UPDATE
Flispky’s new smart antispark is an implementation of the V1.
There is a hardware revision for the V1 to get its current draw from 2.5mA down to 300-400uA, but it will most likely have to be implemented in future production runs. I realized my mistake too late and after the first fab order was put in. Prototype testers reported no issues with the antispark draining the battery. I also left one of these in my board for like two (three? some long ass time) months and it went from fully charged to 36V, It’s a lipo stack, and I suspect that the battery’s internal leakage is actually higher.
Trust me, 2.5mA is nothing. If you charge your board at least once a week it won’t be a problem.
If you need to store your board for more than a week, I would highly recommend unplugging everything regardless of what hardware you have in it.
I didn’t even take my own advice and still didn’t have an issue.
I will open an issues and feedback thread here Flipsky Smart Antispark Issues, Feedback, and Review
Since I keep running my mouth off on other antispark threads, thought I would make my own.
Starting a discussion on this before I roll out prototypes.
Features
- HIGH SIDE SWITCHING << VERY IMPORTANT, NO MORE CANBUS FIRES.
- In-rush current limiting as specified in the ltc7004 datasheet, dV/dt is ~184V/s
- LTC7004/7001 means driving N-Channel MOSFETs.
- 12S operation, swapping the ltc7004 for the ltc7001 and the MOSFETs means 20S operation (there will be an ESC that supports 20S, in time
). - Roll-to-start, works at low speeds, but will also trigger if you give the board a good enough bump. WAIT FOR THE VESC TO BOOT UP.
- Auto-turn-off, detects changes in current through a proprietary algorithm, if no change for 20 minutes, turns off. This also means if you decide to coast for 20 minutes, or apply a constant throttle for 20 minutes, it will also turn off. There is a 10 second delay after it turns off.
- With a heatsink, it will be able to handle 100A continuous without issues.
- OPTIONAL Momentary Pushbutton, 0.3s to turn off, 1s to turn off, holding it down will cycle it on, then off.
-
Easily reprogrammable attiny85. Firmware updates or installing custom firmware is as easy as bodge wiring an arduino uno to the programming pins.Don’t reprogam this thing unless I explicitly tell you to. - Production versions will be around 30x30mm or 30x40mm.
- Very low quiescent current draw, more details on this below.
LIMITATIONS
- Maximum output capacitance of 6,000uF. Beyond this, auto-turn-off gets finnicky, beyond 10,000uF, the MOSFET popcorns.
- NO HOT SWAPPING (this is true of ALL antisparks) when installing, plug in the ESCs first, then plug in the battery, and do the exact reverse when uninstalling (unplug from battery first, wait for ESCs to turn off, unplug ESCs) Try to make sure everything is off when uninstalling.
- No short-circuit protection. This is what a fuse is for. If this device attempted to do that, both the MOSFETs and the gate driver will popcorn from the negative transient.
- Auto-turn-off will not work with messed up ESCs, like some of mine are, namely they get hot for no reason while idling, no clue how they still work on 10S FOC.
Other Details
- There will be two versions made by two different manufacturers. One version will use a high voltage linear regulator to supply the logic circuitry, this one will be low cost. Quiescent current draw is
2.5mA400uA in future revisions at all voltages. The second version will usea buck converter to generate the logic rail, quiescent current draw will be ~20-40uA. The second version will be more expensive. Buck converters are not cheap.
a second LDO to generate an intermediate rail and also implement proper current sensing, this will have a quiescent current of <200uA. In testing, I got it down to 126uA.
To put this in perspective, 200uA will drain a 200mAh capacity in just over six weeks.
The LDO version, a trusted forum member is testing my one and only hand assembled prototype, and I have an OEM prototype. Testing should conclude within the next couple of weeks. I may be able to send the OEM prototype for further testing to another trusted member.
There were five, but now there are three low power prototypes that will be hand assembled, I will be testing one myself, and sending the others to trusted members for testing. I may order up parts to make another five. I am currently resolving issues with the power supply circuit as my dumbass joshineered the buck converter and neglected to include ripple injection that the datasheet clearly mentioned was required.
So now there are two versions, a V1 and a V2.
The V1 will be the cheap and simple version, the V2 will have better power consumption and a couple extra features, namely current sensing.
The difference is, the V2 lasts ten times somewhat longer on battery and will attempt to have short circuit protection, but will cost more and have a more complicated PCB (may affect size).
For now, it appears the V2 will be more expensive, but smaller and slightly lower power handling capability. (Smaller clamping diode to greatly reduce size, I refuse to rate it a current above the diode’s maximum forward current)
The V1 has already been tested, prototype V2s will be available to select testers in late November and early December.
What makes this antispark different from all the others is the use of a discrete high side n-channel gate driver IC. The chip is $5.