"Zephyr" - Finally, the "v4" Fechter/Goodrum/Hecker BMS...

[Zeropointbug]

Yeah, I know the capacity of the battery, which wouldn't be a problem when discharging, that is not an issue as I only ever discharge 75% capacity anyways, and have plenty of safe zone there. The main issue is charging, trying to avoid over charging of any one cell, that is what I want the BMS for, and if you say it's good enough, then I shall just use it then.

I will post some pics of the new battery box I am going to use... never used ammo box, roughly 6x8x12 inches... the only thing I am figuring out is running the wires out, as it seals very tight. This would sit (with rare earth magnets) on a nice rear luggage rack.

 

[Gregb]

Okay, so would you recommend your board vs. the 30A BMS I have from Cell_man? Which I already have, but am hesitant to use it as the in rush current rfom plugging into controller is too hard on it.
)

If you are having any problems with cellmans products why not speak to him. I have found Paul most approachable and helpful with my battery......

 

[intoworldsofuncertainty]

Hi,

I'm interested in purchasing a "Zephyr" BMS printed circuit board but have found the tppacks website to be under construction. Does anyone know when it will be back on-line?

cheers!

I think that is a "how long is a piece of string" question. Due to his provider going belly up with no warning Gary has to rewrite his site......

Yes, I'm still trying to get my hands around this task.

In the meantime, you can PM me if you want a board. It's $79 for a 16-channel version and $99 for the 24-channel variant. Shipping is about $5 domestically, and about $15 for international.

-- Gary

Thanks Gary, I'll do that.

dave

 

[Gregb]

Steveo have you measured the PD across those resistors yet ?? That will let you know if there is actually any shunt current.

 

[Zeropointbug]

Because he has not responded to me in 1 month.

 

[Gregb]

Is this the address you are using? answered me almost immediately
emissions-free_worldwide@hotmail.com

 

[Zenid]

So far I have only been using the BMS as a charger, and have not been using the LVC portion of the unit to act as a cut-off. I wanted to check something first:

Is it okay to use this BMS with regenerative braking? I'm using Lyen's IRFB 4110 12-FET controller with regen set to around 90V. Is it okay to keep it attached to my system while it is motion as it is, or do I need to make any modifications for it to survive the regen?

 

[fechter]

Is it okay to use this BMS with regenerative braking? I'm using Lyen's IRFB 4110 12-FET controller with regen set to around 90V. Is it okay to keep attached to my system while it is motion as it is, or do I need to make any modifications for it to survive the regen?

I think it should be OK. The worst case scenerio is you start out with a full charge at the top of a large hill. In this case, the cells could be charged over the safe limit and the shunts will come on. If there was a way to disable the regen upon HVC it would be the best.

In real life, it's almost impossible to get the cells up to HVC from regen because you drain them some first before hitting them with regen. As long as you avoid the above situation manually, there should be no problem.

 

[PJD]

Gary or Richard,

After receiving the board, I just noticed that the instructions say the CV voltage of the charger must be adjustable. However, this isn't the case on the chargers I have, which are currently being used with my old 16-cell V1.5 boards. These charger's CV voltage is permanently at about 58.8 to 59 volts. The only adjustment on these chargers is an end-of-charge shutoff current adjustment. (although this adjustment on the charger can't quite be turned down low enough to effects a shutoff on the 1.5 boards because of the shunts).

Will this still work? I don't need a full amp of balancing current, 1/2 amp is plenty as long as the charger "throttling" works the same as the 1.5 boards. Can I use higher valued resistors to lower the shunt current and manage the heat? (The charger is only about 8 amp). Might I still get the BMS's charger shutoff to work?

Alternatively, I could change the resistor - pair values for the LM431 reference voltage and raise the shunt voltage up if I can't lower the charge voltage down, huh?

Thanks in advance.

Paul D.

 

[fechter]

That sounds like about the right voltage. If it's a little low, it should be no problem.

End of charge detection is based on current, so when the charging current drops low enough, it will shut off.

Throttling is different. It runs at a much lower frequency to avoid charger issues.

Yes, you could change the resistor values on either the shunts or the lm431 to prevent the shunts from overheating if that becomes an issue. The adjustment on the EOC will compensate for different shunt values, so that should not be an issue.

 

[PJD]

I received my V 4.4.3 boards, Instruction and BOM, and have ordered the parts. FYI,with no MTA connectors, and the slightly cheaper plain aluminum Hammond box, the bill was $94.11.

I have couple questions and comments:

1. As someone else commented way back, resistor R14 is still listed on the BOM, including the Mouser order list, as 1K not 100 ohm as it is supposed to be. The order already went out. I can a find a 1/8w 100 ohm resistor easily enough at home, but this typo should be fixed.

2. I'd like to raise the shunt voltage from 3.6 to 3.7 volts. It's OK if some end up being as high as 3.74 or so. I think this is a better charging voltage for Thunderskys, plus it will work better with the particular voltage of my nonadjustable charger.

How did you arrive at the Rx01 resistor value? A simple no-current calculation gives 3.49 to 3.53 volts for the Rx01=48.7K ohm; so are you also applying a reference current correction (1.5 to 4 uA from the data sheet) to get the value you chose? it looks like to get an average of 3.7 volts, I need to use a 51.1K for Rx01. Would you agree? Alternatively, I'm tempted to add a 0-5K trimpot in series with the 48.7K resistor (solder one end of the resistor to a stub of one of the trimpot leads) so i can be certain about getting the right voltage.

How are the actual shunt voltages turning out in actual use?

Thanks again,

Paul

 

[GGoodrum]

Yes, there's a slight correction applied. The calculated value for 48.7K/120K is 3.612V, but with 1% tolerance, this can range from 3.591V to 3.633V. If you want to go to 3.7V, the best value I see is 52.3K, which puts the center volatge at 3.694V, with a range of 3.672V to 3.717V.

-- Gary

 

[Zenid]

Just to let you know, I finally got around to writing up my Zephyr build on my blog/guide. Here it is in all its majesty...

I've split my build write-up into three sections. So far I've done the first and last sections where I've physically built, wired and constructed it, but I'm also going to do a section about testing and calibrating it (and maybe diagnosing and fixing it when that chip blew), taking on board the issues covered and clarification given in this thread.

Part 1: Building the “Zephyr” Circuit Board
http://zenid10.wordpress.com/catego...ms/4-7-2-1-building-the-zephyr-circuit-board/

Part 3: Wiring and Assembling the “Zephyr” BMS Unit
http://zenid10.wordpress.com/catego...-3-wiring-and-assembling-the-zephyr-bms-unit/

 

[fechter]

Outstanding! Thanks Zenid.

 

[Zenid]

Outstanding! Thanks Zenid.

Thanks

I'm continuing to tinker with the settings to try and get my charger/BMS combination to shut off at about the right time without fail. However there is something I still don't understand about the way it charges.

Sometimes it will charge at a steady current (in my case 9 amps) until the bank is nearly full, the LEDs are all coming on, and the current will slowly start to drop until it reaches a crucial point (about 0.8 A as I have it set up) at which point the EOC will switch off the unit. Other times, though, I will check it when it is finishing, and it will be pulsing in that other mode where the current drawn seems to switch on and off alternately, with the LEDs pulsing. Quite often it simply will not switch off at all when it is doing this. I am trying to fix this by dialing down the voltage on my charger a little, but I'm still testing to see if this has fixed the problem.

I don't understand what these two different modes are for, and what decides whether it charges in one 'steady' mode, and the other 'pulsing' mode (CC vs. CV?). I'd be grateful if you could explain this aspect of its workings as the different modes seem also to slightly affect the peak voltage delivered by my charger, and I think this may be what's causing it to fail to shut off sometimes.

 

[fechter]

When it is pulsing, that means one or more cells is reaching HVC before the cutoff current is reached. This is how the balancing function works when the cells are significantly out of balance. The EOC current detector is disabled during the off part of the cycle and comes back on when the current is restored. There is a time delay to make sure the EOC doesn't trigger prematurely during cycling.
It may be necessary to slightly shorten this delay time to get a consistent turn off point. Decreasing the value of C4 or R7 will shorten this delay time. If the delay is too short, EOC will trigger on the first HVC cycle. You might try placing another 100k resistor across R7, which will cut the delay in half.

 

[Zenid]

When it is pulsing, that means one or more cells is reaching HVC before the cutoff current is reached. This is how the balancing function works when the cells are significantly out of balance. The EOC current detector is disabled during the off part of the cycle and comes back on when the current is restored. There is a time delay to make sure the EOC doesn't trigger prematurely during cycling.
It may be necessary to slightly shorten this delay time to get a consistent turn off point. Decreasing the value of C4 or R7 will shorten this delay time. If the delay is too short, EOC will trigger on the first HVC cycle. You might try placing another 100k resistor across R7, which will cut the delay in half.

Thanks for the suggestions. However I'm stil confused about how it functions with respect to its two 'modes' of charging:

I thought the orange LEDs indicated when each cell had reached its 3.6V HVC limit (and was therefore 'full'), and that the surplus charge was being 'burned off' by the big shunt resistors while the other cells 'caught up'. As they all become caught up, they draw less and less current until EOC cut-off point is reached based on the setting of the 1K pot. Ordinarily this is what I see: The current stays constant until the LEDs start to come on one by one, whereupon the current gradually declines until the EOC is triggered.

So I don't understand why any pulsing is necessary at all. Clearly I'm not understanding something right here, as it's not clear to me why there are these two different modes and what the difference is. Could you clarify this please, and/or link me to a more detailed explanation? I need to understand it better than I do to figure out how best to get it to reliably cut-out when it's done. It would also make my write up more useful if I better understand its workings.

 

[GGoodrum]

That is the normal behavior except when you have the case where one, or more, cells are significantly out-of-balance. What happens in that case is that the shunts themselves can't hold the high cells at the 3.60V point, and they get "swamped". When that happens, the high cell voltages will try to keep rising. This is when the HVC signal kicks in, at the point the shunts are just getting saturated. This causes the charge controller logic to interrupt the charge current for a short time (about 1/2 second...), which lets the shunt circuit pull down on the cell voltage. This cycle basically repeats until the cells get close enough that the shunts by themselves can again keep the high cell voltages in check, which then lets the low cells catch up.

This is actaully a safety feature, to protect the cells from being overcharged. If this HVC cyckling starts happening for a pack where it normally doesn't, it might be an indication that there are some "weak sisters" in the pack. It could also just mean the pack got out of whack, maybe due to taking the pack down to where the LVC starts tripping. I know that the closer I get to LVC, the more the cells start to go out-of-balance, even with higher-end LiPos.

-- Gary

 

[fechter]

Good explaination.
If the charging current was low enough, the shunts would handle any imbalance by themselves, but at higher currents the shunts can max out making it necessary to interrupt the charging current to prevent the high cells from going over.

 

[fechter]

You might try reducing R7 to as low as 10k and see what happens. I'd also be interested in what the cycle timing is like when it's cycling (on time vs. off time). Adjusting the EOC pot higher should make it turn off even when it is cycling, but it would be ideal if it didn't matter. Lowering R7 should help with this. If R7 is too low, it may turn off on the first cycle or may fail to start automatically at the beginning of charge.

 

[Zenid]

If this HVC cyckling starts happening for a pack where it normally doesn't, it might be an indication that there are some "weak sisters" in the pack. It could also just mean the pack got out of whack, maybe due to taking the pack down to where the LVC starts tripping. I know that the closer I get to LVC, the more the cells start to go out-of-balance, even with higher-end LiPos.
-- Gary

Yes. I have had this issue more when recharging after prolonged discharges, like when I did some range tests in which I ran the pack down until the lvc cut-off point. At this point all the cells began to go our of whack and collapsed rapidly, with two weaker cells hitting lvc before the others.

Bank 3 of my pack after 10 miles or so

Bank 3 of my pack after 27 miles across 9 trips around town

My tests show that these cells aren't significantly worse than the others, they just poop out first, and replacing them would probably only buy me an extra mile or so.

I've lowered the voltage a little more on the charger to see if that will help. I noticed that for some reason in 'pulse-mode' the voltage on my charger seems to shift to slightly higher than the steady voltage it delivers in its ordinary operation, making it run hotter. I'll see if this lowered voltage stops it 'sticking' in pulse mode at the end of charge and maybe turn up the OEC pot a bit too. If both of these fail to fix it then I'll have a go at the modifications you suggested, perhaps replacing R7 with another pot to figure out what seems to work best with my charger.

[Edit] Regarding the charger interval as it currently stands, it's about the same as it was when I was testing it with the bank. You can see it in action on the YouTude vid I posted: http://www.youtube.com/watch?v=QN6tYkzuMNM

 

[Zenid]

What might be nice as a safety feature for the 'next generation' is a pot rigged up to a thermistor to cut the unit off if the case/shunts get too hot. Quite often the temperature alone is a pretty good criterion for deciding when the charge cycle is about finished...

 

[steveo]

Hey everyone,

I have yet to run my bms, i did mess up a couple of channels wiring one of the balance tab incorrectly, however to my knowelege ive repaired it but not 100% if it is working correctly.

When im charging the pack say at 50% soc the bms led is orange colour

As all the tiny leds light up and looks like cells are at the fullest the led starts to trip to red and continously tries to change, ive tried to tune the eoc pot but it wont do anything while it is connected

I then let the bms cool off and keep charger disconnected, i then turn the pot fully clockwise and plug charger in and the bms goes to green led indicating charged

Its either something is still burnt out or i cant get the pot tuned correctly for eoc

Thanks
Steveo

 

[fechter]

When you say the LED turns red and continuously tries to charge, do you mean it is cycling on/off?

There may be one resistor that needs tweaking to get a good EOC detection during cycling. Otherwise it sounds like it's working right.

 

[steveo]

When you say the LED turns red and continuously tries to charge, do you mean it is cycling on/off?

There may be one resistor that needs tweaking to get a good EOC detection during cycling. Otherwise it sounds like it's working right.

Here is a video for you to show what is happening

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

Thanks
Steveo