I just got some new RCR123 cells from Manafont. Seems to me that one simple way for me to test them would be to stick them on the lightbox with a light which I know the output of with IMR cells and see how the runtime graph compares. I can measure the total output continually, just not in dimensioned units and can take lux measurements at my usual time intervals. Then do current draw measurements.
Any flaws or idiocies in this idea?
Current draw when the light is complete/closed may not be the same as when measuring using a multimeter. You could check that the output is the same when measuring the current, or just measure the current during the whole test to be sure that it does not provide a false reading.
I can't actually log current continually - at least not at the same time as continually logging output. This is something I'd like to be able to do and I have two meters capable of logging, but no matter what I do, I can only get Windows to see one of them and I can't get Linux to see any of them as the only drivers for the USB-RS232 chip inside the thing are for RPM distributions and the source flatly refuses to compile on Debian or any other .deb based distribution. May just have to throw an rpm distribution on an old box for data-logging purposes. I have more than enough (I may own more computers than lights) old boxes lying around.
What I can do is continually measure current while logging output, but I'd have to manually record it. This doesn't sound too hard. If I take lux and current readings every minute or two, that might work.
Then repeat for another cell.
Better use a light that doesn't run for too long.
Sounds like Mr.Lite LT3 step forward.
Another way of testing capacity is by discharging the cells at a constant current, then just measuring the run time until the discharge device drops out of regulation at a pre set voltage. A constant current discharge device is easy to build and all material you need is a couple 1ohm power resistors and a spare AMC7135 board.
Hobby charger ++
You probably know know this already but if it's a binary-only RPM, you could always give Alien a shot. It's sort of hit-and-miss in my (very limited1) experience but when it works, you can use dpkg just like you would with any Debian packages.
1 (Disclaimer: I'm more of a *BSD guy myself and I don't regularly use Linux.)
So am I. But the options I have are Solaris 8 or 10 - drivers for the UART. Well I'm not coding them.
Alien doesn't like me. I think there's a binary blob of proprietary code that probably needs a specific kernel and library set - at a guess Red Hat. I'll try one of the DeadRat variants to see if I can get it to work. The meters are supported with some rather nice software - much better than the Windows garbage. So it is a case of digging out the oldest box with USB2 ports to see what can be done.
Or MacOS. The MacOS drivers just don't work at all for me. I may try them again as that BSD variant is the one I'm most familiar with. It'd be easiest if I could get 2 meters to work with one of my Macs. Maybe sometime next week.
I did some experimental messing around and here are the numbers from a first attempt. Graphing was right out as the values were so different. Going to have to normalise the data a bit.
This was with a Soshine RCR on Mode 4 in the LT3. The microamps were autmatically logged, the current and lux readings were manual.
| Elapsed | uA |
| 0:00:00 | 74 |
| 0:00:05 | 73.8 |
| 0:00:10 | 73.6 |
| 0:00:15 | 73.5 |
| 0:00:20 | 73.3 |
| 0:00:25 | 73.1 |
| 0:00:30 | 73 |
| 0:00:35 | 72.8 |
| 0:00:40 | 72.7 |
| 0:00:45 | 72.6 |
| 0:00:50 | 72.3 |
| 0:00:55 | 72.2 |
| 0:01:00 | 71.9 |
| 0:01:05 | 71.7 |
| 0:01:10 | 71.6 |
| 0:01:15 | 71.4 |
| 0:01:20 | 71.3 |
| 0:01:25 | 71.2 |
| 0:01:30 | 70.9 |
| 0:01:35 | 70.8 |
| 0:01:40 | 70.6 |
| 0:01:45 | 70.5 |
| 0:01:50 | 70.4 |
| 0:01:55 | 70.3 |
| 0:02:00 | 70.2 |
| 0:02:05 | 70.1 |
| 0:02:10 | 70 |
| 0:02:15 | 69.9 |
| 0:02:20 | 69.9 |
| 0:02:25 | 69.8 |
| 0:02:30 | 70 |
| 0:02:35 | 70 |
| 0:02:40 | 70 |
| 0:02:45 | 70.1 |
| 0:02:50 | 70 |
| 0:02:55 | 70 |
| 0:03:00 | 70 |
| 0:03:05 | 69.6 |
| 0:03:10 | 69.6 |
| 0:03:15 | 69.4 |
| 0:03:20 | 69.3 |
| 0:03:25 | 69.1 |
| 0:03:30 | 69 |
| 0:03:35 | 68.9 |
| 0:03:40 | 68.9 |
| 0:03:45 | 68.8 |
| 0:03:50 | 68.8 |
| 0:03:55 | 68.7 |
| 0:04:00 | 68.7 |
| 0:04:05 | 68.6 |
| 0:04:10 | 68.6 |
| 0:04:15 | 68.4 |
| 0:04:20 | 68.3 |
| 0:04:25 | 68.2 |
| 0:04:30 | 68.2 |
| 0:04:35 | 68.1 |
| 0:04:40 | 68 |
| 0:04:45 | 68 |
| 0:04:50 | 68 |
| 0:04:55 | 68.1 |
| 0:05:00 | 68.1 |
| 0:05:05 | 68.2 |
| 0:05:10 | 68.1 |
| 0:05:15 | 68.1 |
| 0:05:20 | 68.1 |
| 0:05:25 | 68.1 |
| 0:05:30 | 68 |
| 0:05:35 | 68 |
| 0:05:40 | 67.8 |
| 0:05:45 | 67.7 |
| 0:05:50 | 67.7 |
| 0:05:55 | 66 |
| 0:06:00 | 66 |
| 0:06:05 | 66 |
| 0:06:10 | 66 |
| 0:06:15 | 66.1 |
| 0:06:20 | 66.1 |
| 0:06:25 | 66.1 |
| 0:06:30 | 66.1 |
| 0:06:35 | 66 |
| 0:06:40 | 66.1 |
| 0:06:45 | 66.1 |
| 0:06:50 | 65.9 |
| 0:06:55 | 66 |
| 0:07:00 | 66 |
| 0:07:05 | 65.9 |
| 0:07:10 | 65.8 |
| 0:07:15 | 65.7 |
| 0:07:20 | 65.7 |
| 0:07:25 | 65.7 |
| 0:07:30 | 65.7 |
| 0:07:35 | 65.6 |
| 0:07:40 | 65.5 |
| 0:07:45 | 65.4 |
| 0:07:50 | 65.4 |
| 0:07:55 | 65.3 |
| 0:08:00 | 65.2 |
| 0:08:05 | 65.1 |
| 0:08:10 | 65 |
| 0:08:15 | 64.9 |
| 0:08:20 | 64.9 |
| 0:08:25 | 64.7 |
| 0:08:30 | 64.7 |
| 0:08:35 | 64.8 |
| 0:08:40 | 64.8 |
| 0:08:45 | 64.6 |
| 0:08:50 | 64.5 |
| 0:08:55 | 64.5 |
| 0:09:00 | 64.5 |
| 0:09:05 | 64.4 |
| 0:09:10 | 64.2 |
| 0:09:15 | 64.2 |
| 0:09:20 | 64.2 |
| 0:09:25 | 64.1 |
| 0:09:30 | 64.1 |
| 0:09:35 | 64.1 |
| 0:09:40 | 64 |
| 0:09:45 | 63.9 |
| 0:09:50 | 63.8 |
| 0:09:55 | 63.8 |
| 0:10:00 | 63.7 |
Hey Don, where do I attach DMM probes to measure voltage (current?) on a running torch ? I'm guessing it involves dis-assembling the tailcap ?
P.S. have to fess up - I'm an opensuse user. My excuse is it was the first distro that would install on my home built piece of jetway junk back in the day and I just kinda stuck with it (christ, sound more like a winshaft usr every day). Now I admit that I've sold my eternal soul for the best KDE in the business (probably).
Voltage. Across the cell.
Current - between the cell and the LED.
OpenSuSE! Good stuff.
Here it's a mixture of Solaris, OpenSolaris, MacOS X, Winders, and assorted flavours of Linux - usually Ubuntu (AKA, I can't configure Debian). The recently decommissioned server ran SuSE Linux till it killed its third PSU. At that point I shut it down after 11 years of running. Best uptime was more than 3 years. There wasn't a single reusable component after that. When it died it did the job properly!
Cheers, I guess easiest way to measure V will be to put cell in cradle and croc to the driver - do I need to involve the switch (or are they just a switch, will be doing this on kd c8). Similar for A, but between cell + and driver. Have a fluke 115 rms, seems pretty good, read in another thread about short leads, have some good wire, thick and flexable high temp stuff.
My home "server" is an ex corp dell optiplex, 6 months to a year between reboots (usually for kernal update). My claim to fame (but without the actual fame) was patching suse 10.3 to run on a eeepc701 using the sdcard as /home, rewrote the shutdown script to handle the sd as a system drive instead of removable drive. I was well pleased with that one - think it got wiki'd :) My laptops have had suse, arch,gentoo(fail), ubuntu, slax, but keep ending up back with suse. When it works it really is a joy, when it doesn't installing arch teaches me how to fix suse lol. I have even tried the Haiku alpha release and what an absolute joy is is, when I can get my mits on a beta it will go on all laptops/netbooks.