Dan's Data letters #131
Publication date: 16-Nov-2004.Last modified 28-Apr-2012.
Frequency discrimination
A lot of speakers come equipped for bi-amping or bi-wiring. Okay, I can kind of see the use with bi-amping. After all, by using separate amps for mids and highs, you take some of the load off of each one. But what about bi-wiring? The claim is that it helps audio clarity by having a separate wire going to each of the terminals. Yet they're still coming off the same terminals at the amplifier end. The only difference is the point of connection at the speaker end. Is there any real benefit from doing so?
Grey
Answer:
None that makes any sense. Oh, you can get a bit less resistance between the amp and the speakers, but the difference
between bi-wiring with the chunkiest wire that fits and single-wiring with the same wire is likely to be zillionths
of an ohm, so that idea doesn't really hold up.
Some audiophiles insist that bi-wiring with big fat wire to carry the sluggish overweight tobacco-chewing bass and delicate cobweb-like wire to carry the twinkling effeminate treble makes a difference. They then usually either just insist that they can hear it clearly, the Bible says it, they believe it and that settles it, or speak the usual incantations about conductor count and "skin effect" and propagation delay and so on. You can't argue with the first justification, but the laws o' physics are pretty definite about the second one; the phenomena the bi-wirers are talking about are either nonexistent, or exist only at much higher than audio frequencies and/or with much longer cables than they're using.
As pointless audiophile tweaks espoused by people who wouldn't know a blind test if it hit them go, though, bi-wiring is both harmless and cheap (well, as cheap as the cable you use, anyway...). The only down side, besides cable cost, is that the extra wires can be ugly.
Brass monkey weather
Where I live (Moose Jaw, Saskatchewan, Canada) we regularly experience a phenomenon that Aussies probably have a hard time wrapping their minds around. For the sake of discussion let's drop the usual terminology ("it's f*cking cold outside") and use a shorter form: "winter".
Average temperatures (Celsius) for the winter months are
Oct 6
Nov -4
Dec -11
Jan -13
Feb -10
Mar -3
Apr 5What the averages conceal, of course, are the extreme temperatures; cold snaps where it is below -30°C for a couple of weeks, and warm spells of snow melting weather in the middle of an otherwise frigid month.
My little (about 7 cubic feet) freezer keeps my goodies at -19°C, so there should be significant power savings to be had by taking it out of my apartment and putting it in my unheated storage room for the winter months. Assuming, of course, that I can rely on it to survive the cold snaps so that it kicks in as necessary when a warm spell comes along.
I have talked to several appliance manufacturers, including the manufacturer of my freezer, and I can't get a simple "safe" or "not safe" answer from them. All I get is "our product is designed to be used in a house or apartment at normal room temperatures".
Rob
Answer:
Normal consumer fridges and freezers are all straightforward
phase-change heat pumps, but the parts they use
do indeed assume they'll be running at "normal room temperature" on the outside. This means that at (very) low temperatures
you may have problems like plastic going brittle and bearings seizing and so on, but that's not a big deal.
The big
deal is that the compressor isn't designed to pump liquid. If the ambient temperature is low enough that the refrigerant
on (what's meant to be) the high-pressure side of the compressor
liquefies, and if the compressor then clicks on (which it shouldn't, of course, when Mother Nature's keeping everything
nice and cold by herself, but stranger things have happened), it may smash its little valves.
(If the compressor sensibly doesn't bother running, though, this problem will not
arise.)
More probably, it's apparently possible for some refrigerant to condense into the oil in the cooling pipes (these systems all have lubricant that circulates along with the refrigerant), which can create a frothing problem when the compressor starts, and that can result in liquid oil getting into the compressor cylinder and breaking the valves. Modern fridge and freezer compressors are, famously, practically immortal, as long as you don't leave the fridge door open and force the compressor to run non-stop all day. But if there's liquid in the cylinder, all bets are off.
If you've got a freezer that uses the most popular non-CFC refrigerant, HFC-134a, then you've got refrigerant with a native boiling point of -26.5°C. It's pressurised inside the cooling system, though, so its actual boiling point will be higher; I don't know how much higher, but it's entirely possible that most of it will be liquid on many of your Saskatchewan winter nights.
As long as the whole cooling system's warmed up before the thermostat trips the compressor back into action, though - which it really ought to have, since the temperature inside the freezer will of course lag the temperature outside when the sun comes back up - everything should be OK.
I think liquid-in-the-compressor problems only commonly arise in things like split-system air conditioners, where the evaporator and compressor are separated, one indoors and one out, and can therefore be at quite different temperatures. Fridges and freezers should be OK. I think.
Energy is currency
I'm looking to charge batteries via a motor. Say I attach a motor and battery to my bicycle in order to make a motor assisted bicycle (pretend like I already have one and it works). Is it feasible to come up with a circuit to charge the battery for use when I don't need the motor (e.g. flat ground & downhill)? I'm not familiar with battery charging circuits at all, but I'd like to come up with a way to increase the time between charges. Ideally, I'd like to make it so that I'd never need to recharge the battery, but certainly increasing time between charges would be ideal.
More importantly, I want to know whether this design could hurt the battery at all.
Steve
Answer:
What you're talking about is "regenerative braking".
There's no way to get power from the spinning wheels without
braking them, so it's not something you'd want to do while freewheeling on flat ground, but generally speaking regenerative
braking is a brilliant idea, since it re-uses energy that'd otherwise be lost making ordinary brakes hot.
The problem is actually making it work.
Batteries, you see, can't accept charge current very quickly. If you use your drive motor as a dynamo when you're braking you can get a lot of power out of it, but you just can't sink that power into the batteries fast enough. A hard charge for a NiMH battery pack takes 60 minutes; at that rate a ten-second braking event will only let you squeeze a third of one per cent of the battery capacity into the pack. For brief charging events you can push some battery technologies very hard indeed (as long as they're not close to full already), but you're still going to have a hard time getting a worthwhile amount of energy into a battery in the time you spend slowing down.
For this reason, regenerative braking systems in electric cars commonly only add a few per cent to their range. People who do their own electric car conversions usually ignore regenerative braking altogether, because it's not nearly worth the effort.
The only solution to this problem, in the absence of batteries that can accept really hard charge pulses at random moments, is a "power buffer" between the batteries and the motor. A bank of "ultracapacitors" would do the job nicely, and might even be affordable, but they'll require significant extra electronics to support them.
Let's say you're working with a 24-volt motor system - 12-volt is simpler for hobbyists, but 24V keeps the current down for a given power level. The "ultracapacitors" you'd want to use as a power bank only have low voltage ratings, so you'd need to string, say, 14 2.5V caps together in series to make a unit with a 35-volt rating, which should be enough to handle electromagnetic braking from a 24-volt motor system (just in case it outputs rather more than its rated voltage in brake mode, which it probably will; note that high-efficiency brushless drive systems are often very poorly suited to regenerative braking applications, but good old brush motors can be used that way pretty easily).
If you use gigantic 2500-farad caps and charge 'em to 24 volts, you'll be storing 109,375 joules of energy, because the energy stored in a capacitor, in joules, is equal to 0.5*C*V^2, where C is the capacitance in farads and V is the voltage in volts. Each cap in this string therefore gives you 7812.5 joules of storage (in Physics Experiment Land where everything adheres perfectly to its specifications).
2500 farads is an astoundingly high rating for a cap. I just fished an 80-volt electrolytic capacitor slightly larger than a C battery out of one of my widget drawers, and it only has a 3.3 millifarad rating, for a total energy storage capacity of less than 11 joules. No "true" capacitor can come anywhere near the rating of the biggest modern ultracaps without being at least the size of a closet.
Ultracaps manage it by, essentially, using a combination of true-capacitor electrostatic energy storage and battery-like electrochemical storage. Thanks to this, you can get a very useful amount of energy. A joule is a watt-second, so in theory a 109,000-joule-ish reserve can drive a respectable 300-watt bike-assist motor (a typical regular cyclist will have a hard time delivering 200 watts continuously; an Olympic distance cyclist can do 500) for around six minutes.
In practice, though, because these are capacitors, their output voltage will fall as they discharge. So they're not nearly as useful as they seem. You'd need a pretty humungous step-up DC-to-DC converter to use them as a drive source interchangeable with the battery, though a smart motor controller could adapt for falling supply voltage on the fly to mitigate the problem, whenever the user doesn't want full power from the motor. Switching in and out the battery supply would be a further hassle, but all of that stuff can be done in solid state these days without too much heatsinking thanks to the wonders of reliable switchmode power supplies.
The down side? Those huge caps cost something like $US250 each even for volume orders, so 14 of them will set you back $US3500 or so. Twice that if you order whatever the manufacturers consider to be a "small" number of the things.
Oh, and note that shorting out a fully charged 109 kilojoule cap bank would be a quite cataclysmic event - maybe 20% of the power of a hand grenade. Shorted ultracaps can give up energy very quickly, though they don't enjoy doing it nearly as much as true caps do.
More practically, (relatively) little 350 farad capacitors can be ordered online for a mere $US30 or so each, depending on quantity; $US420-odd for a pack that gives you better than 15,000 joule capacity, which would still be useful. Now the total stored energy is only about as much as a 50 calibre bullet carries out of the barrel of a Ma Deuce, which I'm sure you find comforting.
You'd need a motor and drive systems geek to work out the details here, and there may of course be some basic problem that I'm missing, but the numbers look good.
Alternatively, if you just want a dynamo that can (relatively) slowly charge the batteries while you pedal, that'd be simpler, but you'd probably find the benefit to be pretty unexciting. You might, instead, like to look into packing a folding solar panel on the back of the bike, and using some off-the-shelf solar charging gear to hook it up to your battery when the bike's parked (in a secure area...). That won't give you a whole bunch of extra range either (unless you tote around a panel the size of a surfboard), but it also won't require a ton of engineering to make it work.
Note that if you implement any kind of regenerative braking system, no matter how good, you also must have a fallback failsafe mechanical brake, for emergency stops and for whenever the regenerative cache is full.
Find some EV geeks talking about capacitor issues, and alternatives, here.
School toys
I'm in my final HSC year at school and one of my subjects is software design, for which we have a year-long major design project. I've decided to make a robot/small machine/thing that moves and is funky, with the control software being the project itself - but I still need to sort out the hardware.
I'll be using a Zilog Z8 Encore! prototyping board to control everything (programmable in C, with a multitude of IO options), and Dad will be helping me out making the framework (from aluminium, 'cos it's light). But I still need some kind of drive mechanism. The best option I think are tracks, and I thought the kind you might find on R/C model tanks would be best. I'll also need two 12VDC motors and some kind of gearbox I presume, but I'm not too sure about where to go.
The control board is 160 by 180mm, so tracks for a model about 200-300mm long would be appropriate, probably. Any advice you have on what/where to buy would be much appreciated.
Eddy
Answer:
Making your own track-drive chassis just for fun is, of course, an option, but I think you'd do better to buy the
cheap plastic rubber-track R/C tank of your choice - anything big enough to support the electronics would do - and
cut it down to suit. The
going rate on eBay for clones of the Marui tanks I reviewed ages
ago is now around $US30, and you shouldn't pay too much more for shipping to Australia if you find one from
a Hong Kong dealer.
Bigger chassis are more difficult. Doyusha made an excellent not-too-expensive Abrams, but you can't get it any more; apparently this cheap Hobbyzone tank is the same thing, but good luck getting one to Australia for a reasonable price.
(A reader's now pointed out this dealer to me. Australian, and cheap!)
21st Century Toys also make a HUGE M5 Stuart model which is highly hackable and can be found cheap on eBay, but you're again looking at scary shipping fees.
It doesn't have to be tracks, of course. Hit the Trading Post, eBay and garage sales for people's old Tamiya monster trucks and you may be able to find a half-broken Blackfoot (2WD) or Clod Buster/Bullhead (4WD, twin motor, surprisingly large if you've never seen one before) or something for close to no money at all. At 1:16th model tank speeds, there are few places a tank can go that a Clod Buster can't, and plenty of situations in which the Clod's giant tyres make it much more capable - and a monster truck chassis has much better efficiency and reliability than track drive.
Perpetual promotion
Dan, I saw this "free energy machine" on the news page of Overclockers Australia, and immediately thought of you. It looked almost legit until I saw the FAQ page, with its plethora of "We'll let you know soon" answers. Your thoughts?
Luke
Answer:
Been there, done that.
Their updates page now makes clear that the company's keeping itself afloat by periodically selling shares. To suckers.
I'll be overjoyed if the thing finally turns out to work. I'll be less surprised, however, if George W. Bush comes out as a proud atheist.
