connections for? EcityP Smart S5-13 BMS

[rolf_w]

I bought it from bmsbattery.com ... I have 12s lipo pack

Ahh - so you have bought a 5-13s sBMS ; I thought you bought the dual OZ890 BMS shown in the foto above - I am searching for a place to buy thistype...

let me know the outcome of your tests - I'd start bleeding at lower voltages to gain more bleeding time...
r

 

[Mad Professor]

I am also working on a 02Micro OZ890 BMS unit.

I don't have the money to spend out on one of the Aardvark I2C/SPI Host Adapter's.

So I am trying to see what I can do with a simple Arduino Uno.

My topic can be found here: O2Micro OZ890 Based BMS Units.

Best Regards.

 

[krismclean]

I've wired the 14 way plug from one of these units to my battery pack but when I plug the BMS card in it's going to make all 14 connections simultaneously. I can't believe the complicated procedure in the apps note of connecting power then ground then each of the 13 cells is really necessary? What's the point of having a 14 way plug if the connections have to be made one at a time?

 

[rolf_w]

I can't believe the complicated procedure in the apps note of connecting power then ground then each of the 13 cells is really necessary?

it is crucial to energize the OZ890 chip before you connect the balancer wires! Thus connect first B- and then P+ after that you can plug in the JST connector. As all the input lines (B1 - B13) are polarized already after pre-connecting B- and then P+ the connection sequence of the balancer wires isn't that important anymore. What I try to do: I skew the connector a bit so that pin 1 inserts first...
r

 

[krismclean]

Thanks for the heads up on the connection sequence. I'm actually having second thoughts about the whole BMS idea. I've got 39 Panasonic NCR18650Bs (13s3p) charging from a 54.6v 4A charger & they seem fine, no sign of cells getting out of balance. I also dismantled a commercial 10Ah pack, no sign of a BMS inside it either.

 

[friendly1uk]

the 'bleeder' LEDs are very helpful. did you solder them manually? tricky job...

Yes, it was very easy :wink:
The leds are standard yellow in size 1206, but the PCB size is 0603...

Following this, I ordered some 1206 led's and they are way to big. You can't see the solder pads when they are placed on the board. My led's are not square at the end either, they have a D shape cut out which gives you a much better window on to the circuit board below. Still... There is nothing visible to solder too.


If anyone wants 30 1206 led's and some 470 resistors, I can do you a great deal, so I never have to see them again.

 

[krismclean]

I can't seem to get theOZ890X_087 software that came with this card to connect? I've got the card powered from a 24v supply but when I plug it in the little connect icon only goes green briefly & I can't get anything meaningful from the card. The BMS was supposedly setup prior to shipping for 13 cells but the default readout before I connected was 8 cells & it never changes. I think the PC software is finding the USB to RS232/I2C converter (although no USB drivers were supplied). A little green led pulses briefly in the converter box & there is a spike in the card current draw at the same rate. Thoughts anybody?

 

[rolf_w]

I can't seem to get theOZ890X_087 software that came with this card to connect? ... Thoughts anybody?

the OZ890 follows a particular power up sequence: make sure pin BAT4 has the required voltage! from the specs:
"When power supply is applied to VCC>8.5V and the voltage at pin BAT4 of OZ890 is higher than 6.5V, the common bias starts first, followed by the band-gap and 3.3V & 2.5V LDOs. When V3.3>2.4V, the power on reset block generates POR signal to enable the 512K oscillator and initializes the digital section. When power and clock are ready, the digital circuits will read the pin configuration and EEPROM data, which in turn determines the working state.
If RSTN_BYPASS (bit5 of EEPROM register 32h) is “0” (default value), OZ890 will go to the assembly state. In this state, OZ890 does not do ADC safety scan; only after pin RSTN (pin 29) negative pulse (low active reset signal), OZ890 can go to the normal working state. If RSTN_BYPASS is “1”, OZ890 will go to the normal working state directly.
When the voltage at pin BAT4 of OZ890 is lower than 6.5V or VCC is lower than 8.5V, OZ890 is in shutdown status. All LDOs (V3.3 and V12) are disabled and all MOSFETs are disabled.
There is one essential requirement of assembly sequence while connecting the battery cells to the protection board using OZ890. In general, the ground pins and VCC pin of OZ890 should get power before any other pins."
r

 

[krismclean]

it says it's drawing 191 mA but my multimeter says it's only 20mA. Clearly 20mA at 49.8v is making the chip run too hot at 68 degrees. This strikes me as a very cranky design does anybody know of a more robust BMS for 13 cells?

 

[dnmun]

on the other thread i posted up the link to this one. i actually think that 15S is where you should go. you can buy 5 of the 6S hobby king packs and split one in the middle to make two 15S series of lipo so you would have 10A wide of the lipo and then use this BMS to monitor it. the discharge rate is low for most people but i think it can be hacked for the output current by adding more shunt resistors on top of those it has.

http://bestekpower.com/fuelgasgauge/BMS-D141.html

the other thing is that it is less than 63V so all the 36V-48V controllers will work at this voltage. so most speed that the regular controller could provide. plus the inherent safety of having a BMS for short circuit protection and over charge protection too.

irfb3077 that jeremy popularized would work for the output mosfets at this voltage too.

 

[friendly1uk]

Sorry, trying to quote someone but it's missing out the images they posted, and it's all I need to comment on.

Edit: Rolf posted

I have the Smart bms here. The top pic has the led and it's accompanying resistor in the wrong sequence. The led is pointing the right way in both electrical terms and physical positioning, but it is the resistor on the left and led on the right. Poking about with a multimeter I was losing faith for a while. If your capable of soldering in the leds, don't be discouraged by the circuit being incorrect. Just place them with there indicator marks towards the battery. Good luck. I used the right size and it was not so bad a job.

The bottom pic is not like in the instructions. That link should pick up each consecutive cell, not skip to the end. Connecting 8+9+10+11+12+13 although I don't think it matters as the skipped cells have there capacitors shorted anyway.


My board don't work. It came set up for 12s and it appears was built with that in mind. The 13s pin is not in the socket and never has been. CF13 has been swapped for a resistor. Nothing else appears to of been done. This has left me quite miffed, as I have in essence put in the balance wires before the battery appeared across the entire set. Something it was thought might be a requirement. So I may of stuffed it up by not checking china's work properly. However.. the chip was live when the balance wires went in, and it does power the 3.3V test pad. It might be OK. I will know when I link up 12+13 then perhaps need to reset the chip. That looks easy enough, it is marked rstd on the under side of the board, and right by it is ground. Don't let the slightly wrong labeling put you off, it is reset and the ground is both battery ground and the analogue ground needed. The requirement to reset the board(or not)is set in the software so I dunno if I have to or not as yet.

Right now it is just deados. Battery voltage appears to leave my battery box for the controller, but a poor (but digital) multimeter pulls it down to 24v while a good meter see's the full battery volts. It is very weak indeed. Plugging in the charger see's no change. However, using either multimeter on the gate of the main fets turns them on. So I can get full battery power through the parallel fet's, and I can get the charger to light up red like it is actually doing something. Simply by switching on the fets with my multimeter. TBH I did not expect that, but both my meters here do it.

 

[friendly1uk]

bridged the 12+13, no change, hit reset, Full output voltage and some charger activity. Odd though.. 1 second red, 7 seconds green, and repeat... repeat.. repeat...

No balancing light illuminated, I will drain a cell down and see...

 

[circuit]

I've just received my "smart bms" from bmsbattery (outch, that was a pricey delivery). I also ordered their USB-I2C adapter for $100 ( ). Installed the O2 app, but it seems like it does not come with drivers... I get only "unknown device". Any suggestions? Where do I get a driver?

 

[Mad Professor]

circuit: Try the drivers that are within the O2 Software folder.

"C:\Program Files (x86)\O2Micro\CowBoy\Aardvark"
or
"C:\Program Files (x86)\O2Micro\CowBoy\O2USBI2C"

Also if you feel brave please open up the adaptor box and post some pictures of whats inside.

 

[circuit]

circuit: Try the drivers that are within the O2 Software folder.

"C:\Program Files (x86)\O2Micro\CowBoy\Aardvark"
or
"C:\Program Files (x86)\O2Micro\CowBoy\O2USBI2C"

Also if you feel brave please open up the adaptor box and post some pictures of whats inside.

Thanks! I'm now at other PC. Will try tomorrow.

I did open it up (first thing before powering up ). There is one 8051 micro, JTAG port, SO8 package and one missing SO8 (looks like serial memory).
Terrible quality enclosure, hand-cut holes for connectors.
I ordered two BMSes, it seems like they solder in connectors and caps/resistors as per cell count requirement. So it is better to order 13s to get full size connector (if you have a way to reconfigure it).


EDIT: Tried it. First time drivers in those directories do not work due to missing digital signature. I am running windows 8, unfortunately. When disabled driver signature enforcement, it says it can't find a driver for my device.

I guess I will have to contact bmsbattery...

EDIT2: I did. They say it is only compatible with windows XP x86. I wonder if it will work on virtual pc.

EDIT3: okay, It's alive. Virtualbox+windowsXP... BMS ir working from first try. Some measurement around 15mV off.

But I see a problem. I soldered in balancing LEDs and can now see how it balances. LED on cell1 is slightly lit and small balancing current is hoing through shunt even when BMS is in reset (all balancing disabled). Cells 1-3 do not balance. It seems the chip does not output enough voltage to switch on external fets. All upper cells work fine.
Anyone had this problem before? To me it looks like a faulty chip.

 

[circuit]

OK... So I took a second BMS and noticed that bleeding is disabled under Over Voltage condition. And now I see others have the same problem. The hell? I was thinking that this chip will be THE ONE, but no, it's just another useless chinese product.
Anyone have a solution?

Also, my OZ890 BMS crashed during calibration procedure, leading to corrupted EEPROM. Whole memory is corrupted. I have restored it from a backup, but some values don't seem to work. For example, internal temperature reads 117 degrees now (instead of 28) and I can't do anything to change it. But that is just a minor problem compared to disabled bleeding on OV condition.

Crash during calibration:

Error while trying to write to EEPROM internal temperature offset register:

 

[friendly1uk]

Circuit, would you say these boards can work with lipo? Something like a 4.2v hvc. I would of hoped so, but was just told they can't
Edit: Looks like they can to me. Cell range -0.3v to +5v

 

[circuit]

It can in theory, however is completely crippled in practice. Still can't get it working properly.

 

[Knutix]

Hi,
I would like to post that I experience the same as "friendly1uk". After receiving a 10S configured (lipo) version and soldered on the connectors etc.. I connect the battery first, then the balancing leads and then the charger. I see that the charger sets up 42 volts (charge voltage) and in one second 5A (5A charger) is sourced into the board, then the the BMS stops this current (why??) for 7 seconds then it opens up again.. repeat repeat....

Does anyone know why? There may of course be a wrong configuration in the EEPROM.. Need to check.. By measuring on the board I note that the 2.5 (ADC voltage?) is not present...

- Knutix

bridged the 12+13, no change, hit reset, Full output voltage and some charger activity. Odd though.. 1 second red, 7 seconds green, and repeat... repeat.. repeat...

No balancing light illuminated, I will drain a cell down and see...

 

[circuit]

Ah, by the way, I order 2 BMSes. First one I hooked to battery, I guess, the wrong way: balancing wires first, one by one, starting from B-. Result: lower 4 cells can not be balanced, so the chip was damaged. No other BMS is sensitive to that...

 

[friendly1uk]

Hi,
I would like to post that I experience the same as "friendly1uk". After receiving a 10S configured (lipo) version and soldered on the connectors etc.. I connect the battery first, then the balancing leads and then the charger. I see that the charger sets up 42 volts (charge voltage) and in one second 5A (5A charger) is sourced into the board, then the the BMS stops this current (why??) for 7 seconds then it opens up again.. repeat repeat....

Does anyone know why? There may of course be a wrong configuration in the EEPROM.. Need to check.. By measuring on the board I note that the 2.5 (ADC voltage?) is not present...

- Knutix

bridged the 12+13, no change, hit reset, Full output voltage and some charger activity. Odd though.. 1 second red, 7 seconds green, and repeat... repeat.. repeat...

No balancing light illuminated, I will drain a cell down and see...

The 2.5v won't be present. That is correct.

I didn't find the actual cause, although it is one of two things. I was using lipo, wanting a 4.2v hvc but couldn't get past 3.9v per cell, which is the hvc of lifePo indicating the board was not programmed properly. IIRC the flashing slowed and stopped after 10 mins, which would make this the right explanation. As battery voltage increased during these 1 second charge intervals, the interval taken for battery voltage to settle during the off time also increased. So after a short time the 1second/7second gradually increased past 1/10 and 1/15 until it just stayed off as the batteries picked up a tiny bit more charge. If your seeing this sort of behaviour then I imagine you are also charging lipo with lifepo settings.

When did you measure this 42v from your charger? If it is 42v while the charge light is on, your battery is charged, there is no fault.

The other of the two options I considered was an actual malfunction. Caused by them not wiring it up properly. They swapped the caps but didn't bridge the last cell of my pack up to cell 13 to feed pack voltage through R1 to P+ The effect of this could be like plugging in the balance wires before powering up the chip.

I think we both have a perfectly good bms, programmed wrongly.

You could discharge the pack watching you don't go below a safe lvc using your voltmeter. It would then charge properly for a short time before passing lifePo hvc and doing its " I'm finished " behaviour

 

[helgeo]

I have four of these BMS'es, two 12S and two 10S. All preconfigured for LIPO voltages. Have tried one of each, the 12S is working as it should; it charges and balances. The 10S is behaving in the same way as reported by friendly1uk and Knutix.

I have followed the connection sequence carefully on both and have communication with both from some simple software that I made myself.
Both BMS'es works with the software and reports correct number of cells and chemistry type. There are no reported faults on the 10S BMS when it is cycling, and the software switches between IDLE, CHARGE and DISCHARGE in line with what the BMS is actually doing.

On the 12S several components for cell 13 has been removed, for the 10S all components are still in place. Could this be the cause of the problem?

When using my software I could see that it reported some small voltage on cells 11 and 12, perhaps removing the extra resistors would make the extra cells voltage drop to 0 and fix the problem?

Sorry for the bad quality of the pictures:

 

[circuit]

Ah, by the way, I order 2 BMSes. First one I hooked to battery, I guess, the wrong way: balancing wires first, one by one, starting from B-. Result: lower 4 cells can not be balanced, so the chip was damaged. No other BMS is sensitive to that...

OK, so update on my "smart" BMSes. I originally bought two of them: in 12s and 8s configuration.
8s was damaged due to wrong connecting order (not specified in seller's website). 12s was working ok, however it's config crached during current calibration. I have managed to restore it, however internal temperature is still 116°C. I did set OverTemperature protection to 120°C (maximum possible), but if it heats just a few degrees above ambint, it switches off.

So i went on chat with BMSbattery. First they offered me free replacements with my next order. Since I was not going to send more money, they offered to send me free replacement if I payed ~$100 on shipping. No cheaper shipping solutions were available. That was riculously too much for me, so they said thei can't do anything. Then I reminded them that I still have 16 days left for PayPal dispute... They promised they will send replacements and asked to trust them I did'nt of course, but bus out of options... Magically, 2 days before dispute deadline, replacement BMSes arrived. One 8s LiFePO4 and one 12s LiPO.
So.. 8s BMS is operating fine, withstood all callibrations, etc. However with 12s it's the same story: crashed on me during callibration. Restored the config, but same problem with internal temperature...

So now I have four EcityPower BMSes: one operating, two with temperature problems but working otherwise, and one with failed 1-4 channels.


I confirm that all four BMSes stop balancing when OVC condition is detected, which is normal for every charge cycle. So this prevents the pack from being balanced.... That's a shame...

Anyone can confirm this? And ways for fixing this nonsense?

 

[hirnfurz]

Some more DIY stuff for the SmartBMS here: http://www.pedelecforum.de/forum/showthread.php?22585-DIY-SmartBMS-Reader-mit-Atmega

 

[helgeo]

I ordered the usb to i2c adapter from bmsbattery.com. Have just connected it to the bms and could immediately see why it cycles the charging at 2 second intervals. The charge over current limit was set to 10A while the actual charge current was around 11A. Changing the limit to 15A made it work immediately.

I can also confirm that bleeding in idle/over current is disabled by default, but can be enabled in EEPROM.