Double FET wont help, 98% of the FET’s fail in dead short and only during switching.
1 Like
Let me find the guy who blow his VESC because of the powerswitch failure. It was during a braking, not while switching on/off.
A power switch or connector failure will kill the VESC during braking, as there is no backup path for current to flow. In contrast, a MOSFET failure could actually have more redundancy. As @JTAG mentioned, MOSFETs do tend to fail in shorts, and they also have a body diode to allow reverse current. I think he also meant that MOSFETs usually fail only when switching fast (like in an ESC) because the transition period is when it sees the most load. When it’s on or off, it’s just a simple low/high resistance resistor.
1 Like
just letting you know, im loving your project !! If you were able to provide me with an out-of-the-box, ready-to-use set of esc and bms ( no soldering ) then i’d defenatly be down for a set.
keep us updated, love reading this post.
Regards 
1 Like
I have a few more logistics-related updates. BMS progress is going faster than anticipated, and I’m expecting to need beta testers in a month or so. Please message me if you are interested in testing the BMS. I’m looking to acquire around 50 initial units for testing, and the price will be around $110-120 for the initial units (slightly higher because of the lower quantity, exact cost will depend on how much interest).
As for the ESC, I haven’t worked on it much because of the BMS, so it will still be a while. I’ll hopefully have more updates on that soon.
On a side note, if you’d like to further support my development, I’ve created a donation link here. Everything will go towards prototyping costs (I’m basically buying a new BMS/ESC every time I make a new hardware version).
3 Likes
After ~200 cycles for the samsung 30q batteries.
I looked at the spec sheet and it is on a per battery basis. So a 4P pack charged at 4A, would that be 1A per cell or does each cell in parallel get the the full 4A.
I have a few large packs and want to make sure they last as long as possible.
1Amp per Cell. Make sure the voltage matches
1 Like
Great project, already signed up as beta tester
Two things that I remember I saw on the discussion:
- about the limits of the battery (maximum voltage, minimum voltage, etc) I thinks is better to leave everything numerically customizable instead of presets
- the low voltage cut-off be load dependent, it’s something the vesc lacks. Since we push the battery’s to the limit, we have significant voltage sag, so at most times there’s still plenty of energy left but we hit the cut off. The strategy would be the BMS know the internal resistance and calculate the voltage drop due to the current flow, this could be measured or just a crude estimation, would be better than nothing
Edit: And about the battery capacity display, a option with just simple leds, not ws2812, would be interesting, the way boosted uses, compact and simple to design, i don’t know how many pins you have left, but nothing that a multiplexer can’t do
2 Likes
All numbers are customizable in the dashboard. Calculation of voltage sag is possible, but I thought that the purpose of the low voltage cutoff is to keep the voltage from dropping to damaging levels even under load. Another way would be to only measure cell voltages when the current draw is low enough.
There are two accessory pins, but they support I2C so you could use an LED driver.
2 Likes
What would calculating the voltage drop even help with? End of day you should never go below a certain voltage per cell even when under load.
It would make battery SOC more accurate. I think the method is called columb counting or something like that.
Inboard even experiences early battery cutoff because the voltage sag is so high with their battery. Their battery readings are also of because they are simple meters based in voltage readings.
Having your board cut off from hard acceleration is not fun. This would change that.
1 Like
So would it gradually cutoff since it can calculate the sag before hand instead of the instant cuttoff?
Nice, from what i read i thought it would be predefined configuration. So it’s easy to put just a few leds
From all i’ve read across the years, the problem is rest voltage too low, no problem going low during heavy current draw, it’s actually desirable if you want to get most energy out of the battery
watch from 16:30, it shows the diference in booth methods
EDIT: and about measuring only when there is no load will bring more problems than solve, becouse the battery is no a pure resistive load, it take a long time to the battery voltage stabilze, see the first part of the video
Would be a lot of trouble adding one of these?
More work on the hardware side, but less in making a SOC algorithm
That’s it, the problem i see with pure coulomb counting is drift over time and not knowing whats the full capacity of the battery, the series of the video i posted above shows the main advantages of each method
That’s a very fancy algorithm. All the hardware should be there to implement it in software (if I can understand it).
I still think a low voltage cutoff for a cell should be there as a matter of absolute protection. It can be set lower than the 0% voltage, just to protect the cells. SOC tracking is more for user information (as in, a 0% SOC should not shut off the battery, only tell the user that the battery is empty, maybe with a warning).
Maybe that will work as a safe guard, at least on the beta units I wil look on your code this weekend and start to think on how to implement something similar to the TI algorithm
the way we measure SOC on our batteries at work is by measuring power draw every x seconds and integrating over time in order to get an accurate measure of energy used. Depending on the cell, the user could set up maximum capacity through software. For example a 10S4P NCR18650GA pack has a total of 504Wh, so the SOC percentage would be the 504Wh- integral of power over time.
This can be dangerous though. Batteries are not meant to go under a certain voltage level or permanent damage will happen. At higher current draws, the nominal energy of the cell is lower and pulling the rated amount can over discharge and kill your cells. We remedy this by having discharge curve models of cells at different current and temperature levels which we translate to maximum energy tables to reference when and use as our max capacity.
The simple way to do it is add a 10% margin to the rated energy and go by either that or low voltage cutoff.
Cheers
4 Likes
(off topic) what job do you have?
Aerospace Engineer at a small UAV company called Area-I but since we’re so small (15 engineers) we are pretty well versed in several other areas.
PM for more info. Dont want to drag the thread too far off topic.
5 Likes