FLIPSKY new 20s 100A tiny controller (vesc based)

[mxlemming]

Hahaha that's great. So they modified the hardware to be compatible with the 75_300_R2, but still had to slightly modify the 75_300_R2 firmware to make it work. Now when people upgrade the firmware in the VESC tool, it will pull in the unmodified 75_300_R2 and the ESC won't work. If only there was another way...

So to be clear, you are running the 75_300_R2 and just removed this from the header file and it appears to be fully functional?

#define HW_HAS_PHASE_FILTERS

No, just toggled it in the FOC filters menu. No compiling necessary.

I can compile, but that involves opening my Linux laptop and fighting git (ok it's really not that hard...) Fortunately the drop down menu option worked.

 

[tostino]

Small update, I have been tweaking this controller slightly to see if I could make it work for my boat usage, and I think I have found some success.

I am still waiting for the damn bluetooth adapter to be able to do some data logging, but I worked on improving the cooling first and foremost. Simply picked up some thermal epoxy and one of these: https://www.amazon.com/dp/B07TB38W1Y and shoved it next to the motors cooling fan. Also sealed it up with RTV silicone, and filled it with mineral oil with the wires pointed up so we get minimal leaks.

Made an absolutely giant improvement in thermal throttling. I can now plane out on the boat at around 2200w until the 6kwh battery dies without any heat issues in the controller. Whereas before it was throttling in ~45 seconds and would take a bit to get back to more reasonable temps.

Will be messing around with slightly higher power limits now.

I'll get a FLIR recording of the setup along with the data logging here shortly.

 

[MorbidlyObeseKoala]

When you say filled with mineral oil, do you mean you filled the controller case with mineral oil and sealed with silicone?

From the few videos and reviews I have seen on this little guy, the thermals are the main thing holding it back. An affordable cooling solution could really set it apart.

 

[tostino]

Exactly. I flipped the end caps upside down so the mounting tabs didn't get in the way of the flat back side of the case being mounted flush to the heatsink, and put a bead of the silicone around the edges when putting it back together. Did my best at sealing the wires, but it will absolutely leak out around the wires without a better solution.

Filled it up through that little LED hole and then replaced the stopper. I'll do a test with my other controller without the mineral oil to see how much of a difference it makes vs the heatsink alone.

 

[Barncat]

Running a boat is a heavy continuous draw, so good that some coolant tweaks are working.

This thread is a free for all so I'm interjecting another related topic. Finally received and wired up one of these units to run an MY1020 inrunner with AFAIK good halls on 20s. I'm getting a repeated "sensor detection failed" message when running FOC motor detection. Unpleasant surprise. I'm confident it's wired correctly, but there is no temp wire on these motors and so that wire on the comm port is disconnected. That can't possibly be causing this fault?

The firmware is evidently 75 300 r2, and I noticed the 'no phase filters' tip from mxlemming but don't know if that's related. Sort of stumped at the moment...

 

[Barncat]

Or I guess it's the hardware that's 75 300 R2, and supposedly 5.2 firmware... from the screen shot below does that mean the 5.2 firmware is already loaded or do I need to update?

With the older vesc tool version I got the motor to move at least during wizard detection, it was spinning backwards which is not unusual but after the initial spin up it locked up for 2 seconds with a nasty noise which must have been phases fighting each other, then the process stopped and it timed out after about 2 minutes....... ordinarily the motor would cycle back and forth through some partial revolutions to finish. I figured this would be a simple plug and play. Murphy's Law.

 

[Barncat]

I tested my Hall wires and they are all good. There's a simple tutorial on the Grin website for that if you're unfamiliar with that process.

So I reloaded the 5.2 firmware, and I'm using the older vesc tool with the white background. Motor detection was successful this time around, though direction was inverted, which is a simple final step to correct in detection.

ADC throttle mapped fine but motor immediately cuts out at very low rpm with no load. Haven't figured that out yet. I did wire the throttle to the +5v comm pin not the 3.3v pin that is suggested in the Flipsky wiring diagram but it seems unlikely that's the cause. Archer 321 describes exactly this problem recently in the bike troubleshooting section...

(Edit) I tried the 3.3v pin, no difference. Spent a couple more fruitless hours looking at other settings. Must be some firmware issue............

 

[Barncat]

Nothing good to report here after 8 wasted hours. Back to "sensor detection failed" in both vesc tool 3.0 and 3.1. Unit may be defective. I've about had it with (v) esc's.

 

[Barncat]

I'll rephrase my question from above:

Has anyone successfully connected a 75100 to a motor WITHOUT a temp sensor?

This is the last thing I can think of. The common MY1020 inrunner has no temp sensor, therefore the white wire on comm port from the esc is not connected to anything. All their advertising says is supports Hall sensors, it says nothing about temp sensor mandatory.

Also, since my esc is mounted upside down, I just noticed that it has an led status light, which goes and stays blue after plugging in battery and never goes green so that's indicative of a problem right there...

 

[afzal]

Has anyone successfully connected a 75100 to a motor WITHOUT a temp sensor?

Yes, see https://endless-sphere.com/forums/viewtopic.php?f=30&t=113445&start=125#p1689922 & me too

Also, since my esc is mounted upside down, I just noticed that it has an led status light, which goes and stays blue after plugging in battery and never goes green so that's indicative of a problem right there...

Green only when motor runs, otherwise blue, it is normal, IIRC. During fault, red

 

[Ianhill]

Is it the motor turning or throttle signal received, just turn the motor by hand and watch the green light its been a while since i had a vesc but 6.6 give me tons of issues so i never run it, turned a hoverboard motor great foc nice and smooth dualtron hub nope no foc only sensorless and even then it would hesitate get to full rpm and lock up making a awful metallic sound not good at all, really wanted vesc to work as the hoverboard motor showed massive potential out the box but anything bigger its always failed on for me.

 

[Barncat]

Thanks guys. The other two (v) escs I have flash blue briefly then go green
on start up to indicate normal function. So this one is not OK AFAIK... And yes red indicates a fault.

So temp wire may not be necessary. I've asked flipsky directly and waiting for a response.

Either way, this unit will not detect the motor now so I'm not happy.

 

[mxlemming]

The other two (v) escs I have flash blue briefly then go green
on start up to indicate normal function. So this one is not OK. And yes red indicates a fault.

So temp wire may not be necessary. I've asked flipsky directly and waiting for a response.

Either way, this unit will not detect the motor now so I'm not happy.

You absolutely do not need a temp sensor.

You should 100% be running the 5.3 release hardware and VESC tool. 5.2 is liable to self destruct at any moment, especially with that u15 motor, due to the way it handles over voltage situations. 5.3 handles over voltage much better, and the unstable "Ortega iterative" option seems to have gone... Even if your experience is that it runs with 5.2 and does not with 5.3, I would stick it out with 5.3, get that running rather than risk the issues with 5.2.

The phase filters option is here: I am not sure where/if it exists in 5.2/older VESC tool... I don't have that any more...


Vesc has a lot of great features, can work really nicely, has a wonderful setup tool, BUT it is not currently something that can be "plug and played" with certainty in all cases.

I never use the wizards for setup. I always do:
On the FOC page work work across the Read default config ->
Set the motor max current (this defines the current it uses for R L detection) -> ? -> RL -> lambda -> apply, then write the motor config.


Check the values you get - T motor U15 80kV should see (I THINK) 3.4 mWb lambda, and in the region of 10-20mohm and 10-20uH. IIRC, Vedder measured it in his video instruction, but I think that is quite a bit too low L

The implications of wrong lambda are very unstable observer, wrong L and R make the current controller less stable. L can also destabilize the observer, but in my experience less so than wrong lambda.

I never ever set it up with a lithium battery. Ever.

The inrunner you have should be much easier to setup, normally inrunners have much higher L and higher lambda (due to fewer, bigger poles). If you have not flipped the phase filters to off, that is probably your problem (it is certainly one blocking problem).

 

[Barncat]

mx- thanks for the detailed tutorial. I can take another shot at this if I can get the 5.3 firmware loaded. I didn't see any way to disable the phase filters in vesc tool 3.0...

Since this sort of thing is obviously not my forte I need to start with the wizard, then adjust from there. Neither the motor wizard or my attempts at manual detection were working, other than once as described in previous post. No bench power supply on hand.

I don't have a replacement esc for the U15, and may wait for the tech to mature. I figured this MY1020 setup would be a piece of cake...

 

[Barncat]

Well, I'm in WTF computer hell now. For a guy who basically hates this stuff.

I pulled up vesc-tool 3.1 and tried to flash the 5.03 firmware. It failed right before completing and flashed the dreaded red "serial port error " message. So it's apparently "bricked". I don't really know what exact set of circumstances caused this but the normal setup never worked at all for me. I never manually disconnected the power to esc but the software connection repeatedly failed during attempts. And by "don't disconnect the power for 10 seconds " does that mean battery pack power (I assume) or just the USB cable?

Flipsky says the microprocessor is stuck and I need an ST-link, which doesn't surprise me. I have no idea which one to buy, how to connect it or how to use it. I had a much easier time with the previous flipsky and makerx escs, until the latter burned up of course...

The only reason I bought this box was to upgrade from a perfectly functional and dependable big aluminum 55A trap box on the Mongoose Girder build. Will probably just swap it back.

 

[Barncat]

Sorry to bore you developers and advanced users with this crap. Troubleshooting is excruciating for us lay peons...

Amazon will drop off an ST Link V2 tomorrow, that USB thing with 4 wires. One wire goes to GND, but damned if I know where to connect the other 2 on this 75100, since it doesn't appear to have dedicated debug pins like more expensive ESCS. According to the attached pic it may be best to power ESC with battery not 3.3v?

And I'll have to figure out the ST program. Some guy Dr esk8 has a short youtube vid that seems pertinent.

 

[mxlemming]

Sorry to bore you developers and advanced users with this crap. Troubleshooting is excruciating for us lay peons...

Amazon will drop off an ST Link V2 tomorrow, that USB thing with 4 wires. One wire goes to GND, but damned if I know where to connect the other 2 on this 75100, since it doesn't appear to have dedicated debug pins like more expensive ESCS. According to the attached pic it may be best to power ESC with battery not 3.3v?

And I'll have to figure out the ST program. Some guy Dr esk8 has a short youtube vid that seems pertinent.

https://endless-sphere.com/forums/viewtopic.php?f=30&t=113445&start=50

See the plug there, the one with nothing plugged into it.

It's labelled with 3v3 gbd clk and dio. They're the ones you need

Except perhaps 3v3. Don't really need that.

If you can, I'd power it from the USB with a charger while you're programming. Add a battery once you're done.

And buy a USB isolator. They're like 10$. Your laptop is probably 1000$.

 

[Barncat]

10-4 on the pins, thanks mx. Hadn't taken the board out of the case or that would've been obvious... thought I had to work with external plugs.

(Edit) my reading of the text in pic I posted just above is that you either just use the 3.3v pin to power the board during this process OR the battery, as I have no other means on hand. Certainly the former is desirable rather than a 20s pack powering up the whole board it seems. This affects just the microprocessor- someone please correct me if wrong. The ST Link hardware claims to have some built in protection.

I'm also unclear on if this is a two stage procedure with needing some sort of ST software first or just making use of the SWD feature in vesc-tool 3.1 to get the 5.3 firmware installed. Zero experience with this.

 

[jaykup]

I wrote a tutorial on how to flash firmware on this unit with an st-link.

https://forum.esk8.news/t/how-to-update-firmware-on-the-flipsky-75100-foc-esc/61819

Under the "hard method" section

Just make sure you are using the right firmware file for the hardware version you have.

 

[Barncat]

Thanks much jaykup, will study. These sorts of things are always a mystery the first time around. Have the ST Link device now.

As if the ESC problem wasn't annoying enough, I picked up a cheap craigslist laptop yesterday to use specifically for vesc tool etc, running windows 7. Seller said it worked fine and it mostly does but it's missing network adapter driver(s), so no internet......... Will see if a USB wireless dongle with the cd disk driver works around that later today when the Amazon van drops by, and if not, tomorrow a type c android USB flash drive arrives so I'll try driver download from phone to laptop... Lotta gadgets and hoops to jump through just to resuscitate an ESC that shoulda worked right the first time.

2nd update edit: I got the vintage laptop online with the USB wireless, and finally found an old graphics update that runs vesc-tool. Perhaps some light at the end of the tunnel...

 

[gilbertojr78]

Hello guys.
I need a little help.
I have the 75100v202 and with no plausible explanation (nothing connected, no motor, no USB, nothing), only the battery, my vesc just exploded.
Now I see that fets 3 and 4 are shorted and I have this chip burned (attached picture) but I cannot find any datasheet or references for this. If anyone can Help me.
I just know that is a23a, sc 17I or 28640 but can find nothing about it in internet. I think this is a high voltage step down, like the mp9846 but the pins are not the same.

 

[jaykup]

MY1020 inrunner with AFAIK good halls on 20s

The VESC was originally designed for electric skateboards using oversized 14 pole outrunner drone motors. If you are using an inrunner, set the pole count during the motor setup wizard. It's on the advanced tab just before you click "run detection". I think the MY1020 has 10 poles. It's mostly used for speed calcs, but it may help avoid weird issues.

Has anyone successfully connected a 75100 to a motor WITHOUT a temp sensor?

Yes, temp sensor is not needed on any VESC unless you are using some of the new sensorless startup modes like VSS. It is a good idea to have one as it can thermal throttle based on motor temp.

it was spinning backwards

You can swap any two phase leads to make the motor spin in the right direction. I prefer this instead of hitting the "invert direction" switch on the VESC. It's probably just a placebo affect, but the motors seem to run smoother that way.

ADC throttle mapped fine but motor immediately cuts out at very low rpm with no load. Haven't figured that out yet. I did wire the throttle to the +5v comm pin not the 3.3v pin that is suggested in the Flipsky wiring diagram but it seems unlikely that's the cause. Archer 321 describes exactly this problem recently in the bike troubleshooting section...

Use 5v. Use the wiring diagram on Flipsky's website. There was an incorrect version floating around related to the ADC pins.

You can control the motor from the keyboard or by clicking on some of the control boxes at the bottom left of the program. Set I to say 10A and hit the play icon. Be ready to hit the stop button. It will at least help troubleshoot ADC wiring or ADC settings vs motor config.

After running the input wizard make sure App Settings -> General -> App to use is set to ADC and UART

Try different options at App Settings -> ADC -> Control Type. Start with Current and try some others. Remember to write the app config with each change.

I didn't see any way to disable the phase filters in vesc tool 3.0...

VESC tool 3.00 uses firmware 5.2 and firmware 5.2 does not have settings for phase filters. No need to find or disable them if running 5.2

I pulled up vesc-tool 3.1 and tried to flash the 5.03 firmware. It failed right before completing and flashed the dreaded red "serial port error " message. So it's apparently "bricked".

A couple of things come to mind. It always says "SERIAL PORT ERROR" in red on the bottom right after it's finished uploading the firmware. The tool disconnects from the VESC then complains that it was disconnected LOL. It doesn't mean anything failed necessarily. The 3.01 tool also uploads firmware quicker and seems to complete around 89% so that might have thrown you off too.

Second, these units, both V1 and V2 don't come with bootloaders. Why? I don't know. All the other VESCs do. You have to do that first for any firmware upgrades to "stick". It's the "bootloader" tab in the firmware section of the VESC tool. Just upload whatever version it gives you.

Once the VESC tool disconnects after upgrading the firmware, it takes 30 seconds or so before you can connect again. You actually don't have to unplug and turn everything off. In older version you had to... but the device reboots after a firmware upgrade on 5.2 and 5.3. Just keep hitting autoconnect on the Welcome page until it finds the device again and connects.

TLDR: the ESC may not be bricked. Plug it into a battery and USB, wait a minute and try to connect a few times using the VESC tool 3.00 or 3.01. If you still can't connect, then give the ST-LINK a try.

 

[jaykup]

I cannot find any datasheet or references for this. If anyone can Help me.
I just know that is a23a, sc 17I or 28640 but can find nothing about it in internet. I think this is a high voltage step down, like the mp9846 but the pins are not the same.

The white circuit board version 1 of this ESC had a Texas Instruments LM5161 14-TSSOP in that spot. I think you are right, it's a 100v buck likely for 5v and/or 3.3v.

Here is a clearer image of that chip. Looks like a SOP-8? Maybe someone will know. I can't find anything on it either unfortunately.

 

[gilbertojr78]

Really appreciate your time and answer.
Meanwhile Im trying to understand what happened to fry the vesc.

 

[jaykup]

No idea... they usually don't blow up like that. My guess would be a bad mosfet that shorted out or poor QC/soldering from Flipsky. The mosfets shouldn't switch on until you run the motor wizard.