Batteries and New Battery Technologies

Jeremy, the A123 packs and even milwaukee and RYOBI mimic the old nicads in cordless tool applications. Hold voltage and power even under heavy loads then drop steeply at end, but still giving some power until stop. I agree there must be a BMS in there somewhere, probably meant to signal user that power is at the end. They do not start up again after final cut off -until recharged unlike nicads,- and no decrease in capacity on the following recharge. Previous experience with other non cordless tool packs just cut off at about 2.8v per cell or so. Learned the hard way that some packs came without cutoffs and were ruined.

There is a patent dispute between the commercial outfit producing the nano and University of Texas as I understand. Both are big boys that can't be run over easily, so the attorneys will feast on this one.

Throwing out an idea I used small scale with NiMh many years ago for recharging. Might be obvious, but ever know what might be useful to someone, have been surprised before. Used mating multi (12) pin plugs wired so some contacts were series and others parallel for individual cells or packs. The advantage was that only a single charger is needed for charging with the contacts in parallel instead of the many individual chargers. Yet the same pack can deliver multiples of voltage to different devices simply by having the device wired and hot to different and appropriate set of pins that mate with the battery pack. I toyed with the idea of using a "bad boy" charger for short and fast opportunity boost charging directly off the grid and still being able to run my bike setup off the same pack at 24v. The same job would have to be done with high current transformers which were hard to come by, heavy and expensive back in those days. Probably not a consideration anymore with cheap electronic power supplies.

Porta

 

For interest, I've tracked down where I first read about the DeWalt/A123 pack internal construction: http://endless-sphere.com/forums/viewtopic.php?f=14&t=2498

There are some useful hints there as to what's in the pack, it seems that the internal BMS does indeed act to shut-off and protect the cells from over-discharge, as well as maintain balance during charge.

I thought of implementing the multipin connector option into my motorcycle battery, but ran into two challenges. The first was cost, as the connectors had to handle the full discharge current, which meant using big Anderson connectors. The second problem was the force needed to plug and unplug these connectors when configured as a big assembly. I actually purchased a handful of Andersons, bolted them together as a big multipin connector, but found them impossible to unplug!

I could have opted for separate connectors for each cell, but was a bit wary of the potential for "finger trouble" when reconfiguring them from charge to discharge. In the end, I opted to just buy a BMS kit and fit that, as it was the easiest option.

Some of the electric bike people use the "one charger per cell" approach, using (relatively) cheap single cell chargers, connected as a big array and hooked up to balance plug type charging connections. One neat approach has been to use an array of surplus isolated output DC-DC converters fed from a common power supply. The commonly available 3.3V high current converters can often be trimmed up to the 3.6V charge limit for a LiFePO4 cell, and these can often be found cheaply on eBay in quantity.

Jeremy

 

Thanks for the link Jeremy, lots of interesting stuff here. One thing about balancing, at least with larger (automotive) systems is that all cells or packs are not at identical temperatures. So this tends to throw things off, because voltages and capacities of some battery chemistries are quite sensitive to even small differences. The solution was to occassionally overcharge instead of using balancing. This may not be a good idea with Li which appears to be more sensitive to overcharging.

One thing I have not seen addressed is whether undercharging Li causes permanent loss of capacity like in the Pb systems.

Porta

 

Could you explain what you mean by small differences in temperature, please.
TIA.

 

Even one degree or more will have a cumulative effect over several cycles, from what I understand. But then some of the balancers may have thermistors to correct for this as far as I know.

Porta

 

I've not noticed, or read about, any particular temperature sensitivity with LiFePO4. I keep a record of the individual cell voltages for my packs, primarily as an aid to spotting early cell failure, and haven't spotted anything. Having said that, the Konion cells used in Makita power tool packs have a consistent failure mode where the same one or two cells in the pack always fail. Several on the Endless Sphere forum have taken advantage of this, by acquiring dozens of free U/S Makita packs from power tool service centres and stripping them for their good cells. This consistent failure pattern could be for any number of reasons, including temperature differences between the cells in the pack.

One quite noticeable characteristic is the fairly big manufacturing tolerances on cell capacity. +/- 10% of nominal seems quite normal, and does cause charge balancing challenges on smaller packs. Once multiple cells are connected in parallel, to make sub packs for high capacity, batteries, then the problem tends to reduce, as there is an averaging effect between the high and low capacity cells in each sub pack. Several people have suggested that this is a good argument for using a pack with a large number of smaller capacity cells, rather then buy a few of the largest capacity cells and string them up in series.

The float charge voltage for lithium cells is the same as the charge termination voltage for a normal charge, so they can be held at that level for long periods with no harm, unlike lead acid cells that prefer to be float charged at lower terminal voltages. There is a very small usable capacity difference between holding at the absolute maximum float voltage or a figure 100mV lower, but in practice this is less than the normal cell to cell capacity variation. Most choose to charge LiFePO4 to about 3.65V, which gives near maximum capacity and is safe for continuous float charge. The cells will tolerate charging to slightly higher voltages, but there's no real benefit in risking it. My BMS limits each individual cell voltage to 3.65V and seems to give close to full cell capacity. I can use about 39 Ah from my 40 Ah nominal pack before the LVC cuts in on the lowest capacity sub pack, pretty much irrespective of discharge rate (at least over the 0.5 to 5C range I'm running over). This is pretty much what I'd expect for 100% charge capacity, as the Peukert exponent for LiFePO4 is close to unity.

Jeremy

 

Jeremy, don't think temperature would have much effect on motorcycle size packs if they are in a single unit. The temperature from the front to the rear of a car does vary enough as evidenced by "equalizing" charge recommended for lead and nickel systems that can tolerate.
With the trend to fast chargers of 1 hour or less for cordless tools, a lot of heat is generated which might be conducted to the cells unequally according to proximity to the heat source. So if the cells that are failing on the Makita are those closest to the heat source, then heat is a suspect. Otherwise it is probably not heat, would seem to me.

The multi-pin setup I used was cheap and easy to connect and disconnect. I used 4- 12v packs and the wiring was about 18 guage, so connections were flexible. Charged in parallel, the high current loses were small because of the multiple sets of wires. Power delivery in series also had low losses at 5 amps. I can see where quality of connector crimping and mating of connectors would be critical at high power levels. You point out advantages of charging Li in parallel and this might be a way to go. Some connectors of high quality and with a very large number of pins would be required. Using high voltage motors would keep losses low, but with fatal shock danger- especially around water.

Partial opportunity or solar charging might not work well if Li turns out to be finicky about requiring a full charge to reach maximum cycles.

Hope this helps.

Porta

 

There appears to be a dearth of verifiable information on this. The only decent test data I've seen is some testing on BMI (aka LifeBatt) 10Ah LiFePO4 cells undertaken by Sandia Labs. They were specifically looking at partial charge/discharge performance for a typical hybrid vehicle application, with the batteries cycled over a modest depth of discharge around the mid-charge point. The results indicated that cycle life would be well in excess of 10,000 cycles if treated like this. Full depth charge/discharge cycle life was estimaed to be 2,000 cycles to 80% of original capacity, I believe.

From the characteristics I've seen and measured, plus the information garnered from the experience of others, I would expect cycle life to be significantly better for partial charge/discharge operation, pretty much as it is for NiMH cells.

BTW, the motorcycle battery is in four packs, each of a nominal 12V, connected in series. Two are where the fuel tank used to be, one sits partially under the seat and the fourth is just above the motor, so I expect there are significant temperature variations, both from pack to pack and from individual cells within some of the packs. The lower pack, just above the motor, has it's front facing cells exposed to the slipstream, whilst it's rear facing cells are just above a slightly warm Mars ME0709 motor. I've not done any temperature measurement, other than the crude placing of fingers on things to make sure they aren't getting too hot!

Jeremy

 

Making a Super Capacitor

After one week of negotiations, I have decided not to sign the non-disclosure agreement and this means I am on my own again. The contract did not allow me to use privately any outcome of the assistance given and that defeats the object.
I am not a Structual Chemist, thus any support given is welcome in me trying to make some simple super capacitors.

I am considering the following criteria.
1a) 1500 - 2000 Volt capacitor voltage to get the energy.
1b) Some electronics to give a similiar constant output voltage as a battery
1c) output voltage for 180 Volt Permanent Magnet DC motor.
1d) first attempt only 50 x A4 sheets and not like EEstor 32000 sheets (16 bit) according to press reports.


1e) painting with a brush of the oxide onto the dialectrikum, instead of printing or silk screening
1f) initial small capacitance.

I am not worried about EEstor's patent. Years ago I used lots of capacitors parallel to give me 5 Volt 100 Ampere, it is in my opinion not an invention to have today capacitors parrallel.
Secondly as seen below, some 20 years ago thousands of "sheets" bonded parallel was already used in the production of making capacitors. (refer cap01 and cap02)
Thirdly, the use of the oxide and the purity of the oxide used is propably their strong part of their patent. According to press releases.

I like to make it as follow.
2a) Paint oxide on 50 sheets of dialectrikum with 1 cm blank on both edges as per drawing (draw01)
2b) put 100 connection strips between those sheets (draw02)
2c) press all together to a compact block
2d) put 2 stainless steel bolts through the connection holes for the terminals

I need some suggestions:
3a) What should I use for the dialectrikum, Rick any ideas?
3b) I will be violating the patent, but will as a first try, to get hold of barium oxide. Any other suggestion?
3c) Jeremy, any suggestion for a flexible width pulse circuitary.

Bertku

Photo left, Part of a capacitor in the making. Thousands of layers, dilelectrikum, metal foils are wind onto a reel. Thereafter, under vacuum, a tin layer of tin is sprayed onto both sides of the reel. The "wheel" is taken off and cut into small pieces, each having two leads soldered onto the ends and there you are, you have a capacitor.

Photo middle: You can see the layers before it is cut into smaller capacitors.

EEstor would have massive problems in manufacturing 100.000 x 32.000 sheets = 3,2 billion sheets per year. This patented process might be the solution for them. Although the reel has to be a few meters in diameter not to notice the curve too much. But in my view , if they addopt this kind of process, within a few months they could be up and running with minimum manufacturing cost.

 

Bert
Have you calculated the capacitance you will get from the 50 sheets.

How thin do you aim for the dielectric film?
Rick

 

EEstor is getting 1 Farad out of a sheet. If I am getting 0.1 Farad per sheet, I am very happy for the first round. Painting with a paintbrush is not very efficient.

I lost touch with dielectric materials. I like to see it as thin as possible, but still can handle high Voltage. EEstor is apparantly up to 3200 Volt.

BertKu

 

BertKu, the pictures you show in post 324 are Siemens MKM and MKT polycarbonate capacitors. They are made from much thinner foil that you could possible handle, yet their capacity is limited to a few microfarads @ 63 volts. The ones with 400 VDC are under 1 mF.
Their strong points are the metalized foil and large contact area because they are solder dipped.
Buy a large bag of these and make your own supercap; painting will lead you nowhere.

Wasn't the definition of 1 Farad the surface of a sphere with a radius of 9 times 10 to the 9th power? That is quite a lot of paint.

 

Illustration

Hi CDK,

Yes, you are right, they are 1 uF 100 Volt 20 cm long strip out of the production before being cut. The EEstor patent is using, instead of alluminium foil which can indeed hold a limited of energy, a dilectric with Barium-titan printed on each 32.000 sheets per battery. I try to illustrate that the US patent (apart from the oxide) is not new. However they could use the same principle, just larger equipment to make the Super Capitors with the same principle, have 2 printing presses printing the barium-titan onto two rolls and then offset rolled onto the third reel. Thereafter in a vacuum tank sprayed the tin on both sides and cut them up as batteries.
Their method is sheets packed onto each other. They will never get 1.000.000 batteries out of their production line, unless they can fully automate. (32 billion sheets !!)

CDK, to get my electric boat project up and running, I need an energy storige device, I can trust. I trust capacitors, I don't trust lead acid batteries
nor all other flimsy, tricky batteries. Maybe the sealed lead acid ones I am using I have some respect for them, but their life time is too limited to my
likings. I have great respect for Rick, who is making something as light as possible. That is his direction I can only learn from. My direction is to get as much energy on board as I can, thus I can waste some of it in bad efficencies or whatever.

Any suggestions for my dielectrikum on which I can brush the barium-titan on?

BertKu

 

Mylar is a good dielectric. I do not know how thin you can get it. Also not sure how well your conducting material will take to it.

Rick W

 

Mylar

Hi Rick,

Yes, good idea. It does not matter for my initial experiment how thick or thin the mylar will be. I like to get the feel for it on what is possible in the backyard and what not.

CDK, When you are back from Venetia have a look at 252. You need a magnifying glass to see the 8 Farad on the Epcos (nee Siemens) Ultra Capacitor, come on guys, just help me to make a super capacitor. I get enough critisism from my wife.

Jeremy, do you think this will work? The problem is, the Voltage cannot go too high on the super capacitor, because the transistors I can get (afford) are all well below 1200 Volt.

BertKu