BMS modding for more current

Batteries, Chargers, and Battery Management Systems.
User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

BMS modding for more current

Post by ElectricGod » Aug 15 2018 1:44pm

I've posted most of this in another thread, but you have to dig for it. It seems that more than just me would want to see this done.
viewtopic.php?f=14&t=88676

A lot of BMS are decent quality, if they had just a few things fixed about them. I recently got into smart BMS's. These BMS are pretty good, but they have issues.

Buy direct on aliexpress from IC GOGOGO. If you look in their smart BMS section, they come in as low as 7 cells and as high as 32 cells.
https://www.aliexpress.com/store/group/ ... 13d07k6073

These 2 BMS I've modded quite a few of them now and they work great and can handle a lot more current as a result.

I don't care about anything lower than 16S. This is the 20S version. I'll talk about it a lot more.

https://www.aliexpress.com/store/produc ... 3de2mMrlYO

This is the big brother...more amperage, but they come in 20S, 24S or 30S.

https://www.aliexpress.com/store/produc ... 3de2mMrlYO
Last edited by ElectricGod on Aug 24 2018 3:14am, edited 2 times in total.

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

50 amp smart BMS modding

Post by ElectricGod » Aug 15 2018 4:08pm

This BMS is good up to 50 amps. The mosfets are garbage 10 amp parts with very high Rds. The copper busses are typically not soldered down. The copper is too short and as is don't do a very good job. The mosfets are generally NOT soldered down very well either.

This is the BMS as I received it.

Image
Image

I took the covers off one of them and found this immediately. The 3 solder pads for B-, P- and C- have no copper anywhere near them. This is useless. You WANT the copper reinforcement to overlap the solder pads so that you can solder directly to the copper. You want the copper to span the entire length of the power buss, not just cover a portion of it.

Image

Copper doesn't over lap the B- solder pad at all. This is critical for good current handling.

Image

Copper is short by 1/2" and doesn't over lap the P- solder pad. This is critical for good current handling since the EV is powered from this solder pad.

Image

Looking further, I found that on none of the 6 BMS of this version I have has the copper soldered down properly. On all of them the copper would have voids and gaps or lifted off sections. Once I pulled off the copper, I found entire sections of the solder traces with no solder on them connecting to copper and bare copper with no solder on it. I wish I had taken more pictures of this!

This first picture might be hard to see. Both ends and the middle of the copper are bulged upwards and NOT soldered down. There are 2 low places that actually make solder contact with the solder trace underneath. This is virtually useless!

Image

These images show the copper at the B- solder pad completely floating and doing zero good at all.

Image
Image

As soon as the copper was pulled off, I tinned every surface of the copper and tinned all the solder traces. Since the copper was too short to even cover the entire solder trace they are supposed to reinforce, I used the copper strip from C- to extend the P- and B- traces. C- is for charging only and that will probably be at like 10 amps instead of 50-80 amps like P- and B- will see. IE: Copper on C- is much less important.

These pictures show the copper extends the length of the B- buss and over laps the B- solder pad. Later I removed that little piece of overlapping copper so that I could solder down larger shunts.

Image
Image

In this picture , you can see the P- copper buss extends over all the mosfets and overlaps the P- solder pad. The 5.5mm bullet is soldered direct to copper. I later over lapped those last 2 mosfets with copper.

Image

The factory mosfets are HY3410's. They are rated for 17 amps and 105 mOhms. Pathetic in every way! I had quite a few IRFP4110's which are 120 amps and 3.7 mOhms. Not the best mosfet there is, but far superior to the HY3410 garbage. I don't replace the C- mosfets. At most they will see maybe 30 amps so who cares. The P- mosfets however can now handle easily 100 amps. The factory mosfets...most of them were barely soldered down and several were not soldered down at all.

Image


Once the factory mosfets were pulled up, I retinned all the solder pads that they mounted to. I tinned the backs of all the IRFP4110's so that good solder flow between the solder pads and mosfet backs would happen. I pre-tinned the 2 remaining legs. I didn't bother with SMD 4110's since I already had 40 of them in TO-220. I made my own from the TO-220 version. Bending the legs, shortening them and clipping off the center leg is quick and easy.

Image

Soldering them down starts with the 2 legs which I had pre-tinned. The solder pads were pre-tinned as well. Get the mosfets straight, heat the solder on the pad and the leg and it all flows together. Repeat for all the mosfets. Now go back to the first mosfet and with lots more heat, put solder on the tab and solder trace. After a little bit the solder on both items will start flowing together. Press down the mosfet so it lays flat. Notice the right most mosfet, the top of the tab is tinned and the hole is filled with solder. You will use this later when you add the P- copper.

Image

The factory shunts could have been doubled up or replaced with larger ones. I doubled them on my first BMS mod, but later replaced the .5 watt shunts with 2 watt shunts of the same resistance. If I had doubled up the smaller shunts, I would also need to change the value of resistance in the PC app to account for half as much shunt resistance. It is probable the BMS could not successfully measure current with such a low shunt resistance of .1 mOhms. It was better to use larger shunts and keep the total resistance at .2 mOhms like it comes from the factory.

.5 watt vs 2 watt shunts.

Image

.5 watt shunts doubled up on a BMS. Later I pulled them off for the reasons explained above.

Image

Since 2 watt shunts are huge by comparison, I can't put 10 of them in the same space. I still needed 10 shunts in parallel to get the same resistance, but only 8 fit across the space so 2 go on the bottom of the board across the same 2 solder traces.

Image

Image

The final step is to span the mosfet tabs with copper and to overlap onto the P- solder pad. Later I fixed this short missing section. Notice how the copper is thoroughly coated and embedded in solder. As I heated up the copper, I'd push it down so it was pressed onto the mosfet tabs.

Image
Last edited by ElectricGod on Aug 16 2018 11:10pm, edited 2 times in total.

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

80 to 300 amp BMS upgrade

Post by ElectricGod » Aug 15 2018 6:00pm

This BMS is the big brother to the above BMS. It uses the exact same application and the balancing circuitry is identical. The big differences are the smaller versions are 7S to 20S and support a maximum of 8 mosfets at P- and C-. The big brother starts at 20S and goes up to 30S and takes up to 20 mosfets for P- and 20 for C-. IE: LOTS of potential for more current handling. I'm sure it can now deliver at least 200 amps and probably more like 300 amps. I really wish it was designed for 32S.

https://www.aliexpress.com/store/produc ... 3de2HDm94c

This is what I got from the seller. It's about 7" x 5".

Image
Image

It's hard to see in the image, but it's about 2X larger than the smaller version.

Image

This is the top and bottom of the BMS. The copper at P- is floating for the most part and doesn't make it across the entire buss and doesn't overlap P- at all. The 2 B- busses on either side of the shunts are better, but they are the same length as the P- copper and the ends are filled in with solder, NOT copper. The copper is still floating on the B- solder traces, but at least some of the time it is soldered to the mosfets and shunts. Copper or solder to the B- terminals is obviously NOT there at all. When I pulled off all of the copper strips I found spotty solder connection to the circuit traces underneath them as well. Those "QC Pass" stickers are a bit laughable.

Image

Image

The solder used on this BMS is very low grade. It flows poorly, cools rough and clumpy and is rather brittle. I removed as much of the original solder as I could get off the BMS and then used good quality real 60/40 solder to re-tin all the solder pads and traces everywhere. IMHO, good solder is worth the cost! Like in the smaller BMS, All the copper strips got 100% coated in solder.

I pulled up all the .5 watt shunts and crappy mosfets on P-. All solder traces and pads got cleaned off and re-tinned with good solder. The mosfets on this BMS are HY3712's and are good for 250 volts, 9 amps and 450 mOhms...just awful! Since I intended to use it at 20S and I didn't have any 150 volt mosfets (AOT2500 or IRFP4115) on hand, I used IRFP4110's. 100 volt mosfets was good enough and MUCH better than the 3712's that came on the BMS!!!

I didn't care about C- having lots of copper. At most C- will see 30 amps so I didn't bother with copper there at all. I just added some solder onto the C- busses. That gave me extra copper for properly reinforcing B- and P-. I started at the B- busses and soldered down full length sections of copper across them. I made sure the copper was pressed down to the traces underneath them. I butted the copper right up to the 3 B- screw terminals and added a good bit of solder here too.

Next came shunts. THe BMS had 4 mOhm .5 watt shunts on it and there were 20 of them. I have 5 mOohm 2 watt shunts so that meant I needed 25 of them to get the same total resistance. I put 14 shunts on the top of the board and the other 11 on the bottom of the board. As the solder melted, I'd press the shunt down onto the copper. With all that copper and solder and shunts, it took 100% of the power from my Hakko solder station to flow the solder.

Image

Image

Next up came mounting mosfets. Like the smaller version, I started with mosfets with pre-tinned legs and backs and pre-tinned the solder pads where the mosfets mount with good quality solder. I used 20 IRFP4110 mosfets in P- and left C- alone with the factory mosfets. These shots were taken after I put down the copper over the mosfet tabs. The mosfets are 100% soldered down and the copper on both sides overlaps onto P- and runs the full length of the P- busses. It was sloooow going getting this much copper, solder and parts hot enough to flow properly!

Image

Image

The final BMS upgraded and tested. These pics show 5.5mm bullets on the BMS. I later pulled them and used 8mm bullets. I won't be using the screw terminals at all. I secured the 4 temp sensors with thermal glue to the mosfets.

Image

Image

The factory thermal gap filler was obviously too short for all those mosfets and it got torn too. I bought some more on ebay and covered all the mosfets.

Image

This is an important detail for these BMS. Mine is set up for 20S in its programming and because of these solder bridges at the balance connector. From left to right batt- to batt+ or BC0 to BC30. Removing the solder bridges in different combinations and changing a setting in the PC app and it becomes a 24S, 28S or 30S BMS. The smaller BMS work the same way.

Image

Image

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

12-20S dumb BMS upgrade

Post by ElectricGod » Aug 15 2018 6:30pm

These same techniques can be applied to any BMS. I did it to a dumb BMS a couple of years ago. It was originally rated for 80 amps, but I can't imagine how with almost no copper on the power buss. I estimate it can handle more like 160 amps now.

I forget what mosfet this is in TO-247, but they were pretty weak...something like 15 amps each with lots of Rds.

Image

Image

I pulled the factory mosfets and replaced them with IRFP4110's since I had so many of them. The factory shunts were fine so I left them alone.

Image

Not quite done with the copper reinforcing. This is 12 awg copper wire on all the power traces. I added more copper over to the shunts and to P- (red wire). From the factory, they had almost nothing and only partially on the one buss as seen above.

Image

The BMS from the factory was a good bit larger thanks to how the mosfets were mounted. This was a bit harder to work out, but I shaved about 250mm in width off the BMS and folded the mosfets over internally. I laid down several layers of Kapton over the parts that are near the mosfets to avoid any shorts and used taller stand-offs between the heat sink and board. It made the BMS 2mm thicker, same length and 250mm narrower.

Image

Image

User avatar
thundercamel   100 W

100 W
Posts: 216
Joined: Jul 23 2018 2:58pm
Location: Round Lake, Illinois

Re: BMS modding for more current

Post by thundercamel » Aug 20 2018 12:42pm

I just noticed you made this nice separate thread, so I suppose I'll pose the same question here. In practice, what are the downsides of using the chinese BMS units as is? I'm guessing the HY3215 mosfets create more heat because they have more resistance than AOT2500 mosfets?

As your amperage goes up, you risk blowing out a trace without soldering in some more copper properly?

I have two of the 14s 60a BMS units from ailexpress, one of them I use at 25a just fine, and the other I plan to use at ~40a. I suppose I could peek inside before I use it.
My Ebike builds - Existing bikes, affordable motor kits, self built 14s6p batteries - Now with more recumbent!

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

Re: BMS modding for more current

Post by ElectricGod » Aug 24 2018 3:28am

thundercamel wrote:
Aug 20 2018 12:42pm
I just noticed you made this nice separate thread, so I suppose I'll pose the same question here. In practice, what are the downsides of using the chinese BMS units as is? I'm guessing the HY3215 mosfets create more heat because they have more resistance than AOT2500 mosfets?

As your amperage goes up, you risk blowing out a trace without soldering in some more copper properly?

I have two of the 14s 60a BMS units from ailexpress, one of them I use at 25a just fine, and the other I plan to use at ~40a. I suppose I could peek inside before I use it.
I want optimal performance that is reliable. When I mod something those ideals are always in mind. I mod for those reasons. What you do and are willing to compromise on is totally up to you.

With the copper provided from the factory and then installed properly like I do, I'm not concerned about 100 amps blowing a trace out.

More resistance in anything generates more heat due to resistive losses taking away wattage that would otherwise pass on through without losses. So then low Rds mosfets are better than high Rds mosfets.

Duranza   1 µW

1 µW
Posts: 4
Joined: Sep 11 2018 5:38pm

Re: BMS modding for more current

Post by Duranza » Sep 11 2018 5:56pm

....
Last edited by Duranza on Sep 17 2018 9:05pm, edited 1 time in total.

User avatar
ctirad   10 W

10 W
Posts: 96
Joined: May 04 2011 4:50pm
Location: Prague, Czech Republic

Re: 80 to 300 amp BMS upgrade

Post by ctirad » Sep 15 2018 5:14am

ElectricGod wrote:
Aug 15 2018 6:00pm
The mosfets on this BMS are HY3712's and are good for 250 volts, 9 amps and 450 mOhms...just awful!
Are you sure? According to datasheet there http://boorzia.com/download/HY3712.pdf the HY3712 are 125V/170A with Rds(on) of 6.3mOhm. That sounds to me as an exellent choice, especially for the price.
Since I intended to use it at 20S and I didn't have any 150 volt mosfets (AOT2500 or IRFP4115) on hand, I used IRFP4110's. 100 volt mosfets was good enough and MUCH better than the 3712's that came on the BMS!!!
But there are 36 of the mosfets in my BMS (40 when I populate the missing 4), thats 200$ just for the mosfets alone. That is twice the price of the largest smart BMS available.
Member of http://www.cezeta.com team
Erider VK2008 4kW electric scooter
Guewer Classic 2kW electric scooter (modded)

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

Re: 80 to 300 amp BMS upgrade

Post by ElectricGod » Sep 28 2018 2:50pm

ctirad wrote:
Sep 15 2018 5:14am
ElectricGod wrote:
Aug 15 2018 6:00pm
The mosfets on this BMS are HY3712's and are good for 250 volts, 9 amps and 450 mOhms...just awful!
Are you sure? According to datasheet there http://boorzia.com/download/HY3712.pdf the HY3712 are 125V/170A with Rds(on) of 6.3mOhm. That sounds to me as an exellent choice, especially for the price.
Since I intended to use it at 20S and I didn't have any 150 volt mosfets (AOT2500 or IRFP4115) on hand, I used IRFP4110's. 100 volt mosfets was good enough and MUCH better than the 3712's that came on the BMS!!!
But there are 36 of the mosfets in my BMS (40 when I populate the missing 4), thats 200$ just for the mosfets alone. That is twice the price of the largest smart BMS available.
I have a datasheet for the HY3712, I'm quoting specs from it. I have seen this happen before where Chinese mosfets will have very different spec sheets for the same part...gotta wonder which is correct. Maybe yours is, maybe mine is. I have not bothered to actually test the RDS of the pulled mosfets or ran them at 20 amps to see if they self destruct. That would be proof!

Where are you getting mosfets from? Looking on any major electronics distributor, I never see $5 per mosfet even for top grade TO-220 parts.

Also, don't waste your time on the C- mosfets. You use them for charging ONLY. Who cares what crappy part they are. The important ones as seen above in all my BMS upgrades are the P- mosfets. I NEVER bother with the C- parts at all. That will cut your over priced mosfet cost in half.

User avatar
fechter   100 GW

100 GW
Posts: 14053
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: BMS modding for more current

Post by fechter » Sep 28 2018 3:30pm

Nice work.

The difference between a high power, expensive BMS and a much less expensive, lower current version is just the power stage. Everything else is the same, so it makes sense to get a cheaper one and upgrade the power stage.

Here's one I did a long time ago. Originally a 30A rated unit, I replaced the FETs with 4110s and stacked a second row. The second row won't double the rating, but will increase it by maybe 50%. Adding a heat sink helps a lot and the upgraded FETs also help.

After modding, I run this one up to 60A and have had no issues.
6 parallel 4110 arrangement.jpg
6 parallel 4110 arrangement.jpg (103.43 KiB) Viewed 2196 times
Double Stacked 4110s.jpg
Double Stacked 4110s.jpg (52.87 KiB) Viewed 2196 times
"One test is worth a thousand opinions"

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

Re: BMS modding for more current

Post by ElectricGod » Sep 28 2018 3:48pm

fechter wrote:
Sep 28 2018 3:30pm
Nice work.

The difference between a high power, expensive BMS and a much less expensive, lower current version is just the power stage. Everything else is the same, so it makes sense to get a cheaper one and upgrade the power stage.

Here's one I did a long time ago. Originally a 30A rated unit, I replaced the FETs with 4110s and stacked a second row. The second row won't double the rating, but will increase it by maybe 50%. Adding a heat sink helps a lot and the upgraded FETs also help.

After modding, I run this one up to 60A and have had no issues.

6 parallel 4110 arrangement.jpg

Double Stacked 4110s.jpg
It looks like the right most mosfet is possibly C- and why bother with 2 mosfets there.

User avatar
fechter   100 GW

100 GW
Posts: 14053
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: BMS modding for more current

Post by fechter » Sep 28 2018 7:31pm

ElectricGod wrote:
Sep 28 2018 3:48pm

It looks like the right most mosfet is possibly C- and why bother with 2 mosfets there.
Correct!
"One test is worth a thousand opinions"

User avatar
ctirad   10 W

10 W
Posts: 96
Joined: May 04 2011 4:50pm
Location: Prague, Czech Republic

Re: 80 to 300 amp BMS upgrade

Post by ctirad » Sep 30 2018 5:02am

ElectricGod wrote:
Sep 28 2018 2:50pm
I have a datasheet for the HY3712, I'm quoting specs from it.
Me too (see the link above). I'm pretty sure there are two completerly different HY3712 parts that only share the same codename.
And I think even the lowend chinese manufacturers are not complete idiots to use 250V 450mOhm mosftet for the ~100V/100A BMS.
Maybe yours is, maybe mine is. I have not bothered to actually test the RDS of the pulled mosfets or ran them at 20 amps to see if they self destruct. That would be proof!
Or just compare the manufacturer logo with the datasheet one. ;)
Where are you getting mosfets from? Looking on any major electronics distributor, I never see $5 per mosfet even for top grade TO-220 parts.
Mouser, TME, Farnell or R.S. sometimes.
Also, don't waste your time on the C- mosfets. You use them for charging ONLY.
I don't use separete C- input in my installation.
Member of http://www.cezeta.com team
Erider VK2008 4kW electric scooter
Guewer Classic 2kW electric scooter (modded)

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

Re: 80 to 300 amp BMS upgrade

Post by ElectricGod » Jan 14 2019 5:50pm

ctirad wrote:
Sep 30 2018 5:02am
ElectricGod wrote:
Sep 28 2018 2:50pm
I have a datasheet for the HY3712, I'm quoting specs from it.
Me too (see the link above). I'm pretty sure there are two completerly different HY3712 parts that only share the same codename.
And I think even the lowend chinese manufacturers are not complete idiots to use 250V 450mOhm mosftet for the ~100V/100A BMS.
Maybe yours is, maybe mine is. I have not bothered to actually test the RDS of the pulled mosfets or ran them at 20 amps to see if they self destruct. That would be proof!
Or just compare the manufacturer logo with the datasheet one. ;)
You do realize that China copies everything right? I don't trust logos to mean anything. That's just a bit of silk screening or lazer etching.

I was looking at them just last night. Some time back, I needed some cheap mosfets. I got on ebay and bought some 4110's. They arrived out of China. I later bought some legit 4110's from Mouser. The Chinese ones looked identical to the legit parts from Mouser. However, testing them, was very different. It was quite obvious that the Chinese parts were NOT the same thing. AKA...looks don't mean diddly!

Chinese...you are right, they are not complete idiots, but they can be inscrupulous! I've long ago gotten used to this fact. Order from China, expect whatever it is to be 60% what the specs say. China and LED's...OMG!!! Just flat out lies! Same for 18650 cells. Ever see a 9000mah 18650? Right me neither...at least not one that actually exists. BUT lots of Chinese sellers are pushing 1000mah cells and advertising them as 9000mah. I just don't believe what they say anymore. If they can get away with selling something inferior, they will!!! Most people have no ability to double check what they are sold, the Chinese are more than willing to lie, cheat and obfuscate to make a dollar!

If they can get a 3712 that costs 12 cents and has 450 mOhms Rds and it costs 10% of a 3712 with the real Rds, you can pretty well guarantee they WILL do it and NOT think twice about it.

I buy a lot of stuff direct from China. I can tell you this is the case probably 70% of the time.

Buy an LED light out of China today that is all aluminum construction and tomorrow they will replace the aluminum reflector with a plastic one and replace the glass lens with a plastic one the next time you buy. They are always looking for how they can cheapen their products and yet still "look" the same. I just bought 4 LED lights on Aliexpress. Last time I bought this exact light was over a year ago. I'm cringing inside. I know they did something to make them cost less to make!

They have no truth in advertising laws. They have no copyright laws. They have no product safety laws. Literally...buyer beware!

Quality control in China does not mean make a quality product and keep it good. It means make a product that looks like it's quality, but do it as cheaply as possible.

User avatar
MJSfoto1956   10 kW

10 kW
Posts: 612
Joined: Jul 28 2010 9:28pm
Location: Boston, MA

Re: 80 to 300 amp BMS upgrade

Post by MJSfoto1956 » Jan 14 2019 6:01pm

ElectricGod wrote:
Jan 14 2019 5:50pm
Quality control in China does not mean make a quality product and keep it good. It means make a product that looks like it's quality, but do it as cheaply as possible.
I too have purchased many dozens of items from AliExpress over the years. My experience is that 80% of what you purchase is pretty damn good. One of the key things I look for is how long in business they have been. I will pay more for a product from a seller that has 4+ yrs versus only 1 yr. Additionally, there used to be a difference between eBay, Amazon, and AliExpress. Now that so many products are ONLY available from China, you pay more for the exact same product on eBay or Amazon. The only difference would be if the seller has actual USA stock. Then you might actually have a chance at returning the item if found defective or lacking (caveat being: the price is often 3x higher).

SImple rule of thumb to follow with regard to AliExpress: avoid anyone who has only been in business 1 year. 4yrs is very good. 8yrs is outstanding.

M

kdog   10 kW

10 kW
Posts: 587
Joined: Feb 02 2014 9:22pm
Location: hobart, tasmania

Re: BMS modding for more current

Post by kdog » Jan 15 2019 7:42am

Hey eg thanks for this thread. I'm just at the beginning of looking to modify some bms's to make them more reliable. Not so much to increase amps but to better current sharing etc. I've had a few fail bc one or two mosfets basically do 90%of the work.
In fetchers pics above, are the mosfets stacked on top of each other with their respective legs soldered together? Can't quite make it out in the photo... Then the heat sink is sandwiched between the tabs.
What's to stop any bms having mosfets doubled up line this other than space/layout.

User avatar
MJSfoto1956   10 kW

10 kW
Posts: 612
Joined: Jul 28 2010 9:28pm
Location: Boston, MA

Re: 80 to 300 amp BMS upgrade

Post by MJSfoto1956 » Jan 15 2019 7:56am

ElectricGod wrote:
Aug 15 2018 6:00pm
Image
I had been meaning to ask: what is the engineering reason behind having three pads? Is it purely economical (i.e. they can use smaller gauge wire)? Or is there some intrinsic idea behind it like "equal distribution" or something like that?

Just curious.

M

User avatar
fechter   100 GW

100 GW
Posts: 14053
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: BMS modding for more current

Post by fechter » Jan 15 2019 9:06am

kdog wrote:
Jan 15 2019 7:42am
In fetchers pics above, are the mosfets stacked on top of each other with their respective legs soldered together? Can't quite make it out in the photo... Then the heat sink is sandwiched between the tabs.
What's to stop any bms having mosfets doubled up line this other than space/layout.
Yes, the respective legs are just soldered together and the heat sink is sandwiched between the tabs. Just having any kind of heat sink is an improvement over most of the designs I see (which have none). The one in the picture I run at 60A, which is the most the controller can take.

Just by stacking a second layer of FETs, the rating won't double, as the sharing is not really even, but it will be at least 1.5x more. Compared to no heat sink (stock), it will easily handle more than double.
"One test is worth a thousand opinions"

User avatar
ctirad   10 W

10 W
Posts: 96
Joined: May 04 2011 4:50pm
Location: Prague, Czech Republic

Re: 80 to 300 amp BMS upgrade

Post by ctirad » Jan 25 2019 4:09pm

You do realize that China copies everything right? I don't trust logos to mean anything. That's just a bit of silk screening or lazer etching.
I didn't talked about any copies. I just pointed out that for the part name "HY3712" there are two diffent mosfets components from a two different (chinese) manufacturers, that have completely different specs. According to the printed logo, the right one is the Hooyi HY3712, that is 120V, low RDS one.
Member of http://www.cezeta.com team
Erider VK2008 4kW electric scooter
Guewer Classic 2kW electric scooter (modded)

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

Re: 80 to 300 amp BMS upgrade

Post by ElectricGod » May 18 2019 2:43am

ctirad wrote:
Jan 25 2019 4:09pm
You do realize that China copies everything right? I don't trust logos to mean anything. That's just a bit of silk screening or lazer etching.
I didn't talked about any copies. I just pointed out that for the part name "HY3712" there are two diffent mosfets components from a two different (chinese) manufacturers, that have completely different specs. According to the printed logo, the right one is the Hooyi HY3712, that is 120V, low RDS one.
Do you know you are getting the good 3712 or the crappy one? The way to know is to actually measure RDS.

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

Re: 80 to 300 amp BMS upgrade

Post by ElectricGod » May 18 2019 2:46am

MJSfoto1956 wrote:
Jan 15 2019 7:56am
ElectricGod wrote:
Aug 15 2018 6:00pm
Image
I had been meaning to ask: what is the engineering reason behind having three pads? Is it purely economical (i.e. they can use smaller gauge wire)? Or is there some intrinsic idea behind it like "equal distribution" or something like that?

Just curious.

M
I wondered that too. Electrons travel just fine through one or 3 connections assuming the single connection is just as strong as the 3 connections.

User avatar
ctirad   10 W

10 W
Posts: 96
Joined: May 04 2011 4:50pm
Location: Prague, Czech Republic

Re: 80 to 300 amp BMS upgrade

Post by ctirad » May 24 2019 9:16am

Do you know you are getting the good 3712 or the crappy one?
Hooyi HY3712 IS chinese part. And I don't care if it has 30% worse Rdson than the top class part with several times higher price. The 100A BMS cost me 100$, used on 8kW 90V motorcycle and its heatsink is barely warm after a long ride.
The way to know is to actually measure RDS.
Of course, but that is not the point.
Member of http://www.cezeta.com team
Erider VK2008 4kW electric scooter
Guewer Classic 2kW electric scooter (modded)

User avatar
ElectricGod   100 MW

100 MW
Posts: 2663
Joined: Nov 01 2015 5:24pm
Location: 3 OH 3

BMS modding for more current

Post by ElectricGod » Jul 29 2019 10:46pm

These chinese BMS got redesigned a bit. I'll call this v2.0. They clearly looked at the bigger BMS thread and took to heart what we posted.

The board is a bit more compact than before and they added an aluminum cover to the back side instead of the clear plastic sheet.

Image

Image

Image

I'm glad to see this is fixed. The copper reinforcing is properly bonded to the traces. No lifted copper!

Image

The copper overlaps the BATT- pad too! Of course the shunt resistors don't solder up to the copper...so that still needs to be improved.

Image

This is the 20 amp version...which is cheaper since it lacks all the craptastic chinese mosfets. I'll be replacing them anyway...so who cares. Unfortunately P- is still not reinforced. That large middle silver trace is P-. Maybe if this was the 50 amp version there would be copper here?

Image

Image

User avatar
fechter   100 GW

100 GW
Posts: 14053
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: BMS modding for more current

Post by fechter » Jul 29 2019 11:24pm

I haven't seen many that have copper added like that. I usually have to do it myself. Nice to see some improvement in design.
I'm using a similar but smaller one on my Sur-ron pack that seems to be working OK so far. Runs at about 70A.
"One test is worth a thousand opinions"

User avatar
agniusm   10 MW

10 MW
Posts: 2358
Joined: Apr 16 2011 3:06pm
Location: Lithuania, Zarasai
Contact:

Re: BMS modding for more current

Post by agniusm » Jul 29 2019 11:38pm

Can someone explain to me, whats the point of this? I buy small bms'es, fit 10 usd relay, positive from the pack, negative from bms p- to drive coil and i switch controller ignotion to positive.
Discharging through bms mosfets is just another waistfull thing that subtracts from overal efficiency.

Post Reply