Old BMS General Thread

[retepsnikrep]

Greg/ Boelle

Yes sorry my site is down at moment, and ive forgotten my password to re=upload it and i'm 400 miles from where i've got it stored!!

Anyway re the insight see www.insightcentral.net and have a look at my Lithium phev thread.

Rick

Yes I will be using 1amp fuses in the cell leads.

The slaves use a simple npn transistor to turn on the load, yes it could fail on but unlikely, The slaves and the master have the same voltage info and load cut in/out points so the Master knows when any of the slaves should be bypassing. It displays this fact at the bottom of the main screen. The Master can't yet detect a cell being bypassed if not commanded but I could certainly add an imbalance alarm when a cell starts to deviate from the norm.

It might be possible later to detect a stuck load with some clever voltage analysis by commanding the load on and off from the master and seeing if the sensed voltage does drop as it should due to the wiring
.
The master display aready shows the highest and lowest cell V.

You can also look at all the individual voltages.

[retepsnikrep]

Many posts ago I did devlop master and slave software that allowed commands to be sent to individual slaves even when all the slaves had the same software and no id NUMBERS. So it's possible to send a selection of commands to the slaves or a specific slave via the slave bus serial link, the responses can be monitored on the Master data bus as now.

At present the simple digital slave software just responds to an interrupt and dumps it's cell voltage on the master bus in turn.

The basic hardware is simple but the software can be as clever as we can make it.

I we suspected a stuck load the master could enter an analysis mode where it activates each load in turn measuring cell voltage before/during/after load turned on/off.

Any strange variation or lack off could be flagged up. Using fairly thin wire to join the cells to the multi slave boards some sort of voltage drop would prob be evident as loads come on/off. The dig slave software will prob have to be modified anyway to add a compensation value say 10mv to each reading if the load is on, or it will have to turn the load off whilst voltage readings are taken.

[GregsGarage]

Rick,

Regarding the possible stuck on load, I actually had this fault on one of my slaves, due to a poor solder connection. Was probably about 4 hours before I noticed, but because the load is very small, around 250ma if I remember correctly, it will take at least a week before this could kill a cell, probably 2 or 3 weeks for larger cells. The voltage display on the master should detect the cell low voltage long before this becomes a problem, although I suppose if the fault happened right before you went away for a month long holiday it could be a problem. Many systems use larger loads which can cause problems if they fail, but really it is unnecessary to have even a 1 amp load for balancing as long as your cells are in good shape, and if they are not you should replace them.

Greg

[steiner]

Greg,

I am currently using an analog board that I built myself for balancing my cells. It bypasses about 1.2 amps. I have 90Ahr Thundersky batteries. There have been a few times that I wish it had more bypass to keep an individual cell from going to high. My basic problem is that I don't have control of my charger. It is built by Thundersky and has its own charge profile. I have opened it up to see if I could tie into it to control it similar to the Zivan. No luck so far. Due to some intense testing with a few of my cells, those specific ones are always a few tenths of a volt behind the others. Due to them being behind, the others potentially will overcharge. I know this is a unique problem that I have but I can't help but wonder if something similar to this might occur on a standard pack of cells as they age.

Peter,

I like your idea but wonder if a simple delta voltage between the highest cell and lowest cell might work. It wouldn't catch the blown mosfet right away but would prevent it from damaging the cell. The master could simply sound an alarm if the delta V got above a certain value. The display could show which cell has droped in voltage. Then a simple visual inspection of the slave board would show the LED was lit for that specific cell even though charging isn't occurring. Just a thought.

I am really planning on implementing your latest BMS (all of the slaves on one board) as that is how my analog system currently is structured. My simplistic idea of controlling the charger is to simply turn it off and back on at certain cell voltages (I think that is how Greg's works?). My question then is did any of you find a good relay for switching the input power to the charger on and off?

[retepsnikrep]

My basic problem is that I don't have control of my charger. It is built by Thundersky and has its own charge profile. I have opened it up to see if I could tie into it to control it similar to the Zivan. No luck so far.

Rick you need control over the charger output voltage as a minimum. I suggest you invest in a Zivan, I've never had a problem with them. Using a relay to turn it off is fine at end of charge cycle, but it should not turn back on again via the relay unless the bms is running in digital mode. You can control the Zivan output voltage (and current to a point) using ther opto trick on the Zivan adjustment pots.

Greg had an issue which May have been caused by a cell fault resulting in the charger feeding power into a fault scenario. The charger is a critical part of the system.

[boelle]

----

[GregsGarage]

My simplistic idea of controlling the charger is to simply turn it off and back on at certain cell voltages (I think that is how Greg's works?).

This should not be done. The problem you have is when you switch the charger back on it starts its charging profile all over and the charger doesn't know the state of charge of the cells. If you have one cell short or otherwise drop voltage for some reason the charger will go back into it's bulk phase charging and this could cause a fire, don't believe the manufacturers claims of Lifepo4 being 100% safe, even lead acid batteries can catch fire in some conditions. You must find some way of controlling the charger output near the end of charge. If you can't find a way of directly controlling the output of your charger then the next best thing is to add some resistance between the charger and battery pack. This could be as simple as using a relay to put a suitably rated light bulb between the charger and pack. For an other option have a look at the circuit on the agni motors website for a solar charger setup, about 2/3 of the way down on the following link.
http://agnimotors.com/home/index.php?option=com_content&task=view&id=27&Itemid=39
You may be able to adapt this circuit to use with your charger, but the important thing is that the bms must have someway to take control of the charger output during the final stages of charging.

[retepsnikrep]

32 slave pcb now delayed for a month as my man is on holiday. Still he has sent me some pics layouts to check etc. Still too much to do and so little time!!!

[boelle]

well, i'm following this one still, any luck fooling the car to accept the extra charge so that the dash SOC reads correctly?

[steiner]

Peter,

Sorry to hear about the delay. I am very interested in your new approach. If there is anything I can do to help, please let me know. I would be more than happy to put a second set of eyes on the layout.

Rick