stepping voltage up or down and its toll on batteries

rwgast_logicdesign · 2012-11-13 02:06

Ok so ive been 5 steps ahead the whole time ive been building this robot, except for making a desisicion on how to power the thing and its at the point where i need batteries, having 20 ft of a cat5 twisted pair connected to an adjustable supply is lame.

now im pretty sure im going to be using nicad or nimh, probably nimh since the double a cells are 2000mah. after finding out my bot can hall around 15lbs im not to worried about how many cells ill be using with in reason.

i also still need to decide if im going to be running a 12v or 18v supply to my motors, but for the sake of this thread lets go with 18v. Now to male 18v with nimh aa cells im going to need to run 15 in series, so 2000mah @ 18v isnt to bad. I had originally planned on just running the max voltage i need then using buck regulators to power all my logic. a few days ago i started thinking i may want to run a boost regulator to make sure my mmotors always get a constant voltage even when the batteries drain.

so today i was on mdfly and saw a nice little boost unit that is 5-24v adjustable for 5 bucks. this got me thinking if im gonna have to load up 30nimh to get 4000mah at 18v maybe it would be better to just use 15 batteries in a 6v 6000mah configuration then boost from 6v to 18v for the motors. but i dont really understand how boosting works or what kind of toll it takes on the power source. I guess what im wondering is would it be better to use more series batteries and step down or use les batteries but in parallel and step up? how fast do boost regulators eat up your amp hours?

LoopyByteloose · 2012-11-13 02:42

Electric waste is generally observed in the form of heat.
The linear regulators just burn a lot to get the voltage regulated, but the other little switcher devices are more efficient. Still if a linear regulator wastes 30%, I still suspect the other switching devices to waste 10% (read the PDFs). After all, conversion is always wasteful.

Batteries always seem to come in sizes that don't quite fit the 5v or 3.3v logic - but you do have a bit of lee way. 10% tops on the high side and whatever the brownout voltage is on the low side.

For instance 5 volt logic can live between 5.5V and 4.2 or .3 volts and still operate. Regulation just makes a neater package. 3.3v logic will go to 3.6V on the high, maybe 2.7V on the low and this fits well with unregulated 3.0 Lithium cells (not 3.6 or 3.7v ones). And some 3.3V logic is 5V tolerant and even easier to exploit is battery direct power.

Properly designed and separate battery packs for the motors and the logic will work out best.

The other side of this is whatever you are driving. Direct logic driven power MOSfets or superBeta transistors are the best, but harder to come by. If you use a transistor to drive something, there is about 0.7 volts dropped, maybe more - like 1.0v actual. So a 6volt battery is really sending 5.3V to 5.0V to the relay. And batteries are never quite what they are labeled. They can range a bit higher and much lower. The think is that coils in relays and motors are much more tolerant of varied voltages, so you might be able to skip the regulations of those if you take measurements of the actual in use voltage provided after the voltage drops of the control circuit. Avoiding regulation wherever logic is not involved is an excellent way to eliminate a lot of wasted power.

Another thing is to revise all your logic to use much less power. Just because the Propeller can output 25ma per pin, doesn't mean that you have to use that much. insert currect limiting resistors and use higher values in pull up resistors. Often a transistor will saturate at 2-3ma, and the other side of its circuit can be unregulated. I use 100ma regulators in some contexts, 500ma regulators in others, and have the real work done by unregulated coils and motors.

So stop thinking about a particular voltage and think in doable ranges. Find out what batteries really put out when fully charged, not what the label nominally says. Someone recently mentioned that real micro-power designs don't use any regulation at all --- just a lot of real world investigation and planning.

Mark_T · 2012-11-13 03:52

If you boost convert you will lose some power in the conversion - perhaps 10%, perhaps 20% - that's pure waste. If you use the right battery voltage you eliminate this source of loss, so I'd prefer it. You can counteract the reduction in battery voltage during discharge by measuring it and having the PWM to the motor compensate for this. Note that traction batteries will see a voltage drop at high currents due to internal resistance too, so its a good idea to monitor voltage like this anyway to compensate for periods of heavy load dropping the voltage.

Also a boost converter could happily drain your batteries well below the minimum recommended charge state - this usually means battery recharge cycle life will drop catastrophically - you don't want to do this. When you have direct battery supply you'll tend notice the signs of battery discharge early as the peak power fades.

Best practice (a lot of motor controllers do this) is to monitor battery voltage and switch to reduced power if it drops below a soft-limit, then cut-out on a hard voltage limit. If you measure current as well you can both detect over-current and compensate the measurement of "true" battery voltage for internal-resistance losses - otherwise you can't tell over-discharge from temporary high-load voltage drop.

When selecting your batteries you do need to make sure they can handle the current without excessive internal-resistance losses or over-stress to the batteries themselves - capacity by itself isn't enough information to choose.

erco · 2012-11-13 06:32

Bang for the buck: LiPo, Li-Ion or lead acid at the voltage you need. No voltage conversion, big current capacity. Simplest charging: lead acid.

rwgast_logicdesign · 2012-11-13 09:58

hmmm well the reason ive stayed away from lipo is the cost, 18v of lipo cells plus a balancing charger is going to be around 20 dollars. I also dont feel confident that i could make a lipo charger when im ready to build a custom dock. As far as lead acid goes the smallest batteries that come to mind are fairley heavy for a robot that can only handel 15lbs.

So now i guess i need to figure out the best 18v power source i can use.

Mike Green · 2012-11-13 10:33

Here's one vendor of sealed lead-acid and gelled lead-acid batteries. Any combination that gets you to 18V (9 cells) would do. Remember that the open-circuit voltage is somewhat above 2V. This might go as high as 21V total while charging so be prepared for this. A 2.5Ah 18V battery would weigh maybe 4 pounds. Here's a link to an article on charging these.

rwgast_logicdesign · 2012-11-13 11:48

Acually I just tested my motos under load and there only drawing 80mA at 18v connected to two 9vs in series. So if i were to just use 6 9vcnimh batteries at 700mah a piece, seriesing two groups of 3 parallel batteries that would be 2100mah. Assuming both motors are running under load at 18v that would be 160mah draw this mean I could run the motors for 13.125 hours at 18v?? 2100/160=13.125. I plan on mostly cruising around closer to the speeds I get when the motors have a 3v supply but I want that extra power and speed. What would be the best way to go about charging a set up like this, just a charger from scrath or what? I mean im going to have to supply over 18v to recharge the battery pack right?

Mike Green · 2012-11-13 12:02

Don't use parallel batteries. If they discharge at slightly different rates (as they will), one battery will end up trying to charge the others and they'll all run down quickly. To prevent this, you have to put a diode in series with each set in parallel and that results in a voltage drop and wasted power. You're much better off just using a larger battery capacity to begin with.

Yes, you'll have to supply over 18V to charge the battery pack. Figure on at least 24V.

Most 9V rechargable batteries are not anywhere near 9V. They typically have 6 NiMH cells in series internally and supply around 7.2V (6 x 1.2V). If you want a true 9V power pack, you need 7 or 8 NiMH cells in series to yield about 8.4V (7 x 1.2V) to 9.6V (8 x 1.2V).

rwgast_logicdesign · 2012-11-13 12:10

Ya there no where near the amp hour I thought they were either i just looked them up and there 175 to 200 mah, so thats not going to do. So i guess my best bet for off the shelf may just be 15 double a style which will give me 2000 mah and can eaily be charged seperately till I have a full charging solution. That is if my math is right, running 2 motors for about 13 hours seems like along time on a double a

kwinn · 2012-11-13 12:57

If you have an 18V cordless drill you should try using that battery for your robot. If you don't have a cordless drill you might want to consider investing in one rather than buying batteries. It is a very handy tool and comes complete with a charger.

While I have never used my cordless drill battery in a robot (not really into that) I have used it to extend the length of time I can use my laptop when AC power is not available. It works very well for that.

rwgast_logicdesign · 2012-11-13 13:10

unfourtantely i mostly use a corded drill and have a 9.6v cordless drill which ive ripped up, i really started thinking about a cordless drill battery though

Mark_T · 2012-11-13 14:44

Most 9V rechargable batteries are not anywhere near 9V. They typically have 6 NiMH cells in series internally and supply around 7.2V (6 x 1.2V). If you want a true 9V power pack, you need 7 or 8 NiMH cells in series to yield about 8.4V (7 x 1.2V) to 9.6V (8 x 1.2V).

The NiMH PP3's I have are all 8.4V (ie 7 cell).

Whether a battery chemistry is happy to run in parallel depends on the discharge voltage curve - if the plateau is very flat the cells can't automatically balance and its not recommended. If the voltage plateau has a definite slope (like LiPo) then parallel is OK - certainly RC LiPo packs and LiFePO4 cell banks are available in series-parallel combination routinely. Lead-acid batteries have a very flat discharge curve and parallel is not recommended though.

High charge and discharge rates in parallel will act to balance the current because the voltage drops across the internal resistance of the cells - but only while the cell's internal resistances are well matched - older packs will balance less well.

rwgast_logicdesign · 2012-11-13 15:56

ok so i just got 15 double a 2450mah nimh batteries. then i found a 79 cent wall supply at good will that is 30vdc and 1amp. im thinkin that should charge finne maybe a little slow but itll work, do i need to cut the voltage down to 24 or is 30 ok to charge 18v? im planning on using a smart charging ic too

Mike Green · 2012-11-13 16:05

Please pay attention to the information on charging batteries. You can end up destroying your batteries or at least shortening their life. Most NiMH cells will charge at 1/10C fixed current without any controls. With 2450mAh cells, you're talking about 245mA current limited charging. You can set up a cheap readily available voltage regulator like the LM7805 as a current regulator. Look at the datasheet for an example and instructions on choosing the resistor value that sets the current. If you use a more sophisticated charging IC, stick with the examples in the datasheet and application notes for the IC.

LoopyByteloose · 2012-11-14 02:44

I like NiMH, but please study the Wikipedia entry for them. They will behave nicely if charged at 1/10, but that doesn't mean that they will attain a full charge at 10 hours, they have to be on the charger about 150% or 15 hours to get a full charge.

And full charge requires a higher voltage than the rated output voltage, but NOT too high. So you have BOTH, regulate maximum voltage and regulate the charge rate. At least the current required is not too much for a wallwart in most cases. So this is very much a good DIY project.

http://en.wikipedia.org/wiki/Nickel-metal_hydride_battery

In addition to avoiding overcharging on the charger, you have to consider over-discharging when in use. That means a low voltage cut off of some sort.

Peter KG6LSE · 2012-11-14 07:58

NiCd is 1/10 C ..... NiMH is 1/20C
this comes from the datasheets of major brand NiMh cells ..

Attachment not found.

Please see page 8 section 1.6

Trickle charging cannot be used with Ni-MH batteries,
except specific high temperature batteries (please
contact Panasonic to get more information). However,
after applying a refresh charge using a rapid charge,
use a trickle charge of 0.033 CmA to 0.05 CmA. Also,
to avoid overcharging with trickle charge, which could
damage the cell characteristics, a timer measuring
the total charge time should be used.

Every batt app sheet is specing this .....
for a 10 AH pack this works out to 300 mA not the 1/10c of 1 amp . to 500 ma and if my math is right at 7 AM .5 Outta 10 is 1/20th . there for 20/C cause 1/20 of 10 AH is .5 AH

A proper NiMH charger is a set to do Delta V and Delta T every AA NiMH charger I have popped open had a thermistor that some how contacted the batt.

Off Wiki

Trickle charging

The simplest way to safely charge a NiMH cell is with a fixed low current, with or without a timer. Most manufacturers claim that overcharging is safe at very low currents, below 0.1 C (where C is the current equivalent to the capacity of the battery divided by one hour).[15] The Panasonic NiMH charging manual warns that overcharging for long enough can damage a battery and suggests limiting the total charging time to 10 to 20 hours.[14]

Duracell further suggests that, for applications where the battery must be kept in a fully charged state, a trickle charge at C/300 can be used.[15] Some chargers do this after the charge cycle, to offset the natural self-discharge rate of the battery. A similar approach is suggested by Energizer[16] which indicates that the self catalysis can recombine gas formed at the anode for charge rates up to C/10, but as this leads to cell heating, recommends C/30 or C/40 for indefinite applications where long life is important. This is the approach taken in maintained emergency light fittings (which in Europe must remain on for 4 hours in the event of a power cut) where the design remains essentially the same as that used in older NiCd units, but for an increase in the trickle charging resistor value. In comparison, NiCd cells can ordinarily be charged indefinitely at C/10 without damage.

Panasonic's handbook, however, recommends that NiMH batteries on standby are kept charged by a lower duty cycle approach, where a pulse of a higher current is used whenever the battery's voltage drops below 1.3 V. This can extend battery life and use less energy.

To Spec 1/10 C as a "dumb charge" with No feeback or timer Or controls .. may be still safe-ish but it is going to destroy the cells in a much faster fashon then a Real charger .........

This is Poor Engineering advice guys Serious !

On paper you can Do 1/10 C IF Mr end user plays the part of the TImer ..... but would you Really want to Sell or make a product that can go BOOM if the end user fails to do a action ........ can you say legal nightmare......

1/20C is the fastest you can charge with no feed back ... end of story .. and with that you are still not properly useing NiMh the correct way..

Peter..

Mike Green · 2012-11-14 08:02

Thanks Peter.

Peter KG6LSE · 2012-11-14 08:19

I Only point this out as I used a 10/C rate on a large pack NiMH of 28.8V at 10 AH on my E bike...

I needed to get a better charge for me to have enough juce to get to clases so I used a lab PSU a monster of a unit and did 1A Limit at 30V
I came back in 4 Hours and the lab was . Well Yea the packs were oozing hot glue and the smell was badd! a few poped cells .. I ended up just burriung the bike and its batt as one t in the snow out side I figure if it goes cato It will be in a 3 foot snow bank .

I ended up with a post mortim of the charger never went over 1A . the CC mode was still OK . It was Thermal runaway.
there goes $200 in NiMH ...... I ended up on SLA AGMs and that was that .

a simple thing is a crowbar circuit to lock out a hot pack . this way at least one cant just BLow a pack up !
a small opamp and a Mosfet would do welll.

back on topic-ish .

Switching regs are SOP here . unless I cant afford any noise and I need to have a very smooth output , U have adpoted SMPS as the way to go . no more huge heat sinks . the lower loses are a huge plus in my book .
Esp when a batt is involved and you need to milk every mAh out of it .

the Any Volt Micro is a usefull pre made switcher . not cheap ! but works wonders .

I have a student here at OIT whom is doing a Ardiono conmtrolled bike light.

He is :
taking a 6V Alt for a normal bike light .
a charge Reg .
then tons of relays Cause He wants to use the MCU to charge eatch Li Po cell on its own . . then that feeds some LEDs at 3.X volts.....

the relays can series and parallal the cells to slect and charge them one at a time and the MCU keeps track of them

( I would use a PICAXE for such a mundane use . )

the class he is in is for controll systems so he has to use the MCU some how.
...

In the Real world the best way in his case is to take the Bike gen and get the thing set up to take a 10 ish V out Gen

Buck it to 5V ( SMPS ) ( has wide input specs ) use the 5V on a proper LiPo charge chip // then have that run the LEDs

Peter

LoopyByteloose · 2012-11-14 08:48

@Peter
Well, I like NiMH because.........
They come in an AA form factor AND you can buy a charger with them that does it all quite nicely.

If you are using DIY chargers, I keep suggesting a thermal cutout on anything. Using a DS1620 chip (sold by Parallax and can be programed with a BasicStamp) that senses when temperatures are over 35 degrees C and shuts off the AC Mains via a solid state relay will eliminate a lot of hazards.

Charging at 1/20 implies that at 150% of charge, you have to wait 30 hours to go from empty to full.

Trickle charging implies that you are going to leave the batteries unattended and connected to AC Mains. Is that wise without a thermal cut out?

I have had Lead Acid gel cells melt and warp due to charger problems and no thermal cut out. I have had new Lithium Ion packs swell up like a pregnant guppy on their second charge that may have been a fire ball in another hour of charging. All the hazards have one indicator in common as early warning --- excessive heat.

Admittedly NiCds do offer some greater simplicity and are more tolerant of abuse. But NiMH are getting cell densities that are reaching very close to Lithium Ions and still have the AA form factor and consumer charger. There are a lot of good reasons to consider them in battery weight is a factor.

Can we say 1/15AH for conservative NiMH? It is a figure no one mentions, but it is in the middle.

rwgast_logicdesign · 2012-11-14 09:26

well i had planned on using this, or maybe whatever maxim recommends these days since its not recommended for new designs, still available at digikey if i cant get a sample

http://www.maximintegrated.com/datasheet/index.mvp/id/1666

it says it can do 16 cells in series i will be using 18. the thing that confuses me is i think it only outputs 100ma and 1/20 of 2450ma is 122.5. I also have 3 the 1620 chips, ive never used them so im not real sure how to physically mount it to the batteries. im assuming the ic needs to touch the battery pack.. if so ill be using 3 rows of 5 stacked togather, do i only need one chip or should i monitor both sides of the pack using two chips?

LoopyByteloose · 2012-11-14 11:05

Seems not to quite fit. It might be impossible to find something for 18 cells, can you get by with only 16? Less amps output means a bit slower charge. Not sure how that might play out.

But I have notices that an awful lot of this battery charging chips are NOT recommended for new designs and nothing new is available. It seems that the chips didn't quite work as expected. The fact that 16 or 18 cells are in series makes for a lot of variablity that the logic cannot control.

In "theory" all the batteries would be perfectly matched by the manufacturer to assure a balanced charge, but some of the stuff I have read about lithium cells is that they are not easy to match. That may be true of just about any high-tech battery.

Earlier this year I read tons of stuff on how to ideally balance charges in 12 cells and above. It just is not that easy. I wonder how the Tesla auto does it. I think they use NiMH. I think I have read that they actually provide liquid cooling as part of their balancing system.

Yes, the DS1620 would need physical contact with the battery or the battery pack. With 18 batteries you would either need a lot of them or some sort of aluminum heat sink that touches all the batteries.

Of course, you could just get 18 10K thermistors and use 18 ADC inputs to track the temperature of each cell during charging individually. That would be better. I just use the DS1620 with commercial battery packs. And the temperature tigger is quite conservative.

This is a very ambitious DIY project.

rwgast_logicdesign · 2012-11-14 11:17

no no dude im running 18v 15 cells, the mah out is what im concerned about not sure if im reading the data sheet correct but i think it can only output 100mA, which is strange becuase it can do up to 16 cells...

I had also thought about using a heat sink or a thin piece of metal but that might be very accurate and does not account for the middle cells, would the a super high quality 15 thermoresistor or even a 1620 be needed with one of these charge ICs, I thought the point was they managed everything by reading voltage and current. Unfourtantly Im probably going to have to stack three rows of batteries for anything manageable. The reason I asked if 1amp was output on my wart was ok is becuase when I look at consumer nimh chargers they have a 1.8amp output mode for packs over 1500 or 2000mah not sure which but you get the point.

LoopyByteloose · 2012-11-14 11:21

Not really strange when you think about transistors. They all can generallly handle an easy 40V, it is the amps that cause them to get bigger and bigger and eventuallly fail.

So, this is a chip. It cannot disappate much heat in such a tiny package 100ma and many volts is about right.

rwgast_logicdesign · 2012-11-14 11:23

Plz read edit in last post.

Ok so do you think this chip will work or is this not the best solution?

LoopyByteloose · 2012-11-14 11:29

I am sorry, but I distrust all of these recent charging chips. And the fact that they were discontinued seems to indicate they were a big failure, maybe a big disaster for the makers.

The simple fact is that battery charging is about controlling the peak Voltage and the rate of charge in Amps. Both are analog control and the digital logic is an add on feature.

The best I could figure, it you just can't make a chip to do it all and to do it all well.

That PDF doesn't make much sense to me as it is limited to about 12V at 100ma. That isn't going to charge 16 cells at 1/4C.

If you do find a chip, make sure if is one that write a PDF that makes clear sense.

LoopyByteloose · 2012-11-14 11:40

You could use all 3 of the DS1620 chip and have the batteries grouped into 3 groups. The logic from the DS1620 could be wired to have any one of them trigger the solid state relay. That would be a better fail safe set up than just one. Any one of them could locate a hot spot.

The think about charging in series is the cell at the highest end of the chain is likely to get worked a lot harder than the others. When a sudden demand for voltage is created, it get most of the demand, the next one is a bit less worked and the resistance in that first cell protects and on down the line.

And when you go to charge, it is again pushing everything along and most likely gets more wear. More wear is more heat, so it may run hotter and fail sooner. Most of these newer batteries were developed for low voltage arrangements, but now with want high voltages and upset them as they are more fragile than good old lead acid.

Peter seems to thing NiCd tolerate more abuse as well, but I am not sure.

stepping voltage up or down and its toll on batteries

Comments