YGE ESC hukkasi asetukset

[jpietari]

Onkohan YGE ESCeillä joku salaliitto meneillään? Markuksen ESC temppuili viikonloppuna ja Gauini ESCi sai siitä varmaan jonkun tartunnan vaikka oli autossa 10 m päässä. Pienen lumipenkkakosketuksen jälkeen moottori ei enää lähtenyt pyörimään eikä ESCi päästellyt normaaleja undervoltage-piippauksiakaan. Akku irti ja takaisin kiinni - ei piippauksia , BEC kyllä toimii. Laitoin kiinni ohjelmointikortin ja kaikki asetukset olivat kadonneet, ESCi otti niitä vastaan ja alkoi taas piippailla ja asetukset pysyivät tallessa virtojen katkaisemisen jälkeenkin ja moottori pyörii taas

[Vispilä]

Yks sana: Kontronik 

Vitsi vitsi, olisko sinne päässyt vettä lumikosketuksessa ja se olis jotenkin resetoitunut oikosulun takia? Olikos Markuksen escissä sama ongelma?


-op-

[jpietari]

Kyllä tuli muutamaan kertaan puhkuttua lumet pois rungon sisältä ja moottorin ympäriltä mutta ESCi on kyllä aika hyvin kutistemuovin sisällä eikä se vaikuttanut ollenkaan märältä. En tiedä tarkemmin mikä Markuksen ESCissä oli.

[mnentwig]

They don't like humidity at all. I think you got quite lucky.

The problem with mine is now understood - it has a "pilot" supply to charge the input caps, otherwise massive sparks result at 12s.
I left this wire on for too long, it triggered an internal fuse.
The fuse resets slowly, but this time I was too fast, so it booted up a couple of times with undervoltage. You could hear the beeps dying, as it ran out of Volts. This resulted in garbled data in the EPROM.

Now that I am aware of the problem, this won't happen again - I simply connect the ESC, then turn on the receiver. Problem solved.

[jpietari]

So I guess I will have to protect it better from snow/moisture.

I forgot to mention the ESC I'm talking about is YGE 18, it has a 25V 330uF Rubycon ZLH low-ESR capacitor and it doesn't have any precharge wire.

I'm sorry I don't understand your description of the "fuse" operation. If properly designed a reset controller would not allow the ESC to start until the voltage has risen and is stable. In case of power loss there should be some brownout detector that tells the ESC "we are losing power" and finish all writing operations to the EEPROM, by design there should be plenty of time to do that and what the heck is it writing to the EEPROM during normal operation? I thought it would write only when the configuration changes.

Btw. sparking is not the only problem when connecting the wires. When the battery is connected a huge current flows from the battery to the caps. When the caps are fully charged the current does not stop abruptly. There is enough inductance in the wires to cause a short voltage peak at the ESC of twice the voltage of the batteries and thinking about a 12S configuration handling 2 * 12 * 4,2V = 100,8V In hot-plug equipment like CompactPCI systems the connectors have pre-leading pins that make contact first and break last and these as used for pre-charging with limited current.

[mnentwig]

No, the pilot wire - same as your "pre-leading pin" is only needed for large, high-voltage ESCs. First they use higher current (= huge input capacitors), and then the energy is proportional to voltage squared times C. At 6s you barely see the spark, but at 12s it's quite impressive.

If used properly, brownout never happens, therefore I don't want to pay for a brownout detector.
I don't know the internals (and don't care very much about them), but if I run a microprocessor and voltage goes too low, i'm not surprised if it goes crazy. Convinces me.
The problem was that the fuse had triggered earlier and I didn't know there was even a fuse.

Well twice the voltage is the worst case (undamped RLC step response), but we have always losses. So let's dump the .8 in 100.8 

Kontronik is an option. Yes, more expensive, but
a) I don't like their attitude regarding repairs
b) they state publicly, we build features into our products that nobody needs but that are solely to cover our *sses if it comes to court and
c) the YGE ESC gives me AFAIK more freedom to tune the governor response to my particular heli.
d) it's also cheaper
e) since I have by now figured out how to program them, I'm too lazy to do the same for another brand's ESC

The ESC was actually still useable, but I had to program it via stick input. I sent it for re-flashing anyway.
We'll see when I get it back. Maybe, if I conclude
f) Kontronik does just what I want
I may change my mind without further notice

[jpietari]

Many microcontrollers have an internal reset controller that holds the microcontroller in a defined state on power up and power down. With the pilot wire I can imagine that the supplies rise very slowly and this could be problematic for some controllers but I guess this ESC is also using a switch mode supply to step-down the voltage from the batteries and these usually have some power good output indicator which can be used to reset the microcontroller and the output of the switch mode supply probably is not rising that slowly either.

Brownout actually happens every time power is switched off but if properly designed the microcontroller is held in reset-state when the supplies start to drop preventing it from getting lost in the code and doing something unintentional.

In any case there are many ways to keep the EEPROM data ungarbled even if the voltages start to drop in the middle of a write process: you can use checksums and several copies of the sama data and modify only one copy at time and you will end up with at least one copy with the correct data and checksum (the amount of memory cannot be a cost issue, e.g. a 4Kbit EEPROM is > 10c in quantity).