Single-Cell Charging System with DC/DC Converters

[Dolphyn]

I'm working on a charging system for my 1993 City-EL which is powered by 14 ThunderSky LiFeYPO4, 100-AH batteries. I've bought most of the parts, and I've tested a rough prototype. So far it appears to work great, but I wanted to see if anyone has any comments or suggestions before I try to put it all together.

My new charging system is based on DC/DC converter modules, model VSX40MD23, currently ridiculously cheap on Ebay in lots of 24. I've been testing this module (with ORWH-SH-148D relay to disconnect the cell when there is no input power) and it seems to work GREAT as a charger. There is a "trim" feature which I can use to adjust the maximum voltage, and the output current appears to be limited to a tad under 14 amps. The efficiency of this module is rated at 89%, which is better than other similar devices I have seen.

This type of converter works great as an "emergency charger." If one cell is depleted, I can use my pack voltage to give that cell a boost and get back on the road. So, except for minor inconvenience, my range is no longer constrained by the "worst" cell. Yay!

I'd like to take it a step further, because my cells are prone to balance problems and my existing charging system is poor. I want to permanently mount a converter module on top of each cell to use as my primary charging system, perfectly top-balancing my cells every time I charge.

Top-balancing is ideal for my application. My vehicle seldom goes over 1C discharge, so my CellLogs configured at lower limit 2.8V give several miles warning before I have any problem. I'll have a switch mounted on each cell to activate "emergency charging" using pack voltage, as needed, with minimal fuss. Actually top-balancing gives me enough range that I don't expect any need to do that, but it's a nice feature to have available.

Ejonesss has reported that without protection, the cell voltage will damage the converter when its input power is disconnected, so the relay is important. I'm also adding a fuse, and a diode for relay back-EMF, and I've bought some small computer fans and N.O. temperature switches for cooling.

I expect to order custom circuit boards for this project, because my attempts to solder thick wires to my relays and converter modules have been clumsy at best. I've been working with ExpressPCB software to design the board. It's pretty easy to use and kind of fun, so I'm making progress. (Unfortunately ExpressPCB doesn't allow a hole large enough for the battery terminal bolt, but I guess a drill will work well enough.)

My power source will be the transformer from the vehicle's original charging system. It produces ~45-55 VAC, so I'll use a bridge diode to convert to DC. For "smoothing" I think I'll use a 33-microfarad, 0.5-ohm electrolytic capacitor mounted at each module as suggested in the Application Notes, and I figure it can't hurt to use an additional capacitor located at the bridge diode. I don't have experience with this, so any specific suggestions are appreciated.

Some more details and documentation:
VSX40MD23 Data Sheet
VSX40MD23 Application Notes
48V Relay Data Sheet (ORWH-SH-148D)

The VSX40MD23 Application Notes include a discussion about "conducted EMI" and some suggested circuits for dealing with it. So far I've ignored it because it looks complicated and I don't understand any of it. So, if anyone thinks I need to worry about "conducted EMI", please advise.

By the way, is there any consensus on the best maximum voltage to use, for a reasonably complete charge while optimizing the longevity of the cells? I'd like to be able to walk away and leave the charger on a trickle charge without damaging the cells. I think 3.5V will do a pretty good job, but suggestions are appreciated.

Thanks!

 

[GGoodrum]

Yes, these look interesting. Theres been several threads on using DC-DC convertors. This is the first I've seen for actually mounting them on top of the big TS-type cells, though.

For the terminal voltage, you could even just use the full 10% trim, which would put the 3.3V output at 3.63v, but I guess to extend cell life, picking a slightly lower number, between 3.55-3.60V, would be better.

Although ExpressPCB only allows holes up to .25", you could use the "Insert corner" function to make U-shaped terminals.

-- Gary

 

[Dolphyn]

Thanks Gary!

I think I'll use custompcb.com instead, because they offer more options in terms of hole size, board thickness, and copper thickness, and the prices are lower for this project. But the software is a much bigger pain, at least for me. I'm trying to use KiCad and I find its operation to be utterly non-intuitive.

(Hmm, what is the optimum board thickness for a circuit board to be mounted directly on a cell?)

 

[oatnet]

I used Vicor DC-DC's to make a balancer for the cells in my VW Bus conversion, and they put out a massive amount of heat, hot enough to burn, I considered mounting them on cells until I saw them in action. I think mine are only 80% effecient vs 89% for yours, with 20a output, so YMMV.

http://endless-sphere.com/forums/viewtopic.php?f=34&t=8012&start=120#p143941

I used 3.65v/cell because I only use them for balancing, not regular charging.

-JD

 

[myzter]

I was curious about using DC/DC converter modules for charging an entire battery pack
I found a DC/DC converter module in the input voltage range Im using
4 x M1 cells / 13.2v 2300mah

Vicor VI-J0L-CY 12V IN 28V Out
http://cgi.ebay.com/220583784128
http://psearch.vicorpower.com/search/product_specs/?part_number=VI-J0L-CY

I called the company to ask about output current
the agent told me I could get .8amps or 23 watts
He also stressed that these are not used for charging batteries, but for running motors etc..
I cannot use these to charge a battery.. so that kinda sucks..

Is this not what you are doing Dolphyn ??

 

[Dolphyn]

I used Vicor DC-DC's to make a balancer for the cells in my VW Bus conversion, and they put out a massive amount of heat, hot enough to burn, I considered mounting them on cells until I saw them in action. I think mine are only 80% effecient vs 89% for yours, with 20a output, so YMMV.

Thanks, I must admit I've worried about the heat issue. When I tested my prototype, I never saw it go over 70 degrees C, but for the most part I wasn't using the full power of the converter due to voltage loss in the wires. Minimizing that voltage loss is, incidentally, one of the reasons I want to mount directly on the cell. I've done some previous experiments with single-cell chargers, and none of my results made any sense at all until I realized my wires were too thin and/or too long.

I think the heat will be under control because I've bought some heat sinks and I'll have good ventilation with a little fan mounted on top of each battery. To control the fans, I found some N.O. temperature switches on Ebay.

 

[heathyoung]

They are cheap aren't they! Have you considered using Shottky diodes rather than relays etc? Just trim 0.2V higher to account for diode drop.

 

[Dolphyn]

I was curious about using DC/DC converter modules for charging an entire battery pack
I found a DC/DC converter module in the input voltage range Im using
4 x M1 cells / 13.2v 2300mah

Vicor VI-J0L-CY 12V IN 28V Out
http://cgi.ebay.com/220583784128
http://psearch.vicorpower.com/search/product_specs/?part_number=VI-J0L-CY

I called the company to ask about output current
the agent told me I could get .8amps or 23 watts
He also stressed that these are not used for charging batteries, but for running motors etc..
I cannot use these to charge a battery.. so that kinda sucks..

Is this not what you are doing Dolphyn ??

My project is different, because I'm using a separate converter for each battery cell, but it's true that battery charging is not the intended purpose of these modules so I may be taking a risk by doing so.

In my previous research, I stumbled upon Vicor's recommended method of using their converters to charge batteries. You might want to read this:
http://www.vicor-asia.com/documents/application_notes/an_ConstantCurrent.pdf

Granted, my eyes glazed over when I tried to read it, but one paragraph deserves highlight:

Vicor converters have an internal current limit designed to reduce the risk of damage to the module during a fault condition. This limit should not be used as part of normal operation because the converter may be driven into an overpower condition or unstable operation. Trimming down a converter cannot fully protect against overcurrent because there is a limit to the percentage by which the output voltage can be reduced. This requires that an external circuit be implemented for loads that do not have lower bounds on their impedance.

Fortunately I'm not using Vicor converters, so I'd like to think that information doesn't apply to me.

For comparison, the VSX40MD23 Application Notes document says:

Each converter is equipped with current-limiting circuitry designed to provide continuous protection against fault or short circuit conditions. The current limit point is typically 5%-25% above the rated output current. When the short circuit condition is removed, the output voltage will return to its "pre-short circuit" value without recycling the input.

 

[Dolphyn]

They are cheap aren't they! Have you considered using Shottky diodes rather than relays etc? Just trim 0.2V higher to account for diode drop.

I don't have any experience or knowledge of Shottky diodes, but I just did a search for an appropriate one, and the ones I found show "forward voltage" much higher than 0.2V, at least when any significant current is involved. Please let me know if I'm missing something or if you have a particular diode in mind that would do a good job. Thanks for the suggestion!