# Battery level tester.

Discussion in 'Electronic Design' started by Colin Dawson, Sep 25, 2004.

1. ### Colin DawsonGuest

Hi all

I've got a circuit that I use to measure the voltage of a battery. The idea
is that it will give me plently of warning of when the battery level is
getting low. It works great for the actual circuit design, but there's a
problem that I've got when it's in use. First I'll show you the Circuit
design.

<http://www.cjdawson76.btinternet.co.uk/images/astronomy_batterymonitorfulls
ize.jpg>

To explain the application a little futher. The 12V Battery is an 85Ah Lead
Acid Leasure battery (For use in boats and Caravans). I use it for my
Astronomy Hobby to power, my Telescope, Laptop, Dew heater (like a cars rear
window heater) and all the other toys that I need during an observing
session.

Although I've not measured anything specific I think that when everything is
turned on as the same time, I'm pulling about 10Amps. Which means there's
enough juice for a whole nights observing.

I've got a problem with the battery meter, in that as I turn stuff on, and
the amount of current drawn increases, the meter shows a voltage drop, and
since the difference between a full and empty battery reading is about 2V
(12V= full 10V = empty) this shows a significant drop on the readout.

Is there something that I can do to counter balance this high current
voltage drop?

My thought on this is to place a kind of Ammeter into the circuit, that will
adjust the value of VR1, in accordance to the amount of current drawn
through the circuit. Changes in current would would effect this part of
the circuit, and continuously trim the battery meter, so that the readout
remains stable. (and hopefully correct)

Regards

Colin Dawson
www.cjdawson.com

2. ### martin griffithGuest

Hmmmmm.... 85Ah at 10A for a night, not nice. Look up depth of
dishcharge (DOD). You are going to ruin the battery very quickly at
this rate of discharge for a 85A battery.

http://www.varta-automotive.com/eng.../antrieb_beleuchtung/antrieb_beleuchtung.html
http://www.batteryuniversity.com/parttwo.htm
http://www.uuhome.de/william.darden/
http://www.yuasabatteries.com/pdfs/TechMan.pdf
"Deep discharge or prolonged discharge leads to harmful sulfation."

martin

Serious error.
All shortcuts have disappeared.
Screen. Mind. Both are blank.

3. ### John FieldsGuest

---
If it's due to the internal resistance of the battery, no. If it's
due to voltage drop in the wiring to the loads and you've got your
meter at the end of the parallel run, yes. Read on.
---
---
That's cheating.

The first thing you should do is connect the meter directly across the
battery terminals. This will serve to measure the _real_ battery
voltage, not the voltage that's left over after it's travelled down
some length of cable.

The second thing you should do, (maybe the first IMO, is to get
rid of the circuit you're using and buy yourself a cheap little 2-1/2
digit panel meter and connect it directly across the battery with
leads as long as you need to get it to where you can read it
conveniently.

4. ### James MeyerGuest

Measure the current draw with a current sensor. Convert the sensed
current to an appropriate voltage and use the voltage to drive a small, toy
sized, motor with a gear train on its output. Attach pointers to a couple of
the gears and you will have an amp/hour readout similar to the gas or electric
meter on the side of your house. Reset your meter to zero before use, always
start your observing sessions with a fully charged battery, and Bob's yer uncle!

Jim

5. ### John WoodgateGuest

I read in sci.electronics.design that Colin Dawson <>
That shows your 'voltmeter' but it doesn't show the loads you connect to
the battery, and that matters; see below.
Presumably, though you don't tell us, the meter is not connected
directly to the battery terminals but downstream towards the loads. What
you are observing is the voltage drop in the wiring due to the 10 A or
so that is flowing.
Yes; you need THICK wires. Since I don't know how long your cable runs
are, I'll give you a possibly overkill solution. Think 6 mm^2 or 10 mm^2
twin and earth from your local electrician. This is probably the
cheapest; if you tried to buy it from a supplier you would probably have
to buy 50 m or 100 m or pay through the nose at Homebase or B&Q.

You will get *some* voltage drop even at the battery terminals, but it
should be very small at 10 A if the battery is in good condition and
properly charged.
No, all that would do is to alter the calibration of your 'voltmeter' so
that 10 V appeared to be 12 V. I don't think you would be satisfied with
that!

6. ### Colin DawsonGuest

Hi Martin.

Don't worry about that issue. I've not actually tested the ampage in real
life. I only said 10 amp, so that people would get the idea that there
could be a high current which drops the battery voltage.

If you want a more accurate measurement, this should be a little closer....

1. Telescope - 1.5 Amp when slewing, 500ma when tracking (it tracks for most
of the night)
2. Dew Heater - 4.5Amp when at full power but never need full power (it
normally rungs at 1amp, about 20% of full potential, thanks to the PWM
circuit, that controls the heater)
3. Laptop - no idea, but the battery monitor does drop alot when the hard
drive kicks in.

All in all, I can use the battery (it's not a "CAR" battery, it's a deep
cycle version that for use in caravans and boats) for several sessions, and
the chemical indicator still shows that the battery is in good condition.
And then 24h on trickle charge, puts it back into full condition. Trust
me, the battery is quite safe ;-)

The only problem that I've got is that the battery monitor circuit, looks
more like a sound level meter when the laptop's hard drive kicks in. Also
it doesn't give a good reading, when the battery is put under load. This
can be seen when I turn up the heater, as the battery monitor level goes
down. It's quite funny really, but it makes the battery moniter useless.

Regards

Colin Dawson
www.cjdawson.com

7. ### John WoodgateGuest

I read in sci.electronics.design that Colin Dawson <>
Poor monitor! It's doing its proper job! You have been reported to the
Society for Protection of Innocent Circuits. (;-)

8. ### Colin DawsonGuest

LOL. It's well looked after, got it's own box and everything.

Regards

Colin Dawson
www.cjdawson.com

9. ### Colin DawsonGuest

<http://www.cjdawson76.btinternet.co.uk/images/astronomy_batterymonitorf
I see your point. The battery monitor circuit is mounted into a plastic box,
it's connected to a little Fuse board circuit, which the other devices are
connected to.

So basically it goes, Battery down a 3 Meter cable to the "Fuse board".
Then in parrellel to the Battery monitor, Telescope, and Dew heater. The
battery monitor is in the same box as the Fuse board and the dew heater
controller. The heater controller feeds a 5 meter cable that connects to the
heating element. The telescope is connected to the Fuse board via a 5 Meter
cable as well. Something like this...

Battery --------> Fuse Board --> Battery Monitor
--> PWM (K8004) ------> Heating Element
--------------------------> Telescope

It's as close as I can get it to the battery terminals without putting the
circuit on the battery (which I don't really want to do)

That's exactly what I was thinking of. It would be on a par with putting a
spring into a pully system to keep the tension constant.

And ideas for how to do this?

Regards

Colin Dawson
www.cjdawson.com

10. ### John WoodgateGuest

I read in sci.electronics.design that Colin Dawson <>
All these metres of cables are your problem. What sort of 'cable' is it?
I mean, the conductor diameter or area? You snipped my text about using
really THICK cables. You are getting even fewer volts at your heater and
telescope than you battery monitor measures, due to the extra 5 m of
cable in each circuit.
Absolutely not on a par!
You cannot be serious! I explain that it would MISLEAD you to thinking
that everything was OK and you had 12 V when you really only had 10 V,
and you ask how to do it????

11. ### leggGuest

Run two lighter-guage wires from the battery terminals, to run the
battery monitor only. These should be able to shadow the heavier
battery wires in routing, without much trouble.

It's called a Kelvin connection.

Alternately, one lighter-guage wire from the negative battery terminal
could be used to measure cable drop. Inverted (using a single supply
op amp) and resistively summed into your IC's input pin (SIG), it
could compensate for cable drops in the measurement.

In any event, if you are trying to ensure the integrity of the battery
voltage at the point-of-use, then the circuit is already telling you
what you need to know - there's too much cable drop in your harnessing
to supply the voltage that you have determined you need.

RL

12. ### Rich GriseGuest

That's because you've got the monitor monitoring the wrong voltage.

Use two new wires, and measure the voltage at the battery itself, and
let the devices worry about their own cable drop.

And you'll still see droop, because of the internal resistance of the
battery, but that is the droop you actually want to see, from your first
post And measuring the voltage directly at the battery, with two brand
new wires that weren't there before, is the only way to do that without
moving the meter.

Hope This Helps!
Rich

13. ### Tom SeimGuest

The basic lesson learned is estimating battery capacity from battery
voltage is frought with errors. You have battery voltage that is a
function of temperature. And you have the internal resistance of the
battery, which,also, is a function of temperature.

A better approach is to integrate the current drain from the battery.
This will, still, have to be adjusted for temperature because capacity
varies with temperature.

If you want accuracy I would recommend a micro (such as the PIC). If
you don't, why bother?

Tom

14. ### Colin DawsonGuest

I think everyone here has managed to completly miss the point of what I'm

There is enough voltage in my setup to power everything that I want to
power. The cables are thick enough to power everything that I want to
power, and then some. (I can easily double the number of devices without
ricking overrating the stuff that I've used. I've done it)

At the moment my "Battery Monitor" is actually a "VoltMeter". I don't want
a VoltMeter connected to the battery, as is doesn't tell me when it's time
to start thinking about recharging the battery.

What I want is a "Battery Level Meter". Just because I start pulling 10A
from my battery doesn't mean that it's capacity suddenly drops, as a
VoltMeter shows. I don't care what the Voltage of the battery is. I WANT
TO KNOW WHEN I NEET TO START TURNING OFF DEVICES BECAUSE THE DAMMED BATTERY
IS ALMOST FLAT AND MY TELESCOPE IS ABOUT TO LOOSE IT'S ALIGNMENT. (at
this point Colin has thrown his teddy out of the cot)

Get the point now?

Won't anyone give me a straight answer on how the hell to build an Ammeter
circuit, so that I can get the "BATTERY LEVEL MONITOR" to give a correct

Regards

Colin Dawson
www.cjdawson.com

15. ### John WoodgateGuest

I read in sci.electronics.design that Colin Dawson <>
No, I'm afraid that you are asking a question that doesn't make a lot of
sense in the context in which you pose it.
The cable 'rating' may be based on temperature rise, not on voltage
drop. A few volts off a 230 V supply are negligible, but not off 12 V.

The behaviour of your meter, as you describe it, indicates without doubt
that you have voltage-drop problems, either in the cables themselves or
in the terminations.

What size(s) are the conductors in the cables?
Well, it WOULD. Not the best possible solution, but it would. The fact
that the voltage falls is enough to tell you that you need to charge.
Since you have a deep-cycling battery, using the voltage as an indicator
is unlikely to shorten the life of the battery through over-discharge.

The situation is different for 'float' batteries that are intended not
to be deeply-discharged. For them, using the voltage as an indicator is
OK if the 're-charge' point is set at a high enough voltage, but the
required voltage is temperature-dependent, so it's not easy to implement
such a device. In your case, the simple voltmeter solution is an
acceptable solution.
What do you understnd by 'level' in this context? It has a lot of
different meanings in electronics.
That is true. It's about the only thing you have written here that IS
true.
You CAN tell that with a voltmeter. There are more complicated
possibilities, notably an integrating ampere-hour meter, but although
that would be desirable, it isn't necessary.

that is too low**. Your simple voltmeter will tell you when that
condition is being approached.
I would mention that we don't mind ignorance on this NG, and we (many of
us) will work hard to dispel it, but we do not like arrogance.
You CAN'T convert your voltmeter circuit directly into an integrating
ampere-hour meter, if that's what you mean by 'battery level meter' and
'battery level monitor'. You could pull it to pieces and use some of the
parts, but an integrating ampere-hour meter is not a trivial project for
you to build.

But there really isn't any need; your voltmeter will do what you want if
you improve the cables.

16. ### John CrightonGuest

Hello Colin,
I think I know what you are on about. You want some warning
that your battery is about to fail/run out of capacity. Is that it?

I saw a little circuit that was used to test lead acid back up
batteries, that were on constant trickle charge. Every hour
across the battery bank. An op amp was used as a comparator.
One input was connected to the battery using a suitable divider
and the other input was connected to a referece voltage.

If the battery voltage drooped below a preset level when
the heavy current flowed in the load, the comparator's output
triggered a buzzer. The workers then had a warning that
one or more of the batteries in the bank was not performing
well and would be no good for back up power when required.

Well lets say that you made up a similar rapid discharge
tester with a load resistor, in series with a power mosfet
across the battery. Some 555 timer ICs could apply a similar
1 millisecond pulse to the mosfet every 10 or 15 minutes
(or whatever time period you like.)
The 50 amp or so load resistor can be made up of cheap
wire wound types of several watts, it doesn't have to be
super big, power wise, because it only acts as a load for a
millisecond. Trial and error will show you that the load resistor
can be surprisingly small. 1 millisecond every 10 minutes is a
low duty cycle.

and time how long it takes your battery to discharge to
the point where the buzzer on the rapid discharge tester
starts to sound, meaning that the internal resistance of
the battery is rising, meaning the battery is not so good
anymore, nearing exhaustion.

The problem with this idea is that towards the end
of the batterys life, the one millisecond rapid discharge
tester would cause voltage glitches that could upset
your computer. You would have to run a number of trial
and error tests and adjust the thresh hold of the warning
signal to suit yourself.

That is all a bit crude, I know, but it is cheap and
will give you a bit of an idea how much time you
can go on using your battery. You know roughly
how much time is left from your trial and error tests

Regards,
John Crighton
Sydney

17. ### John FieldsGuest

---
I don't think so. What's happened is that you have fixed in your mind
what you want to hear and you don't understand what you've being told
so you've concluded that since it doesn't sound like what you want to
hear it must be bogus.
---
---
If you increase the number of devices demanding current from the
battery it won't "overrate" the rest of the stuff, all it'll do is
reduce the amount of time until the battery voltage drops to some
arbitrary point.
---
---
Yes, it will. There is a voltage below which current shouldn't be
taken from the battery, and once the battery voltage decays to that
point it should either be disconnected or recharged.
---
---
Yes, it does. If you look at the discharge curves for _any_ battery
you'll find that as the rate of discharge increases (as more current
is drawn from it) the smaller its capacity becomes. The capacity of
most batteries (C) is rated in Ampere Hours, but full capacity can
only be achieved if some fraction of the one hour rate of current is
drawn, ususally C/10 or C/20 for lead-acid batteries. That means that
if you have a fully charged 100AH battery rated for C/10 and you draw
10 amperes from it, its voltage will decay to the cutoff point (say
10V for a 12V lead-acid battery) in 10 hours. However, if you take
100A from it its voltage will decay to 10V in substantially less than
1 hour. Also, since the battery's internal impedance will cause its
voltage to fall more and more as more and more current is drawn from
it, that will futher shorten the time until it reaches cutoff.
---

---
Well, you should, and that's precisely why I said that you want to
hear what you want to hear, not what's at variance with what you
believe.
---
---
LOSE its alignment...

Take a look at your drive's specifications and I'm sure you'll find
that there is a VOLTAGE which is called out below which the drive
won't be guaranteed to remain in alignment. For that purpose, you
need to monitor the voltage at the _drive_, not at the battery, since
if you measure the voltage at the battery the drop in the cable going
to the drive won't be accounted for.
---
---
That you're frustrated because of your ignorance _and_ spoiled _and_
petulant? It's starting to sink in...
---
---
Yowzah boss!!!

+-------------------+
/ |
A--+ [FUSE]
| |
[AMMETER] +----------+
| | |
+-----------+ [DRIVE] [VOLTMETER]
| | | |
[BATTERY] [VOLTMETER] +----------+
| | |
B--+-----------+ |
\ |
+-------------------+

Everything else connects (just like the drive with its own set of
wires and its own fuse) to points A and B. That is, directly to the
ammeter and the battery.

But... That's still only going to give you voltage readings and
on battery voltage.

Or maybe you want something to let you know how much charge is still
in the battery or how much time you've got left until it goes flat? A
battery "gas gauge" kind of thingy?

Well, boss, if that's what yuh wants, just ax fo' it an' ahm sho' some
of us ol' niggers'll jump at the chance to serve yuh.

18. ### John WoodgateGuest

I read in sci.electronics.design that John Crighton <>
All this is, is a load, which he already has, and a 'voltmeter' that has
just a 'go and a 'no-go' mark. He's already got a better voltmeter than
that.
There is absolutely no point in all this complication.
He can already do this with the loads he's got. He doesn't need another
contrived test.
Yes. The glitching would create havoc.
It's all totally unnecessary.

19. ### James MeyerGuest

Measure the current draw with a current sensor. Convert the sensed
current to an appropriate voltage and use the voltage to drive a small, toy
sized, motor with a gear train on its output. Attach pointers to a couple of
the gears and you will have an amp/hour readout similar to the gas or electric
meter on the side of your house. Reset your meter to zero before use, always
start your observing sessions with a fully charged battery, and Bob's yer uncle!

Jim

20. ### James MeyerGuest

An integrating amp-hour meter **is too** trivial....

Measure the current draw with a current sensor. Convert the sensed
current to an appropriate voltage and use the voltage to drive a small, toy
sized, motor with a gear train on its output. Attach pointers to a couple of
the gears and you will have an amp/hour readout similar to the gas or electric
meter on the side of your house. Reset your meter to zero before use, always
start your observing sessions with a fully charged battery, and Bob's yer uncle!

Jim