thats what she said 
10 Likes
@Deckoz my bro we use the same G502 mouse ! Itās unkillable and super ergonomic right ?
On topic @evoheyax did you put a fat airflow cooled heatsink on your 3 direct fet version or simply ran it with the small one mounted by @GoldenHusky ?
I donāt know if the FETs go dirty because they canāt cope with inrush speed or if they simply fry due to instant / prolonged heat or get damaged over time by ramping heat ?
Cause if the failure doesnāt come from the FET limited speed, then how about thinking of cooling the anti-spark just like youād cool your ESC guys. I mean all of your battery power runs through it dry and hard like we like it (donāt say you donāt like, else you wouldnāt be pressing that throttle so much).
Just a suggestion, not a solution (I donāt have enough tools to probe failure points on anti-sparks unfortunately).
Edit : Note that I submit this point because you may notice how hot connectors are just after a spark. I suppose a similar reaction can occur even upon connection and shoot the FET down.
Thanks for tagging me @evoheyax Iāve been wanting to find an answer to this for a while now. To start I know almost for a fast the FETās arenāt dying to due to heat or prolonged over current because I ran my board on a single MOSFET switch I made for months and it never even got warm. Now Iāve been running my dual hub motor build on a dual IRFS7730 switch I also built and again, it doesnāt even get warm. Now like I said I have always used Hub Motors so I canāt speak for geared builds. I think the problem lies with the voltage spikes from the inductive load and the high peak voltages from regenerative braking. It could also potentially be high inrush current for the capacitors, but like I said Iām not sure. Oh and on both of my switches there is no added capacitance to the gate for a slower ramp up speed.
I would agree with this.
The FETs on both the vesc and the Antispark donāt know the voltage of the system. They just act almost like a bidirectional buck converter. During regeneration the initial spike of voltage, reversing the flow of energy has a period before the electrons hit the pack after passing from vesc to antispark to battery. As the electrons reach the battery a reverse āloadā is applied to the system. Load always causes voltage drops, so the battery is absorbing and pulling down the voltage spike.
In order to fix this. You need to size the capacitor rail on the vesc for the system. As currently the cap rail is for protection of the vesc only. More capacitance will absorb these spikes, or add a reverse current one way TVS diode to shunt the transients from reaching the antispark.
3 Likes
Can you explain why this is? I was recently experiencing voltage sagging due to capacitors fully draining and then not having any juice left to deal with voltage fluctuation. So Iāve doubled up the number of caps on vescās as a result, problem solvedā¦ What can happen with too much capacitance?
1 Like
He just means the spark will be bigger without an antispark(ie worse). But it should help in regards to transients suppression.
2 Likes
What Iām looking for is an anti-spark circuit that involves an led power switchā¦ I mentioned in the first post the xt90 loopkey is what Iām stuck with for now, but that Iām looking for a better solution. Still donāt see a better solution so far 
Hey thatās fine. You could make a model like mine with an led in the head of the switch and contacts on the backside with an inline SMD resistor to apply the right voltage for the LED
You said it yourself, loopkeys donāt fail. Most peopleās issue is you can loose your loop key, it sticks out, and itās tacky. I simply offered an ideaā¦ A loopkey you canāt loose that looks like a switch. Make a micro pcb to sit in the switch head for an LED. I know itās not a button, but you didnāt say button in the OP, you said switch
ā¦ you could modify it to be a button from the side, with springs.
Add a springā¦
Spring is not strong. Just strong enough to keep it open while open so it doesnāt ābounceā. But really you could easily turn this into a button typeā¦
7 Likes
IF the issue is the voltage spike when turning on due to inrush current to the vesc. Adding additional caps to the ESC would make it worse as it needs more energy to fill up those capacitors and the voltage spike would be larger. Iām almost certain thatās what @b264 meant.
1 Like
On powerup, you stress your antispark electronic switch or your loopkey resistor more, but the ESC should perform better. If you donāt have failing e-switches or loopkey resistors, itās better as long as you donāt overdraw your battery on powerup. Having a longer ramp-up on an electronic antispark switch could remedy that.
But itās not a voltage spike outward from the battery that would be higher then the originating voltage. Voltage from the battery will always be voltage from the battery
Voltage spikes on bldc systems when you reverse torque, can quickly rise if the resistance of your pack is to high to accept the incoming current.
This would show up on Ackmaniaās app, no? Iām looking at my graph for today, and a hard brake shows maybe a 2v rise. Unless itās transient. Very curious about this, mayb enough to put a tiny battery powered o-scope on itā¦
1 Like
Iām not sure the STM32 reads fast enough to capture these. I would scope three different areas.
On a phase pair. After the vesc before the cap rail After the cap rail.
Youāll need load to do this. But it will tell you what is being absorbed by the vesc SMD caps, what is absorbed by the cap rail, and what is making it through the cap rail to the AS or battery.
Whatever @akhlut did seems to work. Iāve had it for at least 6 months or so and itās never had any issues. I think he uses a little board as part of it.
Its just a vedder switch, nothing fancy. Im using loop keys now - less to worry about.
2 Likes
It really depends on how hard you push them. I do high amps at high voltage, and this is the trend Iāve seen with these anti spark failures.
1 Like
How can I add a longer run up? Are there any simple solutions? The problem is not lid cause Iāve blown out 3 different anti spark switches on the bench, and I did not turn them off and on quickly, I always let them drain.
In the schematic @ZackoryCramer posted, I would double or triple the capacitance of C1
(if you have a different schematic, please post it)
This is interesting, have you used it for a while and proven it to be reliable over multiple uses? My worry is the alignment might be difficult or gets sloppy after the moving parts wear down over extended use
1 Like
Not quite. The voltage spike can be up to twice as high as the battery voltage. Iāve killed fets rated upwards of 100V with a 60V battery because the system did not suppress the inrush spike.
Hereās an example:
Yellow line, is scoping before the switch and red after the switch. This is after I already implemented a fix to reduce the voltage spike. As you can see red can go above the input voltage.
Electrical engineering aināt easy or intuitiveā¦
@deucesdown, spikes like these are transient and can be as short as a few 100 nanoseconds, so your ESC would never be able to record them. An oscilloscope is sampling at upwards of 100khz, (I think in the mhz range, not really sure)
This is also why you really should not run 13S on VESC without extensive testing that it wonāt burn your components. Most of the components in it are rated to 63V and there isnāt much overhead for voltage spikes like the ones above.
4 Likes