10S custom skate ESC: testers wanted!

Njakts said:
Any chance I can order just the pcb + parts and do rest of the job myself?

You can! It will cost around $85 in smd components and $39 for PCB's. Another $50 to $500 for a hot air soldering station and $50 to $80 in solder paste and stencils.

Definitely doable but not cost effective, more of a learning experience. You need to buy in bulk to get the build cost down.
 
Njakts said:
Any chance I can order just the pcb + parts and do rest of the job myself?

You can have 3 boards made for $39 through oshpark if you are in the US. The gerbers for the boards are on vedder's github. You could also use a cheap toaster oven or frying pan in place of the hot air station.

That being said, if you are asking this question, you are probably best off buying a professionally assembled board from chaka. It is fairly easy to make mistakes in assembly, and the difference in the cost of parts alone is almost negligible.
 
If you want the satisfaction of soldering it your self, I would got to macrofab in the us they charge $150 to print and assemble the whole vesc no wires tho. But I would get then to just assemble the mcu and drv, maybe fets. then you can solder every thing els your self by hand, or reflow these componants in a frying pan or toaster oven, and hand solder the rest, you can buy a reflow controller for $60 off ebay or a whole toaster over with the controller for $160.
 
vedder said:
Regarding 4.10: I think they are fine, but I have _not_ tested them myself since my boards haven't arrived yet. I think I will get 10 pieces some time this week, but I'm not sure.
The changes from 4.8 to 4.10 are:
* The current shunt traces are moved again in an attempt to get less offset.
* A hardware SPI port is routed to the comm header to support high speed communication for e.g. precise robotics applications.

I'm quite sure that the shunt traces won't cause any issues, but the spi port forced me to move some ADC pins and other pins around. This means that the firmware is not compatible between the boards as before (e.g. the 4.7 fw mostly works on 4.8 an vice versa). It could also cause issues with the ADC configuration since not all combinations of inputs, ADCs and configurations are possible, but I don't think this will be an issue. The analog sampling on the VESC is quite complicated because all three ADCs are used in both regular and injected mode simultaneously synchronized to two different timers to sample currents and voltages at two different places in the PWM cycle. DMA is also used heavily.

Hey Vedder - itching to pull the trigger on some 4.10 PCB's! Any word on testing and any issues?
 
After finally managing to complete and test ride my first e-skateboard, I just wanted to say a big thank you to all of you involved in the VESC. It is a fantastic piece of engineering and the level of refinement achieved even at V4.8 is awesome. The usability of the control of the VESC is really impressive and the design, interfaces, software etc. makes it a superb fit for e-skateboards. Stoked that its open source and community developed also!

I have probably clocked up about 10 km now and am still impressed at how well the board and in particular the VESC goes. I have had an issue with occasional cogging on up hill starts and a few instances where my board suddenly surges really hard at about 20 km/h (under light control input) and has thrown me of at least twice (I managed to stay on the last time). I'm fiddling with a few of the VESC settings, start up boost, and speed and position control gain settings to try and smooth out theseissues. Every thing else is so far working flawlessly.

Thanks again for all the great work.
 
chaka said:
Version 4.10 is confirmed, finished testing the first batch last night. I am placing an order for 50 PCB's if anyone is interested.

I know Vedder is busy, but you've discussed with him? This is confirmed fixing the issues he mentioned which required some trace re-routing and improvements? Just trying to make sure it's the current stable platform as i know firmwares will likely split and want to keep it simple moving forward.
 
I made my oshpark V4.10 VESC file public today for anyone wanting to build one of their own. https://oshpark.com/shared_projects/NNzfrwPG

It can be a very rewarding project for those willing to do it. They look good in purple too!

add1e090cc5a1cf278658629f899ede089a1fe73_1_370x500.jpg
 
chaka said:
I made my oshpark V4.10 VESC file public today for anyone wanting to build one of their own.

That is convenient. I didn't realize that the VESC is 4 layer, I thought it was just 2.

It would be great to have such a public link to order ready-assembled VESC in small batches (ignoring the high price for now). Right now it is very complicated to order assembled VESCs, because the number of configuration settings is totally overwhelming. So currently it seems easier to hand-assemble and SMD solder yourself than going to the configuration pages of various assembly houses in the USA or China.
 
Accorrding to Benjamin the VESC has always been a 4 layer pcb. OshPark's upload system was not unpacking the zip file properly and leading people to believe it was only a 2 layer board. I had to rename a few file extensions and merge the drill files before everything checked out in the preview. I was holding off on sharing this link until I was sure the boards worked. They do!

Have to hand it to Vedder on the quality of this hardware. The default current control settings are a little tame with a max current of 60 amps, once I increased this to 80 amps that sweet acceleration returned!

Really enjoy the customization that is possible, right angle micro jst connectors for wiiceiver users who need a thin form factor or heatsinks and 0.5mOhm shunts for high output systems. I think we are going to have a lot of fun with this.

c12dc14fed651aec6fc35ba683f7bba852194bde_1_690x483.jpg
 
12s is the limit, you need to have a low kv motor or you could damage the VESC. I think 12s allows for a max rpm of 200kv.

This is the limit so care needs to be taken when establishing motor parameters. The VESC turns our boards into a performance machine and as such gives us the ability to drive our hardware beyond its limits.

You can split the signal or run them in tandem using the CANBUS.
 
boisrondevens said:
Can i use two esc on the same motor?
Hummina Shadeeba said:
thats a new question! Definite, possible, probably,..but why would u want to do something like that?
boisrondevens said:
I'm using the vesc to drive a big revolt rv-120 pro for a ebike. and I would like to push more amps...

Only if the commutation of the electrical signal for each phase are in sync for both VESCs will it work, but I totally disagree trying to run two phases in one wire given the timing issue...
 
Recently I have been spending a lot of time on getting FOC running and currently I have both sensored (with pulse encoder) and sensorless FOC working. My code is inspired by this code:
https://github.com/madcowswe/Servo/blob/master/FPGA/software/ServoDrive/Servo.c
and the sensorless observer is the one described in this paper:
http://cas.ensmp.fr/~praly/Telechargement/Journaux/2010-IEEE_TPEL-Lee-Hong-Nam-Ortega-Praly-Astolfi.pdf
Oskar who wrote the Servo.c file helped me with getting started. Here is one of his videos:
https://www.youtube.com/watch?v=7gK_3eZ16RI
I will upload something to youtube soon as well.

I will try to push the first implementation to github today if everything goes well. For now there is no auto detection of motor parameters, so that has to be done manually, something like this:
http://cache.freescale.com/files/microcontrollers/doc/app_note/AN4680.pdf?pspll=1
I would measure the resistance in a different way though. I would run a certain current in the motor from the power supply and measure the voltage drop across one phase and divide that by the current, then divide that by two as described.

Regarding hw 4.10: I haven't tested it myself, but it is awesome that chaka has tested it already and has it working. I have done the FOC testing with v4.8 and the current offset seems to be less of an issue when using space vector modulation. I think FOC will work fine with 4.7 as well and 4.10 should be even better.

Can i use two esc on the same motor?
You can not connect them in parallel since there is PWM output and synchronizing it will be almost impossible. If they aren't in sync they ESCs will blow up for sure. What you can do, if the motor has many pole pairs, is to take the motor apart and connect different groups of pole pairs individually to different ESCs. There are even some motors available that have six wires that go to different groups of pole pairs.
 
boisrondevens said:
Thanks for the respond Vedder.
now your telling me that i can use a 6 phases motor with two controller?
also if a rewire a 3 phase motor that should work too right?

Yes, that should work (unless I misunderstand what a 6 phase motor is).
 
Back
Top