issues finding hall sensor combenations using svmc72150 and revolt 120 pro motor

itsik

1 mW
Joined
Mar 8, 2021
Messages
19
hello!
i had been trying to get this controller to work with 2 motors i have of the same type ( revolt 120 pro)
the issue is that i cant find a hall sensor wiring combination that results in a smooth operation of the motor
i went according to the manual and used the hall test feature, half of the time it results in a fail and the second half it finds a combination that results in a none smooth spin
wired clunking noises at half thrutol while the rpms jumps around, gettery top end, and if you press the peddle quickly, wired electrical noises that i think means the motor and controller went out of sync
i attempted looking for the currect combenation myself, going thorow 36 diffrent hall and phase wires combenations but didnt find one that worked
i also played with the available kp parameter, it affects it but i can't seem to get it to be perfect
if i apply a load to the motor, the motor has some power, but no were near what is supposed to be
the motor and controller do not get hot, and the amp draw from that battery is about 50 amps tops, while the dc limit is set to 150 amps
i uploaded some vids showing the issues
https://youtu.be/kq3GvhGnFF8
https://youtu.be/KzbHwU6juq8
https://youtu.be/M4rLZ5_QPKk
https://youtu.be/N0Ii_4olRig
https://youtu.be/VJI3suWyGBA
https://youtu.be/obRZOTsZqXQ
https://youtu.be/OiLeEfZfKrI

I hope you guys can help me tune this issue out
thank you!
 
i saw this table,
before i got the svmc72150 i was trying to get this system to work with a cheap greentime branded controller (80 amps)
and i tried to find the combination using this table, but it always brougth me to a combination that didnt work :cry:

is there more tuning opthios on the svmc72150 other then the kp and angle?
 
itsik said:
i saw this table,
before i got the svmc72150 i was trying to get this system to work with a cheap greentime branded controller (80 amps)
and i tried to find the combination using this table, but it always brougth me to a combination that didnt work :cry:

You may have a bad controller, bad wiring of that controller, bad hall sensors, bad hall sensor wires, a loose hall sensor.

https://ebikes.ca/learn/troubleshooting.html
 
yea i figured that might be the issue, i left out some stuff i did alrady thou
so both the motors were sent to the manufacturere (they were bougth second hand)
one got rewound to a lower kv (and new hall sensors) , second got his hall sensors replaced (flowing the issues i had)
so both the sensors and the wires for the sensors are brand new, one of the motors have new windings too
regrading the controller
the sabvoton isnt the first controller im trying with
before i bougth it, i tryed geting it to work with 3 diffrent controllers from 2 diffrent comapanies
one of them just didnt work in sensored mode, refused to use them
the 2 others (same type) worked in sensored mode, but only for acouple of seconds and then switched to sensorless
cusing the motor and controller to go out of sync the grind to a halt
after exprinceing all this i decided to get the sabvoton thinking a better controller migth do the trick but nope :|

ive been dealing with this issue for a good month now
thanks for helping btw :)
 
If you want to recheck color combos again, here is what I do;
First, Unless there is already single connectors (like bullets), I use wire nuts for quick change capability.
I write out a chart on a note book (more technical types can use graph paper:)
It looks like this;
Controller...motor
......Y..->......Y
......B..->......B
......G..->......G
That would be for matching colors.
Then try all combos, keep the Halls matched to the phase wires.
Then I X over a combo when it does't work.
I don't think I,ve ever had the Halls end up different than the phases, but can't swear to it.
 
Revolt motors are known to have problems like this (and other problems). This thread
https://endless-sphere.com/forums/viewtopic.php?f=30&t=87305
shows a pic of the hall buried down in the stator tooth gap rather than up close to the magnets, which may be part of the problem with that specific one. He shows the hall signals themselves to be erratic, perhaps because of them being so far from the magnets, or perhaps from interference induced by the stator fields.

There are a few threads in this list of Revolt threads
https://endless-sphere.com/forums/search.php?keywords=revolt&terms=all&author=&sc=1&sf=titleonly&sr=topics&sk=t&sd=d&st=0&ch=300&t=0&submit=Search
that have some info on the problem. I don't know that there are any quick fixes.

Probably the easiest fix is to use an external encoder on the shaft, that is compatible with your preferred controller. AFAICR this is what Larsb did in that first thread. Also, his other thread here
https://endless-sphere.com/forums/viewtopic.php?f=30&t=85534&hilit=revolt&start=250
describes the specific problem as being improper placement of the halls, as them just being stuck in the slots at 120 physical degrees apart, but that the motor isn't physically/electrically setup so that this makes correct position signals. To fix that, without an external encoder, you'd need to locate the correct places for the halls (I don't remember how to do that, but there are a few threads that describe the process), and then make notches in the stator laminations in those spots to embed the halls within them, so taht the rotor magnets correctly line up with them when the phase signals to mvoe those magnets would need to be firing. (which is what the halls are there to tell the controller to do).


This post has links to some issues with Revolt motors; I haven't gone thru them but perhaps they have info you can use
https://endless-sphere.com/forums/viewtopic.php?f=30&t=94975&hilit=revolt&start=100#p1450409
(the thread itself might be useful if there is anything from other posters in it, unfortunatley the OP of that thread hissy-fit deleted all his stuff before storming off in a tantrum).
 
wow, thank you for shedding light!
do you think its posible to mount the hall sensors externally? make a 3d printed mount that is moveable and mount the sensots on the rotors outer skin? well the emf be strong enoth to swith the hall sensors on and off?

or is there an encoder i can get for it that fits on its 15mm shaft?
 
itsik said:
do you think its posible to mount the hall sensors externally? make a 3d printed mount that is moveable and mount the sensots on the rotors outer skin? well the emf be strong enoth to swith the hall sensors on and off?
EMF isn't what switches the sensors. The magnets on the rotor do that.

You can use an external hall sensor array (completely outside the casing) but you will also probably have to use external magnets if you do. You'll have to test to see if a hall sensor in the place you want to put them will be able to clearly and easily detect the rotor magnets passing. If it can't, then you have basically the same problem you probably do now (the one larsb shows).

or is there an encoder i can get for it that fits on its 15mm shaft?
You should first look at stuff like what larsb did; I didn't read it just saw it in his posts.
 
so i did some testing and conected the halls to a scope and the resulting sqware wave is acturly ok
what i did notice thou is that the motor spins faster cw then it does ccw, and it does that regardless of witch of them is considured reverse by the controller,
after watching this vid : https://www.youtube.com/watch?v=_GL5JGfQk8o&t=514s
at 6:40
im asuming the placment of the magnits is abit off cusing diffrent "timing" for cw and ccw oprations
 
That makes sense given the stuff I saw about the sensors potentially being about 30 degrees off of where they "should" be. I don't remember the rest of what i saw about that stuff, though.

If it matters, there are some threads (some recent) about adding hall sensors to a motor, and you may be able to use the methods in those to determine the best position for your sensors if you decide to move them (or add new ones instead, so you can then compare to the old ones for testing).
 
i think what ill do is add an external encoder to the motor like what larsb did here:
https://endless-sphere.com/forums/viewtopic.php?f=30&t=87305

its kinda tricky in my case since my motor is mounted to the shazy from one side only
what i think ill do is ill add a bearing to the side that dosnt have the sproket on it, to that bearing ill mount the encoder using parts ill 3d print
since from what i understod from the manufacturer of the encoder the tolerances are critical, i need the pcb to be exsacly centred with the shaft and at no more then 0.2mm away from the magnit, that needs to be scured to the shaft perfectly
and then in order to prevent the hole thing from spining ill scure it to somthing on the shazy

its abit of a bummer that its now a project within a project but i got the motors for cheap, and as long as itll work and well give good proformnce ill be happy with it
 
itsik said:
its abit of a bummer that its now a project within a project but i got the motors for cheap, and as long as itll work and well give good proformnce ill be happy with it
That's a problem I'm very familiar with; almost all of my stuff ends up being four or five nested projects, each of which takes me far too long to get done, so I don't usually ever actually "finish" any primary top-level project (just get it to a usable state and work on it more on rare "free moments").
 
amberwolf said:
You can use an external hall sensor array (completely outside the casing) but you will also probably have to use external magnets if you do.

Seems like routing the wires would become problematic.
 
i installed the encoder, as sugested in one of the posts you guys linked, and for some reason i get 0 voltage reading betwine the U V and W to the +5V (one should always be +5V) the sensor isnt working ... any ideas?
i wired it currectly, and its about 0.5MM away from the magnit as the spec sheet says it should be
this is siriusly fustrating
added some pictures
 

Attachments

  • WhatsApp Image 2021-05-08 at 18.17.46.jpeg
    WhatsApp Image 2021-05-08 at 18.17.46.jpeg
    31.6 KB · Views: 371
  • WhatsApp Image 2021-05-06 at 14.00.01.jpeg
    WhatsApp Image 2021-05-06 at 14.00.01.jpeg
    153 KB · Views: 371
  • WhatsApp Image 2021-05-08 at 12.56.12.jpeg
    WhatsApp Image 2021-05-08 at 12.56.12.jpeg
    150.4 KB · Views: 371
  • WhatsApp Image 2021-05-08 at 12.56.16.jpeg
    WhatsApp Image 2021-05-08 at 12.56.16.jpeg
    139.6 KB · Views: 371
  • WhatsApp Image 2021-05-08 at 12.56.21.jpeg
    WhatsApp Image 2021-05-08 at 12.56.21.jpeg
    165.9 KB · Views: 371
  • WhatsApp Image 2021-05-08 at 13.34.17.jpeg
    WhatsApp Image 2021-05-08 at 13.34.17.jpeg
    199.8 KB · Views: 371
  • WhatsApp Image 2021-05-08 at 13.37.29.jpeg
    WhatsApp Image 2021-05-08 at 13.37.29.jpeg
    162 KB · Views: 371
  • WhatsApp Image 2021-05-08 at 13.37.29 (1).jpeg
    WhatsApp Image 2021-05-08 at 13.37.29 (1).jpeg
    117.1 KB · Views: 371
  • WhatsApp Image 2021-05-08 at 18.14.53 (1).jpeg
    WhatsApp Image 2021-05-08 at 18.14.53 (1).jpeg
    150.6 KB · Views: 371
  • WhatsApp Image 2021-05-08 at 18.14.53.jpeg
    WhatsApp Image 2021-05-08 at 18.14.53.jpeg
    156.4 KB · Views: 371
Impossible to tell what the error is, this doesn’t look like the rls encoder i got which had a connector - did you crack it loose from it’s case and desolder the connector? Then i’d guess anything could have happened during modification.

Do you have 5v between + pad and ground?
Do you get any signal?
 
hey, thanks for the replay!
its the same encoder you got its just shellless, the guy at rls told me i should go with that since itll be easyer to make a new enclosure for it then itll be to modify the original one
regardless, i solved the hall erorr isuue, i soldred to the large pads (as the guy from rls told me to) but in the end i should of soldred to the smaller pads that were for the mulex conector(that wasnt included in my package)

regardless, now the controller spins the motor using the encoder, but the proformance is as bad as before
and when im runing the self learn feature (sabvoton 72150) it always fails
i think i didnt zero the encoder currectly becuse when i switch from forword to reverse at the controller, the motor proformance changes (bad in both directions but diffrent)

in order to zero the encoder i conected 2 random phase wires(is this currect?) to a constent current power supply at 5A and 45V (the max the power supply can do) and then shorted the zero pads
the motor spins to a surten position, but i can easly move it acouple of dgrees off that position and it wont go back to were it was (it well just hold the new position)
do i need to give more power on this 2 phase wires? do i need to give a spasific polarity at spasific phase wires? if so, how do i determine witch phase wires do i need to energies?

i think theres hope to this, since the suspected issue was the position of the sensores and now we have control over it, just need to get it rigth
thanks!
 
It must be the correct phase wires, fed in the correct direction for zeroing to work. Do they have colors or abc/uvw marking? Then just follow the zeroing instructions on the correct combination.

It’s a static alignment so you don’t need high current. Keep it at 5-10A, zero pads, done!
 
i dont have any markings on the phase wires, i guiss ill just try all 6 opthions
the thing is im limited to the power of my power supply and it can only do 5A, and at that current i can easly move it by hand
somtimes it even sets on a sligtly diffrent position, how acuret does it need to be?
ill see if i can get a stronger power supply
 
I don’t think you need a stronger power supply. It doesn’t take much for the rotor to align correctly.

How do you intend to keep the position of the sensor? It cannot be allowed to move after zeroing and your pics look like you’ve only locked it temporarily (and not so securely)
 
i took 3 power supplies that can do 5A and conected them up in peralel for 15A, now it holes position .. ill try later the 6 opthions


the motor itself is fixed to a plate that can slide in order to tention the chain, so ill need to make some sort of an arm that is bolted to that plate and then bolt it to the 3d printed part (the 3d printed mount has holes for bolts)

overall this arm just needs to keep the mount from spining with the motor, the bearing takes care of centring it and keeping it in place and the mount keeps it at the currect distance from the magnit

im adding the stl files and some more usefull stuff in case some one else well need it. overall the design isnt application specific and i think it can be usefull for other peaple with the same issue ... ill update in acouple of hours how well it acturly works :D
 

Attachments

  • pic3.PNG
    pic3.PNG
    15.3 KB · Views: 299
  • pic1.PNG
    pic1.PNG
    17.3 KB · Views: 299
  • PIC2.PNG
    PIC2.PNG
    12.9 KB · Views: 299
  • STLs.rar
    169 KB · Views: 5
  • ‏‏לכידה.PNG
    ‏‏לכידה.PNG
    25.2 KB · Views: 298
  • spec.PNG
    spec.PNG
    32.7 KB · Views: 297
update
i went throw all the 6 combinations(plus on one phase wire and nimus on another and so on)
all 6 gave poor results
i basically zeroed the sensor, then did self learn with the controller (acouple of times.. changed some parmaters)
then i compared the motor behaviour runing forword and reversed to see if theres a diffrence
5 of the opthions the self learn failed consistently, on one it passed, but the reverse speed was faster then the forword speed then i asumed its not the currect one
moved on to the next, and when i came back to it i couldnt get the self learn to pass again

i really dont know what todo at this point, any ideas?
 
Ahem. This controller has autolearn? Then it should have worked with any zeroing as you did initially. Didn’t think about that.

If you get it to run well (any direction) then i guess you could also switch running direction in program?

Can you see the hall readouts in the program?
 
larsb said:
Ahem. This controller has autolearn? Then it should have worked with any zeroing as you did initially. Didn’t think about that.
idk how good is his autolearn to be honest

If you get it to run well (any direction) then i guess you could also switch running direction in program? -yes

Can you see the hall readouts in the program? - yes

so i ended up finding 2 phases that seemed to give an ok result
the motor spins kinda smoth, it has a bit of a jumpy top end but no miss fires
so i took it for a test raied https://www.youtube.com/watch?v=zbXIT3yEgQI

it was alot better then before, but still weak, and the motor got hot (current limit is set to 150A at batt side and 350A phase)
the thing is, im not really sure to witch direction do i take this now, i have acouple of opthions

1 limiting to 80A on the batt side in order to keep it from heating up
2 add the second motor, see if double the power well be enoth (well need to buy another controller)
3 change the gearing .. (as you can see i have a variator gearbox, the motors are conected to its input using sprokets)
currently i have a 28 tooth sproket on the motor and a 13T on the variator input shaft, i can put 15 or 16 on the variator input
shaft and that well change the ratio from 2.15 (=7450RPM on the input shaft) to 1.86 1.75 (6500 , 6000 RPM on the iput shaft)

i can also do all 3... the qustion is well it be enoth to be fun, or should i just scrap all of it and get a motorcycle engine :|
 
Back
Top