Successor to the infineon clone controllers in the 1000-2000w power range?

True, the solution is easily found in PC parts world.
I don't think i need an extensive heatsink, anything is a big improvement than a thin flat plate.

Exposing the controller to the air is a problem because it's the opposite of weather sealed, unlike the phaserunner.
We have dust, rain, massive winds, and snow in my locale so... yeah.. i think any kind of heatsink would be an improvement
 
Nep if you come up with something custom, resurrect this old thread:
It's been a while, but I wanted that thread to be a place where DIY cooling mods for controllers can be documented.

Cheers
 
Will do, i think talking about cooling more is pertinent in this era where people are using tiny FOC controllers instead of infineons. These controllers don't tend to come with the extra 2lbs of aluminum they used to. 😅
 
Thanks to this my regen is working and i'm SO happy!
Thanks a gazillion, i would have never thought of that.

Even better - the regen works all the way down to 0RPM and with my tune, tends to get stronger the closer you get to 0RPM. In that way it acts very much like a bike brake.

This is the best regenerative braking i've experienced.

Also, i bumped up the power to 40A batt, 100A phase, 10A of flux weakening. Nice boost, a 6 mile ride finally started to get the controller warm-hot in the bag, So, this is approaching or already at the limits of what the controller will tolerate without any kind of ventilation.
Glad to see you got it working. Same experience here: vesc fine tuned adc input beats phaserunner throttle and regen in my opinion. You can even hook it up to a torque sensor and use pedal pressure to control motor speed. Though implementing power levels like on classic ebikes is a hassle with vesc.
 
The ewheel v3 adapter has some provisions for making a 3 speed switch work, and it should work on any VESC. haven't tried it, but very interesting.

1718890497494.png
Side note,
 
I got notification that the flipsky ebike controller will be here monday. Less enthusiastic about testing it since i'm so happy with my VESC, but in the name of science, it will be put through the ringer!
 
The waterproofing is an annoying factor of many VESCs, although some are the opposite and they are totally potted and water cooled for jet boards. So far my solution is to take them apart and silicone where the case comes together and around the power and phase wires then make a 3D printed cap that covers the signal wire plugs with ports for cables to leave and apply gaskets and silicone where needed.

I think though for some applications, particularly with controllers like the spintend that have a flat thermal plate and no built in heatsink a different strategy could be employed by just putting a thermal passthrough into a waterproof/water resistant container like a battery bag or box. So for a bag just cutting a hole in the bag and installing an aluminum plate, perhaps clamped to the edges of the bag with screws at the perimeter then mount the ESC to that and you can install a heatsink on the plate if you needed or maybe just the plate being outside would be enough.

I wonder if anybody makes a variable hall sensor brake lever, of course a lever could be modified quiet easily as well. Thing is then you could set the brake voltage range in the VESC so that the regen is maxed right before the pads make contact so you only use the pads if you need a ton of braking.
 
Most vesc controller are sold without heat sink for cuteness factor(deceptive marketing), but is necessary for full power operation.
Same with with the phaserunner.
 
After you do this, I recommend turning on the traction control, assuming the motors are identical, it is a nice bonus. My bike with a very similar config, fatbike, 2WD, flipsky 75100 dual via canbus benefited nicely from it. While certainly not perfect the traction control does keep the front wheel in control when conditions are loose.
Unfortunately they are not identical. Rear is from XOFO and front is from Zemake. They also have different kVs, main reason I'm moving to FOC control. In that case does traction control still works?
 
Unfortunately they are not identical. Rear is from XOFO and front is from Zemake. They also have different kVs, main reason I'm moving to FOC control. In that case does traction control still works?
As far as I know probably not. The kVs don't matter, all that matters is the ERPM to ground speed so if they have the same poles, same gear ratio, same size wheel and same tire it should work. Unless it was changed in a newer version I think the traction control only uses the ERPM and not the gear ratio and tire size you input so they do kinda have to be identical motors.
 
I wonder if anybody makes a variable hall sensor brake lever, of course a lever could be modified quiet easily as well. Thing is then you could set the brake voltage range in the VESC so that the regen is maxed right before the pads make contact so you only use the pads if you need a ton of braking.
I made these myself about 8 years ago for my Adaptto on my Magura MT5/MT2 brakes. Just used variable voltage hall sensors glued to the lever housing + magnets on the lever. Add/remove magnets to adjust braking force/point.
Had it fine tuned to the point you could not feel the difference between regen and pads.

AFAIK, none of the off-the-shelf offerings work with hydralic systems...so a DIY solution like above is still needed.

Cheers
 
Cutler Mac is telling me they can't sell me a controller
This doesn't seem to have changed to date unfortunately. I was able to negotiate a couple of motors last week and had hoped it would be a good opportunity to buy a few cheap controllers but no luck. No details given but they just couldn't/wouldn't sell me any size controller.
 
This doesn't seem to have changed to date unfortunately. I was able to negotiate a couple of motors last week and had hoped it would be a good opportunity to buy a few cheap controllers but no luck. No details given but they just couldn't/wouldn't sell me any size controller.

Strange that they're also out of controllers during the busy season.
If you can wait a moment, i should be doing a review of the flipsky ebike controller to see if it's a good budget pick. Hopefully it's a good controller for the dollar, and being a high eRPM capable controller, it should be able to drive MACs even on higher voltages.
 
Strange that they're also out of controllers during the busy season.
If you can wait a moment, i should be doing a review of the flipsky ebike controller to see if it's a good budget pick. Hopefully it's a good controller for the dollar, and being a high eRPM capable controller, it should be able to drive MACs even on higher voltages.
That's great, I'll be very interested. I currently have two phase runners (V2 and V6 I think?) which I've found good overall, but they do have some weird bugs (V2 regularly resets to sensorless for example).
 
I wonder if anybody makes a variable hall sensor brake lever, of course a lever could be modified quiet easily as well. Thing is then you could set the brake voltage range in the VESC so that the regen is maxed right before the pads make contact so you only use the pads if you need a ton of braking.
Yup, as noted grin makes an analog signal brake lever:

And I also have repurposed a Xiaomi replacement brake to do the same.

My current dual motor build (with a flipsky vesc, coincidentally) uses this setup and I have basically zero need for the physical brake at all. In fact, I only have a single hydraulic disc on the front (right hand side) for keeping the bike still while at a stop light (as I don't have "plug braking", I think it's called, enabled in VESC to hold the bike still).

AFAIK, none of the off-the-shelf offerings work with hydralic systems...so a DIY solution like above is still needed.

This is true; nothing I've ever run across in hydraulics is more than on/off.
 
I made these myself about 8 years ago for my Adaptto on my Magura MT5/MT2 brakes. Just used variable voltage hall sensors glued to the lever housing + magnets on the lever. Add/remove magnets to adjust braking force/point.
Had it fine tuned to the point you could not feel the difference between regen and pads.
Can you post photos of them?
 
Can you post photos of them?
I did it on my Fighter also, but could only find photo's of it from the Alpha and from when I was still test fitting it with hot glue.

I later re-did this with epoxy and made it look better. As ghetto as it looks, it worked well!

Cheers
 

Attachments

  • DSC_4509.JPG
    DSC_4509.JPG
    2.4 MB · Views: 16
  • DSC_4510.JPG
    DSC_4510.JPG
    2.1 MB · Views: 16
I'm glad you are giving VESC a try, in the ebike community not a whole lot is out there on vesc ebike stuff but I really believe it's what most people will end up moving to eventually. A lot of the features currently missing can be added if you know how to write lisp scripts. For example this is how my lunacycle ludicrous v2 controller for my bbshd has PAS, FW, MTPA (these merged with the current software to give it to all but was added by luna / marcos) bafang display support, wheel speed sensor, multiple power / pas power levels. Other notable companies that have moved to VESC I know is Torp and EBMX for their surron controllers and CYC for their ebike motors all are on vesc now. Also including features like a regen brake throttle and fully custom apps.

As you can tell though the biggest hurdle is that from a end user if you just buy a vesc you end up wanting some of these features (that are not currently found stock) and because these companies haven't merged them into the official vesc stuff they will not get them unless they do the programming themselves which not a lot of people know how and there just isn't really any easy how to for them. Which in my opinion is holding people back from testing / recommending vesc controllers. If the software came with native support for a good display, wheel speed sensor, better pas / tuning for pas and levels on said display we would see a lot more people test it out. With my bbshd vesc I've been super happy with it though, for sure beats the stock controller in every way!
 
Ride report on the Spintend VESC.

It was 85F, i rode up a and down a 1.1% grade with lots of stop and go and pedaling for about 3.5 miles.
Got the Grin all Axle and spintend box both to 'kinda hot' in this relatively short ride.

This is very close to an Infineon 12FET, with no heatsink and in the bag.

..but it doesn't have ~1lb of aluminum bolted to it. My guess is in reality it's more like a 15FET with a nice heatsink, or at least some good air flow.

The Grin all axle also got kinda hot and i can tell i'm pushing it at 2.2kw peak, ~1.2kw continuous, especially in this heat. I could add ferrofluid, but installing the RH212 will be a better choice for torturing controllers.

More later!
 
Yeah, lisp is a weird choice of language, if they used something more mainstream like python, i would be contributing to code myself. Just don't want to learn lisp for just one application.

As you can tell though the biggest hurdle is that from a end user if you just buy a vesc you end up wanting some of these features (that are not currently found stock) and because these companies haven't merged them into the official vesc stuff they will not get them unless they do the programming themselves which not a lot of people know how and there just isn't really any easy how to for them.

Hmm, and none of these features made it to open source?
..maybe we should start collecting whatever code we can and making an ES VESC package ^_^

If the software came with native support for a good display, wheel speed sensor, better pas / tuning for pas and levels on said display we would see a lot more people test it out. With my bbshd vesc I've been super happy with it though, for sure beats the stock controller in every way!

I can see that. there's a thread on here from a guy who made a DIY display, but no notes about it getting produced. Apparently trampaboards has just started to produce a display: VESC DISPLAY

I think the support will come in time!
 
Yeah i can say i'm a VESC fanboy RN, i love how good the sensorless is and how i can tune everything, i think it was just intimidating to jump into from zero. I'm glad i know it now.

We really should write a VESC for newbies article and see if we can popularize it.


On that note, the Flipsky FT85BS v2.0 arrived today.

It was totally 'as described':
1719289543300.png

Initial impressions:
- Came very well packed, each item was bubble wrapped individually and this package would survive being chucked.
- The smaller connectors are cheap and flimsy, the pins want to come out the back a bit when they are connected and the tiny things require a ton of force. Good thing i have a JST male and female kit ( thanks to @chuyskywalker for the tip )
- The connectors look laid out like a VESC, so i am thinking this Flipsky is a 'fork' of the VESC project.
- The fact that you have to solder on the phase/battery connections is a negative for newb friendliness.
- brake handle is kind of weird but will work with a mechanical brake thankfully.
- Don't like the tacky gold accents on the throttle/other handlebar, but the throttle looks and feels otherwise well constructed.
- I like the larger case and bet this controller sheds heat better than the VESC.

More later!
 
and how i can tune everything
The big disadvantage of the VESC is, that it is not made for a bicycle. The support of PAS is only rudimentary. For a torquesensor there is no solution at all. (Except the fork of Luna Cycle, but it only implements the Bafang torquesensor sensor via CAN) If you just want to ride moped with the throttle it's OK.

regards
stancecoke
 
Hmm, and none of these features made it to open source?
AFAIK, the features though are not in official VESC repo, the sources are available in open.

VESC, being GPLv3, vendors have to provide the source to their customers or rather to any one whom they provide the binary. They need not make it available to every Tom, Dick & Harry, but once the customer gets the source, they can even share it with the aliens ;)

GPLv3'ed VESC is a double edged sword for the vendor, on one side they get a big market (relatively) advertising it as VESC capable, at the same time they are obliged not only to provide source, but also needs to make it clear on how to update the software & it is possible that end users end up damaging the board going beyond the limits intended by the vendor

And I feel, vendor adding features to the official VESC would be at a disadvantage - after putting the effort, other vendors would reap the benefit for free.
The support of PAS is only rudimentary
The last time I checked, official VESC supports 2 wire PAS, cadence controlled mode. But one has to find a suitable PAS signal pin & update it in the config file of the relevant hardware, unless taking the Lisp route.

I did add the simple, but commonly found torque based PAS control to official VESC around couple of years back, one day when I checked it vanished, I think it was inadvertently removed when the torque sensor support was added, though it can be retrieved with git archeology. I was able to get it work with 1 wire PAS, though didn't get time to add it's support to official repo, but the source for it is available in the public domain.

For a torquesensor there is no solution at all. (Except the fork of Luna Cycle, but it only implements the Bafang torquesensor sensor via CAN)
Seems above is merged to official repo, but only supports Luna Cycle hardware.
 
Last edited:
We really should write a VESC for newbies article and see if we can popularize it.
This, times a million.
I don't have the time to create the whole thing from scratch, but I have almost 10 years of bookmarks and studying VESC hardware/software, and 5 years of hands-on hardware/software.
Would love to contribute to a structured article.
 
Back
Top