LCD Instrumentation for Cougar Controller

[Freyguy]

Originally Posted by harlequin2

..... Resetting SOC on power up gives me 57% not 100%. Any clues?

harlequin2,
Did you change the battery capacity constants? If so, maybe you want to put it back to my defaults (115Ahrs) and verify it shows 100% after resetting SOC. I know I tried different battery capacities and it always showed 100% on reset. Very strange.

[harlequin2]

Yeah, the problem was finger trouble. I calculated my 135 AHr battery capacity as 0x76A70 and entered that as 76 and A7 instead of 07 6A. Silly error, when I entered the correct values it worked OK.
Chris, does that mean if I had a 2160 AHr battery, it would not ever show 100% SOC?

[jackbauer]

I did a little video. Couldn't help myself

http://www.youtube.com/watch?v=_zg7CiqjsOo

[Freyguy]

Originally Posted by harlequin2

Yeah, the problem was finger trouble. I calculated my 135 AHr battery capacity as 0x76A70 and entered that as 76 and A7 instead of 07 6A. Silly error, when I entered the correct values it worked OK.
Chris, does that mean if I had a 2160 AHr battery, it would not ever show 100% SOC?

Ok, whew!! Thought it had to be something like that!

If you were able to afford 2160 Ahr worth of LifePO4 and then somehow?? fit it in your BMW I'd be glad to make changes to support that range

[Freyguy]

Jackbauer,
Thorough video as usual for you
Nice to see the battery amp match the motor amps within 1A, when you consider the 2A steps of controller and rounding errors in firmware.
A little surprised to see the 2C error between controller temp and the other two. Mine only shows 1 deg F difference which is less than 1C. Controller does
use a different thermistor type and lookup table though.

-Chris

[jackbauer]

I'm not actually using the termistor specified for the controller. Its a 10k ntc sourced locally. Probably why its a bit off. Far from a problem in any event.

[Freyguy]

Hey guys.
I want to make you aware of two issues discovered that I’m correcting, which will result in new firmware.
1. MTE/KTE display fields are only setup for 10’s and 1’s. Values over 99 will not display properly. This is a simple fix to insert 100’s field since there are two spaces between SOC “%” and MTE “10’s”. Don’t think will affect many people. (Thanks to harlequin2 for finding this)

2. I discovered an issue with transmitted profile parameters “not taking”. I have the controller transmitting at a 200ms rate and it appears that due to this, some of the profile data sent from the LCD are ignored. If I disable controller transmit (rtd-period=0), I have no problems with it accepting all new values. Explains why I didn’t catch this, since during testing I disabled transmit to make it easier to verify updated values (typing “config” for parameter dump).

I need to update the code so that before sending profile changes, it sets rtd-period to 0. This gets tricky, since I really should do it immediately after receiving a string from the controller (i.e after I detect a CR) when there is guaranteed serial com dead time. After profile changes are made it will turn it back on.

I have most of the coding done but won’t be able to test it till next weekend (at the earliest)

-Chris

[jackbauer]

I installed the module in the car today just sitting on the center console and it works great. One observation regarding the fuel gauge. The driver on the lcd module and indeed the commercially available "zeva" fuel gauge driver from evworks i think all assume the gauge has a +12v supply and ground is via the tank sender. Perfectly valid for most older cars. My bmw is a 1996 model but uses a computer in the dash to control the instruments. Its not can bus or anything but the fuel gauge is not directly driven by the tank senders.

It uses 2 tank senders. One for each side. Both standard resistive elements. Approx 220 ohms on each side equals a full tank. 20 ohms is empty. Two problems. First , its not ground referenced. Second , although i can rapidly "fill" the tank by increasing the resistance , any attempt to "drain" it fast results in the gauge staying put. I would say a lot of modern and modern-ish vehicles will have this sort of system or someting similar. I happened to note that Jack Rickard's 2009 mini cooper (made by bmw) has the same two tank sender setup.

I'm looking at a possible fix using a digital potentiometer but all that i have checked so far are 5v only. Sender works on 12v. Any thoughts?

[harlequin2]

I wonder if another linear opto-coupler might be able to do it? I have a 1988 BMW and I think it works like that too.

[Freyguy]

Yeah, sorry if it wasn't clear in all my previous info, but the LCD brd is intended to drive an analog gauge directly and assumes a ground referenced sender.
Analog Devices makes digital pots that work up to +30V. The AD7376 comes to mind.
The problem is going to be finding a digital pot with such a low resistance. 10K~100K is the normal range of these devices. You can't really parallel something like this with a 200~300ohm resistor and expect to get somethig useful. It will be terribly nonlinear with the pot only impacting overall resistance near the end of its overall step range. Maxim makes a DS3906 for low ohm applications, but its 5V, so stuck again.

I suppose you could build your own low resistance ladder network with several analog switches (silicon not mechanical) and resistors. Need at least 6 bits (64 steps), but now you need a 6-bit parallel interface to the micro. I'll think on this some more........