Has anyone built a battery charger from scratch?

[auraslip]

With all the trouble we've been going through to find the correct meanwells for bulk charging, and with the limited options we have it seems like the next step. Why spend hours of time and $$$ trying to order a product of questionable quality and utility from half way around the world, when you could build something that will fit your needs perfectly and be of unquestionable quality.

I personally don't need a high power unit, but rather a small 2a-5a unit capable of 18s-24s voltages. That should be easy right? Just a tiny, mobile trickle charger. There is probably a million ways to go about this. Switching units. High voltage linear regulators. Linear regulators in series. Regulators made from discrete components. Really, what ever will be simple, cheap, and reliable.

 

[neptronix]

OK, sorry to derail but.. have you considered the kingpower units?

 

[Philistine]

Ausraslip I would love to have a variable charger, which could charge from 10s through to 24S at say 10 or 15amps. That would be awesome, that is why I have been watching what you guys have been doing with the Server PSUs, but to be honest I barely understand half the stuff you guys are talking about, I would definately buy a boxed unit, with display, showing voltage and current, that could be adjusted for different battery sizes within that range. I realise I could buy a lab supply, but that is thousands, and I realise I could build it myself from a big charger with a pot, but that is the point, I would much rather pay someone else to do it and know it isn't going to kill me or burn my house down.

I was just thinking the other day that of all the things this forum has homebrewed really well, charging solutions are somewhat lacking (for big pack charging).

 

[gensem]

OK, sorry to derail but.. have you considered the kingpower units?

Its KingPan, not kingpower, and from what I understand you can regulate the voltage from outside on the units with the display that are 900w or above.

http://www.kingpan.cn/en/Product-23.html

 

[SamTexas]

With all the trouble we've been going through to find the correct meanwells for bulk charging, and with the limited options we have it seems like the next step. Why spend hours of time and $$$ trying to order a product of questionable quality and utility from half way around the world, when you could build something that will fit your needs perfectly and be of unquestionable quality.

I personally don't need a high power unit, but rather a small 2a-5a unit capable of 18s-24s voltages. That should be easy right? Just a tiny, mobile trickle charger. There is probably a million ways to go about this. Switching units. High voltage linear regulators. Linear regulators in series. Regulators made from discrete components. Really, what ever will be simple, cheap, and reliable.

I have built a simple and worry free charger that has no need for balancing. It consists of 14 individual cell chargers for my 14s pack, but it can easily be changed to any configuration from 1s to infinity. Each individual charger is a Sony AC-LM5 adapter, 4.20V - 1.5A.

The main task was to reduce the number of wires from 28 down to 15, and to wire them into an easy to use connector. I used the ubiquitous 25pin parallel cable for connector (female on the battery side, male on the charger side.) I also added 2 fans (taken from PC power supply) to keep the 14 adapters cool.

For rapid bulk charging, I built a 6.5A charger by serializing 3 HP/Compaq PPP017H adapters [18.5V (specs), 19.0V actual].

 

[Philistine]

I have built a simple and worry free charger that has no need for balancing. It consists of 14 individual cell chargers for my 14s pack, but it can easily be changed to any configuration from 1s to infinity. Each individual charger is a Sony AC-LM5 adapter, 4.20V - 1.5A.

That is genius, so each cell gets its own 4.2v charger? So each cell reaches charged state individually? Would be awesome if you could post some photos or a tutorial/thread on your setup. I would be interested even to see what it looks like all put together.

 

[auraslip]

I was just thinking the other day that of all the things this forum has homebrewed really well, charging solutions are somewhat lacking (for big pack charging).

My thoughts exactly.

I have built a simple and worry free charger that has no need for balancing. It consists of 14 individual cell chargers for my 14s pack, but it can easily be changed to any configuration from 1s to infinity.

I actually have been looking into this. It's been done with life chargers from voltphreaks. The problem is the cost. One thing I think is worth looking at is using $3 ebay phone chargers and modifying them to 4.15v. I priced out building them last night, but the price came out to $10 each! The main price is in the transformer which is required for isolation. If there was a cheap way to isolate power supplies, this would be super easy.

About king power chargers: they're fine. Just fine. Not exceptional quality, and for the price and lack of customer support and warranty support.... I should be able to build a charger for less than $20.

 

[Philistine]

Its KingPan, not kingpower, and from what I understand you can regulate the voltage from outside on the units with the display that are 900w or above.

http://www.kingpan.cn/en/Product-23.html

Thanks for posting that Gensem, because people had always called them Kingpower I was always looking at the wrong charger. I looked at the one you linked to, if you can adjust the voltage on the front panel to whatever voltage you like, then that is exactly what I am after (probably ideally I would like a 1200W model). I have emailed them to get some details. Thanks.

 

[SamTexas]

I have built a simple and worry free charger that has no need for balancing. It consists of 14 individual cell chargers for my 14s pack, but it can easily be changed to any configuration from 1s to infinity. Each individual charger is a Sony AC-LM5 adapter, 4.20V - 1.5A.

That is genius, so each cell gets its own 4.2v charger? So each cell reaches charged state individually? Would be awesome if you could post some photos or a tutorial/thread on your setup. I would be interested even to see what it looks like all put together.

Thanks. Yes, each cell reaches full charge individually. I have told myself many times to make it pretty but I'm too lazy for that. It's still somewhat messy right now but fully functional. I will post pictures later.

There isn't much for a tutorial thread. Here's a quick overview for a 2s charger:

2x Sony AC-LM5 adapters. Each adapter has 2 output wires, 1 positive and 1 negative. Connect the positive wire of the first adapter to the negative wire of second adapter. You are now left with 3 wires. Connect those 3 wires to a 3 pin male connector. Do the same on the battery side with a 3 pin female connector and you are done.

For each additional cell, you need one more wire. So 15 wires for a 14s charger. Let me know if you have additional questions.

 

[Toorbough ULL-Zeveigh]

-- THE SINGLE CELL CHARGER POST --

 

[auraslip]

I asked for help building a switching power supply over in one of the reddit electronics sub-reddits, and now I've spent an hour explaining that it's perfectly safe to charge lipo with a power supply. Ack... "lithium is scary!!"

 

[amberwolf]

Some schematics for "lab power supplies" that could probably be adapted to your needs:
http://www.google.com/search?q=lab%20power%20supply&um=1&hl=en&tbm=isch&source=og&sa=N&tab=wi


Regarding the celphone chargers as single-cell units, just make sure that they are current-limited (or add this to them), or else you may damage them when charging up cells (or cell groups) from a low state. If they aren't limited, they'll end up trying to output more power than their design can take, and either die, or age quickly. Possibly dramatically, at some point.


The only charger I actually built from scratch was for SLA, just a typical trickle charger, built from parts from salvaged equipment with a regular transformer, various diodes, transistors, etc. Was a couple decades ago, nearly, but I think it did about 1/4A max at about 14-15V, IIRC, and was current-limited. Since it was a linear supply, it made a lot fo waste heat and was way bigger than it had to be--probably the size of two paperback copies of Dune stacked on top of each other. I don't ahve it anymore, though.

 

[Philistine]

Thanks Toorbough, I wasn't aware of that thread at all, much appreciated.

On that thread EMF gives a really helpful diagram, which might help any dummies like me reading this trying to picture SamTexas's explanation. This was how I understood what Samtexas was saying, but seeing the diagram confirmed it.

by EMF » Mon Nov 05, 2007 2:10 am



GGoodrum wrote: Each charger is completly isolated from the others, so it doesn't hurt to basically wire them in series, just like the pack. The - side of the 2nd charger is connected to the + side of the first, and the - side of the nth charger is connected to the + side of the nth-1 charger. For 10 chargers it takes 11 wires. One each for the - of the 1st cell, one for the + of the 10th cell and nine more for all the junctions between all the rest.

-- Gary


Took a while for this to sink into my thick head. Hopefully, I understand. Is this what you are saying?

 

[SamTexas]

Regarding the celphone chargers as single-cell units, just make sure that they are current-limited (or add this to them), or else you may damage them when charging up cells (or cell groups) from a low state. If they aren't limited, they'll end up trying to output more power than their design can take, and either die, or age quickly. Possibly dramatically, at some point.

Very good point amberwolf. How do you ensure that a charger is current limited? Here's how I tested and verified mine: I charged a very large empty (3.6v) LiCo cell pack. The cell pack capacity was 20Ah, which was 13 times higher than the max current of the charger (1.5A). At the beginning of the charge, I constantly measured the current going into the battery and made sure it never exceeded the max current.

Is there another way to test for current limiting?

 

[bearing]

I would measure the waveform of the current to see if it is a normal continous DC/triangular signal, or if the charger is in "hiccup"-mode, that is, charging for a short time (maybe a few milliseconds or microseconds), then pausing, then trying again and so on. The overcurrent and/or overtemperatur protection logic in many DC/DC IC is of the hiccup type.

 

[Philistine]

I would measure the waveform of the current to see if it is a normal continous DC/triangular signal, or if the charger is in "hiccup"-mode, that is, charging for a short time (maybe a few milliseconds or microseconds), then pausing, then trying again and so on. The overcurrent and/or overtemperatur protection logic in many DC/DC IC is of the hiccup type.

This sentence (for a non-science type like myself - and by that I mean my formal science education stopped at year 10), this is a sentence of pure beauty and genius. I always wondered what "hiccup mode" meant, and had always intended to look it up. I now understand. Thanks bearing.

 

[auraslip]

If you could build a a 4.2v charger, and then isolate the individual outputs, I think that might work. Each out put wouldn't be regulated though. I'm not sure how tightly regulated it would need to be though.
If anything, using 6 individual "cell phone" chargers setup for lipo chargers would be a much quicker way to balance than using battery medics.

Another thought I had was using a single high powered single cell charger and having it switch it's outputs between every cell in the pack. Like every second it would switch cells. So all the cells would be getting pretty much even charge. Not sure that'd be much simpler because each cell would require switching components.

 

[auraslip]

Here is a quick schematic of an idea I have

transformer and rectifier not included. Basically it uses the TI783 controlling some power transistors. I actually copied it straight from the TI data sheet for the regulator. It includes current limiting. I have no idea if it'll actually work, and I don't know how to simulate it in LT spice because they don;t have the tl783 in the model library.

Another thing I'm not sure about is weight of the transformer.