Doctorbass: Single Cell Charger for M1

[drewjet]

I have used similar power supplies. They are isolated. They will put out a bit more than the rating without complaining. However a dead A123 or a few in parallel will probably draw more than the 20 amps, a long zipcord could be used to control amps.

 

[tailwind]

I am going to record some charging graphs for the 25w and 100w power supplies.
I will record amps, volts, watts verses time to give people an idea of their charging profile.

The amps should spike at the beginning of the charge when the voltage of the cell(s) is the lowest.

For example for a 100% effecient powersupply.

One cell at 2.0v, connected to a 25w PS should start charging at 12.5A.

Two cells in parallel at 2.0v, connected to a 100w PS should start charging at 50A or 25A per cell.

Both are over the 10A recommended fast charge set by the manufacturer.

Greg

 

[Doctorbass]

What is important with these power supply as charger is that you NEED to know the equivalent load that represent one cell when it is fully discharged at let say 2.5V / 3.6.

I am finishing my large 12s multi DC-DC charger and i needed to calculate that to avoid the dc-dc to overload.

Also you need to know what is the voltage of a cell when charging at different current. (if you discharged a cell down to 2.5V, after that if you apply let say...10A on it.. it should rise around 3.0V (AT THE CELL TAB)

so if your charger apply 3.6V and apply 10A and your cell is 3.0V, the 0.6V will be somewhere for sure!...in the wires, resistors or voltage connections loss...

OR you can also limit the current of the power supply and then tranbsform it in a CC-CV. Some power supply that are rated 20A like the one showed above and that i've been talking about in "the single cell charging post" can deliver more and their current limit will only limit the INPUT CURRENT! so that mean if they are rated 20A at 5V(that you re-adjust to 3.6) that's 100W.. so if your voltage is readjusted to 3.6V, that current limit become 100/3.6 =28A... and the output component will not hold that for long!..

what you need if you want to keep the cc-cv feature and not use resistor to dissipate the vooltage and only have a CV charger... is to control the current limit!

That'S what i've been working on for few weeks. Right now, i'm mooding the current control resisitor of the PWM IC on the DC_DC board

I am limiting it to 15A instead of 30A.. for two reasons: the first is that i dont want to have 24 conductors of 10AWG between the charger and the cells! 30A is too big current than i need from now and the cells will not like that for their cycles lifes!

Second, the main power supply that give the 48 or 72V on each of their input would not be able to suply all 12 dc-dc at the same time if they give 30A at 3.6V or 4.2V !

I will also use the voltage sensing option they have.. that option ensure that the charger will not be affected by the voltage loos in the long wires between the cell and the charger. so each of the dc-dc will have 4 wires at their output. 2 for the high current main output and two for the voltage sensing that will be connected directly to the battery tab to ensure that i have 3.600 or 4.200V on the cell.


As you know, it take me long time to finish this charger that i've been talking from many months but i will finish it in the next 2 weeks... that will be a universal li-ion battery charger!

Doc

 

[Johnbear]

Would 3.9v be too high for lifepo4. Yesa sells chargers that cut off at 3.9 - he has them on ebay now but from what I have read this seems a bit high? Anyone know for sure?

 

[Doctorbass]

Would 3.9v be too high for lifepo4. Yesa sells chargers that cut off at 3.9 - he has them on ebay now but from what I have read this seems a bit high? Anyone know for sure?

LiFePO4 cells can be charged to 4.2V.. BUT THEY WILL NOT HAVE THE SAME CYCLE LIFE

THEY WILL HEAT MORE

Best is 3.65V

Fast, great capacity and great cycle life.


Doc

 

[tailwind]

Heres the two charging graphs as promised.

For the 25w PS
I discharged one A123 cell to a cutoff voltage of 2.7v at a rate of 2.3amps (1C).
When it was finished I waited 5 minutes by which time the voltage was then 2.85v.
Set the PS voltage to 3.60v.
I connected it up with a WattsUp meter inline.
And heres the result.
View attachment 25w-1s.jpg

At the beginning of the charge the amps vary between 0.95 and 1.98 amps at a rate of twice per second. At the same time you can here a faint ticking noise coming from the PS.

For the 100w PS
I discharged four A123 cells that are wired in parallel to a cutoff voltage of 2.7v at a rate of 9.2amps (1C).
When it was finished I waited 5 minutes by which time the voltage was then 2.85v.
Set the PS voltage to 3.60v.
I connected it up with a WattsUp meter inline.
And heres the result.
View attachment 100w-4p.jpg

There was no noise coming from the unit when it was outputing 24amps.

It looks like a protection circuit in each PS is kicking in when the voltage is below a certain level. But each unit does it in a different way.

25w
Output Voltage Adj. Range: ± 10%
Output Overload Protection: 110-150%
Output Over Voltage Protection: 115-150%

100w
Output Voltage Adj. Range: ± 10%
Output Overload Protection: 110-130% Constant Current, Auto Recovery
Output Over Voltage Protection: 115-150%

Greg

 

[Dee Jay]

This is exactly what I needed to know! Very cool of you to provide this info. A hearty thanks and a deep Bruce Lee bow. *whaaa* That's an honorable bow.

A dishonorable bow is when you bow soo deep that your head touches your anus.

Either that, or a finger comes off. Ha ha!


I've been working on my rotating charger and making much progress but to be honest, 50% of the time I'm just sitting here watching parts on the table as if they'll miraculously float together and assemble all by themselves! :wink:

Stay tuned, I'll have pics soon whether it's working or not.

Thanks again, tailwind!

J

 

[Jeremy Harris]

Very useful stuff indeed. Thanks for taking the time to post it.

The first power supply seems to have foldback current limiting. When it sees an over-current it cuts off, periodically trying to restart (the ticking noise) to check if the overload is still present. Each time it attempts a restart, it kicks out a hefty current pulse (probably too short in duration for your logger to record). Gradually, these very short pulses put enough charge in the cell to reduce the current demand down to a level that the supply can deliver, which is when the current rises to it's peak on your plot.

The 100W supply looks as if it has linear current limiting, rather than foldback, hence the different characteristic.

For a CC/CV charger, a linear current limit supply seems better, as it will deliver maximum current from the start of charging. I'm going to add an external current limit to my switched mode charger I think, as I can't see an easy way to accurately adjust the rather crude current limit circuit in the supply I have.

Jeremy

 

[fechter]

Outstanding! Great test.

I wonder if there's a way to mod the current limiting on the 25w one to prevent the oscillation at the beginning?

Somehow I missed where you got those from. Can you post a link?

 

[tailwind]

Thanks guys,

Both powersupplies are from powersupply1.com.
They distribute through mpja.com and CircuitSpecialists.com / web-tonics.com(Intr orders)
http://www.web-tronics.com/253siouposup.html

Doctorbass originally found the link to the 25w PS.
http://endless-sphere.com/forums/viewtopic.php?f=14&t=2586&st=0&sk=t&sd=a&start=60#p37816

If you are ordering from outside the US,
mpja has a minimum order of USD100
web-tronics minimum is USD200 and that doesnt include postage and handling. :x

I am going to make up a parallel harness for my 4p x 13s setup and do the round trip ride to work. I will then plug it in to a single 100w supply and see how long it takes to charge.

The 25w supply could be a problem if you do public charging.

Concerned citizen, "Yes Officer it has wires coming of it and its making a ticking sound"

You might come back to find the bomb squad has destroyed your bike.

Greg

 

[tailwind]

The round trip ride to work is 25km. Used 5.0Ahr.

Connected each of the 4p packs in parallel (52 cells).
Plugged it into one 100w powersupply and went to bed.
Got up this morning and found out that it took 8hr to charge.
All cells balanced and fully charged to 3.60v.



I monitored the voltage by connecting up a voltmeter to the pack and then setting up one of my cameras on a tripod. The camera (Canon S60) has a intervalometer, so I set it up to take a picture every 5 minutes. In the morning I did a rapid scan through the images to find out what time it reached 3.60v.

Used the same technique to record the data for the graphs above, but in that case I used the wattsup meter and my other camera (Canon G9) which can take an image every 1 second in timelapse mode.

Greg

 

[fechter]

LOL!
Poor man's data logger. Good use of existing equipment.

 

[GGoodrum]

I remember McGiver-ing an RC servo to the shudder button of a Minolta SLR to take oscilliscope shots once every ten minutes. It was driven by a buffer card that was connected to the PDP-8 that was the heart of a test set that we were using. I wrote a small program (I was a software weenie back then..). This was way before computerized instrumentation of any sort. Actually, it was even before McGiver. Back then it was a "Rube Goldberg contraption".

Christ I feel old... I think I need a nap...

 

[Dee Jay]

Not quite done but getting there.

It's pretty self-explanatory: a simple rotating switch for 14 cells/15 tap contacts expandable up to 30 cells. Still trying to work out a way to mount the 6 volt solenoid to pull the ratcheted disc when the cell reaches full charge.

Oh! and it can he used to check individual cell voltage. Hooked it up between A Pack connector and a voltmeter. Results: Not too happy with using domed bronze nail heads as contacts. Fling the dome the nails' heads down flat caused inconsistencies and intermittent continuity. I might have to fab flat copper contact plates.

And I may not use the ratcheted teeth and end up using magnets and iron tabs to insure precise positioning of the power contacts. Still working out a way to use a spring to apply pressure to the contacts.

I know this is a Flintstones' device but it's an EV necessity! And I'm sure it can be made solid state.

Any EE out there willing to improve and market my idea, go for it! But be sure to kick me some freebies, ok?

Cheers and Happy Friday!
J

 

[Link]

Nice. Love the translucent green plexiglass. That needs to be shown off. 8)

Still working out a way to use a spring to apply pressure to the contacts.

Put a ring of sheet steel on the bottom. Put a neo magnet on the copper strip.

 

[Doctorbass]

I remember McGiver-ing an RC servo to the shudder button of a Minolta SLR to take oscilliscope shots once every ten minutes. It was driven by a buffer card that was connected to the PDP-8 that was the heart of a test set that we were using. I wrote a small program (I was a software weenie back then..). This was way before computerized instrumentation of any sort. Actually, it was even before McGiver. Back then it was a "Rube Goldberg contraption".

Christ I feel old... I think I need a nap...

lol.. McGiver! A great TV show it was!

what would look like a 2008 version!!!


Doc

 

[lawsonuw]

Nice rotary switch Dee Jay. U know people make those right? http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=451-1038-ND would work IF the charger is off whenever you switch positions. This switch is two "decks" of SP24T with non-shorting contacts and would support connection of up to 24 cells. (one deck would have all the posative terminals while the other deck would have the negative terminal of each cell. would look sorta like this => '////////. )

How are you going to sequence this rotary switch? Have a timing relay drive the indexing solenoid at set time intervals? Or roll something more sophisticated using a Basic Stamp or something similar?


Marty

 

[Dee Jay]

WASSUP GUYS!

Link:

Yeah, it would be cool to show it off some how. I imagine enclosing the whole thing in a clear plastic box. MY SON ALMOST SPILLED JUICE ON THE PS WHILE CHARGING!

These were precut plexiglass pieces 100mm x 100mm and 100mm dia disc for less than 2 bucks each at my local DIY shop. You might find them art supply shops.

Although plexiglass can be melted with a soldering pen, soldering on top of it wasn't a problem, it retained its shape really well. I guess silicone glue helped too.

Sweet! I found some tiny neo magnet tabs at amazon! Will be good for precise positioning as well as contact pressure.


lawsonuw:

Thanks for the info! Yeah, only after I decided to call my creation a "Rotary Switch" did I even think to do a search for rotary switches! duh... I was looking at the pdf http://dkc3.digikey.com/PDF/T082/P2030.pdf How do you hook up these rotary switches to a powersupply/ single cell charger leads to distribute power to one cell at a time? ... BTW, it's only 1.5 amp. I need at least a 10 amp switch.

Since my Power Supply doesn't have full charge indicator LED, I'm not sure how I'll trigger the mini solenoid other than using zero current sensor/relay, and it is expensive!


If I don't experiment with a voltphreaks single cell charger that has the LED indicator, I may just use some sort of timer relay. For a cell at 3.20 volts charged to 3.65 volts 2 hours seems to be just right. S L O W. Voltage settles down to 3.60/61volts

Happy Friday!
J

 

[lawsonuw]

lawsonuw:

Thanks for the info! Yeah, only after I decided to call my creation a "Rotary Switch" did I even think to do a search for rotary switches! duh... I was looking at the pdf http://dkc3.digikey.com/PDF/T082/P2030.pdf How do you hook up these rotary switches to a powersupply/ single cell charger leads to distribute power to one cell at a time? ... BTW, it's only 1.5 amp. I need at least a 10 amp switch.

Since my Power Supply doesn't have full charge indicator LED, I'm not sure how I'll trigger the mini solenoid other than using zero current sensor/relay, and it is expensive!

If I don't experiment with a voltphreaks single cell charger that has the LED indicator, I may just use some sort of timer relay. For a cell at 3.20 volts charged to 3.65 volts 2 hours seems to be just right. S L O W. Voltage settles down to 3.60/61volts

Happy Friday!
J

Yes, the switch I linked is only rated to switch one amp, but the data sheet says it's fine to conduct twenty-five amps. This means that whenever the rotary switch changes positions your charger needs to be off. The switch I linked has two "decks" each deck has a single terminal for the moving contact and 24 static contacts. As the switches knob is moved the moving contact moves between each static contact. Each "deck" of the switch has it's own moving contact and all "decks" move in unison. http://www.electro-nc.com/rotaryus/c7.pdf bottom of page one shows the layout of a one "deck" 24 throw switch.

Wiring is fairly simple. Lets label one of the switches "decks" "top" and the other "bottom". The charger's positive terminal would go to the moving contact of the top deck while the charger's negative terminal would go to the moving contact of the bottom deck. Next the positive terminal of the first cell would be connected to the "1" terminal of the top deck and the negative terminal of the first cell would be connected to the "1" terminal of the bottom deck. This pattern is then repeated for all the cells in the pack. Since the cells are hooked in series inside the battery pack the positive terminal of the first cell is also the negative terminal of the second cell. To save wire this connection can be made right at the switch. I.e. the negative terminal of cell one connects to terminal "1" of the bottom deck, the positive terminal of cell one connects to the "1" terminal of the top deck and terminal "2" of the bottom deck, finally the positive terminal of cell two connects to terminal "2" of the top deck. (bha, that was as clear as mud!)

Page 885 of the McMaster.com catalog has a description of what types of timing relays are available. McM# 6964K2 multi-function relay with McM# 7122K21 socket would be a good start. I'd use the "repeat cycle" mode and one of the DPDT contacts to turn the charger on and off. I'd rig the other contact to turn on a time delay relay when the charger is off. This time delay relay would then activate the indexing solenoid in the middle of the time the charger is off. McM# 77055K73 would work well. If I wanted to do anything more than that I'd get an easy to use microprocessor like the Basic stamps I linked above and do it all in software with minimal circuitry.

Marty

P.S. McM# 7630K23 might be a better master "repeat cycle" relay. It looks like the "on" and "off" times are independently adjustable. Needs a McM# 7122K19 socket.

 

[Dee Jay]

(bha, that was as clear as mud!)

:lol:

Sounds complicated, but thanks for all the info! I'll check out that timer relay.

I think my switch is more direct and to the point. Simple enough for non techies *like me* to follow

*Imagine a quality, precise version (no neo magnets required)*

*mass produced specifically for the ebike market *

*multiple I/O for charger, DMM, & drain bulb*

*bare wire clamping port holes like on stereo speakers (no soldering required)* EDIT: Probably not a good idea. Wouldn't want to accidentally yank out live wires! ~ ~ ~ ~

*A manual version*

*Option to add a solenoid*

A simple rotary switch for up to 30 cells . . . $20? $30? I'd buy it!

HF!
J

 

[jeffkay]

How about an old dot matrix printer connected to a computer? Contacts (or switches) could be arranged along the path. The printer travel would be precise and repeatable and the computer could turn of/down the supply amps via some kind of FET during the relocating at each cell. To take this a step further, why waste time "charging" the full cells, just have the computer measuring each and skip cells that have met the required top-of-charge? This may sound complicated but these are very simple PC programming tasks.

Also, each cell could have an associated light bulb current limiter so when switched with say a small micro switch, it does not arc...

Jeff K. "Deep Cycle" project