LOW VOLTAGE ALARM

[ex-Gooserider]

If the voltage alarm / meter won't give a useful indication of how discharged you are during use - i.e. 3/4 full, half full, 1/4 full, etc... then is there any sort of status indicator you can use for this, assuming that you are staying with a nominal 24V pack, and one of the usual crop of controllers - not something with the logging you get from Roboteq??

At least on the Redman, it looks like I will need to stick with the existing P&G setup in order to keep the actuators happy, and I'd like to still have a "gas gage" of some sort, at least as good as the status LED's on the controller (a pretty low bar to be sure...)

ex-Gooserider

[Fulliautomatix]

Or connect via a capacitor / resistor that acts to smooth pulses.

I'll do that then, thanks.
But one is puzzled...
With an 80A controller + 8S/9P, I'm drawing 10A per 9p cell = 1.1A per individual blue cell.
They are rated at 10C = 100A/cell.
No load V after a bit of use = 26.5 ish, for the alarm to go under load V = 20.8 ish...
Why is there a 6V drop on such a relatively teeny current draw?

[Burgerman]

Why do you get voltage drop on batteries? Internal resistance. And wiring resistance. And end plate connector resistance.

Its why you get better performance with thicker cables and odyssey batteries compared to stock thin cables and gel batteries for eg.

Odysseys drop to about 18 volts or less, and it keeps on dropping fast after this at 100 amps due to resistance and then then peukert. Gel even more, getting worse as they discharge too.

Its also why I keep saying you need high C rate cells, AND a big pack, and why BMS systems keep incorrectly cutting off power...

High C rate usually also means lower voltage drop under load.

[Fulliautomatix]

If the voltage alarm / meter won't give a useful indication of how discharged you are during use - i.e. 3/4 full, half full, 1/4 full, etc... then is there any sort of status indicator you can use for this, assuming that you are staying with a nominal 24V pack, and one of the usual crop of controllers - not something with the logging you get from Roboteq??

At least on the Redman, it looks like I will need to stick with the existing P&G setup in order to keep the actuators happy, and I'd like to still have a "gas gage" of some sort, at least as good as the status LED's on the controller (a pretty low bar to be sure...)

ex-Gooserider

The only thing that strikes me as useful is the idea of integrating current draw so you have an idea of how much you use out of a full tank.
With some current sensor - preferably not a shunt - an arduino, a display and programming, one would have a reasonable idea.
All I've done so far is drive for a while and fill it up again and figure the battery should last 50km with my sort of usage, which is further than I would ever go in 1 day so it'll be right. Now to shut the bloody alarm up!
In the long run I'd like a gauge so I can confidently recharge every few days to extend the battery life.

[Burgerman]

There are lots of hobby ah and watt metrers

[Fulliautomatix]

Heeeelp! Electronics people...
This was easier before I started thinking about it.
Having one of these;
http://www.wheelchairdriver.com/lithium-ion-phosphate-battery-alarm.htm Which does seem to be the only choice.
And wanting to hold the voltages up for 10? 30? seconds to avoid it going off under load...I need 8 different? capacitors from the common (negative) across to each reference wire??
0 - 3.5V
0 - 7V
0 - 10.5V
0 - 14V
0 - 17.5V
0 - 21V
0 - 24.5V
0 - 28V
Current in the common is 22mA so each little circuit should be ~3mA.
What capacitors size(s) to use?

[Burgerman]

There isn't a common. Only the volts on each individual cell. Each pin is:

neg0 cell1 cell2 cell3 cell4 cell5 cell6 cell7 cell8

You need a cap across each cell, fed by a resistor, then to the meter. And possibly a diode also in series to drop it from 2.7 to approx. 2.2v so it doesent go off too early.

[g7pvs]

Hi,
I think you will find that some of the inputs draw a great deal more current than other inputs.
To provide a time constant you need to feed the input through a resistor to charge a capacitor.
I think the best way to wire the resistors and capacitors would be to connect a resistor from each battery plus to the input to your battery monitor and connect each capacitor from that input to the negative of that battery.
You could try something like 10 ohms and 10,000uF to give a time constant of 0.1 Second (time to charge from 0V to 2/3 of full Voltage) - a capacitor that size will have one end marked with a minus symbol, make sure you get it the correct way around. The capacitors will need to be rated to more than the maximum cell voltage and the lowest prefered value above that is probably 6V3.
Even using 10 ohms, as one of your inputs is probably drawing 10mA or more (the supply to the LEDs) and 10mA * 10 ohms gives a voltage drop of 0.1V, so to get a longer time constant you will need to increase the value of the capacitor, not the resistor.
Another problem is the internal resistance of the capacitor - you may need to go for a low ESR capacitor.
Best Regards
Trev

[Fulliautomatix]

There isn't a common. Only the volts on each individual cell. Each pin is:

neg0 cell1 cell2 cell3 cell4 cell5 cell6 cell7 cell8

You need a cap across each cell, fed by a resistor, then to the meter. And possibly a diode also in series to drop it from 2.7 to approx. 2.2v so it doesent go off too early.

Yes...neg0 is the common. If you disconnect it everything stops, if you mess up the sequence the cells are not recognised - same with Hyperion.

Generally - what a mess to have to do this alarm damping and shit. Be nice to mention it on the alarm page - you don't just get one of these and whack it in, works perfect. Yes, I know we are experimenting to an extent and this is fairly minor, but I'm ready to drive it and all the wiring is tied up and tucked away, now I have to cut it up again and put in a bunch of other stuff.

I agree, the nicest way is to damp the volt drop so it has to exist for 30? seconds before the alarm goes off. I'd appreciate some guidance from someone on values and connection as electronics was never my strong point. Electrics, yes.
No way are they going in the battery box and, yes, some isolating diodes will probably have to go in so the charger is not messed around, which will drop the alarm volts.

Watt meters: http://www.hobbyking.com/hobbyking/store/RC_PRODUCT_SEARCH.asp?strSearch=watt%20meter
I don't like the idea of inserting these into my 8ga cables. They do use a 'clamp meter' sensor so it might be relocatable.
However, they go to 65Ah. Not enough.

[Burgerman]

Realistically, I think the only sensible use of the alarm is to set to well below the usable end voltage as a warning. Say 2.7v - .5 or .6 from a diode. That will warn you before damage on discharge but wont bug you on slopes.

If you want a fuel gauge you need to look at Ah meters. Or watch the Ah back in on the chargers display. And learn to judge distance. The range is huge so its not really a problem. I can set the roboteq to stop or put a red warning light on, if it drops below a set voltage - say 2.2v per cell.

[Burgerman]

Watt meters: http://www.hobbyking.com/hobbyking/stor ... tt%20meter
I don't like the idea of inserting these into my 8ga cables. They do use a 'clamp meter' sensor so it might be relocatable.
However, they go to 65Ah. Not enough.

Its to do with memory or something. The hyperion software also stops counting at 65Ah for a similar reason. Hoping for a fix soon... The charger itself goes on, the software stops! But MrMel knows about this.

You could fit 2 hobby ones to count higher. They use a hall sensor I think, no reason not to add to the battery cable.

The tiny hobby alarms are untested. But I absolutely expected them to go off on slopes or under load. I don't see how any volt measuring device could not. The other solution is to use a tiny ssr and a sensor, or whatever, and only read volts when not driving (zero throttle) position.

[Fulliautomatix]

Hi,
I think you will find that some of the inputs draw a great deal more current than other inputs.

Yes, I think so, unless the input resistors on the alarm are staggered.

To provide a time constant you need to feed the input through a resistor to charge a capacitor.
I think the best way to wire the resistors and capacitors would be to connect a resistor from each battery plus to the input to your battery monitor and connect each capacitor from that input to the negative of that battery.

I really don't want to pull the battery out and add more wires now!!
This is the only way to provide same voltage across RC networks though.

You could try something like 10 ohms and 10,000uF to give a time constant of 0.1 Second (time to charge from 0V to 2/3 of full Voltage) - a capacitor that size will have one end marked with a minus symbol, make sure you get it the correct way around. The capacitors will need to be rated to more than the maximum cell voltage and the lowest prefered value above that is probably 6V3.
Even using 10 ohms, as one of your inputs is probably drawing 10mA or more (the supply to the LEDs) and 10mA * 10 ohms gives a voltage drop of 0.1V, so to get a longer time constant you will need to increase the value of the capacitor, not the resistor.
Another problem is the internal resistance of the capacitor - you may need to go for a low ESR capacitor.
Best Regards
Trev

Thanks Trev.
That's a big chunk of capacitors.

[Fulliautomatix]

Its to do with memory or something. The hyperion software also stops counting at 65Ah for a similar reason. Hoping for a fix soon... The charger itself goes on, the software stops! But MrMel knows about this.

It'll be an integer register then! Counts to 65535 in most languages.

[Fulliautomatix]

The tiny hobby alarms are untested. But I absolutely expected them to go off on slopes or under load. I don't see how any volt measuring device could not.

I assumed that the C rate had so much headroom there would be so little volt drop at the currents being drawn that it wouldn't be an issue.
As per the intended use on RC stuff.

The other solution is to use a tiny ssr and a sensor, or whatever, and only read volts when not driving (zero throttle) position.

Not bad, but they make a racket when turned on as well. I originally thought to turn it off with the chair to save power - gave that idea up.
For the moment I'll tape over the piezos. Might put some 1N4001s in the...

Hang on! It only needs 1 diode in the neg0 wire to drop the whole lot. Much easier.

I'll give a bit of thought to a programmable alarm, voltage AND time, then it might as well have a current sensor and a display...

[Burgerman]

I assumed that the C rate had so much headroom there would be so little volt drop at the currents being drawn that it wouldn't be an issue.
As per the intended use on RC stuff.

With LiPo batteries, at 50 and 70C and above its not! They are so low resistance that one as big as a pack of cigs can fire up your car. But LiFe are not like that. Its why going for 3c or 1c batteries like CALB or the ready to go bricks with a BMS built in are useless.

Its why I am always going on about C rate, resistance etc. It matters!

Its why I cringe when I see yet another post looking at small 50Amp lead replacement monoblocks with low capacity and unknown cells and bms inside... Or ready built lithiums that are not utilising enough Ah.

[robnnorthaustin]

Ex gooserider, The tiny combo volt/amp meter which I posted a pic of in an earlier post has accomplished the objective I had of wanting to have an idea of my state of charge when using my scooter. It is not an exact science but I have something to gauge by, other then just "feel". I can tell you from use that it is more accurate gauge then the stock Pride SOC meter. I always worried when I was spending hours on the scooter and the meter has at least taken away the worry. I would have used one of the hobby watt meters but I didn't want to run my power wires up to my tiller so I ended up using a very beefy shunt and that allowed me to make the entire installation look a bit more professional.