ELECTRIC TRANSMISSION DELTA /WYE and SERIE/PAPALLEL stator

[neptronix]

Just saw this video ( i'm new to all this )

INSANE O_O
Reminds me of Honda's oldschool VTEC. Sudden burst of horsepower, and no torque.

Okay, go on with your intelligent discussion

 

[aaannndddyyy]

Just finished wiring up my conhismotor 350 48v mini kit http://cgi.ebay.co.uk/350W-ELECTRIC-BICYCLE-E-BIKE-RETROFIT-KIT-MINI-MOTOR-/330441744921?pt=LH_DefaultDomain_2#ht_12876wt_930 in the star delta mode using a 3DPT 30a 24vdc http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=180545672184&ssPageName=STRK:MEWNX:IT#ht_500wt_721
and managed to get the 3 wires through the axle without disconnecting and removing the harness here is a before and after image

and a small video of the result.
http://www.youtube.com/watch?v=1Ro2eD88StM
I still need to wire up the dc-dc converter and a switch. http://img841.imageshack.us/i/b07yklq2kkgrhqzhqew5jdj.jpg/
Just like to say what a great thread this is by following what has been posted managed in one day to inprove the top end nls speed by 73%

Tue Sep 14 2010
Just finished converting my 1000w 48v motor with pics. http://img138.imageshack.us/g/wirediagram.jpg/
As a couple of hull wires had been damaged and repaired at the axle a while back, a complete removal of the wires was necessary, I am also planing on running at 72volts so beefed up the wires inside the motor and ran 10awg from the axle to the 3dpt switch.

 

[rkosiorek]

by having your relay outside of the motor it should be subjected to less vibration and shock. that may resolve some of the relay failure issues.

rick

 

[steveo]

so who will be the first to run 12 wires out of there x5 motor

i agree with you rick... when i did star delta @ 150 amps .. 100v .. i never had any relay failures.. and to top i of i used a 3" tire in the rear .. so the bike rode very smooth ..

-steveo

 

[Doctorbass]

Oh.. I must sa... YES I had relay failure in the serie parallel setup.. contact of the relay blasted and alot of carbon on them... some was stick to close position!!!.. making short when i swtiched them to serie to parallel...

High power + relay inside = BAD idea!!

The first to put 12 wires IS MARKCYCLE with the MH602. He posted pictures on his blog.. impressive stator !!!

Doc

 

[markcycle]

Oh.. I must sa... YES I had relay failure in the serie parallel setup.. contact of the relay blasted and alot of carbon on them... some was stick to close position!!!.. making short when i swtiched them to serie to parallel...

High power + relay inside = BAD idea!!

The first to put 12 wires IS MARKCYCLE with the MH602. He posted pictures on his blog.. impressive stator !!!

Doc

Did you switch the relays under load?

I'm bring the throttle to zero then switching so far about 60 miles and no problems. I'm working on a circuit that when I switch it brings the throttle to zero first waits 100ms or less switches then returns the throttle. Switching with current in the phase wires will destroy 12 volt auto relays really almost any small relay.


Mark

 

[Doctorbass]

Oh.. I must sa... YES I had relay failure in the serie parallel setup.. contact of the relay blasted and alot of carbon on them... some was stick to close position!!!.. making short when i swtiched them to serie to parallel...

High power + relay inside = BAD idea!!

The first to put 12 wires IS MARKCYCLE with the MH602. He posted pictures on his blog.. impressive stator !!!

Doc

Did you switch the relays under load?

I'm bring the throttle to zero then switching so far about 60 miles and no problems. I'm working on a circuit that when I switch it brings the throttle to zero first waits 100ms or less switches then returns the throttle. Switching with current in the phase wires will destroy 12 volt auto relays really almost any small relay.


Mark

I always released the throttle. What caused that is the vibration of the motor and the spring loaded NC contact that had not enough force in the spring i think.

Doc

 

[fechter]

Oh.. I must sa... YES I had relay failure in the serie parallel setup.. contact of the relay blasted and alot of carbon on them... some was stick to close position!!!.. making short when i swtiched them to serie to parallel...

I want pictures!

If there is too much current upon closure, the contacts can get welded. They can fail in other modes too.

Ideally, there could be a circuit that measures the current and does not allow the relays to switch unless the current is near zero, like a zero crossing detector. Things could get pretty complex though.
Getting 12 phase wires out of an axle could be tough.

 

[Arlo1]

So could this give us better regen?

 

[amberwolf]

Ideally, there could be a circuit that measures the current and does not allow the relays to switch unless the current is near zero, like a zero crossing detector.

Problem is, vibration doesn't ask if it's ok to switch. :lol: I'm pretty sure that's what caused the relay issues for him (and possibly others if they've tried this with similar problems).

If it vibrates or bumps in the right direction at the right moment at sufficient amplitude while the motor is under power (or regenerating during braking), one or more of the relay contacts could bounce, arc, and burn.

 

[rkosiorek]

I'm pretty sure that the relay failures are due to the mounting location. in the motors with the relays mounted internally the relays take quite a pounding as the wheel bounces up and down.

on the older "Dual Mode" Crystalyte motors (408/4011 for example) i had the relays mounted on the frame. no problems there. but as soon as i switched to relays inside the motor... problems. on the frame the relays are isolated and protected from the worst of the shocks and vibration. inside the ub they are not.

now the problem of how do i get 12 decent size wires out of the case is the next item to solve.

rick

 

[John in CR]

now the problem of how do i get 12 decent size wires out of the case is the next item to solve.
rick

Rick,
That's the easy part. Find a bearing with a significantly larger ID than the current one. Then make a steel bushing with an ID to fit the axle, and OD to fit the bearing. Drill holes in the bushing to pass the additional wires. Be sure to secure the bushing to the axle in a manner so it can't spin.
John

 

[markcycle]

now the problem of how do i get 12 decent size wires out of the case is the next item to solve.
rick

Rick,
That's the easy part. Find a bearing with a significantly larger ID than the current one. Then make a steel bushing with an ID to fit the axle, and OD to fit the bearing. Drill holes in the bushing to pass the additional wires. Be sure to secure the bushing to the axle in a manner so it can't spin.
John

Larger bearing ID same OD = hard to find may not exist.

I wonder if I offered for sale my side cover for one side with a bushing for the shaft as you said John if it would sell

Mark

 

[rkosiorek]

I think that you would find at least one customer for your side cover Mark. I'm thinking DoctorBass.

Mark, Could you post a couple of pictures? or a link to some pictures of your modified cover and bushing?

Rick

 

[oldswamm]

Hi,
Doctorbass,
Are you using the motor with the spin on brake? Are you aware that the bolt on brake motors use a 5mm larger bearing (at least according to ebikes)?
Also, if you fly cut the wire channel in the shaft, instead of what appears to be a ball cut channel in the oem, you could bring out more and/or bigger wires without significantly weakening your shaft. Might have to modify or replace the plastic thing.

(edited out an idiotic idea)

From what I've read, I don't think the relays in the hub are ever going to be dependable enough tor short term (racing) or long term. AND, I don't see a reasonable way to do it electronically, (back to back 'intelligent', floating (optoisolated), mosfet circuits, back to back floating SCRs?). Wish the guy who earlier claimed to be electronicly switching delta/wye would say wheather he was using SCRs, or what.

(edit) Doc, I wasn't trying to insult your relay in the hub idea, I just don't think it would be practical for most delta/wye or s/p setups. If you're going for records, you only have to keep it together for 1 run after all. I intend to run a rear suspension, which will just make it worse for hub bouncing, so it definitely won't work for me.

Bob

 

[rkosiorek]

Hi,
Doctorbass,
<snip>
Isn't this motor wound as 12 windings in a series for each phase (as you progress around the stator)?
Couldn't you cut the wires and seperate them into 2 halves (on the bottom, opposite where they normally come together), which you could series/parallel switch, so a 5304 could be the equivalent of a 5304/5302. Or cut them in 2 places for a 5304/5301-1/3 (if you could bring out 16 wires ).
(And as a disclaimer, I've learned as much motor tech in the month I've been lurking on ES, than the previous half century, so I DON'T necessarily claim understanding.)

From what I've read, I don't think the relays in the hub are ever going to be dependable enough tor short term (racing) or long term. AND, I don't see a reasonable way to do it electronically, (back to back 'intelligent', floating (optoisolated), mosfet circuits, back to back floating SCRs?).

Bob

bob,

yes the 53XX motors have 12 coils in series. each of these coils is made up of XX number of turns. for example the 5304 motor has 4 turns around each stator pole, repeated 12 times. the only way to reduce the number of turns would be to cut each individual coil somewhere in the middle, say after 2 turns on each individual pole. that is a lot of cuts. there would be one cut through half of the strands on each of the 36 total poles. physically this would be practically impossible. each coil is made up of a number of strands that are randomly wound around the pole. because of this randomness it is impossible to see which of those individual stands is on its 1st, 2nd or whatever trip around the pole. that is why it is only practical to deal only with the ends of the windings and multiply the number turns in the series/parallel scheme.

rather than try to identify which individual strands is what and which ones to cut or leave alone it would be more practical to just cut/strip off the old wiring completetly and re-wind it anyway you choose.

rick

 

[oldswamm]

Hi Rick!

I understand all that. Obviously we wouldn't be reducing the number of turns per pole, but we would be halving the number of turns presented to the controller. We get 1/2 the voltage for any given current, or twice the current for a given voltage. Another way of saying it is that we halve halved the impedance, which is what we are trying to do if we go to the work of rewinding for 2/4.

Bob

 

[oldswamm]

Me again, still thinking.
What got me started thinking on the 5304 to 5302 etc conversion was a previous poster who said that you could convert a 5302 to a 5304, etc, but not the other way. (parallel to serial, but not serial to parallel) Naturally, that got me to trying to think of a way. (Don't like to be told I can't do something.) I even (still) occasionally try and defy the laws of physics.

I can't see why one winding (for one pole) cares whether it's neighbor is in parallel or series with it. All windings from one phase are, well, in phase (or 180 out), after all.
As an extreme, I can't see why all 12 couldn't be wired in parallel. Be a real bitch to wire, and I don't know how you would get the 2-O cable into the motor, but should give 120 mph on 30v with the 'right' controller. :lol: Sorry, can't help myself sometimes.
Biggest problem I could see with trying to make the 04 to 02 mod to an already wound motor, would be if you couldn't avoid a 1/2 turn difference between the 2 halves.
Guess I'll try and shut up till someone convinces me I'm missing something (again).

 

[Toorbough ULL-Zeveigh]

(Don't like to be told I can't do something.)

not trying to tell u that u can't...
but let me know if u ever figure out how to un-bake a cake.

same difference.

 

[oldswamm]

Ok, all the experts have checked in. It CAN'T be done. What I asked is WHY.

Cutting 3 wires and paralleling 2 sets of coils for each phase is NOT the equivalent of unbaking a cake!
(I drew an image of a 6 slot motor with the coils rewired from series to parallel, but can't figure how to insert it.)

If I cut the connection between 2 coils that are in series, and wire them in parallel for 1/2 the inductance, why won't it work? If I feed those 2 coils 1/2 the voltage at twice the current, they see exactly the same 'signal' as they did in series. If each coil has exactly the same current and voltage flowing through it as it did in series, WHY would ANYTHING be different other than the V/I ratio?

If all you big brains agree that paralleling coils that were formally in series wont cut the impedance by a factor of 4 PLEASE tell me why. Or is there some OTHER reason why paralleling the coils WONT work.
Don't just tell me it won't work, say WHY.

 

[oldswamm]

I don't think anybody has understood what I'm trying to describe here. Certainly not the 2 who replied.
Please take my word for it that I am not an idiot, at least long enough to read and try and comprehend what I'm saying!!!

Each phase consists of a string of 12 coils, If we cut them into 2 strings of 6 coils, then connect the 3 cut ends of the sets of 6 coils whose other ends are the phase inlets, to the wye, then connect the 3 other cut ends, which are already connected to the wye, through their respective sets of 6 coils, to the proper phase inlets, we have cut our inductance and resistance by a factor of 4, the same as if we rewired an 04 to an 02. 'All 12 coils are seeing the same current and voltage as if we had done that rewire. They are all phased the same as before.

Can anyone who understood the above paragraph tell me why a 5304 rewired as described wouldn't be functionally the same as a 5302? And I KNOW it will still have 4 wraps around each pole. And NO you do not have to separate any strands from each other!

Look at it this way.
A 5302 has 2 wraps of 10 strands of 21 ga wire on 12 stator poles.
A 5304 has 4 wraps of 5 strands of 21 ga wire on 12 stator poles.
A 5304 that has had the described mod, has 4 wraps of 5 strands of 21 ga wire on each of 2 strings of 6 stator poles.
Each string has 4 wraps times 6 poles for a total of 24 wraps, the same # as a 5302 has for a string of 12, but with only 5 strands. Now when we connect them in parallel, the 2 strings will have a total of 10 strands, the same as the 5302.
Can't anyone see that that would be functionally the same? The phasing HAS to be the same. If the motor switches phase 180 on each pole, were still cutting it at an even # of poles.

The only reason I see for the factory, or hand winders for that mater do it in series is convenience. Fewer wires to deal with.

I enjoy learning. If you can explain WHY my idea won't work, I will thank you for enlightening my ignorance that slight amount.

Bob

 

[oldswamm]

Well, some of the most respected names on this forum have read my posts and decided I wasn't worth explaining it to.

I guess I'll erase all my messages and apologize to the forum for wasting your time, but I really wish just ONE of you would take the time to explain WHY the idea is so idiotic it's not worth commenting on..

You don't have to go into detail here, just post a link that will explain WHY a motor won't work unless all the windings of any phase are in series. It's been over 30 years since I studied motor theory, and I honestly don't understand why the windings can't be paralleled. I always just assumed it was because they didn't want to make 2 splices for every winding, which you have to admit would make motor assembly WAY more difficult.

It seams SO obvious to me that it would work, and I've given it a lot of thought and can't see why a motor won't run unless all it's windings are in series. The only 2 readers who have bothered to take the time to tell me I'm an idiot, obviously KNOW why it won't work, but for some reason don't want to give me, or point me to, the explanation.

Bob

 

[Arlo1]

Well, some of the most respected names on this forum have read my posts and decided I wasn't worth explaining it to.

I guess I'll erase all my messages and apologize to the forum for wasting your time, but I really wish just ONE of you would take the time to explain WHY the idea is so idiotic it's not worth commenting on..

You don't have to go into detail here, just post a link that will explain WHY a motor won't work unless all the windings of any phase are in series. It's been over 30 years since I studied motor theory, and I honestly don't understand why the windings can't be paralleled. I always just assumed it was because they didn't want to make 2 splices for every winding, which you have to admit would make motor assembly WAY more difficult.

It seams SO obvious to me that it would work, and I've given it a lot of thought and can't see why a motor won't run unless all it's windings are in series. The only 2 readers who have bothered to take the time to tell me I'm an idiot, obviously KNOW why it won't work, but for some reason don't want to give me, or point me to, the explanation.

Bob

Im thinking it is a combination of things for one when you run them in paralel it will be more load on the controler and if you go from 12 in series to 6 in series then paralell the other 6 it is 4 times the load on the controler!
Another thing is it doesnt matter the winding style in the motor it will still have the same acceleration! The only thing winding determines is top speed!

 

[oldswamm]

Im thinking it is a combination of things for one when you run them in paralel it will be more load on the controler and if you go from 12 in series to 6 in series then paralell the other 6 it is 4 times the load on the controler!
Another thing is it doesnt matter the winding style in the motor it will still have the same acceleration! The only thing winding determines is top speed!

Thanks for the comeback Arlo.

To the first, yes, it would present the same load to the controller as a motor with half the winding count (and double the strands). A way to get 5302 speeds from a 5304.

To the second, I've decided a 5304 in a 16" MC rim, with no winding mods is what I need. There's pretty much no way I can go over 50mph where I live anyway, (and then I would have to use the brakes right away ) This bothers me, mostly because I never leave ANYTHING stock.

I was just presenting it as an alternative parallel/series scheme, which SHOULD be in this thread. (And I've read most of, and been convinced by, LiveForPysics's 'current = torque' rants)

It would also allow someone with an old 5309 to convert it to a 5304.5 , or 5303 (by making 2 cuts, and paralleling 3 strings of 4 windings). And, I realize that not just everyone could get good short free connections to the short ends that would result.

 

[oldswamm]

I've searched. Is there another thread where electronic delta/wye and/or series/parallel switching have been discussed?

I would like to see what ideas have been suggested, or even tried, and (obviously) rejected.

Triacs/SCRs would hit the efficiency at lower voltages, but wouldn't they work for you guys running 100v+ at least for competition? Would still burn lots of energy if/when cruising. A floating circuit with back to back mosfets could be built, but would be fairly complicated.

Just curious. (I've got 50yrs experience with relay failures, and avoid them if at all possible!)

Thanks,
Bob