Accidentally left Cycle Analyst on overnight, drained my batteries to death?

Baron

100 W
Joined
Jun 15, 2015
Messages
209
Location
Philadelphia, PA
Here's the situation: I rode my ebike over the weekend, used about 75% of the battery and then stored the bike away, but i forgot to unplug my battery / turn off the ignition to the controller, which left the Cycle Analyst screen on. When I checked the bike about 24 hours later, the screen was still on but it was blank, which was strange. The backlight was still on and I couldn't cycle through any of the menus (everything was blank), but I could enter setup. When I checked on my batteries (12s lipo 10ah) the cells had all drained to less than 2V, some less than 1V. Since these are lipos I am definitely not messing around trying to revive them from this low voltage, I just consider them trashed at this point.

The only thing that could have been draining the battery was my controller or the cycle analyst, I have nothing else electronic connected to them. I don't have a bms, which probably could have saved the batteries. The controller is a 12 fet Powervelocity bought from a member of these forums. But I think the culprit was the Cycle Analyst since the screen was left on for over a day. I am just amazed that the cycle analyst could have drained that much battery in just 24 hours. There was still about 2.5ah remaining in the batteries when I stored the bike, I had only used about 7.5ah out of 10ah total capacity during the ride. Is there something wrong with my cycle analyst??

I realize it's my fault for forgetting to turn off the ignition switch to my controller, but shouldn't there be an automatic shut-off function for the cycle analyst and controller so this doesn't happen and drain your battery to an unusable state?
 
Sorry for your loss.

I have done this. But not in one night. Usually it takes longer.

The Cycle Analyst doesn't draw much power. But the controllers draw quite a bit to power all their logic, and of course the combination was left on here.

The CA has a switching regulator that is quite efficient. But the controller logic power regulator may not be efficient, some of them are linear regulators down from battery voltage to 12 and 5V. Some are switchers but they still have to generate enough power for all the controller logic and FET gate controls.

An automatic shut-off would be nice and could be done but it has to operate at the full battery voltage and contain a timer.
 
Most of the controllers I have use anywhere from 50 to 150mA and up, just sitting there idle.

I forget what the CA draw is, but it's less than that (though if you have a torque sensor or any other device like headlight that takes it's power from or thru the CA, it may read higher).

Together all the things that have idle current draws on a bike could be high enough to make a difference to an already-mostly-discharged pack.



An auto shutoff could use a timer made from a 555 chip, reset by pulses from the speedometer. When there aren't any pulses for a certain amount of time, say, 10-30 minutes, it stops being reset. The output of the 555 could control a small FET that switches battery positive on/off for the keyswitch/ignition line on the controller, and the battery positive input of the CA. Those are both low enough current that the FET wouldn't have significant heating, so it may not even require a heatsink as long as it's in open air. Or it could be bolted to the controller's casing, from inside or outside.
 
Baron said:
The only thing that could have been draining the battery was my controller or the cycle analyst, I have nothing else electronic connected to them.
...
I don't have a bms, which probably could have saved the batteries.
...
But I think the culprit was the Cycle Analyst since the screen was left on for over a day.
...
There was still about 2.5ah remaining in the batteries when I stored the bike,

The CA draws about 15ma.
The controller draws about 100ma in a quiescent state. Simple math - over a 24 hour period the consumption is:

CA: 0.015A * 24hr = 0.36Ah
CTLR: 0.100A * 24hr = 2.4Ah (about all the charge you claim was left...)

>>> The controller is the culprit...


Baron said:
I realize it's my fault for forgetting to turn off the ignition switch to my controller, but shouldn't there be an automatic shut-off function for the cycle analyst and controller so this doesn't happen and drain your battery to an unusable state?
You don't mention what type of CA you have, but the CA3 has configurable Shutdown Voltage and goes to sleep below that threshold. However, this is really to save data integrity. System-level power management is beyond the capabilities of the CA as you see above - it doesn't control the power to the rest of the vehicle - that's the job of the BMS - or in this case with no BMS, you.


Alan B said:
The CA has a switching regulator that is quite efficient.
CAs use linear regulators.

EDIT: okay - a cross post with AW - same conclusion, though...
 
I understand now, I guess that is the risk of running without a bms. Not a huge loss on my end, just around $120 worth of batteries. Lesson learned I suppose.

Idea: would it be possible to wire the controller ignition wire to a CA3 so that it can switch the ignition off when you reach the configurable Shutdown Voltage threshold? Such a feature could have saved my batteries in this case.
 
Directly, no. The CA doesnt' have any output to do that with.

If the CA shuts off the 10v torque/PAS sensor power line, or the 5v output for the throttle, you could use that to drive a transistor that turns a relay or FET on that turns on teh controller ignition/keyswitch wire. But I don't think it does actually shut off it's voltages; I think it just puts the MCU in a low power state (sleep). Can't remember, and couldn't find the info in a quick search on Grin's site or the forum.

If the Throttle Out pin goes to 0V when it's in sleep, but normally it's above that, you could make a comparator circuit that detects this and drives teh relay or FET to turn the ignition/keyswitch line of the controller on.
 
You could wire a BMS up to your LIPOs. If you used a smart BMS you could have active cell level monitoring on a smartphone.
 
I don’t let RC lipo plugged to anything without direct surveilance. No one should. Sleeping voltage of RC lipo is 3.8v/cell, that should be a concern to anyone using them.

You don’t need a BMS for RC lipo. If you feel that you do because you can’t manage them yourself, what you really need is a safer chemistry.
 
MadRhino said:
I don’t let RC lipo plugged to anything without direct surveilance. No one should. Sleeping voltage of RC lipo is 3.8v/cell, that should be a concern to anyone using them.

You don’t need a BMS for RC lipo. If you feel that you do because you can’t manage them yourself, what you really need is a safer chemistry.

Yes, but you could use a BMS for the same reason you would with any other chemistry.

In this situation the BMS would have cut the power at the set minimum voltage and it would have been happy days. Charge back up and off you go.

Humans do make mistakes and a BMS can help when/if you do. It's not really feasible to just say, 'never make a mistake'.

It should be pretty easy to make a harness to connects the BMS to the lipo balance leads and discharge leads.
 
That's why I put a key switch to turn on the CA and controller logic on every ebike I build. An ebike left on is dangerous, and without a real key switch it's simple and common to forget. With a key switch it quickly becomes habit, and the habit should be to turn the bike off before getting off the bike.
 
I'd like to see a key switch and a kill switch in series on any ebike, removing power from all the low power battery voltage items. This includes the CA, lights, and controller at least. Being able to remove power without reaching for a keyswitch is a safety requirement for motorcycles. There's no reason an ebike should be different, especially when just moving the pedals or bumping a throttle can bring on power suddenly (without the warning of an idling engine).

Note that the key/ignition switches may not be sufficient for long (or even short) term storage with all controllers, many of them have bleeder resistors of around 10k which will draw 8mA at 80 volts which can damage a pack in a fairly short time period. This resistor should be removed, and the drain reduced to microamps. Only the leakage of the power FETs and Capacitor bank should remain. For longer term storage the circuit breaker should be opened so the only drain is battery internal leakage (and the BMS if present).

Even BMS's draw a little power all the time (amount varies), so leaving a battery connected to a BMS over the winter will draw power from it (that could be damaging). The amount of current draw is a characteristic of the BMS quality, a good one will be really low. Others may be a lot worse. It is preferable to disconnect the BMS for extended storage.
 
No Shame, I have done this before as well, just dumped the lipos and bought a new set, as others have said a BMS or battery murdering system as I normally refer to them is not required, back in the early days of lipo testing 12 years or so ago all of the problems came from BMS failure, ive been running with no BMS for over 10 years with no problems other than the one time I left the bike on versus the 4 or 5 failures I had with BMS fets failing when cutting off under load, I have a 4QD illuminated battery gauge on the handle bars now which makes it almost impossible to not notice and leave the bike switched on, If I dont notice that then I consider it my own stupid fault.

I have nursed lipo back from 2V though, very slow and low charging or single cell charging can sometimes work, Lipo is so cheap though its not really worth the fire risk.
 
Back
Top