kjb@cs.brown.edu (Ken Basye) (03/15/90)
I'm in need of answers to a few questions about batteries. I don't
know if there's enough general interest in either of these groups to
justify using follow-ups; I'll leave that to the individual
responders. If there is interest, I'll summarize all the e-mail
responses I get.
To introduce the problem, I'm working in mobile robotics, and looking
into various ways to power some VME boards with batteries. The
manufacturers of these boards (reasonably enough) aren't designing
them with mobile robots in mind, so they aren't too concerned about
power. They also generally use serial and video standards which
require +/- 12V in addition to +5V for logic. Between a Moto 147SA-1,
3 vision processing boards from Itex, and a few other small devices,
the approximate power requirements as taken from the manufacturers
specs are:
+5V - ~25 Amps = 125W
+12V - ~3 Amps = 36W
-12V - ~1 Amps = 12W
We may also be powering some other small devices and motors as well,
so I'd prefer to have more available when it's needed. We'd like to
ba able to run all this for at least a couple of hours before
recharging, and it would be OK if the charging took overnight (We're doing
research, not making a product someone would want to take home).
Our robots are built on bases that are 12" in diameter and are rated
to carry about 45 lbs, although we have put as much as 80 lbs on them,
at which point they have problems going over small bumps in the floor.
We may also be getting some larger bases with much larger payloads,
perhaps capable of carrying a standard-size deep-cycle marine battery.
Ok, now on with the questions:
1) Assuming that we use high-power, triple-output DC-DC converters to
supply the 3 required voltages, is using higher voltages (say 48V as
opposed to 12V) a big win? If so, is putting 4 12V batteries in
series a reasonable way to do this? Is there a better way to provide
3 different voltages?
2) Manufacturers of Gel-Cell lead-acid batteries (like Panasonic) rate
their batteries by amp-hours, but using a rate of discharge that
drains the battery in 20 hours. Thus, a battery with a 10 Ah rating
can be expected to produce .5 amps for 20 hours. What's a good rule
of thumb for converting such a rating to higher discharge rates? My
experience has been that if you draw 10 amps from such a battery, it
won't last anywhere near an hour, and may not even be able to hold its
rated voltage. Given that the Ah rating over 20 hours gives a
reasonable current, how far above that current can you go and still
hold rated voltage?
3) I've found 12V batteries for wheelchairs that come in two flavors.
The sealed gel-cells are rated at 30Ah, while the wet batteries of
about the same size and weight and are rated at 23A for 54 minutes.
Does this mean that wet cells are better at providing higher current
over shorter times?
4) If you've read this far, perhaps you're interested or altruistic
enough to say something about how you might tackle this problem?
Thanks very much,
Ken
Internet/CSnet kjb@cs.brown.edu U.S. MAIL Ken Basye
UUCP uunet!brunix!kjb Box 1910
Dept. of Computer Science
Brown University
Providence, RI 02912dick@hpcilzb.HP.COM (Dick Lucas) (03/16/90)
Why not use deep cycle batteries for the +/- 12 volts? (one for each polarity) For the 5 volt supply you can still buy 6 volt auto batteries. Most 5 volt requirements have a 10% tolerance on them and you could drop the required voltage with a large diode. This may not be a very elegant solution, but it would be cheap and easy to accomplish. If you assume a full volt drop across the diode (or two) the efficiency of this method is 83% which compares favorably with most switch mode power supplies.
raoul@eplunix.UUCP (Nico Garcia) (03/16/90)
In article <32769@brunix.UUCP>, kjb@cs.brown.edu (Ken Basye) writes: > > +5V - ~25 Amps = 125W > +12V - ~3 Amps = 36W > -12V - ~1 Amps = 12W > > We may also be powering some other small devices and motors as well, > so I'd prefer to have more available when it's needed. We'd like to > ba able to run all this for at least a couple of hours before > recharging, and it would be OK if the charging took overnight (We're doing > research, not making a product someone would want to take home). OK, these are your keys. Since you need a couple of hours use, and you probably want your voltages to remain high, and your batteries to be multiply rechargable, and to be idiot proof (for lab safety) gel cells are probably the way to go. But, geez, 25 amps at 5V? For two hours? That's 50 amp-hours times a fudge factor of three to prevent deep discharging your cells. That's 150 amp-hours in a, say, 6V battery. *ouch* Yuasa makes a 10 A-hour 6 Volt sealed lead-acid. Using that as a standard, they're 5 lbs. a piece, at 6" by 2" by 4". That's 75 lbs., 720 cubic inch. Your other requirements pale in comparison to this. Your 12 Volt supply can use only 18 A-hours, and your -12 can use 6 A-hours. Yuasa makes a 12-Volt 24 A-hour one, 7" by 5" by 7", 17 lbs. And a 12-Volt 6 A-hour one that's 4" by 6" by 3", 5.28 lbs. You probably want to actually use batteries rather than DC-DC converters. They're noisy and generate considerable heat, and they are not efficient. Anybody heard of one with more than 50% efficiency? Save yourself the trouble: use separate battery systems. -- Nico Garcia Designs by Geniuses for use by Idiots eplunix!cirl!raoul@eddie.mit.edu
mikec@wheaties.ai.mit.edu (Mike E. Ciholas) (03/16/90)
In article <861@eplunix.UUCP> raoul@eplunix.UUCP (Nico Garcia) writes: >In article <32769@brunix.UUCP>, kjb@cs.brown.edu (Ken Basye) writes: >> >> +5V - ~25 Amps = 125W >> +12V - ~3 Amps = 36W >> -12V - ~1 Amps = 12W >> >You probably want to actually use batteries rather than DC-DC converters. >They're noisy and generate considerable heat, and they are not efficient. >Anybody heard of one with more than 50% efficiency? Save yourself the trouble: >use separate battery systems. > Well, let's see... I can get a DC-DC converter from Mechler (Ashland, MA) that takes anywhere from 8 to 80 volts, produces 5V at 25A, and does this with a minimum of 82% and as high as 94% efficiency. It does not generate any audible noise and has only 30mv P-P ripple (electrical noise). As it is high efficiency, I cannot see how it will generate lots of heat. I do NOT reccomend using separate batteries for each voltage required. First, some sort of voltage regulation will be required to maintain the proper voltage. You will need a battery voltage that is a minimum of 2 volts higher than the desired voltage for a linear regulator (and if you use a switching regulator, you are using a DC-DC converter without knowing it). At 5 volts, this ammounts to only 71% efficient. Second, as the batteries will discharge unevenly, the system may be subject to complete absence of one of the voltages possibly resulting in system failure or destruction. Third, if you have multiple batteries, they will never undergo uniform discharge/charge cycles with the result that one battery may wear out before the others. Fourth, charging multiple batteries can be a pain. With one battery you need only one charger. Mike Ciholas email: mikec@ai.mit.edu snail: 289 Highland Ave #108/Somerville, MA 02144 phone: (617) 623 3563 air: N1909C, 1954 Cessna 170B
raoul@eplunix.UUCP (Nico Garcia) (03/17/90)
In article <7308@rice-chex.ai.mit.edu>, mikec@wheaties.ai.mit.edu (Mike E. Ciholas) writes: > In article <861@eplunix.UUCP> raoul@eplunix.UUCP (Nico Garcia) writes: > >You probably want to actually use batteries rather than DC-DC converters. > >They're noisy and generate considerable heat, and they are not efficient. > >Anybody heard of one with more than 50% efficiency? Save yourself the trouble: > >use separate battery systems. > > > > I can get a DC-DC converter from Mechler (Ashland, MA) that takes anywhere > from 8 to 80 volts, produces 5V at 25A, and does this with a minimum of > 82% and as high as 94% efficiency. It does not generate any audible noise > and has only 30mv P-P ripple (electrical noise). As it is high efficiency, > I cannot see how it will generate lots of heat. Hmmm, then I take it back. When I looked at DC-DC converters, I couldn't find anything with the efficiencies or noise limits I needed. On the other hand, it was for sensitive multi-channel medical equipment, so they also had to be really small and almost utterly noise free. (1 mV was too much.) > proper voltage. You will need a battery voltage that is a minimum of > 2 volts higher than the desired voltage for a linear regulator (and if > you use a switching regulator, you are using a DC-DC converter without > knowing it). At 5 volts, this ammounts to only 71% efficient. Second, How much regulation is needed? For the 5 Volt digital supply, he may need that much regulation and the resultant power losses. For the +/- 12 Volt supply, I suspect precise regulation to be unnecessary. So the batteries would still work there. Your reliability and failure concerns are valid, but I trust batteries more than I trust DC-DC converters. It takes a *lot* to fry a gel-cell, and I've seen some DC-DC's destroyed in the lab by accidents. So what about using a lot of 12V batteries to power the 5V DC-DC and maybe a 12V DC-DC, or use one of the 12V units reversed for the -12V? You'd still only need a 12V or 15V power supply to recharge them. -- Nico Garcia Designs by Geniuses for use by Idiots eplunix!cirl!raoul@eddie.mit.edu