home made discharger help?

[ripit]

I want a home made discharger for nimh aa and aaa batteries. I have researched it but don't know much about building electronics. Will this setup work ok?

I am building it in the shell of an old battery charger. It will go from the positive terminal of the battery, to a in4003 diode (as soon as I figure out which way it has to go), to a 1.5v light bulb (not led), to a cheap watch for a capacity timer, to a 10w/2 Ohm resistor, and back to the negative terminal.

Does that all seem about right? I assume that it all has to be in series so that I don't get two resistive devices changing the resistance?

how would I calculate the capacity of a battery off this?

Fyi, I chose these components because I'ts what I had laying around (except for the light bulb).

 

[uk_caver]

That arrangement shouldn't end up discharging the cell below ~0.6V, though ideally ~0.9V/cell would be better. Maybe adding a Schottky diode (1N5820 or similar) in series with the regular one would work to get a better final voltage.

Measuring capacity from it could be pretty difficult - the current would vary quite a bit with changes in battery voltage, and I'm not sure how much variation there might be between one component and another.

However, thinking about this just gave me an idea that might (should?) work, though without an auto-timing element, and since I need to break in a load of new cells for battery packs, I think I'll hack that design together and check it out, and see how it works.

Will report back soon.

 

[ripit]

I built a prototype, but apparently I messed something up. Voltage is about 1.3v at the battery, and 1.05 just past the diode. Just past the watch, there is no voltage. The battery is not discharging (had it hooked up all day). I guess either I put the diode backwards, or the watch cannot pass the current through it (I was worried about that, but If I did it parallel instead of in series, wouldn't that change the resistance of the resistor)? Oddly enough, the watch intermitantly runs even though there is no voltage past it (and therefor no ground?).
I guess I'll try to bypass the watch and check voltages?

 

[Mr Happy]

I built a prototype, but apparently I messed something up. Voltage is about 1.3v at the battery, and 1.05 just past the diode. Just past the watch, there is no voltage. The battery is not discharging (had it hooked up all day). I guess either I put the diode backwards, or the watch cannot pass the current through it (I was worried about that, but If I did it parallel instead of in series, wouldn't that change the resistance of the resistor)? Oddly enough, the watch intermitantly runs even though there is no voltage past it (and therefor no ground?).
I guess I'll try to bypass the watch and check voltages?

OK....I think you should put everything down and step away quietly. You seem to be a bit confused when it comes to this electrical malarkey and you are going to hurt yourself.

We were all with you when we thought you wanted to use the watch to time the discharge. But when you started trying to pass electricity through it...well that's got to be the most original thinking in the history of CPF... :huh:

 

[ripit]

OK....I think you should put everything down and step away quietly. You seem to be a bit confused when it comes to this electrical malarkey and you are going to hurt yourself.

We were all with you when we thought you wanted to use the watch to time the discharge. But when you started trying to pass electricity through it...well that's got to be the most original thinking in the history of CPF... :huh:

I did want to use the watch to time the discharge so I could compare cells. Thats why I asked in the first place about how to wire it (because I only partially know what I am doing). I am researching online as much as possible (I have spend many hours reading on how to make it but everything I find seems to be about rc car nicd battery packs). I ran from the positive end of the battery, to the diode, to the watch, to the resistor, and to the negative battery terminal. Since it was all in series, wouldn't that mean that what ever power the resistor was drawing, is being drawn through the watch?

I kind of based it on a post here but cannot find it again (there was a list of usefull threads that contained one about making a discharger but the instructions were vague).

Anyway, I figured that since electrons flow from negative to positive, if I put the watch on the positive side of the resistor, the current that the resistor is using might not be flowing through the watch (kind of guessing so I used what seemed the safer guess).

I found another plan (this one had a drawing, but no diode as it was for nicd). It had the watch directly off the battery on a separate circuit (not sure why I didn't think of that).

So anyway, I temporarilly removed the watch. Its running from the positive side of the battery, to the diode, to the resistor, to the negative side of the battery. So Before the diode, I'm getting 1.3v, and after the diode, but before the resistor, I'm getting 0.5v. After the resistor, I'm getting no voltage (because there is no current in that part of the circuit maybe?).

I know a resistor can change voltage, but it still doesn't seem right?

I am basically trying to make something like this
http://www.nickhill.co.uk/nimh-capacity-tester.html
but with a diode to stop discharge at a particular voltage. 0.9v would be nice but I don't know how to do that. I was under the impression that common diodes all cutoff at 0.7v (but I'll look into the diode uk_caver mentioned).

Does the voltage drop I'm getting after the diode seem right (I'm still thinking the diode may be backwards or maybe it was overheated during soldering)?

Also, I never put in the light bulb (the 1.5V one I got at radio shack barlely glowed at 1.3v). If a find a low enough voltage bulb, can it go inline between the diode and the resistor? I want the bulb to turn off when the diode cuts out (so when its done discharging, the light goes out)?

Sorry for being such a noob. I have done several small projects with electronics like this, but it has just never been my strong point....

 

[uk_caver]

Schottky diodes have a rather lower voltage drop than regular silicon diodes.
The voltage where they start to conduct is around 0.2-0.3V, rather than the 0.6V of a regular silicon diode.

However, the greater the current flowing through either type of diode, the higher the voltage, or alternatively, it takes a higher voltage to get greater currents to flow through a diode.
The diode acts a bit like having an 'ideal' diode in series with a resistance,

Voltage/current graphs are available in the datasheets for diodes that electronics suppliers typically give links to, and are well worth looking at.

In the diagram you linked to, the watch was 'seeing' the whole battery voltage, and stopping when the battery was flat.
With the watch in series, the only current flowing would be the tiny current the watch consumed, which would mean the cell would likely take years to flatten.

 

[ripit]

Schottky diodes have a rather lower voltage drop than regular silicon diodes.
The voltage where they start to conduct is around 0.2-0.3V, rather than the 0.6V of a regular silicon diode.

However, the greater the current flowing through either type of diode, the higher the voltage, or alternatively, it takes a higher voltage to get greater currents to flow through a diode.
The diode acts a bit like having an 'ideal' diode in series with a resistance,

Voltage/current graphs are available in the datasheets for diodes that electronics suppliers typically give links to, and are well worth looking at.

In the diagram you linked to, the watch was 'seeing' the whole battery voltage, and stopping when the battery was flat.
With the watch in series, the only current flowing would be the tiny current the watch consumed, which would mean the cell would likely take years to flatten.

I'm glad you asked about the watch as I now realize that hooking the watch on a separate circuit wouldn't work. I now wonder what use the watch was in the link I gave.

The way I want it to work, is that you set the watch to 12:00. When you insert the battery, the watch stats to run. When voltage drops enough that the diode will no longer allow power to pass, it cuts off power to the resistor as well as the watch and the watch stops running. You can look at the watch hands and see how long it ran for before it quit because the diode cut power.
Do it with another cell (assuming both are fully charged), and in theory, two cells that run for the same amount of time, have similar capacity. Of course since I am trying to recondition/recover many old batteries, the would need to be cycled several times to stabilize them first.

If this project works, I was thinking of just a board, with several diodes and resistors (no watch, light or anything else), to discharge several cells at once (each battery would use a separate diode and resistor on its own circuit and maybe just two bent pieces of sheet metal to hold the battery). The one with the watch would be for a final test of capacity.

Here is the link that first gave me the idea of the diode and resistor.

http://www.candlepowerforums.com/vb/showthread.php?t=104123

I have read several sites/threads where people use a watch to test capacity. It would seem many are with battery packs for rc cars though. It does bring up the dilemma of what voltage the watch stops running at and will it stop running before the diode cuts power (I'm using a watch that cost 2.50$ new). I guess thats what testing and experimenting is for. I may have other watches I could use but I really don't want to trash a more expensive watch.

 

[uk_caver]

The way I want it to work, is that you set the watch to 12:00. When you insert the battery, the watch stats to run. When voltage drops enough that the diode will no longer allow power to pass, it cuts off power to the resistor as well as the watch and the watch stops running. You can look at the watch hands and see how long it ran for before it quit because the diode cut power.

The problem is that if the diode is used to limit the discharge voltage, the voltage across the resistor with any given current flowing will be the remainder ('resistor voltage = cell voltage - diode drop', at whatever current is flowing).
That'd mean that with a cell that had been discharged even slightly, you'd be into the ~1.25 plateau of cell voltage, and if you were discharging the cells at 200-400mA, with a 1N400x diode, you'd be looking at a diode drop of 0.8-0.85v, so a voltage across the resistor of only ~0.4V.

Would the watch actually run off 0.4V?

 

[ripit]

The problem is that if the diode is used to limit the discharge voltage, the voltage across the resistor with any given current flowing will be the remainder ('resistor voltage = cell voltage - diode drop', at whatever current is flowing).
That'd mean that with a cell that had been discharged even slightly, you'd be into the ~1.25 plateau of cell voltage, and if you were discharging the cells at 200-400mA, with a 1N400x diode, you'd be looking at a diode drop of 0.8-0.85v, so a voltage across the resistor of only ~0.4V.

Would the watch actually run off 0.4V?

No the watch won't run which is the exact problem I have just ran into. I don't think I will find anything that will run off .4v (light bulb, watch, or even a relay to trigger power to the watch off a second battery).

I guess that I'll just have to go with the diode and resistor only design (cannot check capacity but at least I can discharge to try and recover batteries), unless I can come up with a solution to that.

I don't suppose there is a simple, cheap way to increase voltage to the watch?

 

[ripit]

I don't suppose an op amp might work?

Another thought... Dont nimh keep high voltage till they are 80-90% depleted? If I connected the watch to the high power side (direct to the battery), wouldn't it run till the battery was mostly depleted and the voltage suddenly started dropping? I think I'll experiment with that to see what happens. It won't give an exact capacity figure but it might give a useful ballpark. I think I might also change to a larger resistor (I didn't calculate for the voltage drop from the diode).

 

[uk_caver]

I don't suppose an op amp might work?

You could certainly use one wired as a comparator to trip out at a set voltage.
It'd need independent power, but that wouldn't be hard to arrange - just 2 or 3 AAs in a holder.

Probably best to have it sensing the actual cell voltage.
With a little modification, you could have the op-amp turn off the discharge, *and* have a constant current discharge, which would make the time reading meaningful
A very simple circuit for that would throw away the diode, and use an AMC7135 LED driver chip as a constant current driver, kept running by the op-amp driving its enable input high, and a suitable resistor as a load for the 7135, in place of the regular LED.

The op-amp would be wired as a comparator, set to turn off (output low) as a selected cell voltage, turning off the supply to the watch as well as the discharge current, and possibly also turning on a 'finished' indicator LED, or turning off a 'discharging' LED.
The op-amp would need to be configured to have enough hysteresis that it didn't turn back on as the cell voltage recovered, and have a mechanism for making sure it turned on when a full cell was connected, but that wouldn't be hard.
Also, if the threshold voltage was set with a trimpot, the circuit could easily be set to have whatever trigger voltage you wanted.
It would need something to give the correct voltage to the watch when the op-amp output was high, but that'd be easy to sort out.

 

[Apollo Cree]

Connect the resistor to the battery. Throw the rest of your electronics stuff away.

Only connect one battery into the circuit at a time. If you want to test multiple batteries, use multiple resistors and multiple circuits.

Get a voltmeter. Measure the voltage every 5 minutes or so. Write the time and voltage down.

For each time you check the voltage, write down

mA=voltage/2000.
mAh=mA*T/3600 (T is elapsed time in seconds from the last time you checked.)

Stop when the voltage reaches 80% of a full charge.

Don't try any of this with any kind of lithium batteries. You'll blow yourself up.

If you want to do this automatically, you have to essentially do all the above things. You'll basically need a computer.

It would be an interesting intellectual exercise to design a non-computer device to do this. You'd need a circuit that waits until the voltage drops to 80% of full voltage, and then opens a switch to stop the current draw. You'd need a timer that stops when you reach the stopping voltage. In theory, you'd need some sort of integrator circuit to add up the currents involved, but you'd be within 20% or so if you simply assumed the voltage was 90% of full voltage for the full time.

 

[csa]

Yeah, building a little computer to do that would be nifty. Not super simply though....

 

[Mr Happy]

I am basically trying to make something like this
http://www.nickhill.co.uk/nimh-capacity-tester.html

OK, I was about to draw you a circuit diagram, but then I looked at the link you gave and my circuit diagram was right there! So I don't need to draw it again.

Instead of trying to make "something like", why don't you make exactly what is in that circuit diagram? It will work. It was described by someone who knows what they are talking about. Trust me on this.

A slight improvement you could make is to replace the 4R7 resistor with your 1.5 V bulb. This will approximate a constant current drain and will give you better linearity.

Also, did you note the important specification that it must be an analog quartz watch, not a digital one?

Here is how the circuit works: The watch is powered by the battery. While the battery has sufficient voltage, the watch runs. The resistor drains the battery. Eventually the voltage drops enough that the watch stops. When that happens you note the time and disconnect the battery.

This works because of the discharge curve of a NiMH cell. When the cell is discharged the voltage drops very sharply towards zero and makes the watch stop at a fairly predictable time.

 

[ripit]

as far as building a computerized one, there are actually plans online but I think its a bit too expensive and definitely too complicated for me (right now anyway).

http://www.instructables.com/id/Arduino-True-Battery-Capacity-Tester-Li-IonNiMH/

For now, I think that I will just use a diode and resistor. I already have a working model after all and it should be pretty cheap to build several on a board. I will of course use a separate diode and resistor for each battery on their own circuit. I left a battery in the one I made while at work today and it stopped with the battery at 7.01 volts (no power on the other side of the diode so it did cut it off at 7v). I had read that going down to 7 volts instead of 9v can help break up large crystals if a battery has lost a lot of its capacity. This is mostly just to recover capacity on old batteries (I have a box full of them). I'll charge them and then discharge them several times. I might play with getting the watch to work. The reson I don't follow the diagram I linked to exactly is it doesn't have the diode. Without the diode, I would have to monitor it so the battery doesn't discharge too far killing it. I have young children and distractions are far to frequent (I can see myself forgetting and killing a lot of batteries). If I put the watch on the high voltage side before the diode, it should still work the watch should run till the battery is mostly discharged and the voltage drops.

I'm going to look into the amc7135 and op amp idea too. Just looking quick, you can get 20 of the amc7135 for 15$ shipped (or 1 for 5$), and op amps start at a couple of dollars. I really have no idea how these parts work or what op amp to use, so I'll have to do a little (or actually a lot) of reading. In the mean time, perhaps I can manually time and check capacity/time with a multimeter or perhaps the watch on the high voltage side will work.

Thanks a bunch for all the help everyone. I'll post when I come up with anything new that works.

 

[Mr Happy]

Do read my post above.

If you want to do what Apollo Cree suggested in a computerized way, you can take a look at this thread. You can see from the picture that no great construction skills are required.

 

[ripit]

The watch is already installed and wired (I just unsoldered the wires) so I decided to solder it back in on the high voltage side tonight (basically directly to the battery before the diode). It will only give me an approximate capacity, but I would think it would be consistent, so at leas I can match up cells (in other words, two cells that ran for 2 hours each should have the same rough capacity, even though I can only approximate what that capacity is).

 

[ripit]

Do read my post above.

If you want to do what Apollo Cree suggested in a computerized way, you can take a look at this thread. You can see from the picture that no great construction skills are required.

I did read your post (thats what mad me decide to wire the watch back in and try it). Thanks a bunch for the link. A multimeter with a computer interface is a bit out of my price range right now but I can do it manually to a point (log it myself). Fyi I read your post after I posted so I edited it a bit.

 

[Mr Happy]

The watch is already installed and wired (I just unsoldered the wires) so I decided to solder it back in on the high voltage side tonight (basically directly to the battery before the diode). It will only give me an approximate capacity, but I would think it would be consistent, so at leas I can match up cells (in other words, two cells that ran for 2 hours each should have the same rough capacity, even though I can only approximate what that capacity is).

Right, but you should take out the diode. There is no diode in the original circuit diagram. Including a diode in the circuit will make the time the watch stops unreliable (if it stops at all).

When you are given a recipe, you should:

1. Follow the instructions exactly as given and check that it works;
2. Study it carefully and learn how it works;
3. Only after you know how it works think about trying to improve upon it.

Don't do step 3 before step 1 and then start asking why it doesn't work. It will likely be because you didn't follow the instructions.

 

[ripit]

Right, but you should take out the diode. There is no diode in the original circuit diagram. Including a diode in the circuit will make the time the watch stops unreliable (if it stops at all).

When you are given a recipe, you should:

1. Follow the instructions exactly as given and check that it works;
2. Study it carefully and learn how it works;
3. Only after you know how it works think about trying to improve upon it.

Don't do step 3 before step 1 and then start asking why it doesn't work. It will likely be because you didn't follow the instructions.

I already have it running with the diode so I'll let one battery discharge and see what I get. It shouldn't be too hard to bypass the diode after that so that it is the same as that diagram, and I'll test it out that way.