bluetooth BMS?

The 120A verson of the bms i habe here has like 30 mosfets on them. But it simply makes me sleep better if i know the power stage is good. Virtually all power stages of any bms discusses here has utter crap mosfets. If you can simply replace them for 30 bucks, use half the amount of mosfets and make the spare current handling capacity and reliabillty 10x better it should still be considerd a worthwile upgrade. It would also massivly reduce the need for cooling the board and make the setup more efficient.

link to the bms in question:

https://www.aliexpress.com/item/16S-to-32S-2018-new-DIY-Lifepo4-li-ion-50A-80A-100A-110A-120A-smart-bms/32847205125.html

does look the part, nice beefy bus bars on the power board. only the crappy mosfets wich i am scared i will blow their areses out 20 miles from home.
 
I finally received and wired up my 16s 60a Bluetooth BMS from here: https://m.alibaba.com/product/60619821174/Wholesale-bluetooth-bms-RS485-60v-60a.html

Unfortunately I am unable to connect it to my phone. I can see it's Bluetooth name as 'minew' but it refuses to connect to it. My phone is Oreo, but I have tried am old lollipop version phone with the same result.

Also the pc program can read the settings and individual cell voltages but it refuses to write/adjust any of the settings.

It's so frustrating as I can't get any help from the seller due to Chinese New year.
 
shorza said:
Also, can someone upload a recent version of the android app and pc program please?
i'm not sure it's the app's version, but maybe the bluetooth stack used on oreo maybe?
i even can't SEE the bms! also tried different BT scanner apps, but they can't see it, whereas my "old" android has no problems connecting to it.
 
anyone got any information about STANDBY drain? i realized that the 12v rail for the BT module is "always on". so the processor and other stuff is "on" as well. i read that standby current should be around 0.2mA w/o BT connected. but at what voltage?
voltage of the pack? can't be. those BMS range from a few cells to 30 cells in series. if they all had 0.2mA (multiplied by voltage) the wattage would differ a lot. if it was 0.2mA of the whole pack it would drain the whole 30Ah pack in 30.000 * 5h = 150.000h = 17 years. i could live with that ;)

but how could i measure the voltage drain? the BMS itself is powered by the balance leads, and i guess it takes the PACK's voltage to power it's logic board, as the cells drain evenly. i could just connect the first and last balance wire and see what happens. that way i could measure the current going over one of the wires. would this fry the board? it should NOT do any harm, but who knows?!
 
flippy said:
ElectricGod said:
The mosfets are HY3410's which are really low grade...100 mOhm, 17 amp...meh. I'll be replacing them with something better.
With what mosfet would you replace them? I got 3 bms here with shitty mosfets.

Hi,

I have loads of IRF4110's that I pull from controllers and replace with AOT290's or CSD19536KCS. I upgraded both of my BMS with IRF4110's. It was easy to cut off the tops of the mosfets so they were "SMT" and then bend the legs too. I snipped off the drain leg since these BMS use the tab for this connection. I've also ordered more .005 ohm shunts to solder down next to the existing ones. They are 1 watt each. I'll see if they are enough already. With there being 10 .005 ohm shunts in parallel, they see very little wattage absorption. I haven't measured the voltage across the shunts to be sure, but I bet the existing shunts are enough for 100 amps like the BMS ought to handle now. Worst case the BMS has 4 temp sensors, put one on top of the shunts and see if they heat up or not.The BMS can be set for more amperage...or at least both of mine let me change the default 50 amps to whatever I wanted. Adding a few more shunts will "fake out" the BMS since it will measure a lower voltage drop and therefore less current. Doubling the number of shunts will make the BMS "think" it is measuring 1 amp when actually 2 amps is flowing. It's not like the 7 IRF4110's in full saturation can't handle it. I didn't swap out the charge side mosfets (top row) since I don't use them. I connect to P- not C- for everything and I have very little use for adding a separate circuit for charging anyway.

16-20S%20smart%20BMS%203.jpg


I didn't take any pictures of it, but something I discovered when pulling up the copper reinforcing is that it is poorly soldered down. That strip of copper between the mosfets and after the shunts was soldered at either end, but nothing in the middle on BOTH of my BMS. Before I put down the copper again, I tinned the entire copper strip on all sides and then made sure it actually soldered down. It takes a lot of heat, but I have a Hakko solder station and it can deliver the heat needed to make this happen. If you only have a 25 or 35 watt soldering iron, you may not be able to get the HY3410's off the board. Even the mosfets were not soldered down very well. They had large gaps underneath most of them where solder never got into. At least half of the solder pads for the mosfets were making no actual electrical contact with the mosfet except at the end of the tab. I have a couple more of these BMS coming. If I remember, I'll take pictures of the poor soldering under the copper and under the mosfets.
 
izeman said:
anyone got any information about STANDBY drain? i realized that the 12v rail for the BT module is "always on". so the processor and other stuff is "on" as well. i read that standby current should be around 0.2mA w/o BT connected. but at what voltage?
voltage of the pack? can't be. those BMS range from a few cells to 30 cells in series. if they all had 0.2mA (multiplied by voltage) the wattage would differ a lot. if it was 0.2mA of the whole pack it would drain the whole 30Ah pack in 30.000 * 5h = 150.000h = 17 years. i could live with that ;)

but how could i measure the voltage drain? the BMS itself is powered by the balance leads, and i guess it takes the PACK's voltage to power it's logic board, as the cells drain evenly. i could just connect the first and last balance wire and see what happens. that way i could measure the current going over one of the wires. would this fry the board? it should NOT do any harm, but who knows?!

I think you might be making a few wrong assumptions. Since this BMS runs at 16S or 20S, that's realistically 66 volts or 82 volts. I think you are going to see many years of in place use before the BMS drains the pack.

The balance connections make up a voltage divider. I think you need to connect all the elements of the divider or else you will blow up the BMS. I haven't looked up any of the parts on the BMS, but you have 2 charge controllers, a CPU and a BT board. What do the spec sheets say they consume?
 
Poking around a bit more...

They use an atmega 328 for the CPU. There's no DC-DC converter on the BMS that I can identify. These 3 CZT5551 NPN transistors and 3 more on the other side of the board all look like they are in parallel. Perhaps they make up the power component for a regulator somehow? I see no obvious inductor for a DC-DC converter and there's nothing else here that gives itself away as a high voltage linear regulator. The 328 needs 5 volts so that's more or less 2 cells in series for a low voltage regulator. The CZT5551 is a 180 volt NPN transistor. As you can see in both of these pictures, there's a test point labeled VREG. There has to be some kind of regulator here that powers the CPU and BT board.

2018-02-19%2010_42_27-16-20S%20smart%20BMS%20regulator%20close-up.png


On the left you have the Atmel 328, then some 4 legged square component and then far right 3 CZT5551 transistors. It's an odd place to put it, but is that square box the inductor for the DC-DC converter? DC-DC converters usually have a few large caps off their outputs to clean up any ripple. The only place I see anything relevant is just above and a little to the left of the 3 transistors. Maybe that smaller 3 legged part is a high voltage linear regulator and it's output current is bolstered by the 6 NPN's? If this isn't a high voltage regulator circuit that is powered from the entire pack, then that means the CPU and BT modules are powered from the lower few cells which would be bad.

2018-02-19%2010_51_57-16-20S%20smart%20BMS%20board%20close-up.png
 
which assumptions are wrong? i don't see you say anything different than i do.
my BMS is 15s so around 50-60V. and something has to power the 12V for the BT module, and i guess the atmel is always on as well (i should be in sleep mode though). i will need to measure that.

still i see some voltage drain over 24h. i'll do some long-term test now. user merlin stated, that his bms (the very same type we two have here - different cell numbers, but otherwise it's the same i guess) drains the pack by 100mV over 2 weeks time.
this would mean that a fully charged pack is empty quite quickly.
 
shorza said:
I did try my old phone which is on lollipop. Same result :(


Hi folks,

The latest software can be found here.
https://www.lithiumbatterypcb.com/smart-bms-software-download/

I'm running Lineage OS 14.1/Android 7.1.2 and the app works fine on my phone.

The phone app is a bit odd at first until you can get GPS turned off. It flat out won't run if GPS is not turned on and it can't access your location. It will put you through an endless loop of demanding GPS access. So initially turn on GPS, start the app and then in the App Setting page, you can turn off GPS. I have no need for my BMS app knowing where I am located or what my speed is. The dashboard is pretty useless IMHO. I'd rather use strava or a real GPS app for a bike computer.

16-20S%20smart%20BMS%206.png
 
izeman said:
which assumptions are wrong? i don't see you say anything different than i do.
my BMS is 15s so around 50-60V. and something has to power the 12V for the BT module, and i guess the atmel is always on as well (i should be in sleep mode though). i will need to measure that.

still i see some voltage drain over 24h. i'll do some long-term test now. user merlin stated, that his bms (the very same type we two have here - different cell numbers, but otherwise it's the same i guess) drains the pack by 100mV over 2 weeks time.
this would mean that a fully charged pack is empty quite quickly.

I think the electronics are powered by the entire pack voltage and you said that can't be the case. I think those 6 transistors and some other components make up a linear regulator that runs at pack voltage, but I'm not 100% sure...yet. A regulator would operate at a wide range of input voltages and still output 5 volts.

Last night, I permanently connected a BMS to a pack. It was at 65.4 volts last night. It's made of 4 4S Graphene 10,000 LIPO packs. I'll check it later and report if I'm seeing any voltage drop.

Which version of this BMS do you have? What was your starting voltage? How long has it sat that you see a voltage drop? Any other loads on the battery pack besides the BMS?
 
flippy said:
i am looking for a mosfet that can do more then 75v. my next battery is ~115v charged.
and takers for a solid mosfet choice for these voltages?

Best mosfets for anything...

100 volts:
This is the best 100 volt mosfet existing...CSD19536KCS
This is the second best by a close margin...AOT290

120 volts:
IXFP180N10T2

150 volts:
AOT2500 destroys the IRF4115 by a good margin!
 
ElectricGod said:
The balance connections make up a voltage divider. I think you need to connect all the elements of the divider or else you will blow up the BMS.
That's not the case, you can connect them in any order so long as you connect the correct tap to the correct cell input port.
If you connect a higher voltage cell (say cell 8 at 30V) to a lower input position (say cell 4 - which expects about 15V), you will burn out the 150ohm precision resistors that form the voltage divider on the inputs.

izeman said:
this would mean that a fully charged pack is empty quite quickly.
Check out this video where he measures the standby current. The BMS goes into sleep mode at which it only draws 0.5mA, it jumps up to 5mA periodically to scan for activity on the UART. There are some errors in other parts of the video however.
http://www.youtube.com/watch?v=ClGmMY0kXvQ&t=11m44s

The minimum input voltage appears to be above 20V.
I have one version that will power up once the ground wire and cell 7 are connected (about 23.5V input with my battery at current stage of charge). However I have a later one that needs Cell 12 connected to stay powered up. There is over a year between production dates for the two boards however.
 
izeman said:
YES!!! i works! :)

just connect a latching switch to K1, enable SWITCH in eeprom (windows software) and voilá! you can switch the BMS on/off with it.
if you DISABLE the SWITCH checkmark in the software it will be ALWAYS ON.

edit: it's a bit strange: i can easily connect with my redmi note 3 android, and i could connect with my mi mix2 an nougat firmware, but it's not working with oreo. any ideas why this could happen?

Where is this K1? I have the 16-20S BMS and I don't see it anywhere. Can you post some pic's please?
 
flippy said:
izeman said:
first to mind would be IRFB4115 capable of 150V, but resistance is very high (~10mOhm)
would the IPP048N12N3G be a better alternative? (4.8mOhm)
https://www.infineon.com/cms/en/product/power/mosfet/20v-300v-n-channel-power-mosfet/120v-300v-n-channel-power-mosfet/ipp048n12n3-g/

i still plan on putting like a dozen on the board to replace the chinesium mosfets to get decent current capabillty.

Ignore the charge side mosfets. You only care about the discharge side mosfets. I wouldn't bother with mosfets that are rated higher than 100 volts. The BMS can't handle more than that anyway if yours is 20S or less. The IRF4110 or AOT290 will work well for this BMS.
 
ElectricGod said:
Hi folks,

The latest software can be found here.
https://www.lithiumbatterypcb.com/smart-bms-software-download/
....
Thanks ElectricGod. That is the same version of the app that I am using. I will try turning GPS off and on. Otherwise I'll have to wait for the seller to come back online.
 
ElectricGod said:
Where is this K1? I have the 16-20S BMS and I don't see it anywhere. Can you post some pic's please?
It may not be present on your board, see below pic posted by izeman which shows it on his board variant.
izeman said:
 
Fionn said:
ElectricGod said:
The balance connections make up a voltage divider. I think you need to connect all the elements of the divider or else you will blow up the BMS.
That's not the case, you can connect them in any order so long as you connect the correct tap to the correct cell input port.
If you connect a higher voltage cell (say cell 8 at 30V) to a lower input position (say cell 4 - which expects about 15V), you will burn out the 150ohm precision resistors that form the voltage divider on the inputs.

Are you saying I can connect cells 3, 11 and 20 and NO other cells and the BMS will survive? That seems unlikely. Each channel on the board is part of a much larger voltage divider. No one channel "sees" more than it's small portion of the whole pack voltage (the voltage of the one cell it monitors). Connecting random cells and leaving others disconnected would mean that the voltage at those connected channels could be significantly higher than that of a single cell.
 
Fionn said:
Check out this video where he measures the standby current. The BMS goes into sleep mode at which it only draws 0.5mA, it jumps up to 5mA periodically to scan for activity on the UART. There are some errors in other parts of the video however.
http://www.youtube.com/watch?v=ClGmMY0kXvQ&t=11m44s

I thought I had a lot of scrounged laptop cells. He's definitely been at it a while longer than me!

2018-02-19%2012_25_30-25%20Smart%20BMS%20review%20-%20Cheap%20choice%20for%20those%20with%20low%20budget%20-%20YouTube.png


Mine...

Laptop%20cells_zpst3j3hilp.jpg
 
I've only tried it with a 12S pack but yes it didn't suffer any ill effects from connecting and disconnecting cells in a random fashion so long as they were connected to the correct port.
 
Fionn said:
I've only tried it with a 12S pack but yes it didn't suffer any ill effects from connecting and disconnecting cells in a random fashion so long as they were connected to the correct port.

Well cool...teach an old dog a new trick! I'm always careful to plug in from lowest to highest or to disconnect from highest to lowest.
 
Izeman...

I have the 16-20S version and there's no K1 connector. Can you figure out where that goes to? Maybe it's just to the gates of the mosfets or to some control pin on a charge controller?

K1%20connector.jpg
 
Yes, and I think the instructions warn to do the same. However I inadvertently smoked two sets of resistors when I made a mistake so I said I'd take a few chances as I thought the board was toast anyway. Luckily for me I was able to move 4 unused ones from the top channel down to replace the ones I destroyed. Despite my ham fisted & ill equipped SMT rework effort it worked fine again afterwards.
 
Back
Top