Sensorless Sync Issues (was Phase Combinations, Aotema)

So I'm in the process of upgrading my 'Basket Case' cruiser with a new Lyen 12 fet sensorless.

The motor is an Aotema front hub, nominally 750 volts. The stock no-name controller supplies 35 amps. I've been running 12s lipo, just about 50 volts off the charger. That should be about 1750 watts. (There's no CA plug, and no manufacturer's info on the controller, so I'm going on hearsay.) It tops out at about 30 mph and pulls pretty well in my hilly area.

The Lyen sensorless is set to 52 amps. CA is showing just under 2500 amps constant, with peaks to 2700. Here's the catch: It feels much slower than the stock controller, with less torque. When I give it full throttle, the power seems to oscillate, making the bike surge slightly.

Something doesn't seem quite right.

Rather than start right in bugging Lyen with (more) noob questions, I decided to try some different phase combinations. I had initially connected them Y:Y, B:B, G:G, because, well, it seemed to make sense.

I just spent a half hour going through all the combos (only six for a sensorless setup, thank you jesus). Long story short: three combos make it go, and three make it go backwards.

The three forward combos are indistinguishable in performance. So what am I missing? Is there one 'right' phase combination, or does it not matter?

For posterity, here is the 'data' I have collected thus far.

Lyen 12 fet sensorless to Aotema hubmotor

Cont:Motor Result

Y:Y, B:B, G:G Goes
Y:B, G:G, B:Y Goes
Y:B, G:Y, B:G Goes
Y:Y, G:B, B:G Goes backwards
Y:B, G:G, B,Y Goes backwards
Y:G, G:Y, B:B Goes backwards

The right combination would use the least amps or watts at no load, so an ampmeter or wattmeter would help. Sounds like it may be losing synch at the higer rpms.

Cool, so just spool it up full throttle no load, and watch the CA? Lowest power draw wins?

*slowly* up the throttle. Do it suddenly with the wrong combo and you could blow FETs in the controller. Not likely but possible.

Try it without the CA plugged in to be sure it's not a setting in the CA overriding the system.

If it is runnin g, it may well be correct..you may have halls and phases wires correct and it could be the CA limiting..

BUT

just to check your wiring it is a good idea to try all combinations
Use the spreadsheet

http://endless-sphere.com/forums/download/file.php?id=77863

and run through all combo's//making notes int eh Notes section, then choose the best.

And another but...the pulsing sound more liek a CA limit kicking in and out.

check the CA manual, go through the current, LVC and Speed limit settings settings etc, and set them so they can't be limiting..so
Current 99 amps
Speed 99mph
LVC...20 volts

OR just unplug the CA as previously suggested

If you still get the pulsing, ( and assuming the motor/controller phase wiring is correct) then it is an LVC or current limit on the controller.

You say a Lyen controller, so,did you get it with the programming lead? Maybe you need to adjust controller settings too

NeilP said:

....just to check your wiring it is a good idea to try all combinations Use the spreadsheet
http://endless-sphere.com/forums/download/file.php?id=77863
and run through all combo's//making notes int eh Notes section, then choose the best.

Click to expand...

Sorry Neil, but it's never a good idea to try all 36 combos. The file encouraging and enabling that silliness needs to be deleted from the forum, because it only leads to confusion and frustration, and makes a simple task seem daunting. Every phase combo has 1 good hall combo out of 6 possible, and every hall combo has one good phase combo out of 6. You just don't no till you try whether that good combo is forward or reverse. Whichever is easiest to change (hall or phase) try all 6 leaving the other set static. If you get a nice smooth reverse, just swap 2, any 2, and only 2 of the static set to get forward, and go back to trying the remaining 5 (the current one on a valid reverse won't work) and get your valid forward.

Many motors will spin roughly with an invalid combo, which I call a false positive, but every false positive I've seen is wrong direction. This enables me to always get the right combo within 9 tries, and I average about 5, no notes or anything. I've done it with so many motors I could do it blindfolded no problem.

John

Well we are all different John

For me trying to do it without some form of notes as to what I tried already is doing it blindfold

Just because you have the sheet in front of you with all 36 combos already written down does not mean you have to try them all.

What you say is correct in that within a few tries you will usually find a combo that works. But I usually like to try them all in the hope of getting a combo with as many matching colours as possible

We all work in different ways. If you can mentally remember which combos you tried, that is great, I can't, or get interupted
The more experience you have and the more times you do it yes, I am sure it is something that can be done without notes, but i think to someone who has not done it before, having a logical set of combos written down to try is less daunting than swapping and remembering

Luckily this motor is sensorless, so in this case I only have to deal with six possible combos. More testing tomorrow with speed and current draw should hopefully give me the correct combo. Seat of the pants testing seems identical for the three forward combinations.

Neil, I had followed all your recommendations above except for removing LVC from the CA. It's set at 42 v for 12s lipo, so I suppose sag could be a factor. Will try setting it lower for my test runs tomorrow. Thanks everyone for your help. I think this is just a matter of tuning the controller and finding the sweet spot. Gotta get my PC out of mothballs to run the controller programming software.

Actually there are only two different wiring combinations on a three phase sensorless setup, all others are rotations of the two different ones. Rotations don't matter, no need to test them. Swap any two wires to reverse direction. Get direction right, you are done with the wiring.

I think the problem is you may have a geared motor running sensorless and the rpm is too high for the sensorless controller to keep up.
There is probably nothing wrong with the wiring unless you have a really loose wire somewhere.

NeilP said:

Well we are all different John

For me trying to do it without some formmof notes as to what I tried already is doing it blindfold

Just because you have the shert in front of you with all 36 combos already written down does not mean you have to try them all
What you say is correct in that within a few tries you will usually find a combo that works. But I usually like to try them all in the hope of getting a combo with as many matching colours as possible

We all work in different ways. If you can mentally remember which combos you tried, that is great, I can't, or get interupted
The more experience you have and the more times you do it yes, I am sure it is something that can be done without notes, but i think to someone who has not done it before, having a logical set of combos written down to try is less daunting than swapping and remembering

Click to expand...

Neil,
I have more faith in you than you do. I wouldn't have the patience to watch you go through that process, much less do it myself. If you want color on color, then by all means start that way with your static set of wires. You've got a 50/50 chance that it's forward rotation. Regarding remembering the combos tried, I don't even try to remember them. I just pick 1 color or 1 plug position as an anchor point. eg I start with my favorite color, blue. First blue on blue, and try the 2 possibilities of the other 2 wires. Then green on controller blue, and try the other 2 possible. Then yellow on controller blue, and try the final 2 possibilities. Blindfolded I'd have to go with positioning, not color, but it's the same.

The whole wiring simplicity thing didn't click for me until I heard that sensorless operation was just swap any 2 wires to change direction. I thought that it should be the same for motors with hall sensors, and it is. That means you only have 3 wires to mess with unless it's reverse.

I know what you mean John, about the patience thing...but I am mucking around with various motors and controllers here..some found in the dump etc.

I have a set of leads to connect to motor and controller, all with big easy to remove and swap bullet connectors on them..both the phases and the halls. So I am not having to muck around with the little pins on the multi pin connectors.

having them as easy to swap plugs means the testing of each combo is very quick..certainly less than 10 mins, more like 5 to try all combos. marking up the sheet with a tick for good, cross for bad and an F or R for forward or reverse, it does not take long

Alan B said:

Actually there are only two different wiring combinations on a three phase sensorless setup, all others are rotations of the two different ones. Rotations don't matter, no need to test them. Swap any two wires to reverse direction. Get direction right, you are done with the wiring.

Click to expand...

Thanks Alan. That explains it. I'm trying to make this more difficult than it is. Sticking with the same-to-same lineup for simplicity's sake.

The motor is not geared. It's an Aotema DD front hub, 750 watts nominal and unsensored. About as simple as it gets.

The surge was indeed from the CA. I had LVC set at 42 v, and it seems that I was possibly bumping up to that due to sag, even with a fresh pack at 50 volts and 10 ah. Runs great with the CA disconnected. Honestly, wondering why those things have such a fan club on ES, because mine has only managed to slow my bike down, clutter up my handlebars, and fail to give me speed readout or any associated functions like wh/mile. So far, the only thing it's done as advertised is cost me 120 bucks.

I'm sure this will turn out to be operator error in the end, but at this stage in the game it's frustrating.

Sorry, big apology from me for trying to confuse you with the spread sheet and phase combos hall combos. My fault completely for not spotting the word " sensorless" multiple times in your first post

As for CA not reading speed. Have you got the remote speed sensor type ? The one that has the seperate sensor and the magnet clipped to the spokes? If so you need to set the Pole Count in the setup menu to 1

If you are not reading speed and therefore miles travelled, then wh/mile will not give correct or any results

Did you buy the CA. From Lyen or a seperate purchase?

Love my CA's, but I only have two of the non DP ones. I never saw any real need for the other functions on most bikes. They can be handy, for instance to keep a weak battery from being whacked by a strong controller. But I have lipo for my strong controllers.

I've been hanging back, watching this thread, wondering what the real problem was besides the phase order. For sensorless, usualy it's just try any random combo, then see if it's forward. If not, swap the 1st and 3rd wires.

dogman said:

wondering what the real problem was besides the phase order. .

Click to expand...

just in case you missed it..the real problem was the CA..probably voltage sag and a to high CA LVC...and me leading the OP in the wrong direction because I did not read his first post correctly and started t talk about hall wiring

Sorry to hear you are having problems with you Cycle Analyst. I've had three of them on different bikes, and the first large screen CA I got has had some issues. It is still acting weird so I'll have to send it back and get it repaired. Apparently a few of them had a problem of some kind.

The other two work great. Give all the featured readouts and allow me to protect the batteries with Low Voltage Cutout and max current settings.

If the batteries or wiring have too much resistance it will cause excessive sag which will make the CA's low voltage cutout trip early. I don't consider that the fault of the Cycle Analyst. That's a problem in the system and the CA is the messenger bringing the bad news.

dogman said:

I've been hanging back, watching this thread, wondering what the real problem was besides the phase order. For sensorless, usualy it's just try any random combo, then see if it's forward. If not, swap the 1st and 3rd wires.

Click to expand...

The real problem was operator error, Dogman. I had limits set on the CA that I didn't think would affect performance, but they did. Simply unplugging the CA solved the problem. I then plugged it back in and lowered the LVC to 20 and maxed out the amp limit, which was set at 50.

Took a test run up to the local hardware store, and all of a sudden I'm hauling ass and have the data to understand why--I'm pulling ~3,800 watts at WOT.

The phases were never a problem. It's like my eBike went to the doctor for a sore throat, and I got a rectal exam.

That trip to the hardware store was for solder and shrink wrap. I spent the next couple of hours soldering 4 mm connectors to 10 awg silicone. Took a break to put my daughter to bed, then wired four bricks of lipo in series.

Plugged it in, and the CA flashed 99.8 volts. Just lifting the wheel off the ground and touching the throttle, I knew that I'd created a monster. I sat there for a minute or two contemplating my next move, and finally chickened out. I rewired the pack at 18s. Cycle Analyst showed 74.9 volts.

I wanted to see how the extra voltage affected my top speed, so I grabbed my GPS and rolled. I could tell this knew configuration is a beast before I even got out of the driveway. Way more torque than I'd expected. Wattage was nuts. First time I opened the throttle it fed 5,000 watts to my little Aotema hubby. I cruised along at super low throttle till I got to the first long, flat straight.

At my regular 12s/49 volts hot setup, I max at around 29 mph. I was still accelerating at 32.4 when it happened. BBRRRAAAAAPPP. A load grating noise, like grinding gears. It also felt like the motor lost power, but that could have been because I came off the throttle so fast when I heard it. I turned around and tried again, this time easing into it a bit. At just over 30 mph it did it again. BRRRRAAAAAPPP.

Everything up to 32 mph feels great, so I cruised home, poured a drink, and logged on to the Sphere to report my latest noob adventure.

Anybody have a diagnosis for the speed burp?

When I first read this I though ..Axle startin to spin...but then I though no , can't be. If it happened at high acceleration rather than high speed I would have suggested the axle slipping but as it is at high speed..umm..mechanical or electrical cause do yo recon?

How does the motor feel, can you easily spin it by hand? any play in the bearings? maybe balance issue and or harmonic vibration that only happens at a certain speed.

Sounds like someone has a death wish. For either the rider or the motor.

Small geared motor in front wheel. High power, high speed.

Perhaps Motor RPM is too high for the sensorless controller which is losing sync?

Gulp...gerread motor forgotten about that

Alan B said:

Sounds like someone has a death wish. For either the rider or the motor.

Small geared motor in front wheel. High power, high speed.

Perhaps Motor RPM is too high for the sensorless controller which is losing sync?

Click to expand...

No, it's not geared. It's Direct Drive. Lots of people running these and similar at 18s lipo/75 volts and beyond. The high volts/amps was a test ride situation, just short bursts to gauge performance with the new pack configuration. I wasn't WOT for more than 30 seconds, if that. Motor has been cool to the touch at every juncture, including much longer pulls uphill, like a half-mile 8 percent grade at 48v/3,700 watts. I'll likely dial down the controller programming when/if I'm ready to move full-time to 18s. Looking to move to a rear hub soon on a different frame, so this one will go into semi-retirement in a couple of months anyway.

Harmonic imbalance--could be. It has a bit of drag when spun by hand. Nothing you'd really notice while moving under pedal-power. Bearings, dunno. Will have a look.

I think my next step is to limit current while running 75v/18s to see if it'll go past 32 mph at more reasonable wattage. And though the axle didn't budge, I'm going to put a second torque arm on too.