility to get any stray light onto the sensors and also the disk will remain clean they can be a bit of a problem if mounted on the outside due to light and dirt.
 
ordered!

will let you know the results :)

any particular controller choice or any sensored should work?

thanks!
 
After much swapping of wires and gnashing of teeth, I was able to make the motor spin the other (correct) direction. Twisting the end of the can to change the timing makes a big difference in top RPM, so I'm going to adjust for max RPM.

Burtie suggests holding the timing board in place with silicone glue. I'll give that a try.

-Warren.
 
Warren, I recommend that you don't set the timing for maximum unloaded RPM. This may yield very poor efficiency, and may not work well when the motor is loaded.

Instead, you want to set the timing for minimum no load current at WOT, then very slightly advance it from there, so the current increases only a small amount,
(just as Thud has described above).
This will be the point at which you will get best performance from the motor when it is loaded.

Cheers,
Burtie
 
Ha, Burtie I was just going to ask for for clarification on that.

After opening the motor and adjusting the timing about 30 times:
When timing is adjusted for max RPM, no load current at max RPM is 650 watts
When timing is adjusted for min Current, no load current at max RPM is 50 watts
I advanced the timing so no load current at max RPM is 70 watts, does that sound about right?

The max RPM at 650 watts is way higher than when adjusted for min current, but the motor gets hot...
The low speed performance seems similar across the range of timing adjustments that I tried.

-Warren.
 
If 50W is the lowest, then 70W Sounds reasonable, I might try and get it a bit lower, perhaps 60w would be ideal.

Perhaps you could use the motor with the back temporarily taped on, so you could find your ideal timing point before fixing anything more permanent ?
 
Here are the sensor wire colors needed when using a Lyen / Infineon controller, with the motor running clockwise (looking from the output shaft side). To run the motor counter-clockwise, flip the green and yellow wires.

sensor_wiring.jpg


Burtie, is there more involved in finding the ideal timing point than adjusting for minimum WOT current and then advancing a bit?

-Warren.
 
Thanks for the excellent photos and diagrams Warren.

www.recumbents.com said:
Here are the sensor wire colors needed when using a Lyen / Infineon controller, with the motor running clockwise (looking from the output shaft side). To run the motor counter-clockwise, flip the green and yellow wires.
-Warren.

Note the wiring scheme will vary slightly depending on how the encoder pattern lines up with the magnet poles of the rotor, so some trial and error wiring exercise may still be necessary, unless you can adjust the angle of the sensor board through a full 90degrees (360 e degrees) to compensate.


www.recumbents.com said:
Burtie, is there more involved in finding the ideal timing point than adjusting for minimum WOT current and then advancing a bit?

-Warren.

Thats all that is needed :)

However, if we wanted to try and improve on this, things quickly get a bit more complicated: We would need some sort of dyno to set it up, and/or a timing controller that dynamicaly shifts the commutation timing depending on the motors operating conditions
http://endless-sphere.com/forums/viewtopic.php?f=2&t=19054

Burtie
 
I buttoned it all back up today with the new 6/32 screws to replace the screws in the motor can that were lost or had the hex drive stripped. Testing showed that it's working fine and drawing about 65 watts at WOT. The next step is to put it on my bike and test it to see if it really does improve the low speed performance under a load. I have a video in the works to show the comparison.

-Warren.
 
what controller have you been using with sensorless solution please? thx
 
Been following sensored builds upto this thread and it seems to have died ,,recumbrance did you ever do the comparison vids ?
Reson i ask is i was thinking of building 3220 or c12090 with sensors through the gears ,cant seem to find any long term results any wear?
Just cant convince myself of what to spend my limited budget on ,,even thinking of just poping for a big hub (i like smoth controlled starts )


Dusty
 
Back
Top