Feeler: New economy battery analyser/charger chip

Hello folks,


This is basically a feeler to see who might be interested in an economy
battery analyser and charger chip that would interface with the PC
computer and provide information about the cells under test.
The chip would require a minimum amount of support parts: a few resistors
and a small 5v regulator. The circuit would come in two main
varieties: either isolated or non isolated, where the isolated version
completely isolates the computer from the circuit or cells being tested
(and requires a small 9 to 12v wall wart).

The chip is capable of testing up to four cells at the same time,
using two to four inexpensive power resistors for the load on the cells.
It can also charge NiMH cells but you have to supply your own external
power supply that can put out the current you want to charge at,
like 100ma, 1 amp, 2 amps, 4 amps, etc., but you can choose any
current level you like so you can use a wall wart.

Also possible is temperature measurement using inexpensive 60 cent
thermistors available almost anywhere. Each temperature measurement
requires one thermistor.

Tested so far:
Charging two NiMH cells and monitoring their temperature and the
ambient temperature, and discharging two cells using inexpensive
power resistors (which you can select to adjust the discharge current
to whatever you wish).
The software does the measurements, capacity calculations, and just
about everything else.

Computer requirements:
Just about any speed processor (probably even old computers) so
you can use your old computer as a stand alone tester.
At least one serial port, or a USB to serial cable.
Windows 95 or up, but a dos version could also be available.

If this sounds interesting to you, let me know.
Oh yeah, the cost of the chip (which i will be producing) will be between
4 and 5 dollars (probably 4.50 each). The cost of one tester including
external resistors and maybe 78L05 regulator and DB9 connector should
total under 20 dollars. That means test up to 4 cells at one time with
a unit that costs under 20 dollars (software comes with the price of
the chip).

Also, if anyone is interested in building boards for these chips please
contact me via PM.

Interesting?

Where do I get them? How do I pay? International shipping?

As long as I don't have to solder anything count me in for one. My shaky hand's are no good at projects I'm afraid but a completed type unit I can just toss on my USB-Serial Cable and go I'd be happy to buy.

Yes - I'd also be interested in this.

It would be good if the charge and discharge rates could be controlled by software.

Another thing that would be good would be if the software provided an API so that we could write our own software for doing our own custom testing, etc...

Here's a screen shot:

http://upload3.postimage.org/171576/photo_hosting.html


koala, i guess international shipping to Au. wouldnt be a problem,
however i dont know how much it would cost offhand, do you?

Macdude, i'd have to look into building them up too then i guess. It may
be possible or perhaps someone else would be interested in building them.

PowerMeUp:
Controlling the charge and discharge current via computer would be possible,
but in trying to keep this low cost it doesnt work out that well. I could
add that feature, and even though the cost of the chip would not increase
too much the total cost to build would because a few more parts would
be involved, like power transistors, op amps. I guess it wouldnt be too bad
though. Right now the discharge rate is controlled by the choice of
power resistor (sometimes 1/2 watters work just fine) and this keeps the
cost down.
The API can easily be supplied.

MrAl said:

Controlling the charge and discharge current via computer would be possible,
but in trying to keep this low cost it doesnt work out that well. I could
add that feature, and even though the cost of the chip would not increase
too much the total cost to build would because a few more parts would
be involved, like power transistors, op amps. I guess it wouldnt be too bad
though. Right now the discharge rate is controlled by the choice of
power resistor (sometimes 1/2 watters work just fine) and this keeps the
cost down.
The API can easily be supplied.

Click to expand...

Although fixed currents would probably be OK for a lot of people, having the option to control the current through software would be quite handy - with the capability to control the current using custom software, it would make such a charger/discharger a lot more capable than the Maha C9000 - I recall quite a few people saying that they wanted a computer controlled charger/discharger.

So this is a computer controlled analyser? That means the computer has to be on for the analyser to work? I would like to propose a logging system, so the computer can be disconnected while the analyser runs. I don't know if the chip is capable of doing the controls without the computer.

Hello Al,

Sounds interesting...

I would like to try one out. Count me in.

Tom

Hi again,


PowerMeUp:
I agree that software control of the current is a good idea,
no doubt, but that just going to have to wait for a future
version. Right now this is the most basic analyzer but
again it has four channels so it can do all 4 cells (which
usually come in one set) at the same time, and all it takes
is four low cost resistors. This means someone could rig up
a 4 cell battery holder and four resistors and connect it
to the board and be able to fully test the capacity of the
cells.
Also keep in mind that if you find that there is a feature
that would be very handy to have i would be happy to add it
to the software, sort of a custom service The way i
see it is if you like it then there will be others that like
it too.

koala:
Yes, the computer controls the board and does some of
the work, and yes it has to be turned on to do the analysis.
I havent found this to be a problem however, because i
can run most tests in less than 4 hours.
The non-isolated version only has one ic chip too, so it's
very inexpensive to build, that's what makes this technique
so attractive. I've found other models on the web that do
less and cost 120 dollars (US). Im not even sure if you
can go to total isolation with their models either as easy
as you can with this chip.

Tom:
Ok sure, i'll send PMs when the chips and/or circuit boards
are ready. I would like to be able to deliver a fully
assembled circuit board but im not sure yet if that is going
to happen so i thought i would at least offer the chip
plus the software so everyone could build their own circuit.
The circuit can also be customized to fit other applications
too by the user without any difficulty.

Would this work with 12v batteries or other battery packs?

I'd still go for one even if it didn't have the capability to control the charge/discharge current via software.

Something that I think would be important if you haven't already considered it would be for it to have some sort of timeout system implemented so that if the circuit didn't receive any input from the PC within a set period of time, it would automatically shut down - just in case the PC were to crash, or the cable fell out, etc - basically a fail safe system.

It wouldn't be too important for when doing discharge testing - the worst case scenario would be over discharged cells that had to be disposed of. However, if the circuit was charging cells at a high current, there could definitely be problems. Have you heard about the LaCrosse BC-900 meltdowns?

Hello again,


Thanks for the replies from everyone. I should have more screen
shots and a manual ready soon too which will answer many
questions like those already being asked. I do appreciate
the input anyway though and any ideas which might make it
a better product.
It might also be possible for me to provide a kit of all parts
except for the circuit board too, for those interested. The
other parts are cheap and i would sell those at cost too.
The resistors are so cheap in quantities they almost arent
even worth mentioning, and the regulators are like 50 cents.


BatteryCharger:
Yes, the instructions will show how to use it with just
about any kind of battery or pack. It should work up to just
about any voltage with the addition of two inexpensive 1/4 watt
resistors to scale the input. All the calculations will be
done with the software, so all you would need to do
is enter the values of the two resistors on the 'Setup'
page. It will also be possible to monitor different
battery types at the same time too, say one Li-ion, two
NiMH's, and one Lead Acid for example.

PowerMeUp:
Ok great, i'll add you to the list and PM you when the
chips are ready.
The timeout idea sounds good too, i'll have to give this
some thought.
No i havent heard about the LaCrosse meltdowns, is there
somewhere i can read about this?

MrAl said:

PowerMeUp:
Ok great, i'll add you to the list and PM you when the
chips are ready.

Click to expand...

Great - thanks!

The timeout idea sounds good too, i'll have to give this
some thought.

Click to expand...

Good

No i havent heard about the LaCrosse meltdowns, is there
somewhere i can read about this?

Click to expand...

There are numerous threads here on CPF - I just did a quick Google search and found these two:
https://www.candlepowerforums.com/threads/103802
http://www.candlepowerforums.com/vb/archive/index.php/t-129250.html

I'm sure you'll find quite a few more if you do a search.

I'm in.

It would be nice if it employed four wire methods for more accurate battery voltage measurements. Also, would it be possible to see real time voltage plots during testing?

Hi again,

Power Me Up:
Thanks for the link. Im looking into protection for this sort of thing and
i think it can come cheap. It may require the addition of a buzzer or
other alarm however to warn the user in case of a problem.

heavy duty:
Ok you are on the list too now.
Im not sure if the real time voltage plot will make it to the first
version of the software, but perhaps with version 1.02 it will.
I intend to add this at some point, so it's just a matter of time.
Of course anyone who uses the software will receive a new copy
with the added features as soon as it becomes available.
I've wanted this myself too so i know what you mean
The first version will provide a box for each cell that updates as
the program runs to notify the user of the applied ampere hours
(or discharge ampere hours)
up to that point in time, so it will update something like this:
0.100 Ah
0.101 Ah
0.102 Ah
etc., ect.
Until the charge or discharge is done.
Eventually this box will also have an efficiency factor, which can be used
to estimate the Ah capacity of the cell from the time it took to charge.
The efficiency factors for your various cells can be estimated after one
or two charges, then saved and reloaded later when you go to charge the
cell again. Perhaps each cell could have it's own serial number so you
can load the data for that cell, i'll have to see how this works out in the
long run.
As far as the 4 wire method goes, when the user discharges the battery
they supply a resistor to discharge the cell(s), and since the voltage
measurement already has 2 wires it is inherently a 4 wire system
as long as the user connects the load close to the battery terminals.
Same goes for charging.

MrAl said:

Hi again,
Power Me Up:
Thanks for the link. Im looking into protection for this sort of thing and
i think it can come cheap. It may require the addition of a buzzer or
other alarm however to warn the user in case of a problem.

Click to expand...

I think it's more important for it to simply shut things down automatically rather than relying on the user to hear an alarm and come running over to shut things down - just in case it's left unattended...

I have a battery analyser with LM35 as an external sensor. I would normally use blue tack to attach the sensor to the battery/cell. The blue tack wraps around the sensor and battery not between the battery and the sensor as it will cause inaccuracy. The blue tack method works very well, I use my multimeter to measure the temperature and verify with the analyser info.

The LM35 is quite costly, since we're having an economy analyser it can be replaced by a low cost resistive thermistor. It works as well as the LM35. I don't believe LM35 is better. The low cost resistive thermistor requires only 2 pins to work, whereas LM35 has 3 pins.

If the controller has enough I/O pins with A/D converters, it can be implemented at cost of 1-X(1 for each cell) resistive thermistors and some additional software(flash space) that reads the thermistors at certain intervals. This will may lead to enabling dT termination.

An alternative is to build some external analog circuit that uses a relay. The analog circuit will also have the 1-X number of resistive thermistor that monitors the temperature of each cell. If one of the cell goes hot the circuit will trip the relay that cuts the power to the charger or cell. The circuit will also include variable resistor for tuning the trip temperature. I like this idea because we can also use this circuit on different chargers but this will require 'playing with the mains' if the charger does not use a wallwart. Most Maha chargers is 12V but not all of the brands.

I'm in... prefer complete board. thanks,

Hi again,


PowerMeUp:
As i was saying, im looking into the various ways of doing this.
Im not sure yet what is going to make it into version 1, but
i'll see what i can do. I mainly want it to be able to work
as a monitor more than anything else, but adding protective
features is a good idea too. I have the feeling it's going
to require an external chip however to be sure everything
is going right. For example, a chip that monitors the
temperature of all four cells and shuts down the power.
More simply, a simple circuit that monitors the transmit
pin and shuts everything down in the event the cable
becomes unplugged or broken. The shut down feature will
require the user to use a MOSFET or something similar to
shut down the power to the cells, so that would be additional
circuitry so i guess this would come as an option for
the user to decide. For myself, i dont have any problem
with the way it works right now as im at the computer
for 4 hours easy or at least in the same room.
It's never a good idea to leave Li-ion cells alone to
charge anyway with *any* kind of charger, so keep that in
mind too. Discharging isnt a problem (to test capacity in
the 'usual' way).

koala:
Yes, the LM35 is a bit more expensive than a thermistor i think,
so i made the Monitor/Analyzer to work with thermistors which
can easily measure the temperatures typical of cells when
charging. I did think about your external circuit idea too and
i like it best because it would basically be a circuit that
runs independent of the measuring board and the computer too,
so it would not depend on either connections or computer crashes.
This would require of course another chip that would have maybe
four thermistors connected to it and it's sole purpose
in life would be to constantly monitor the temperature of the
cells, and in the event that an over temperature condition
developed, would trigger a relay or MOSFET that would turn off
everything. Of course this requires more hardware which
means more cost, so i'll leave this up to the option of the
end user. Im trying to keep the cost under 20 dollars for
a complete board (assemembled by the user) but the more hardware
you add the more the cost goes up. Right now, with a little
care it works just fine as is I'll still look into this
however because it seems a lot of people would like some sort
of backup system in case things get too warm.
For the 'backup' chip, i am also looking into using a push
button switch to set the trip temperature...push 4 times for
40 degrees C, 5 times for 50 deg C, etc. This saves one pot
that doesnt have to be purchased just for that.
And yes, the monitor circuit should work with other chargers too
because it mainly measures voltage and sends the readings to
the computer, but i havent tested it with any commercial chargers
yet. I would like to be able to monitor my chargers too just
to see how well they work and answer questions like..."Do they
really detect the minus delta V soon enough?"

cy:
I am not sure at this point if i can produce complete boards or
not yet. I can supply a complete set of parts however, but
then there arent that many parts anyway
In the simplest version, there is one monitor chip, one cheap
78L05 regulator (50 cents), five 1/4 watt resistors, and
one 0.1uf cap (10 cents).
Depending on what packing material i have to use to ship the
parts, the price for the chip is around 4.50 each, so that's
a total of about 6 dollars for one circuit, not including the
board or wall wart or battery.
There are many options the user can apply, such as
an activity LED, going to total electrical isolation, etc.
Total electrical isolation will add about a dollar or two
to the circuit cost so im leaving that as an option.
I'll have schematics drawn up for all the options as well
as the main circuit so the user can pick whatever they want
to build. If i can build the circuits myself, i can probably
offer the options as an order option so the board would
come with the option built in.