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

A pair of IRFB4115 should handle that but will be close on the heat dissipation. If the FETs were mounted off-board on a heat sink (or even still on the board but standing up with a heat sink), it should be fine. Heat dissipation of the FETs is the limiting factor.
 
Ok , so I understand that if i put the FET in another box with copper or aluminum heatsinks it would be enough cooled.The only thing that still intrigues me , is how i can transform the throttle limitation , that reduces throttle when too much voltage sag into another thing or that can be compatible with my vectrix?

Thank you for your responses and advice.
 
If the FETs have even a minimal heat sink area that gets cooled, they should be fine.

Interfacing the LVC is another story. I don't have much experience with a Vectrix.
Is there a schematic or wiring diagram of one somewhere? Somebody must have reverse engineered one by now.

Plan B is to simply use an audible/visual alarm.
 
Regarding the vectrix throttle , i found some patents , And i had read that it works by PMW signal(not analog).

I can't attach the file so , here are the links:

http://www.google.com/patents/US6724165.
 

Attachments

  • image.jpg
    image.jpg
    18 KB · Views: 6,498
I've had my zephyr bms going since it was released years ago and its been going ok so far with 16s headway lifepo cells from years ago, which seem to last forever. Up until now I've been using a meanwell with the 'hiccup' current limiting board to charge.

I just bought a cycle satiator charger and wondered if anyone has had any problems with it fooling the charger or causing strange charger faults.
I seem to be getting charger faults right away when I hook it up with it set to 7 amps.
When set down to 4.5 amps it seems to be ok, i'm slowly experimenting and seeing how high I can go or if I can change some charging parameter to get it back up to 7 amps.

Anyone else using the this bms with the satiator?
 
I wouldn't expect any errors until the Zephyr board starts hitting HVC on a cell and starts throttling.
You might need to contact Justin to see if there are any software tweaks that can fix it.
During throttling, the charge current will drop to zero, then come back up to normal in a fairly fast pulsing mode.
The charger is looking for the current to drop below some threshold to indicate end of charge. If the charger logic is set such that the charging current needs to drop below the threshold for some period of time longer than a few seconds, it should handle the pulsing.
 
midwest mayhem said:
Awesome, thank you. One other question - in the charge control section, it states:
Use a current limited bench power supply set to around 100mA. Set the End-of-Charge (EOC) cutoff pot about half way. Connect supply to Charger + and Pack – terminals. Slowly turn the voltage up to around 20v. The master LED should not light up and current should remain near zero.
I've done this without any cells connected and the master LED faintly lights up (red, I believe) around 8v, but goes up to 20v without drawing any current registered on the power supply. The board passes all the other tests. Is this bad or can I ignore it as long as everything else functions correctly?

I seem to have the same problem with my charge control section:

Connect supply to Charger + and Pack – terminals. Slowly turn the voltage up to around 20v.
[20V, same with 20.5V]
The master LED should not light up [It does, red] and current should remain near zero [yes].

Next, connect supply to charger + and – terminals on board. You should see the master LED light up orange for about ½ second, then turn green. [stays orange all the time]Measure the voltage across the HVC terminals. You should see 10V-12V.[seeing 10.9V] Short the HVC terminals with a jumper. LED should turn red with terminals shorted, and then go back to green when jumper is removed.[It turns red, but goes back to orange]

I checked Q3 as above and see 11.0V (negative lead on shunt). Any tips?
 
It's been a while since I've looked at the schematic. The first test connecting the supply to the pack connections may be in error. I'll have to look at that part more.

The behavior with shorting the HVC connection looks correct. It should toggle red, then back to orange.

It should quickly change to green, though, and that part definitely has a problem. This is controlled by U1a. You might try measuring the voltage on pins 2 and 3 against charger (-). It's real easy to have your probe tip slip and short the chip, destroying it, so be very careful doing this. Pin 3 should be zero volts. Pin 2 should be near zero, but a few millivolts.

If pin 3 is high, Q2 may have a problem or the EOC disable jumper has continuity.
 
I took a longer look at it and the first test with power applied to the pack terminals is correct. It should not light up in this mode. U1b compares the voltage on the charger input to the pack voltage. Anytime the charger is higher than the pack, it enables U2, which drives the LEDs.

With power applied to the pack terminals, U1 pin 5 should be slightly negative compared to the charger (-) input.
Make sure R3 is 1M ohm. You might also try cleaning the board around U1 to make sure there is no conductive flux residue.
 
Thanks fechter.

fechter said:
It's been a while since I've looked at the schematic. The first test connecting the supply to the pack connections may be in error. I'll have to look at that part more.

The behavior with shorting the HVC connection looks correct. It should toggle red, then back to orange.

It should quickly change to green, though, and that part definitely has a problem. This is controlled by U1a. You might try measuring the voltage on pins 2 and 3 against charger (-). It's real easy to have your probe tip slip and short the chip, destroying it, so be very careful doing this. Pin 3 should be zero volts. Pin 2 should be near zero, but a few millivolts.

If pin 3 is high, Q2 may have a problem or the EOC disable jumper has continuity.
EOC doesn't have continuity
Pin 2 is at 5.5mV, Pin 3 ~20mV (fluctuates a bit)

not sure if helpful
Q2 against Pack- (C/B/E) - 0.14 / 0.46 / 0.5V
against Charger- 0.02 (LED flickers green/orange) / 10.76 / 11.1V

fechter said:
I took a longer look at it and the first test with power applied to the pack terminals is correct. It should not light up in this mode. U1b compares the voltage on the charger input to the pack voltage. Anytime the charger is higher than the pack, it enables U2, which drives the LEDs.

With power applied to the pack terminals, U1 pin 5 should be slightly negative compared to the charger (-) input.
Make sure R3 is 1M ohm. You might also try cleaning the board around U1 to make sure there is no conductive flux residue.
Pin 5 is -390mV
R3 measures at 985kΩ, 11.1V & 0.11v on the ends

I appreciate your help, mate - at this point just tell me in which order to replace the components / which to order :mrgreen:
 
Unless I specify otherwise, all voltage measurements are against the charger negative (ground).

The emitter of Q2 (and the connected EOC jumper terminal) is the switched 12v line. It should be around 11v or zero depending on the mode. 0.5v seems a bit odd, but for sure not enough to power U2.

Try feeding power to the pack terminals and measure U2 pin 6. this should be the same as Q2 emitter, since they are connected. If U2 6 is low, there should be no way the LEDs can light up.
 
fechter said:
Unless I specify otherwise, all voltage measurements are against the charger negative (ground).

The emitter of Q2 (and the connected EOC jumper terminal) is the switched 12v line. It should be around 11v or zero depending on the mode. 0.5v seems a bit odd, but for sure not enough to power U2.

Try feeding power to the pack terminals and measure U2 pin 6. this should be the same as Q2 emitter, since they are connected. If U2 6 is low, there should be no way the LEDs can light up.

Spot on:
U2 pin 6 is at 11.15V
while probing around I found that while the LED is red(~1Hz), it turns off occasionally when I touch Q2's emitter.
Still 0.5V against Pack-, 0V against Changer-; area looks clean, will check back later with more details.

Edit: closing EOC disable causes LED blinking red, difference between on and off is just a few mV (I can "leave" it on or off by timing EOC disable open again) :?
tried removing R6 & D9 - no change
 
Might need a picture of your board to see exactly what layout version you have.

I'm seeing a discrepancy between the schematic and layout file I have. There may be an improperly routed trace on the board. I need to see the area around U2 and U3.
 
Board Correction!

I don't know how I managed to do this, but there is a trace that's improperly routed.
Luckily, it won't be too hard to correct. U2 pin 6 is going to the wrong place.
You can cut the trace with a sharp blade or the end of a tiny screwdriver. Just scratch across it until there is no more connection. The wire can be added on the bottom of the board, it will probably be easier there but on top will work too.

See the picture below:

Board top correction.jpg

The way it was wired before, U2 would be on all the time and the LEDs would never turn off.
You may still have another problem, but let's try this first and see how it tests.
 
fechter said:
I don't know how I managed to do this, but there is a trace that's improperly routed.
Luckily, it won't be too hard to correct. U2 pin 6 is going to the wrong place.
You can cut the trace with a sharp blade or the end of a tiny screwdriver. Just scratch across it until there is no more connection. The wire can be added on the bottom of the board, it will probably be easier there but on top will work too.
Did that, easy enough.
fechter said:
The way it was wired before, U2 would be on all the time and the LEDs would never turn off.
You may still have another problem, but let's try this first and see how it tests.

True again: now the LEDs are on all the time, Round 2:

Connect supply to Charger + and Pack – terminals. Slowly turn the voltage up to around 20v.
The master LED should not light up [It does, orange] and current should remain near zero [yes].

Next, connect supply to charger + and – terminals on board. You should see the master LED light up orange for about ½ second, then turn green. [stays orange all the time]Measure the voltage across the HVC terminals. You should see 10V-12V.[seeing 10.8V] Short the HVC terminals with a jumper. LED should turn red with terminals shorted, and then go back to green when jumper is removed.[It turns red, but goes back to orange]

I checked Q3 as above and see 11.0V again, same on Q2.

fechter said:
With power applied to the pack terminals, U1 pin 5 should be slightly negative compared to the charger (-) input.
-0.25V

Then I saw that R5 had a cold solder joint (or maybe I knocked it loose at some point, some of my components "stand" rather high) - fixed that:
no LED on C+ & P-
green on C+ & C- , red withHVC shorted, orange with EOC disable shorted

Happy days!!!

playing now with my 4s test pack (old cellphone li-"ions") :mrgreen:
PM me your PayPal mate - I owe you a pint!
 
I suspected something else wasn't right, but glad you found it!
I'm kicking myself for not catching the routing mistake. You did a bunch of people a favor by finding it.
 
Let's call it another happy customer - especially because he knows now what "all this shit does" 8)
my crude test setup, charging with a 1300mA LA charger in series with a linear regulator off a 12V LA battery:
SDC18292_zpsp4ivthyv.jpg

My resistor selection works out to be spot on 4.00V (give or take 20mV). Anyone else charge cellphone batteries through the Zephyr BMS? :lol:
it's now happily balancing away
 
Does this track correction need doing on all boards?

My board is from Zenid purchased last year version 4.4b, I have built it but not properly tested it yet.


Regards,

Chris
 
chris1968 said:
Does this track correction need doing on all boards?

My board is from Zenid purchased last year version 4.4b, I have built it but not properly tested it yet.


Regards,

Chris

Yes, I believe so. If you aren't sure, you could send me a picture of that area on the board.
 
Looking at the board I think it does need it.

I've done the modification now anyway, it seemed to test as per instructions.

Still need to set the charge voltage and EOC but can't do that until I've fitted it to the pack.


Cheers,

Chris
 
I'm looking to build a Zephyr BMS for 12 Headway LiFePO4 15A cells, but the through hole version of the 2.1V TC54 is obsolete (well you can get them from MicroCHIP Direct after 3 months). So two options:
1) substitute with 579-MCP111-240E/TO, but this has a cut-off voltage of 2.32V (I'm okay with protecting the batteries a bit more), only drives 10mA (instead of 50mA) and operations up to 5.5V (instead of 10V). Will it work, and would I need to lower the protection Zeners to 4.7V or 4.3V?
2) try to solder the surface mount versions of the TC54s onto the pads. Has anyone tried this?

Also can the 512-1N4004 (also obsolete) be substituted with 512-UF4004?

Thanks,
Ian
 
IN4004 is still available from other manufacturers. 833-1N4004-TP The 512-UF4004 is equivalent and should work fine too.

For the voltage detector, 579-MCP111-240E/TO looks like it should work but I didn't check the pinout to see if it is the same. If not, you can bend the legs around to make it fit. 10ma output should be more than enough to trigger the optocoupler. Using lower voltage zeners is probably a good idea. 5v ones should be OK.

579-MCP121-240E/TO looks like another option but still 2.3v.

I haven't tried mounting the surface mount TC54 on one of those boards, so not sure if that would be easy or not. If the legs reach the pads on the board, it would be easy, but again, need to check the pinout to see if they are oriented right. Sticking them on the bottom of the board is an option if that makes the pins line up. Eventually all the through hole parts are likely to go obsolete.
 
As luck would have it the pinouts of the MCP1x1 are the same; as is the pinout of the SOT23 version of the TC54. Looking at the discharge curves of the Headway cells,cutting off at 2.3V will only lose a few percent of capacity at room temperature. I won't be cycling when its too cold!

Your diode choice is cheaper, so wins.

Thanks for help,
Ian
 
Saludos
I am about to test a couple of the last version BMS boards and I have a question about the Cell Circuit Test, page 10, Zephyr BMS Instructions:

texto.jpg

It says it should measure 169Kohms but the test is not taking into account the rest of posible parallel paths (facing the 169K divider).



In my case I just get 7Kohms in all individual cells. The only thing I have done diferently is using the SOT-23A package type (wraped in epoxy glue) for the Voltage Detector since the TO-92 was impossible to get.

Am I missing something here?

Thanks for your comments, -apg
 
Back
Top