Loaded Dervish/dual6355-190/dualVESC/90mmFlywheels/nunchuck

[bbcane78]

Second test ride today.
This time on flat ground (23m total climb) and pretty steady speed using the cruise control

Distance: 9.36km
Avg speed: 18.3kph
Peak power: 1035W
Peak current: 22,5A
Charge used: 2019mAh
End voltage: 45,9V

Max speed on this run was 30,5kph, which corresponds nicely to the theoretical speed I calculated using a max ERPM of 30000 (which corresponds to approx. 4286 mechanical rpm on a 14pole motor) I had set in the VESC to protect me from myself

[strike]The batteries suck, they should be giving me double this range. It was only 12°C but I don't think it's the temperature. I think I got what I paid for.[/strike]
EDIT: I have my voltage cutoff way too high and I'm leaving half the battery capacity unused. I'm a dumbass. /EDIT

No nunchuck dropouts.

I did manage to trip the VESC hard (ie not soft) twice, each time it happened when I tried to overtake a cyclist and accelerated hard.
I have no idea why this happened. Any ideas ?

What was the voltage cutoff originally and what did you change it to?

 

[Trbt555]

I initially had the cut-off start set at 3.8V/cell and cut-off end at 3.7V/cell but in the meantime I've learnt a thing or two:

My Zippy Flightmax Lipos arrived but the weather was bad and I was bored so I decided to see if I could take an objective approach to setting the low voltage cutoff and whether or not I would need different settings for different Lipo packs.
So I had some good clean geeky fun with my Arduino and did a discharge test.
I set up the Arduino to measure the pack voltage over the balance plug while I discharged the Lipo over a 35W halogen bulb. I had the Arduino send a timestamp and the measured value over the serial port to my PC, from where I could import data into Excel.
I had a cell voltage monitor/alarm set up at 3,6V so I could leave the setup alone until stuff started to get interesting.
The 35W bulb drew a current of approx 2,8A. Unfortunately I didn't have anything more powerful, it would have been fun to see the effect of a larger discharge current, especially a current close to the max discharge rating.

My setup:

The results:

Some observations:

- 3,6V per cell will be a good cutoff limit for both packs. Under 3,6V the voltage drops VERY fast. It took both Lipo's about 15 min. to reach 3,0V coming from 3,6V at approx 2,5A, imagine how fast the voltage drops when you're drawing 10 times that current.
- Both curves show a small dent at 4,0V, I don't know what is going on there, it could have been someting in my setup.
- The Multistar takes 13% longer to discharge to 3,6V while it only has 4% more capacity, while the current in both tests was identical. I don't know how accurate my test was but one thing for sure, range with the Zippy's will be shorter.

If I ever decide to repeat this kind of test, I'll put my wattmeter in the circuit as well and maybe even have Arduino monitor and log individual cell voltages, and have it shut off automatically at 3,0V

Now I'm waiting for the weather to clear so I can go out and do a real-life test.
I'm also brooding on the idea of having my Arduino log values from the VESC's over CANbus and store them on an SD card, while also logging accelerometer data, which would give me useful data to analyse after a ride.

 

[WeeChumlee]

Great stuff man!
A graphical depiction always really puts things into perspective.
Would be great to see if you could log the values onto a SD card, that would be really interesting to see graphically.

 

[benj]

...I'm also brooding on the idea of having my Arduino log values from the VESC's over CANbus and store them on an SD card....

This is a great idea I have been trying to get values from VESC UART over Bluetooth to a mobile app (with no joy yet!) but I was wondering if that was even that useful. Bluetooth transfer might not be fast enough and it might not log quickly enough (ie, it might miss spikes and important data).

An SD card logger for VESC would be super...if the SD Card write speeds are quick enough to keep up?

And great build BTW 8)

 

[Trbt555]

I have been trying to get values from VESC UART over Bluetooth to a mobile app (with no joy yet!) but I was wondering if that was even that useful. Bluetooth transfer might not be fast enough and it might not log quickly enough (ie, it might miss spikes and important data).

You haven't been able to get the data over Bluetooth or you haven't been able to get it from the VESC altogether ?
Anyways, I'm running a Nunchuck so using UART is out of the question I think, right ?

Good point regarding the write speed for the SD card, I'll have to look into that.

 

[benj]

I can control the VESC, and ask it for values, and get data, but the return values don't make any sense I posted about it on Vedders forums http://vedder.se/forums/viewtopic.php?f=8&t=14 but we are all drawing blanks.

I'm not sure about the UART / Nunchuck thing, I use 2.4ghz RC remote. Let us know if you find out anything on the SD card speed!

 

[Trbt555]

I had a quick ride after work on the Zippy's, on flat terrain.
Distance 17km / 10,5miles
Avg speed 24kph / 15mph
End voltage: 3,6V per cell
They definitely don't get me as far as the Multistars, so contrary to my earlier conclusions, for a mellow riding style on flat terrain the Multistars really give bang for buck.
I hope the weather stays OK so I can try the Zippy's on some hills.

 

[Trbt555]

After having raided the kitchen drawers for my lunchbox enclosures, I raided the sock drawer and found this silicone rubber swim cap:

Which is perfect for some extra waterproffing of my battery box, which wasn't sealing perfectly due to the thickness of my lipo's:

Aftere scalpeling the top off, I now have a perfect secondary seal for those particularly wet days, which also functions as a safety to keep the lid on tight:

 

[Trbt555]

My swim cap did a good job of keeping the gunk out today. It was pretty wet out. Great for sliding.
Not so great for carving.

 

[haimindo]

Great solution with swimming cap

 

[sl33py]

Genius idea on the swimming cap.

One question remains - why is your swimming cap in with your socks?! :wink:

 

[206monkey32]

This is a beautiful build man congrats. Love the science behind your lipos as well ......the ardrino measure and data to SD would be awesome. I'm curious, why the fuse after the key loop switch in your wiring? I havent seen many fuses used in builds here -what is it rated and what is it for? Protecting the VESC?

 

[Trbt555]

Thanks.
It's an 80A fuse which in hindsight is way too high, it will probably maybe only protect the batteries.
It's not in the way so I guess I'll just leave it there.
My max amps going uphill never get any higher that 35A, so a 40A fuse might be more suited.

 

[Trbt555]

Genius idea on the swimming cap.

One question remains - why is your swimming cap in with your socks?! :wink:

Don't you keep your swimsuit in your sock drawer ?

 

[Trbt555]

I've been making some progress on my on-board DIY battery status / VESC data display.
I finally found the time to draw a housing in Sketchup and have it printed via 3Dhubs.com, which by the way is an awesome service.

I solderded a JST connector to the OLED display, and fixed it in place with some hot-snot:

Finished product after I pimped it with some carbon-vinyl:

Next steps: migrating my prototyped Arduino Uno setup to a dedicated Arduino Nano board.