Will this work?

I'm just lookin for some feedback on this idea and whether or not it will work and what problems I may encounter.
My aunts work is looking to develop some sort of automated battery tester which will test AA batteries to insure they make at least 1.5 volts. Right now they pay people to manually test several million batteries every year. Me and my dad have come up with a way we think will work. We plan on using an analog input card to test for enough voltage. We plan on using an Automation Direct DL06 to do this. (Neither of us know AD at all, mostly only AB but we can't afford the AB).
So....a couple questions:
How hard will it be to learn to program the AD plc? My dad is a maintenance man with about 4-5 years experience and I'm a coop with about 6 months experience.
Will the 0-5v analog input card be accurate enough for this application?
Any other potential problems with this way of testing?

If you guys need more info let me know and I'll be glad to post it.

Thanks,
Josh

Let me start off by saying that I don’t have any AD experience either. That being said, let me offer some thoughts. 1. Go for it! You will learn more from this small project then from your next 6mo of coop. 2. I’m sure that anyone’s analog card would will work for reading a battery. 3. Some of the hurdles that you will need to cross are (a)to ensure that you are testing only one battery at a time, (b) allow a settle time for the card to read the battery. (c) a method of capturing the voltage reading at the end of the settle period and hold that value, or at least a reject signal until the battery reaches the reject point. (d) a method to tracks the defective battery to reject.

I assume that there will be more then one battery in the machine at a time, the key will be tracking both the battery and the reject signal through the machine to the reject point.

Good Luck

My dad assures me he can get only one battery in at a time using a couple of different gates, I dont understand exactly how 100%, but if I saw it I probably would. We plan on having the battery stationary for a 3 second period right now. (We may or may not have to change this, we'll have to do some testing) There will be a trap door which will open for all defective batteries. I've already done ways of holding an accept/reject signal at work so that wont be a problem. I had also planned on keeping some running counters to keep track of the number of batteries which passed and failed.
There will be a bunch of batteries in the machine at all times. The company has a tumbler which the batteries will be loaded into. This tumbler will sort the batteries by polarity.

In my opinion, if you can program A-B you can program AutomationDirect no sweat. I've used both, and I prefer the DirectLogic products and software.

As far as resolution, the analog inputs are 12 bit, meaning you have a resolution of ±1 part in 4096. For a 5 VDC input that is ±0.001 Volt. I suspect that is plenty good enough. The accuracy will, of course, be lower than the resolution, but the total error should still much better than ±0.02 VDC - advertised accuracy worst case is ±0.3%.

Depending on hte model you select, you should think about using a bipolar (±5 VDC) input. This wild eliminate the need to sort by polarity.

Be sure to get detailed specs from the battery factory. The battery probably needs to maintain 1.5 volts at some specified current. Even a nearly dead battery will show 1.5 volts potential when connected to a very high impedance.

You may need to include a fixed resistance in your test circuit and use the analog input to measure the voltage drop across that resistance.

That sounds like a Fun Project!

It has all the facets of any large project. That is, there are the technical questions as well as the financial questions.

I agree with Tom about the PLC... AD is a piece of cake! Don't get yourself all knotted up looking for AB-type functions in the AD. Just look for the functions you need.

Steve's point about Voltage under Load/Time is Critical! As he indicated, instantaneous voltage readings can be pretty useless. What you are really looking for is the Voltage Curve while applying that voltage to a fixed load for a particular amount of time.

Whether or not the battery is good (at least in terms of the available power - visual inspection is another issue) is determined by...

• Starting Voltage
• dv/dt - that is, the change in voltage over the change in time.

If Starting Voltage is too low, the battery fails.
If Starting Voltage is OK, but dv/dt is too large, the battery fails.

The battery passes only if Starting Voltage is OK and dv/dt is within limits.

You indicated that this device is to replace/reduce "several million" manual measurements a year.

I highly recommend that you find a way to make the device read multiple batteries at the same time. If you can read 10 at once, then you can reduce the number of cycles by that factor. That will extend the cycle-life of the device.

You can have the "trap-door" built right into the tester, under each battery-under-test. If a battery fails, the trap-door drops that battery down to the junk bin - otherwise, all remaining batteries move sideways(?) to the next stage (packaging?). (Don't try to slide the good batteries across the open trap-door.)

Whether or not this plan is cost effective depends on how much it costs to do the several million measurements manually. It also depends on how much the automated system will cost. Don't forget potential maintenance costs.

If the cost per manual measurement is measured in a large % of a dollar, then the return on investment should occur very soon!

If the cost per manual measurement is a very small % of a dollar, then it will take much longer to get a return on investment.

AD Analog Hardware is very low priced - comparatively.

The rest depends on your programming abilities. As a PLC Process, this is very simple!

I think you'll find yourself adding features that you haven't even thought of yet...

• tracking failure rates...
• tracking failure types...
• identifying failure rates by batch...
• ???

Please keep us posted on your progress!

Be sure to get detailed specs from the battery factory. The battery probably needs to maintain 1.5 volts at some specified current. Even a nearly dead battery will show 1.5 volts potential when connected to a very high impedance.

Click to expand...

Very true point and one we have thought about. As far as I know, they only check for 1.5volts right now using a Radioshack type meter (which we would still leave on the machine so the operator can double check). I haven't had a chance to talk to the company to determine their exact requirements yet. These batteries are used in those throw away type cameras. They get about a million batteries a year from one supplier which they need to check and my understanding is they check all the batteries coming in from cameras to determine whether or not they are still usable also.

I highly recommend that you find a way to make the device read multiple batteries at the same time. If you can read 10 at once, then you can reduce the number of cycles by that factor. That will extend the cycle-life of the device.

Click to expand...

We had originally considered making several of these machines but it would probably make more sense just to make a couple of tracks and make the most out of our one PLC



I'll ask my dad to get some more specifics about exactly what they want us to look for and I plan on keepin everybody posted!

btw....from what i hear this company has virtually no automation at all in there factory so I may have more of these little projects coming along....

If they were only testing with a Radio Shack Meter like you said then you may not need to do the tests Terry and Steve were talking about in order to satisfy the customer. But I think you should do them anyway for two reasons. 1. Its a good problem and opportunity for you to do some learning. 2. If the customer likes it they will have a better product because of it. If they don't you can always take it out. Little projects like these are the best way to start learning because you have a goal in mind and its not overwhelming.

Good luck and keep us posted with your results.

You can have the "trap-door" built right into the tester,

Click to expand...

I like it!
Maybe you can have the power from the battery keep the trap door closed?
(probably not practical?)
Then if the battery isn't strong enough, the weak battery falls off into the abyss.
Maybe no PLC is needed then?

Maby obvious, but be careful in your handling of the batteries not to accidentally short the (+) and (-) terminals even for a short time as this will drain a battery rather quickly.