KT motor controllers -- Flexible OpenSource firmware for BMSBattery S/Kunteng KT motor controllers (0.25kW up to 5kW)

[tenten8401]

Also, related to post above, what part number is the phase current sensor for the 48v kt controller? Can barely see it in the pics.. If someone with a sine wave controller could pop theirs open and get a part number it'd be greatly appreciated

Edit: Nevermind, found it.. ACS758LCB-100B-PFF-T I think

[KonKey125]

So, since posting last time I've gone ahead and ordered an ebike kit with a KT controller. I've been informed it's a square wave controller, but the wiki says it's possible to retrofit phase current sensors onto most boards, but I can't find anything on it so far. Does anyone have anything that could point me in the right direction here? Would love to use the existing cheaper controller and install some parts than just throw it out and buy a sine wave model..

In addition to this, is the retrofitting only available for the open source firmware? I'd like to run it but after I ordered I realized it said that LCD8H isn't compatible at all due to a missing clock or something.. If I retrofit a phase current sensor onto a square wave controller will the proprietary firmware that comes with it pick it up and use it? Is there anything I can do to get LCD8H working on the open source firmware?

Sorry for all the questions.

It's either going to be an ACS711 or ACS712 depending on the controller. The chip is also tiny. If you have the LC package (8 contacts), you might be able to solder it on with a normal soldering iron. If it's the EX package, it's going to be a nightmare to solder it on without a hot air station. I've tried to solder on the EX package with a soldering iron and a tiny conical tip. It's a nightmare and the first time I tried I ended up shorting the motor phase to the STM8 IC (40v tends to break stuff that doesn't expect more than 5).

Here's the datasheet for the ACS711: https://www.google.com/url?sa=t&rct=j&q ... 1zeS3EAjUf

[tenten8401]

So, since posting last time I've gone ahead and ordered an ebike kit with a KT controller. I've been informed it's a square wave controller, but the wiki says it's possible to retrofit phase current sensors onto most boards, but I can't find anything on it so far. Does anyone have anything that could point me in the right direction here? Would love to use the existing cheaper controller and install some parts than just throw it out and buy a sine wave model..

In addition to this, is the retrofitting only available for the open source firmware? I'd like to run it but after I ordered I realized it said that LCD8H isn't compatible at all due to a missing clock or something.. If I retrofit a phase current sensor onto a square wave controller will the proprietary firmware that comes with it pick it up and use it? Is there anything I can do to get LCD8H working on the open source firmware?

Sorry for all the questions.

It's either going to be an ACS711 or ACS712 depending on the controller. The chip is also tiny. If you have the LC package (8 contacts), you might be able to solder it on with a normal soldering iron. If it's the EX package, it's going to be a nightmare to solder it on without a hot air station. I've tried to solder on the EX package with a soldering iron and a tiny conical tip. It's a nightmare and the first time I tried I ended up shorting the motor phase to the STM8 IC (40v tends to break stuff that doesn't expect more than 5).

Here's the datasheet for the ACS711: https://www.google.com/url?sa=t&rct=j&q ... 1zeS3EAjUf

I should've specified, I ordered the 48v 12 fet version which appears to be this chip for phase after doing some digging on the website... https://www.ebay.com/itm/ACS758LCB-100B ... 4012326992
https://opensourceebikefirmware.bitbuck ... .html#h1-1

Would this be something I can just swap and it'll work, or will I need the open source firmware to take advantage of? From what I see LCD8 doesn't work at all on open source firmware so I'd prefer to just stay with what comes with it if I can. My soldering skills are pretty good so I think I can get it, have a rework station and lots of flux

[sylvain_wm]

In addition to this, is the retrofitting only available for the open source firmware? I'd like to run it but after I ordered I realized it said that LCD8H isn't compatible at all due to a missing clock or something..

Hi tenten8401,
Where did read that the LCD8H doesn't work? Could you send the link?
I have a LCD8H with KT controller, and it seems to work except the 6kph assistant without PAS.

Have a good one

Regards

[Iambuilderman]

Hi guys,

I've been reading through this forum; the work everybody did is absolutely great. But I cannot find a definitive answer on the following question: what exactly do you need to modify to make a 18fet 48v kt controller with a 48v KT3 display and a 48v bluetooth dongle work at 20s 84v max?

Currently I have the following components (see pics):
- k150e09ne mosfets (Vmax = 85V)
- 470uF 63V capacitors
- 104J 100V metal film capacitors (red capacitors)
- LM317T voltage regulator + 260 Ohm resistor

I want to enable the use of a 20s 84V battery. Are these the correct upgrades to do, or do I need to do more:
- Change mosfets to IRFB4110 (Vmax = 100V)
- Change caps to 1000 uF 100V (see pic)
- change red capacitors to 104J 400V 10mm ones (are these the correct ones (V and J)? does size matter here?)
- change LM317T + resistor for a DC-DC buck converter from 20V -> 15V (where does the 20V input come from? does it stay 20V when I provide 72V instead of 48V? Where do I need to connect the 15V output of the buck converter?)
- use a 72V -> 48V buck converter to supply a seperate 48V power supply to the KT-LCD3 and the Bluetooth dongle (is this really necessary? and is this enough to protect both the display and the dongle?)
- edit: do I need to do something with the 78L05 voltage regulator as well (did not find that one yet on the PCB), or is this one already safe if I supply 15V through the buck converter instead of the output of the LM317T + resistor?

Are there any other components that I need to swap or change? Is the rest of the controller capable of handling the 72v/84V input voltage?

I see many people with similar questions about voltage tune-up. But not a complete overview of all necessary steps and components. So a little help would be highly appreciated!

Thanks in advance!

[tenten8401]

In addition to this, is the retrofitting only available for the open source firmware? I'd like to run it but after I ordered I realized it said that LCD8H isn't compatible at all due to a missing clock or something..

Hi tenten8401,
Where did read that the LCD8H doesn't work? Could you send the link?
I have a LCD8H with KT controller, and it seems to work except the 6kph assistant without PAS.

Have a good one

Regards

It works fine on stock but it doesn't work with the open source firmware this thread is about

[stancecoke]

It works fine on stock but it doesn't work with the open source firmware this thread is about

is this your personal experience? I know, that there are new "L"-parameters, but I don't know how they are implemented in the communication protocol.
http://www.szktdz.com/upload/file/20181 ... _12550.pdf
Can you post a log of the bytes sent by the display?

regards
stancecoke

[tenten8401]

It works fine on stock but it doesn't work with the open source firmware this thread is about

is this your personal experience? I know, that there are new "L"-parameters, but I don't know how they are implemented in the communication protocol.
http://www.szktdz.com/upload/file/20181 ... _12550.pdf
Can you post a log of the bytes sent by the display?

regards
stancecoke

I don't actually have the kit with me yet, but I remember seeing a forum post at some point that said it didn't work at all on the open firmware for them.. I have no idea on the specifics yet.. Is it working for you on the open source firmware or anyone else that you can find? I just don't want to take the one-way trip only to make my LCD8H useless

I don't even know where I saw that it didn't work and the entire thing does sound kinda confusing to begin with since it connects to the same header on closed firmware, so I'd be willing to give it a go (when my kit arrives a couple weeks from now) if I can just get some confirmation it works.

I'm getting a newer firmware revision because the layout looks different from what I see online:

[sylvain_wm]

It works fine on stock but it doesn't work with the open source firmware this thread is about

Hi, my LCD8H works with the open source firmware this thread is about.
Everything seems to work.
But, as I said, when I hold the "DOWN" button, the meter assist function logo flashes, but the function doesn't work. The motor isn't powered...
I don't have the newer firmware revision on my display.

Regards

[stancecoke]

But, as I said, when I hold the "DOWN" button, the meter assist function logo flashes, but the function doesn't work.

Yes, but this is not a matter of the LCD8H, it doesn't work with any display, as the walk assist function is not implemented yet.

regads
stancecoke

[sylvain_wm]

Ok,
thanks stancecoke.
So the LCD8H works well with the firmware...
I'm glad to read this, because I was trying to pore over every single ligne of the program to figure out where I had to change the code....

Have a good day.

Sylvain_wm

[tenten8401]

Oh cool, might give it a try then.. Might like to keep the walk assist but otherwise sounds like I'll give it a try
Thanks for confirming it all works

[stancecoke]

because I was trying to pore over every single ligne of the program to figure out where I had to change the code....

I've implemented the walk assist in the new branch "Walk_Assist":
https://github.com/stancecoke/BMSBatter ... alk_Assist

It's just quick and dirty and not tested yet. When the walk assist bit is set, the current target is simply set to 1 amp (for the 6FET calibration).

Code: Select all

if (ui8_walk_assist) uint32_current_target = 10 + ui16_current_cal_b;

If someone can confirm that it works properly, I can push it to the master branch.

regads
stancecoke

[tenten8401]

because I was trying to pore over every single ligne of the program to figure out where I had to change the code....

I've implemented the walk assist in the new branch "Walk_Assist":
https://github.com/stancecoke/BMSBatter ... alk_Assist

It's just quick and dirty and not tested yet. When the walk assist bit is set, the current target is simply set to 1 amp (for the 6FET calibration).

Code: Select all

if (ui8_walk_assist) uint32_current_target = 10 + ui16_current_cal_b;

If someone can confirm that it works properly, I can push it to the master branch.

regads
stancecoke

Seems like a pretty good idea, but wouldn't it be better to set a constant speed target rather than amp draw so it'd assist more uphills and less downhill?

[stancecoke]

but wouldn't it be better to set a constant speed target

feel free to implement it in your suggested way and share your work here!
The whole firmware doesn't know a speed control yet, just a ramp down at the speed limit. So you have to tune a PI-control for the speed, similar to the PI-control we are using for the battery-current at the moment.

regards
stancecoke

[tenten8401]

but wouldn't it be better to set a constant speed target

feel free to implement it in your suggested way and share your work here!
The whole firmware doesn't know a speed control yet, just a ramp down at the speed limit. So you have to tune a PI-control for the speed, similar to the PI-control we are using for the battery-current at the moment.

regards
stancecoke

̶s̶u̶r̶e̶l̶y̶ ̶t̶h̶e̶ ̶P̶I̶ ̶c̶o̶n̶t̶r̶o̶l̶l̶e̶r̶ ̶w̶o̶u̶l̶d̶ ̶b̶e̶ ̶u̶s̶a̶b̶l̶e̶ ̶f̶o̶r̶ ̶a̶ ̶3̶m̶p̶h̶-̶i̶s̶h̶ ̶s̶p̶e̶e̶d̶ ̶l̶i̶m̶i̶t̶ ̶w̶o̶u̶l̶d̶ ̶i̶t̶ ̶n̶o̶t̶?̶ ̶I̶'̶m̶ ̶c̶o̶n̶f̶u̶s̶e̶d̶ ̶o̶n̶ ̶h̶o̶w̶ ̶t̶h̶e̶ ̶o̶f̶f̶i̶c̶i̶a̶l̶ ̶K̶T̶ ̶s̶o̶f̶t̶w̶a̶r̶e̶ ̶w̶o̶u̶l̶d̶ ̶d̶o̶ ̶i̶t̶ ̶i̶f̶ ̶t̶h̶e̶y̶'̶r̶e̶ ̶n̶o̶t̶ ̶u̶s̶i̶n̶g̶ ̶a̶ ̶P̶I̶ ̶c̶o̶n̶t̶r̶o̶l̶l̶e̶r̶ ̶t̶o̶ ̶t̶r̶a̶c̶k̶ ̶t̶h̶e̶ ̶t̶a̶r̶g̶e̶t̶ ̶s̶p̶e̶e̶d̶
Edit: I see what you're saying now, there's no PI controller for speed yet

[stancecoke]

If someone can confirm that it works properly, I can push it to the master branch.

After a little bugfix, a german user confirmed that the walk assist works now. I've pushed the Walk_Assist branch to the master!

regards
stancecoke

[tenten8401]

If someone can confirm that it works properly, I can push it to the master branch.

After a little bugfix, a german user confirmed that the walk assist works now. I've pushed the Walk_Assist branch to the master!

regards
stancecoke

Awesome, greatly appreciate the work being done on this

[sylvain_wm]

Thank you stancecoke,
I'll try your code as soon as my controller is fixed.
I burnt a couple of transistors during my last trip.
I'm waiting for MOFETS from aliexpress.
I'm loorking forward to trying your program.

Regards.

Sylvain_wm

[Corrida victim]

I've implemented the assistance from standstill in torquesensor-mode, after a user in the german forum asked for a similar function in PAS+THROTTLE mode, too. Now the throttle works as push assistant up to 6 km/h in all ride modes. For correct 6km/h the gear ratio has to have the correct value.

Hi stancecoke!
I was wondering if this is still a thing, I have been using this software on my daily commuter for a couple months and 3000km. It's all flawless!! Except for the fact that start from standstill is impossible (especially uphill) with my 60-11 gear ratio (when I'm not able to down-shift).
The feature mentioned above would greatly improve my daily commute

Best regards,
Inse

[Corrida victim]

Walk assist turns out to be a nice solution to my 'departure from standstill' issues!

I set the walk assist current target to 60 instead of 10 in ACAsetpoint.c. If someone is interested, my code change will set the target power for walk assist to ±500w and limit the speed to 20km/h (to make it legal in Europe for speed pedelecs).
Change this line:

Code: Select all

if (ui8_walk_assist) uint32_current_target = 10 + ui16_current_cal_b;

To this:

Code: Select all

if ((ui8_walk_assist) && (ui16_virtual_erps_speed < 57)) { uint32_current_target = 60 + ui16_current_cal_b;}

This line can also be used to limit the speed to another value. For example, set "ui16_virtual_erps_speed < 18" to limit walk assist to 6km/h.
Credits to stancecoke for adding walk assist!

If anyone could guide me towards activating torque sensor "throttle" from standstill (without PAS detected) I would be super happy

[geofft]

@stancecoke,

I've recently switched from Xnyle's 'x4throttle' branch to your 'Torque_from_X4' branch and glad to say all is now working very sweetly.

Sorry to put another request on you, but just wondered if you have any plans to add walk assist to the 'Torque_from_X4' branch?

[stancecoke]

my code change will set the target power for walk assist to ±500w and limit the speed to 20km/h (to make it legal in Europe for speed pedelecs)

but just wondered if you have any plans to add walk assist to the 'Torque_from_X4' branch?

You could have made the few changes yourself, but here you go (not tested):
https://github.com/stancecoke/BMSBatter ... 8db04c547d

regards
stancecoke

[geofft]

but just wondered if you have any plans to add walk assist to the 'Torque_from_X4' branch?

You could have made the few changes yourself, but here you go (not tested):
https://github.com/stancecoke/BMSBatter ... 8db04c547d

regards
stancecoke

Aaww... thanks, you're a star.

Tested and all seems to be ok. The walk assist is a bit racy for my weary old legs, I'll need to turn the current down a little. That much I can manage myself...

[eLsc]

Well regarding the strange behavior of my controller described by my last post, I managed to get rid of it by flashing the firmware again (still not sure what caused this but I will keep looking if this happens again).

One thing that I tried back then was to add a brake to stop the motor from spinning at max eros, but this seemed to only cause the same issue, even accelerating the motor from standstill to the max erps with max power.
This still happens after re-flashing the controller and i'm starting to wonder if this is because my setup ( Q100CST, BMSbattery S06S controller at 36V) is having some sort of problem with the regen (or more specifically the controller)?

Otherwise this open source firmware is really great and has been very reliable in the past 5 months of using it, so great thanks to stancecoke, casahino and all the other contributers of this awesome firmware.