Controller standby power

[SpeedEBikes]

All of the controllers I have used have fairly high standby power requirements, ranging from about 2 to 12 watts depending on the supply voltage. If one forgets to turn off power the standby drain can completely discharge a pack in a day or two, especially if the pack was fairly deeply discharged in the previous use. Even with a BMS with per cell LVC, trickling a lithium battery down to the cutoff is hard on it, especially if the cutoff point is set aggressively low as is typical. Without a cutoff lithium packs can be destroyed and most other chemistries will be somewhat degraded by such a deep discharge. Even if no harm is done it is still annoying to find ones battery significantly more depleted than one was expecting. I usually leave my lithium batteries standing at around 3.7 v per cell which is sufficient to do my typical shorter errands on a moments notice without any charging.

My guess is the large standby power draw is due to sloppy controller design. It goes up fast with higher input voltages and I expect most of it is drawn by crude voltage regulators for the low voltage supplies.

Are there bldc controllers that do this right and only use a few milliwatts of standby power regardless of input voltage?

Are there any external devices that could be placed between a battery and the controller that could accurately detect an 'idle' situation and cut power? Can this be done by a drain brain / cycle analyst?

At worst, could a timer switch be used such as are put on heat lamps in bathrooms? I'd be concerned about inadvertantly switching off power while under load. Are there timer switches available that can cleanly disconnect a 20 amp draw at 75+ DC volts and would the switch see much of an induced voltage spike? Is abruptly disconnecting power hard on the controller and could it cause controller failure under any circumstances? I've had BMS units repeatedly cut power at full throttle without controller damage. But I've never had a wiring failure or BMS disconnect while blasting down hill when motor back EMF is super high. Are there timers with long duration settings, say 4 hours which would be sufficient to minimize risk of an inadvertant shutoff while in use?

How hard would it be to mod a controller to reduce standby power? Are there ICs available that can do efficient voltage conversion/regulation for the low voltage supply?

If standby power could be dropped to a fraction of a watt and with the addition of a flashing power LED and perhaps a piezo beep every few seconds after say 5 minutes of 'idling' I'd be pretty happy.

 

[fechter]

I don't think you can do a whole lot about standby drain. With clever microprocessor design, you could make it go to sleep, but that would require a complete software hack. Some kind of auto-off feature could be done. You really only need to switch the input to the voltage regulator.

 

[Russell]

Put a disconnect switch between the battery and controller...or remove the battery completely when charging or not in use.

-R

 

[spinningmagnets]

edit: silly response deleted so as to not confuse others.....

 

[EMF]

Here is something *external* that might at least give you some ideas of how to shut down your pack and even start it back up when it is needed. This device would perform the function you want. But maybe you could google around and find something else, but good luck on DC current amp sensing relay. This was the best one I could find. As long as you could tap 24 volt DC from your pack to power this DC current sensing relay, they sell one that runs on that voltage. But, these are not cheap at about 85 bucks.

Having said that, sometimes to get what you want and to play around with something, you don't care. You can select the amperage range and also the time delay and they are both adjustable.

I think you need the "LL" model. You could put a pushbutton on your handlebar or some handy hidden spot to reset the relay by cutting the DC supply to the relay. Actually, this is kind of a nice safety feature as if you left your bike unattended for a couple minutes, ( or for however long you set the timer) then no one could turn it on again unless they knew to push the button first.

http://www.crmagnetics.com/newprod/ProductView.asp?ProdName=CR5395

You can build just the one you want if you look at the PDF: http://www.crmagnetics.com/pdf/5395a.pdf

I use mine a lot and it works great. May not be exactly right for your application, but it may give you some ideas....

Here is my application for this device: http://endless-sphere.com/forums/viewtopic.php?f=14&t=3783&p=56169&hilit=current+sensing#p56169

 

[Drunkskunk]

I found on mine it's the voltage regulator for the IC thats burning up the power. its usuing a resistive voltage devider to pull the voltage done before an ic regulater in a t220 package drops it down to the main board voltage. I believe it was sucking up 5 watts in standby.

A switching regulater or DC/DC converter would be far more efficent, but not as bullet proof.

 

[SpeedEBikes]

Fechter & Dunkskunk, if most of the standby load is coming from the low voltage circuit / regulator, perhaps a quick and dirty solution would be to power that from a second low voltage battery. Say a 12 volt nicad that enjoys a deep discharge now and then. The regulator wouldn't have to bleed off much voltage so the standy current would be much lower too. I suppose one could rig this up with a bypass switch to still take power from the main pack for emergencies when the low voltage pack is dead.

Drunkskunk, rethinking what you said about the resistive voltage divider, I need to open up my controllers and look for them and find the resistors selected. I bet they were picked to allow operation at fairly low voltages, say 24 volts or maybe 36 volts and are likely feeding much too high a voltage to the regulator when a pack is at 90 volts. And since I never use less than a 48 volt pack I could probably significantly mod the divider ratio and also at least double if not quadruple their series resistance. Ought to work much better then. Still bypassing the voltage divider altogether and maybe even the regulator too with an appropriate low voltage pack might be cleanest. I wonder if there could be any problem having independent shut off for each of the low and high voltage circuits?

Russel, in my post I wrote "If one forgets to turn off power the standby drain can completely discharge a pack in a day or two, ..." which implies that there already is a switch or way to disconnect power. The problem is a switch is only as foolproof as the user. Car manufacturers realize that people can sometimes FORGET to do the easiest things. That's why it's hard to leave car lights on by accident and if you leave your keys in the ignition after turning the engine off you get those annoying beeps prompting you to take your keys out. It took car manufacturers decades to address this. Hopefully it won't take as long for ebike controllers.

Spinningmagnets, I'm not sure why you are talking about chargers in this thread I started about controllers. If you want to do timed charging there are off the shelf solutions for it. There are also chargers and power supplies with switched outputs making them appropriate for use with an external timer. Some use relays, some use transistors, but either way the result is the same. If you cut power to the charger the output will be disconnected and remain so even after power is restored. At least until you use the controls to initiate a new charge cycle or switch the output back on.

EMF, thanks for the product tip. Not exactly what I was looking for but good to know about. A shame its so expensive.

 

[deardancer3]

Put a disconnect switch between the battery and controller...or remove the battery completely when charging or not in use.

-R

I did find that Giant and the EVGlobal/Merida NIMH ebikes had extremely low standby power. The EVglobal full size was horible. I suspect bionx is a little above average, but it is hard to measure.

But like said above, just remove or cock the battery for no connection. That is the simplest.

d

 

[fechter]

On the controllers I've measured, the control circuit draws about 50ma. If it uses a linear regulator, then it draws the same from the entire pack. A switching regulator could reduce that amount, but it would still be more than you'd want to leave on for a long time. You really need some kind of automatic switch to disconnect the input to the regulator when it's idle for a given length of time. It wouldn't be too hard to make a circuit that does this.

An second battery would work, but would be sort of a pain to maintain. It should be around 12v to supply the FET drivers and the 5v regulator.

 

[SpeedEBikes]

Of the controllers I've used, increasing the input voltage increases the standby current. Not linearly like with a purely resistive load but still pretty bad. The controllers I'm currently using draw about 7 watts at 80 volts which would about be about 87 ma. I should hook them to a variable supply and determine exactly how the current changes with voltage.

I'm not crazy about the idea of a second battery but it's a solution within my skill level that I could implement fast. I suppose I could design a circuit that could say watch the throttle and if the the throttle remained off for more than 5 minutes it would cut off power to the low voltage regulator until a reset button was pressed. But I'm not terribly experienced or good at that sort of thing and it would take me quite a while and the end result would likely be suboptimal and more expensive to to fabricate and install them then it should be.

As for a second battery option, I could see it being fairly simple to use. Imagine buying a cheap nicad 12 v cordless drill kit complete with battery and charger. Take an angle grinder with a cutting disc and chop the handle off the drill. Mate the handle to the controller case. Rewire the controller to take the 12 volt supply from the drill handle. Instant 'cordless' bike controller! Plop in the drill battery to make it go. Plop the battery in the charger when you get home. Forget and leave the battery in the controller and not too much damage is done other than if you have to immediately go on a trip you either have to have a 2nd drill battery or a super secret override switch to reconnect the low voltage regulator to the main battery input. Forget to turn off the super secret override switch and we are back to square one... But at least that requires a double screw up instead of a single screw up.

On the controllers I've measured, the control circuit draws about 50ma. If it uses a linear regulator, then it draws the same from the entire pack. A switching regulator could reduce that amount, but it would still be more than you'd want to leave on for a long time. You really need some kind of automatic switch to disconnect the input to the regulator when it's idle for a given length of time. It wouldn't be too hard to make a circuit that does this.

An second battery would work, but would be sort of a pain to maintain. It should be around 12v to supply the FET drivers and the 5v regulator.

 

[TylerDurden]

The used Schwinn/Currie S1000 I just bought has an auto off feature in the controller; so manufacturers are aware, to some degree. It is for a brushed motor. http://www.monsterscooterparts.com/cu36v30a6co6.html

Clothing irons have motion based auto shutoffs. You might find one at the secondhand store and hack it.

For a very simple solution, a tethered deadman key (found on treadmills) can be put on a retracting keychain (now commonly used for access badges): the deadman key can close a relay in the battery path.

 

[biohazardman]

The 48v controller I recieved with my 600BMC kit is at .038A at idle. Measured it a couple days ago. Sticker says Brushless Motor Corp. USA 36-48v 30A max. I don't think the controllers should use that much at idle. Hope you get it figured.

 

[317537]

I use a switch, although its not what the op is looking for. I rig my batteries up to the middle pole of a dpdt switch so when the controller is switched off it connects the battery to the charger plug. Makes sure the standby is off incase the mains power goes out or if the charger any of the chargers plugs become unplugged.

 

[kZs0lt]

The problem is a switch is only as foolproof as the user.

I seriously need to address this problem, once I've ruined my battery this way.
I am using a konion pack with no BMS, charging with a RC balance charger and using the controller LVC. Have a main on/off switch.
I had almost completely drained my battery after the ride. Next day was checking something on my bike, and left the switch on, no recharge prior to that, battery was already depleted(and was pretty out of balance turned out).
About 2 days later I find my 18s(66v) battery sitting at 6v, most of the cells less that 0.5, a couple at 3v, some reversed at -0.3v

I had a hard time reviving some of them, replacing the totally screwed ones, fortunately had some spare cells from the initial build.

At the moment I'm trying to do the transistor mod, to be able to use the controller with a wider battery voltage.
So the separate battery for controller logic seems a good idea, would resolve both issues at once, however I don't like the idea of a separate battery pack to charge and maintain. NiMh-s need complex charger, can't use LiIon as I'll just kill the controller logic battery the same way.

The ideal would be some sort of circuitry that would use only uAmps, watch for throttle input, turn off controller after 5 min inactivity and turn back when used again. That could be powered from main battery, and take months for the same disaster to happen, or using a separate low voltage batt (like 5-10v) that could last years.
Without this a 1-5w constant controller draw is hard to feed with additional battery, that would be needed to be recharged every trip and would totally deplete every time if used only weekly and switch was left on.

Do you have any ideas for such a circuitry, or some other good method to the problem above?

 

[SpeedEBikes]

I haven't dealt with it yet, but it is still high on my list. One potential solution would be to have the battery connected to the controller through a relay or solid state switch. One could use a resistor voltage divider to set the low voltage threshhold that would keep the relay on.

I currently use Junsi celllog battery monitors. I have them set up to trigger a very loud piezo buzzer if any cell drops below a programmable threshhold. I might try using one to control a switch to cut power too. If I use multiple celllogs, I could set one at a higher threshhold to sound an alarm and one at a lower threshhold to completely cut power. I suspect that if it is possible to rig a separate battery to power the logic level voltage of a controller it ought to be possible to wire things up such that there is a switch or relay that cuts the power to the logic side only. This would require a much smaller switch or relay than one that would be needed for the full output power.

It does raises the question of what would happen if power was cut to the logic side while under load at speed. My guess is the output transistors and associated circuitry should switch off cleanly, but I'm by no means certain.

 

[magudaman]

How about a dead man switch? Something maybe on the handle bars that requires you to hold onto or like a treadmill as soon as you step away it pulls the plug.

 

[kZs0lt]

I suspect that if it is possible to rig a separate battery to power the logic level voltage of a controller it ought to be possible to wire things up such that there is a switch or relay that cuts the power to the logic side only. This would require a much smaller switch or relay than one that would be needed for the full output power.

It does raises the question of what would happen if power was cut to the logic side while under load at speed. My guess is the output transistors and associated circuitry should switch off cleanly, but I'm by no means certain.

My idea was to use a relay or circuit that would switch off the logic side of the controller from the main battery. The power FETs would be still on the battery, but their current leakage is very small, probably would take years to deplete the battery only by FETs that are off.

How about a dead man switch? Something maybe on the handle bars that requires you to hold onto or like a treadmill as soon as you step away it pulls the plug.

I will have to think about implementing somehow the dead man switch. I just don't want to set up restrictions, I like to drive without hands, without pedaling, or having the legs on the pedal, without sitting on the seat, so there are not many choices left... I also don't want to use big powerful relays, small electronics are better and they should handle my 65mA controller logic current easily.

I'm not sure what specific device are you referring to by treadmill switch. TylerDurden mentioned motion based shutoff clothing irons, well I'm not familiar with those either. Those could be pretty rare here. Could someone point me out a product using this technique or some description about these "tethered deadman keys" found on treadmills. (I have to admit my english is still deficient, and I'm not even sure what is a treadmill). What I know we have here and is becoming pretty popular are motion sensors coupled to outdoor lamps 220v. I'm not sure how could I use one of those.

Forget to turn off the super secret override switch and we are back to square one... But at least that requires a double screw up instead of a single screw up.

At least while I find a decent solution to this problem, I will make a habit of recharging the battery as soon as possible after a ride, besides my habit to switch the battery off However I still need to check my battery often, as it could last only 3 days on a full charge, but still more chances to catch it.

 

[SpeedEBikes]

At least while I find a decent solution to this problem, I will make a habit of recharging the battery as soon as possible after a ride, besides my habit to switch the battery off However I still need to check my battery often, as it could last only 3 days on a full charge, but still more chances to catch it.

Unfortunately, if you are using lipo or liion, they age more rapidly when sitting fully charged. I like to leave mine at around 3.7 volts per cell and just charge them as needed before use.

 

[kZs0lt]

Yes, they are aging more rapidly under full charge, but my charger is slow, I make a full charge overnight. I tend to keep them fully charged to have them available most of the time. Definitely not good for them, but with my stubborn usage I don't think they'll last many years anyway

 

[SpeedEBikes]

Yes, they are aging more rapidly under full charge, but my charger is slow, I make a full charge overnight. I tend to keep them fully charged to have them available most of the time. Definitely not good for them, but with my stubborn usage I don't think they'll last many years anyway

Many lithium chargers are set for 4.2 v per cell, but if you can set it down a bit it will significantly reduce the aging. 4.1 v seems to be pretty good compromise of extended lifespan while still getting a decent amount of capacity per charge.