Fast Charging a Ping

[themadhatter106]

So...charging a 20Ah battery with a 5A charger is great but who wants to wait over 4 hours for a full charge when they are traveling far distances or touring? PING's batteries are basically 1C charge and 1C discharge. The main problem with charging fast is the BMS. Does anyone know of a LiFePo4 BMS which can take 20A charge current?

I saw this one:
http://www.bmsbattery.com/bmspcm/323-17s26s-24a-max-discharge-current-bms.html

not sure of its quality though.


I'll bet I could push a bit more current though the PING bms if I added a heat sink and a fan to the charging FET. Wouldn't come anywhere near 20A though.


Do you think another option would be charging at 20A bypassing the BMS and then when the charge current drops to 5A or so I switch it over to go through the BMS for balancing?


I am considering this charger:

http://www.bmsbattery.com/alloy-she...ifepo4li-ionlead-acid-battery-ev-charger.html

Any comments?

It has a nice review here: http://endless-sphere.com/forums/viewtopic.php?f=4&t=33405

and is much cheaper than a Meanwell.

 

[999zip999]

Go for 6amp. and let us know. Meaning it the Stock bms wouldn't like it. I would just ask Ping. The 20ah size is more in your favor.

 

[johnrobholmes]

A less crappy charge fet would help a lot. Otherwise you will have to bypass it or build a comparator that controls your current path based on amp or voltage.

 

[themadhatter106]

I know PING sells a high rate BMS which supports charging rates up to 10A and he also said that if you put a heat sink on the charging MOSFETS it can support 15A. All of the MOSFETS for charge and discharge are the same...they are all HY1807.



On the 40A board there are 4 MOSFETS for discharge and 1 for charge. That means that for discharge each can take an average of 10A. If this is true why is the max charge current only 5A? The MOSFET should be able to take 10A. R_ds is 7.5 mΩ which sucks. A better MOSFET would help or even adding another charge MOSFET but I wonder why the board can't take 10A of charging current in its current state.

 

[dnmun]

if you wanna go 10A and it burns up the mosfet then just put two mosfets back and bolt a copper tab on for a heatsink between them.

 

[themadhatter106]

I think I want to go straight for 20A. I will probably need to beef up the trace from C- to the source of the MOSFET. Any recommendations for a good couple of charge MOSFETS for a replacement?

 

[themadhatter106]

If I replaced the crappy FET with an IRFB4110 and added another one on the backside of the board in parallel do you think that would work?

 

[dnmun]

if you have a piece of copper to solder onto the tab as a heat sink and connection to the other drains then you might be able to push 15A through just one of the 4110, but the current mosfet may do that also. they overheat when the voltage on the gate drops too low and the transistor is only partially turned on so it gets really hot. this is a problem the signalab has when the pack voltage drops and the voltage of the circuit current sags down

 

[wineboyrider]

Using copper heat sinks is exactly what Mr. Ping told me to do when I charged my 36v 20ah battery at 10 amps! Use copper 1/4 copper tubing like what we use for swamp coolers and make sure the fets can breathe the heat!

 

[dnmun]

if you have a shop in your town where they make custom gutters or do sheet metal work for home furnaces they will likely have small pieces of scrap copper they have cut off of a larger piece while making something. go to their shop and ask if they will give you a small flat piece, 22 gauge thick would be perfect. size of a quarter is more than enuff and you don't even have to solder it down onto the BMS pcb and can allow it to stand up and then run a 12 AWG solid copper wire from the copper soldered to the tab down to the tabs of the other mosfets next to it that are the output mosfets which is how the juice gets to the battery from the charging mosfet.

 

[999zip999]

Those cells don't like to be charghed at to high a rate for long life. As don't boil the cell chargin.
Plus not enough info.
12v20ah or 60v20ah pack ?

 

[themadhatter106]

For the record its a 16s (48V) Ping pack 20Ah.

Most of the time I would charge it slowly but for those occasions when you need it fast charging is nice. I haven't read of anyone doing this before. I'm surprised.

 

[999zip999]

So what is Ping's 10 fet 16s 60a. signalab bms rated at for charging ? I used one at 5a. and have one in a draw. Don't like 3.7-3.8v for it to kick on to drain. 3.6V is what I like for lifepo4, but don't have a pack for it.

 

[themadhatter106]

Its rated at 10A but 15A is possible if you add a heat sink.

 

[999zip999]

So what does Ping say ? He would say 5a. Maybe 6a.

 

[dnmun]

the high current v2.5 signalab has two mosfets. i think they may be T470 from trying to read the ground off label

 

[wineboyrider]

So what does Ping say ? He would say 5a. Maybe 6a.

That's what he recommends, but I did tell him that I wanted to opportunity charge at 10 amps he said it will degrade the cycle life of the battery, but it is not dangerous.

 

[999zip999]

Thanks winerboyrider. You said what I wanted to hear and that always sounds good to me.

 

[themadhatter106]

Well I replaced the crap charging MOSFET with two IFRB4110 MOSFETS and I added a nice heat sink by soldering on a piece of copper pipe.



It was a good thing too, I discovered that the original charging MOSFET on there was shorted out. Ugh...I've had my BMS kill a cell group by overcharging before. I hate it when components fail on these things. Now my BMS is working hard to balance my pack. I had reduced battery capacity before thinking that I had a weak cell group. I will need to test it again after it is newly balanced.

I'm just waiting for BMS Battery to ship my charger. I ordered it about a week ago so knowing BMS Battery it will probably be another few weeks until I can get the charger and test out the mod. I will need to take some temperature readings to see if my heat sink is effective enough.

 

[dnmun]

i thought this was a new BMS you were hacking. did you try pushing hi current through it already that may have burned the mosfet?

 

[themadhatter106]

No, up until now I have only had a 5A and 2A charger. I also had another BMS board which had the same thing happen to it. For some reason the charge MOSFET randomly gets shorted with no signs of heat damage.

The bad thing about it is that it is impossible to tell that you have a shorted charge MOSFET until either your battery gets overcharged or you test it with an ohmmeter.

 

[dnmun]

it is easy to find a shorted mosfet because they get really hot. a functional mosfet doesn't really make heat, only when it is shorted.

 

[themadhatter106]

Hmm...when I was thinking about using a heat sink I decided to go down to my bike while it was charging and feel the MOSFET to see how hot it was getting. This is when I still unknowingly had the shorted MOSFET. It was only a little warm to the touch with my 5A charger. All three terminals of the MOSFET were shorted together though so the gate voltage was 0V yet it was charging which throws up a few red flags.

On another note I measured the gate voltage to be about 5V when the charge MOSFET is on which is much less than the discharge ones which if I recall correctly were at about 13V.

 

[dnmun]

if the gate was 0V then the comparators may have turned off the mosfet. you can test a mosfet with the diode tester to verify if it is functional. it should show the body diode from source to drain and then be open circuit from drain to source. for an nchannel mosfet.

 

[themadhatter106]

Hmm...well something is wrong. I just blew another stock charging MOSFET on one of my unmodded boards which was on the battery for less than a week. I came down to check on my bike charging and I noticed a bad smell. The plastic cover over the BMS was melty near the charge MOSFET, the PCB was slightly warped under the MOSFET and a multimeter check confirmed a shorted MOSFET. I wonder why this is happening. Could my charger be messed up? For now I will see how an ifrb4110 fares as a replacement.