That diode looks more like an inductive kick protector, such as you see across the leads of inductive loads such as relay coils or coil gun coils. A full wave rectifier looks like it will do the same job, as well as providing some additional sources of power sink.
I would recommend using one or the other, now that you have pointed it out. Without the protective diode to drain off the kick-current, the inductive kick may bounce the voltage high enough to damage the transistor.
In the project text, the diode is specifically called out as protection from reversed polarity.
[INDENT]Power diode MR751 simply protects the load from the accidental application of reverse voltage.[/INDENT]
This dummy load will be purely resistive, and since I don’t plan to use it in conjunction with inductive loads, I don’t think inductive kick will ever be an issue. That being said, a FWBR or series rectifier provides both protection and improved dissipation at the acceptable expense of an increased minimum voltage.
In my case, I’ll have a lower minimum voltage anyway because I’m going to run the LM10 from a separate supply (I need a small supply to run the fan and the LCD ammeter I’m putting on it).
Now all I need to add is a comparator circuit to automatically drop the load as soon as the battery hits minimum voltage, and I’ve got the makings of a fairly decent automatic battery discharger. Probably use an LM339 driving a latch (to avoid oscillations at the end-of-discharge)…
Cheers,
- Dean
My bad, for skimming past that part.
I can’t decide if that description is
- simply wrong,
- well intentioned but dangerously ineffective (as you pointed out, the diode will short the source if source is connected backwards),
- shows the author knows something that we dont,
- or just a simplified explanation for the inductive kick protection.
Good thinking about latching the end of load oscillations for auto-shut off.
My theory is that it is an iffy adaptation of a standard crowbar protection circuit. That would look exactly as the schematic is drawn, except that a fuse rated at the maximum allowable load should be added in series with the power input.
The MR751 is rated at 22A average (if kept cool), or 400A surge, so a fuse in the 20A range would allow the circuit to work within spec, and would blow if a high-current source was incorrectly attached.
Simple battery packs are only limited by their internal resistance, so they can often drive very high currents when shorted… Perhaps the author naively assumed that since this dummy load maxes out at 20A, you would never attach it to a supply capable of sourcing more than 20A.
Cheers,
- Dean
Actually, the best <$20 method I have seen is to use a 12V Car Headlight.
Simply connect the head light to the battery, monitor the voltage and disconnect when it reaches 1V per cell. Works every time.
… and they are only about $12.
And you can use the head lamp as a rough indicator for how much it has discharged …
Yes, this technique works very well. I’ve done the same with a standard 2-pin 12V halogen bulb.
Cheers,
- Dean
OK, the dummy load is finished. Well, v0.1 is finished 
This project is made primarily from junk box parts and is based on this schematic, but with a few upgrades:
[INDENT][LIST=1]
*]An LCD panel meter to monitor Volts & Amps
*]A fan to keep air flowing over the heatsink
*]Regulated 5V and 12V power to run these upgrades
*]Input for a DC wall wart to run the regulators
[/LIST][/INDENT]
[ATTACH]3337[/ATTACH]
OK, a quick walk through the controls:
The knob on the lower left corner sets the current load. This is a 10-turn control, so each CW turn increases the load by about 2A.
The hole next to the Amps knob is for a voltage limit knob to disable the load below a set voltage. Not yet implemented.
The black and red buttons will enabled and disable the load. The LED by the black button tells you when the load is enabled. Not yet implemented.
The LCD is a 3½ digit meter to monitor the current load or voltage. The top toggle switch by the meter switches between V and A.
The bottom switch by the meter lets you select between monitoring the actual V/A or seeing the V/A setpoints. Not yet implemented.
It seems to work very well, and works all the way down to about 0.6V. I haven’t yet let it run full-power to see how hot it gets - I think I’ll work up to that. I did let it draw a full 20A load from a Vex NiCd pack for a short while. The pack voltage fell to about 1.2V and the wires got hot, so I didn’t keep that going for long
As for the insides - this was a pretty quick job. I’ve got room on the protoboard for a CMOS chip or two to implement the voltage limiter.
[ATTACH]3340[/ATTACH]
Cheers,
- Dean
Here is the discharge curve from my first test run. This was taken using an integrating power meter with one sample every second.
The battery under test is a 9.6V NiCd transmitter pack (1000mAh) and the test load is a continuous 1A. The battery is fairly old and has not been specially cared for. It was fully charged using the new Smart Battery Charger in “Safe” mode, and then allowed to rest for about 10 min to cool to room temperature.
[ATTACH]3343[/ATTACH]
You can see the voltage starts at about 11V (about 1.4V/cell), which immediately falls to about 10.5V when the 1A load is applied. Calculations suggest the internal resistance of this battery is about 0.543Ω.
The discharge graph is a perfect textbook curve, with the battery discharging to 8.5V right at the 50 minute mark. (8.5V is the point that the transmitter starts beeping). Exact measurements show this battery produced about 836mAh, or 84% or its original design capacity. This seems reasonable for a battery of this age and condition.
You can see that the voltage falls off rapidly after 8.5V, as you would expect from a depleted battery.
Also, note that the midpoint of the discharge (25 min) shows 9.6V, which is the rated voltage for the battery pack.
This all tells me my approach to battery analysis seems to hold up to some data. Over the next few evenings, I’ll run the full line of Vex batteries under increasing loads until I get a good set of graphs, or until the dummy load burns out ;).
Cheers,
- Dean
WOW those are some really nice numbers and i am sure your method for discharging will be a useful guideline for other people looking to discharge their batteries. I am not an electrical guy at all and you made all of this make perfect sense, thank you so much. Can i assume that you will also test on the NiMH battery packs at some point?
~DK
Glad I was able to help! Batteries are a very tricky subject and could probably do with a full episode of MythBusters to straighten us all out
I will be running tests on all the Vex batteries (NiCd & NiMh) as well as a set of 6 AA eneloop batteries in the Vex battery box. I realize the transmitter battery isn’t very interesting as a test subject, but I wanted to start with a battery that I didn’t care quite as much about to get my methodology down.
Cheers,
- Dean
Here is a graph of the new NiMH batteries being discharged at 1A. Note that these each represent a single run of a single battery, so this is more anecdotal than statistically sound data.
I’m posting this quickly since folks seem interested, but I do plan to post more complete data up on the wiki after I’ve collected everything.
[ATTACH]3346[/ATTACH]
They have a wicked-steep voltage falloff at the end of the discharge. It can drop from a sensible voltage down to the danger zone (below 5V) in a matter of seconds. This means I need to get the voltage limiter circuit working before I attempt the higher-current runs.
You may notice that they don’t deliver their rated capacity, but that is expected. The internal resistance of any battery makes it impossible to extract the full charge. According to Wikipedia, NiMH batteries are usually rated assuming a 0.2C discharge rate, which would be 0.4A for the 2000 mAh pack and 0.6A for the 3000 mAh pack. Since I’m discharging them at 1.0A here, I would not expect to get the full rating, and the effect will get worse as the current load goes up.
Cheers,
- Dean
Nice work Dean.
Even at just 1A, the 3AH (new large) battery shows smaller internal resistance and a higher voltage than the 2AH (new small) battery, as expected.
Try not to burn down the house as you crank up the load current.
Yep - it’s nice when theoretical and experimental results line up.
BTW, I’ve calculated the internal resistance as 0.109Ω for the 3000mAh battery, and 0.264Ω for the 2000mAh battery. I’m sure those numbers are off by a bit (±10%?), but they should serve as a ballpark figure for those that care.
Always good advice
I keep a fire extinguisher on my workbench, but luckily I haven’t ever needed to use it (knock on wood).
I’ve got the parts for the voltage cutoff circuit now, but it’ll probably be this weekend before I can get it all put back together and start testing in earnest.
Cheers,
- Dean
I just realized I have access to an “Accucycle Elite” RC charger.
This can do charge and discharge at constant current, time them, and provide ingoing and outgoing capacity measurements.
The discharge current is limited to about 2A, but possibly I could gang the two channels to get 4A. It doesn’t provide a voltage curve, but I could plot the voltage with a RadioShack DVM with a serial port output (last used to watch temperature curve of Thanksgiving turkey).
I have several new 3AH batteries, but no 2AH to check.
I’ve read somewhere that RC boater people keep their batteries in an igloo freezer before competition. Does any NiMH battery information site mention if cold batteries perform better?
I highly recommend giving it a try. This exercise has taught me quite a lot, and solidified all the things I already knew about batteries. I feel like I’ve got a much more practical handle on their ins-and-outs now.
Once I finish the full sweep of Vex batteries at various loads, I’m going to subject my collection of NiCds to a deep charge/discharge to see what their health really is.
From some of my reading, freezing (or cooling) batteries will reduce their self-discharge rate. For Alkaline batteries, this is apparently not worth doing because they hold their charge well over time. However, NiCds & NiMHs can discharge a few percent per day, so it can keep them from going flat as quickly. On the other hand, NiMH batteries apparently have a reduced capacity when cooled.
As for race-day advantages, all I can think is that it would slightly reduce the internal resistance allowing an initial burst of high-current, but that would only last for a few seconds since the batteries will heat up quickly under load.
Cheers,
- Dean
After trying it last night, it looks like this charger doesn’t actually do constant current discharge.
The display toggles slowly between battery voltage and discharge current.
The discharge current limit can be set at 2A, but the actual value shown slowly ramps up to 1.4A, and then declines to less than 1A over a period of a few minutes, so it is not a constant current drain. The case gets quite hot, so maybe the internal temperature monitor is limiting the current.
Summary: not useful for a constant current drain plot.
I’m looking forward to an update from Dean with a higher current, like 4A.
If we get > ~10minutes at 8A on a 2AH battery, then a battery Y cable is a viable idea; PIC + PowerExpander, although that only equals one Cortex.
A Cortex plus powerexpander could be up to 12A, although you are not likely to be at that max for the whole match.
OK, I got the voltage cutoff implemented, and upgraded a few other parts that were a bit iffy. Everything seems to be working well now.
I tried it out up to 8 amps and the transistor gets hot, but not too hot to touch, so I think I’m ready to go now. I’ll save the 12A runs till after I’ve finished the 8A runs, just in case
I also powder-coated it a nice “Vex battery” blue:
[ATTACH]3397[/ATTACH]
I’ll post results as I get them; probably a run or two each evening this week.
Cheers,
- Dean
For those of you who don’t have Quazar’s mad-scientist build skills you can try the CBA III tester. I have the CB II and it does a good job of doing testing and also the full discharge cycle. The CBIII allows a higher discharge rate. You can find CBII’s on Ebay for about $100.
I’m looking forward to seeing his results to see if they match ours.
Couldn’t you just take a spare cortex, and hook up lots of motors to it to drain the battery. You could then program the cortex to turn the motors off once the batteries reach a certain voltage. You could do a second battery for each cortex by using a power expander and the status port in the expander.
Yeah, but that wears out your motors. Dissipating all the energy as heat doesn’t really wear anything out… plus, it might help heat the house in the winter…