[TruEnergy]

As long as your contactor cuts out at the right voltage, I believe your charge method will work fine. My guess is the lithium will consume the majority of the charge current due to its lower IR.

Originally Posted by Ziggythewiz

..., and I decided I need some more meters before I'm ready to go.

Not sure what you have in mind for meters, but I have one of these Doc Wattson Meters, and its pretty neat for the price. It doesn't log, so you have to manually record data at regular intervals. But for tests like this, you'll probably be monitoring it anyway.

http://www.powerwerx.com/digital-met...dc-inline.html

I ended up finding one on eBay a while back, for slightly cheaper. If you google it, there seems to be more online stores carrying them now.

[Ziggythewiz]

I have a JLD404 for measuring amps through the CALBs, and I'll compare that to the crappy westach reading the full pack. I'll also have a bunch of cheap DVMs on individual batteries/groups.

[TruEnergy]

Nice.

Glad you are able to do this experiment. Thanks, and looking forward to the results.

[Ziggythewiz]

System's been good to go for about a week. Just need the rain to stop or find the time to get my wipers fixed so I can start getting data. Some areas got 17" yesterday.

[Ziggythewiz]

So...I've finally managed to squeeze a few commutes in between the thunderstorms.

Before starting the experiment, the floodies (HM29) measured at ~12 mΩ. The CALBs measure at 6 mΩ. So far, current sharing seems to be close to expected based on those numbers, with the CALBs passing 65%

With the batteries typically within 1/5th of a volt when connecting or coming off a cruise, current between strings starts around 10A, and decreases to ~5A within a few seconds, then tapering down to 2 within a minute or two. I never gave it a chance to fully equalize, will have to see how long that takes.

With my current charge level the CALBs were almost always feeding the floodies, though it was running backwards at least once. I need to tweak the charge level to try to minimize that.

Voltage sag for the boosted batteries was dramatically improved.

I really need another JLD or two to get better data, and will look into accessing the data port for some automated logging as well.

[PStechPaul]

I like experiments like this. The idea may go against conventional wisdom, but you have made some thoughtful approximations and it seems like it may work for you.

My idea was to have complete battery modules containing any number of cells of the same chemistry, and including all the BMS and charging and protection circuitry in the module, and also boosting (or bucking) the voltage so that the output is a consistent value, such as 144V. Each module would be individually rated for a certain kW-Hr and could have built-in displays showing the approximate state of charge and cycle data and expected end-of-life. But such a device would need a very efficient DC-DC converter or buck-boost circuit. It might be useful for small vehicles where you might be able to use one 2 kW-Hr pack with 320 VDC output for a 2 HP riding mower or small tractor with a 3-phase motor.

Keep up the good work. Your monitor panel is impressive!

[Ziggythewiz]

Thanks, I'll have to rework my monitor panel when I get another JLD. Will need some better tools so I'm not having to 'cut' the plastic with my grinder again.

[PStechPaul]

Have you considered using a PIC or other microcontroller with multiple A/D channels to do the monitoring? It could interface to an LCD display and/or it could connect to a Windows program vis serial port or USB to display the data and log everything to a data file for later analysis. Here is what I have done:
http://www.mytractorforum.com/showthread.php?t=239700

It uses quad op-amps as differential amplifiers so it can read voltages and currents without worrying about the ground reference. I don't have the circuit in good form but I'm willing to share details and source code. The PIC code is written in C and the Windows GUI uses Borland Delphi, but you can also use Hyperterm or any other serial port terminal application. However, for a nice display, you will want something a bit better.

I wasn't sure what the JLD was but I found this:
http://evtvshop.projectooc.com/prodd...?prod=JLD404AH

That is a nice meter but for $175 it's a bit much. If you don't mind a few hours of breadboarding you can build something like I made for less than $50 or so.

[Ziggythewiz]

Thanks. I do need to look into that or a celllogish method to store the data. My meters are great for watching, but it's a a bit of work to take samples from a video.

The JLD404's awesome, after having it for months I don't know how I ever drove without it. Mine's a bit gimpy, so I need to send it in for replacement, but I may get another one as well, haven't decided. It's only $70 from the vendor, $100 if you count the shunt and a DC-DC if you want that. It can do logging as well, but I haven't messed with that yet.

[Tedktis]

Awesome! Love the expirement, regardless of the outcome. Think out of the box and take risks, that's what got us here in the first place isn't it?