Build thread: submicro brushless motors

[shagrug]

Hi Keith-


      I can't get your vid to play, sorry. I've read thru your post a couple of times and I've studied the pics also, but I'm still not clear on the design.

      I understand a bit on the direction of rotation problem, and your question on startup. I will refer you to the two photos below-

      The 1mm with the chip mounted 90 degrees to the windings runs- but has a hard time deciding which way to go. It will run in both directions.

      The bigger 2.5mm motor with the adjustable ring mounted chip does this- with the chip at the top (0 degrees) it runs in one direction. Rotating it past the 90 degree point (like the photo) will change the rotation direction. Spot on 90 degrees it can go either way, same as the 1mm above. So by moving the chip to the other coil pole side changes direction. I hope that helps?

      Is it possible to use two chips and two separate coils in your design? I would consider trying two chips wired to the same coil with reversed output connections, it might work. (if only one chip is powered at a time) It would help to know more what you are trying to make. I can't tell how your motor is set up inside from your pics.

      You are also right about Allegro discontinuing the A1442 chip this year. I saw a post by nachman at RCG about it a few days back. I am sure there are other suitable chips that can be used, I've been looking at other datasheets this winter anyway, looking for more power output and better H-bridge connection. The Melexis chip (datasheet attached) is very close, Ivan says they are in the pre-production stage now.  (having some on order already) I'm hoping to get some to test soon as they are released.

[Keith55555]

Okay well that gives me some confidence knowing that sensor position will definitely cause a change of direction. Tonight while driving, I had thought about what you just mentioned; two chips and two coils. However I was wondering if I needed the coils to be on top of one another and thus neither being really centered. How do you see the two coils positioned? Not one each parallel to the magnet, opposing each other, right?
If that's all it would take, the second sensor and coil causes me no problem whatsoever.

From what you say in your two examples, I'm thinking that the inclination of direction is directly related to the security of one pole overwhelming the other (larger magnet's percentage of exposed pole versus smaller magnet's percentage) in intensity in the sensors 'decision making' (I'm not sure if it is this sensor is the one I was reading that did a pulse 'shaking' for a stalled state).

The motor in the pics isn't the motor I'm design now. That was for a magic trick. Those pictures don't show the brushes commutator (if you look closely you'll see there is however a milled out section for their placement) or battery contact or the 'lid' which looks like a nickel face that you can see in the movie. I had it on a magic site and didn't want to give it completely away. The correct URL for the movie is: http://www.lightlink.com/keith/NickelSpin.wmv
Safari deemed to remove the capital letters in 'NickelSpin'.

What I'm doing now is basically what you've done except being that since  it isn't for flying, I am trying to get it 'torque-y' by creating a longer version (more magnet length and more coil 'umph' but having the height remain tiny-ish.
Trying to get a custom tube magnet with some length (about 40mm) but with a diameter of 2mm. Might not be possible because of drilling that long given this diameter. I have the call in and waiting for a reply about that. Considered stacking some but lining up the poles could be daunting. If they're not lined up, destructive interference would kill the efficiency. Have to create a test jig and a way to secure them in position on a shaft. 
I should just start experimenting and doing this empirically. So thanks for giving me the confidence that direction can be predicted with a little magic.
When I get something. I'll post it.-Keith

[shagrug]

      The new url worked fine, and I have seen the video now. Interesting project, and I understand what you are doing much better now. It should be possible I think, and probably several ways to do it. (various motor types-styles) I'm already curious about how you will power it, and switch it on-off-direction.

      I forgot to mention one more thing on the chip position relative to the coil... On all of my motors you see the chip on top, and the coil connection pads-ground pad facing toward the front-prop end. If you turn the chip around 180 degrees with coil-ground connections rearward, the motor runs in the opposite direction. The chip doesn't have to be attached to the coil either, it can be quite some distance away. (as long as the magnet still triggers the sensor)

      The "pulsed shaking" you mention can happen if the chip is way out of place or if there is interference with metal (steel) near the coil//magnet. It's just the chip trying to restart the motor over and over.

      I had a look at the JinLong website at the new 3mm diameter motor you mentioned- thanks for the info! I'll check into getting a few to test. The page is here for others interested:
http://www.vibratormotor.com/cyl-table.htm

      That reminds me to mention the Shicoh 2.8 mm pager motors I received last month from Palomo, another reader of this thread.

[shagrug]

      I had mentioned searching for the 2.8mm Shicoh pager motors back at the very beginning of this thread, and to my surprise I have three new ones now. Many thanks to Palomo for this recent trade, I am very happy to finally see one up close- soon as I can find the time I'll test run one (or two) and post the results here.

      I've been searching online for details on them- two different case versions were made and I'm not sure which I have yet. The later ones were said to run better. I've checked the ohms on one, which read 24.6 or so. I may need to attach wires for a good reading, it varies with shaft position. Also from what I have read it needs a 6:1 gear ratio for flight, which sounds right with the high ohms.

      I'll keep looking online for more detailed info on them- and some photos of one apart. I sure don't want to open one of these new ones up since they are so rare. Anyone reading have the datasheet by chance?

[Keith55555]

I have a lot of horizontal and vertical 'real estate' for what I'm doing and as I've said, this is just about 'height' (sounding like another magic trick yet?). So I'm powering it with a LiPo from PowerStream (http://www.powerstream.com/thin-lithium-ion.htm) and that can go beside it to maintain the height needed.

As for 'switching'; as lovely as it would be to get away with a one chip solution with an Allegro type I.C., I want to experiment using a linear hall effect device (panasonic makes some low voltage sot-23 package devices. This way, if there's any trouble with 'direction control' I might be able to play with intensity from pole position as a last resort to get reliable direction out of it. I'll feed that to an A/D input on a PIC. I'm hoping to get away with a 10F part (sot package) but that I think is unlikely and may go to a 12F part because my compiler is limited with the 10F parts. I'll use a full H-Bridge for coils and there are a few parts that have duo-full bridges in SMD. I don't think I'll run out of I/O on the 8-pin PIC since it's getting a wireless, serial link to control command things. I tested an ANT version of the Nordic wireless chips and they are hearty and only have a few 'gotchas'. See the attachment for the test prototype. I CNC all my boards so I have a lot of leg room for starting over if I mess something up.

Off topic a little,  I'm told the only way to get one of those 3mm motors we mention, from stock, is to get one with an eccentric weight on it. You can get them shipped from China but the last time I did that, they charged my Fedex account for $60 because they ship them fast. So I'm getting them with an eccentric weight on them and I'll have to cut it off.

Lucky dog, you; re: 2.8 Schicoh. I've tried for years to get those 2.8's. Indoor Airplane had them for a blink of an eye. They still list them on their site.-Keith

[Keith55555]

So I received the JinLong motor. Out of stock, it comes with a 'carrier'. That's welded on and quite well. Will have to inspect how to remove. I could grab an area that allowed me to subtract the carrier's addition to the size. Indeed it's 3mm high but it's 4mm wide. So I think the Schicoh still has it beat. Did not weigh yet. Have to remove that weight as well. Here's a pic. -Keith

[shagrug]

Hi Keith-

      Great stuff, thanks for posting.

      I found the Power Stream lipo's interesting, and nice to know they have samples available to test. I can't help but wonder if they would be easier to cut down smaller than a Full River cell. (for sub gram planes)

      On your project, wow. I'm impressed with all the work you are putting into it, it's much more complex than I imagined. Wireless control would be cool and keep everything well hidden. I am wondering if you have considered stepper motor designs and controllers? (something I am now learning more about)

      The case on the JinLong does look way overdone in the photo. I'm guessing the big weight is just a press fit and easy to remove. I really can't tell if the case can be removed from what I can see in the drawings on the datasheet. I'd sure like to see more of it if you can get it apart. Also to know if it is 30 ohms (+-6) as stated, across the terminals.
http://www.vibratormotor.com/vib/Z30C1T8219651.html

[shoutchen]

here is the company web sire. these pages have a simple drawing
and some specs on some of the the 2.8mm motors

http://shicoh.com/vibration_1.htm
http://shicoh.com/vibration_2.htm


SteveH

[shoutchen]

I guess Keith already posted one pix of the 2,8 mm motor a couple posts ago..
Sorry to interrupt..

SteveH

[Keith55555]

I was able to remove the 'mount' on the "3mm", JinLong motor without any damage. Turns out it's not a 'weld' but a very aggressive epoxy( you can see some remaining in the attached photo). I started grinding off the apex of the sides to the bottom and once I saw some fissure, I put a fingernail under the top lip and one side popped off. Heat of grinding was, I believe, the major influence.
The eccentric weight is crimped on. If you grind the small area crimped, it's just a matter of restraining the weight and tapping the shaft in the hole in the opening of the weight witha fine tool end (I used a broken end mill with a diameter less than the hole).
Haven't done any measurements save for dimensioning it after the 'surgery'. 3.03mm but only in height. The other side is about 4mm.-Keith

[shagrug]

I have received four samples of the JinLong motors  from their Seattle warehouse. I'll post a few more photos-details on them along with test run comparisons, comparing them to the 2.8mm Shicoh and 3.2mm F7 Shicoh.

Weight (with weight/clip removed) is almost the same as the 3.2mm F7, .about .284g and .296g with the F7 being lighter. It has a obround case (3mm x 4mm) and the shape sure seems to add to the weight. The 2.8mm Shicoh weighs .163g or almost half what the other two weigh. It could probably be less than half if some of that long shaft was cut off. The shaft diameter is .023" on all three motors.

The ohms are a bit hard to read on all three motors (normal for pagers) with each movement of the shaft changing readings. Best I can do for each is about 24-26 ohms on the 2.8mm Shicoh, 36-38 on the 3.2mm Shicoh, and 34-36 on the 3x4mm JinLong.

You can get the JinLong 3x4mm samples from this lady for $5 each, plus a bit more for shipping (by mail). I tried to get them for free but no luck, seems they don't have many on this first production run.
claire.swe@jinlong-machinery.com
 
Keith's advice on removing the weight works well- just grind off all of the crimped area and then push or drive the shaft out. I used a bit of .020" drill rod for a drive pin, and it came off easy. I found the clip even easier to remove- just by using a razor blade to pry it off, starting at one corner.

The two solder pads on the bottom could be ground off to make it just a wee bit lighter and smaller, but I won't do that for now. Next is to wire it up for a test run.

[shagrug]

The usual test runs with 32mm prop (plantraco), three runs each with best one posted. The 3.2mm F7 Shicoh did a just little better than the 3x4mm Jinlong, with more rpms and less current draw. The 2.8mm Shicoh did better than I expected considering it's size and weight.

Starting off with the 3.2mm F7 Shicoh-

3.2mm Shicoh motor test run

Next the 3x4mm Jin Long-

3x4mm JinLong motor test run

And last the 2.8mm Shicoh-

2.8mm Shicoh pager motor test run

[shagrug]

      I had a quick look inside the 3x4 mm motor after test running, as I wanted to confirm that it was a normal "can" type motor and not a coreless design. The two crimps that hold the brush cap on were easy to open with a x-acto blade tip and it went back together just as easy.

      I liked the quality of construction, the bearings appear to be bronze and the brushes look thick enough to last for a while. I'm thinking the armature could possibly be rewound with thicker wire for more power, as the solder connections on the commutator are easy to get to.

[shagrug]

      I think I've already mentioned that the A1442 chip is soon to be discontinued by Allegro, the reason said to be "low volume sales" in the following document:
http://media.digikey.com/pdf/PCNs/Allegro/Pre-EOL.pdf

      Tonight I see that Digikey is down to 936 of the A1442's left in stock, so they may all be gone before long. I bought some more a couple of weeks ago, and the stock showed 2,160 at that time. Someone has made a big order recently, from the look of it.

      Lately I have been busy testing a couple of new chips from Diodes Inc. that will be good (mabe better in some respects) than the A1442. Both of these new chips are available in "legged" versions as compared to the smd A1442 type- this makes soldering the connections very easy. Note too there is no reverse polarity protection as with A1442, so be sure to double check pos-neg before powering up.

      First is the AH5795 (digikey part # AH5795-WU-7DICT-ND) that is rated at 500ma on their website. When I first received one I did a quick test to see if it had more output than the 350ma A1442- and it does. I tested this by hooking it up to a 4x5mm test motor (no.4 to be exact) and changing out the prop bigger and bigger to increase the loading (up to a U-80). I saw it go to 450-500ma before overloading and shutting off, so this pleased me. I have to mention here too, that it takes a bit longer to reset-restart than the A1442. (a 1 second or less delay) But otherwise it runs the same. You can see in the photo I used a jumper between the positive leg and pwm leg for speed control on the quick test, as detailed in the datasheet here:
http://www.diodes.com/datasheets/AH5795.pdf
      I'll eventually do a detailed test on using the pwm input of this chip with a rx pwm signal (or two), as I'm very curious to see the difference. The "FG" leg is not connected to, being for tachometer purpose on pc fans. This AH5795 is also available in a slightly smaller smd version (AH5795-FDC-7DICT-ND) if you should want.

      The AH5798 is next to test, different having a lower website power rating of 300ma and no pmw connection. I've noticed the datasheets list a higher rating than this, so I'm not sure which is true yet. (these might be for a direct short) the datasheet is here:
http://www.diodes.com/datasheets/AH5798.pdf

      It is also available in the legless smd version, but I didn't get any of them to test. With the lower output rating than the AH5795 I'm thinking it won't test nearly as well, so I only got 4 of the legged version to test.

[shagrug]

      This is something I have gave much thought to this winter- a way to get the perfect-optimum coil for your motor with a "rewindable test frame"...

      It is based somewhat on Ivan Pepe's excellent idea of a fiberglass box frame that can be rewound (posted back on the bottom of page 9 here, reply #134), but this frame is not meant to be the actual motor.

      What it will do is allow you to use the given size magnet in a frame that lets you adjust and fine tune the diameter gauge wire used and length, matching it to the model's need.

      Since the weight and power of the coil go up "hand in hand" when increasing, the weight of the coil can vary quite a lot. If the power goes up past certain points then the lipo weight goes up too- feeding the bigger coil. Ok, so it's a normal thing and to be expected- for example my recent 2.5x4mm test motors run well with the coil weight one to two times the magnet's weight. The motor might weigh say .28g up to .42g, all depending on the wire's weight.... But you will want no more weight than needed if flying minimal weight indoor or submicro planes, and it would sure be nice to test the motor beforehand with the same prop/rx /lipo you plan to use.

      Below are photos of a test frame and how it works- remember the coil wire is not glued anywhere
but just wrapped around the frame. Once the desired running characteristics are reached you can unwind the wire to measure the exact length. Note too that you can tap into a uncut wire length (on the spool) to check ohms or even test run, just by tinning at a spot. If it doesn't run well you can just unsolder the chip leads and wrap on more (or less) wire. It will also let you feel the heat of the coil if it is getting hot- always important not to melt the glue used to hold it all together.

      This frame design can let me swap the new controller chips out easy when testing, using the brass screw. I actually have two frames built now, one using the 2.5x4mm magnet and one with the 4x5mm magnet. (I'm busy making some more motors of both sizes now) Next will likely be a smaller frame for the 2.5x2.5mm magnet. The chips will work on either frame, having the same hole pattern. I'll have more chips to test later on (like the Melexis mlx90283, and some from Rohm) along with the h-bridge boards that Ivan made. So there is some length to the garolite strip the chip is glued too, to allow for the bigger boarded parts.

      The garolite parts are slightly thicker than what I use in a motor, and I used lots of CA glue to hold it together since the weight doesn't matter. The plywood handle is about 3.5" long and will need to be the exact width of the magnet (plus clearance) I used brass screws for the chip hold down since they are nonmagnetic. The bearings are double the normal length, as the sloppy overlapping winding will take up more space. (or if you need more than expected length) I also use a shaft protruding at both ends, thinking it would be easier to wind plus mount props. (tractor or pusher, in either rotation direction)

      If you are using a bigger prop and gear reduction, the frame sides could be longer and the shaft/bearings/gear added into the tip end. This would not be too much added work, and I think would pay off with the most accurate test results.