Trinary Hard Drive

• 05-08-2005
nickname_changed
Trinary Hard Drive
(My understanding of the technical side of things may be wrong, if so please correct me. But the principles should still remain)

We all know magnetic hard drives are made up of lots of little switches that can be on or off. I think on is north and off is south or something like that. Given this, an 80GB hard drive would have 687,194,767,360 of these switches. Each switch is a bit.

I believe the reason we are limited to just using north and south is because it is the simplest way to measure mechanically.

But I was thinking. What if I invented the Trinary Hard Drive? Instead of using bits (binary), it would store data as tits. Each switch would either be magnetised north, south, or not magnetised at all. Rather than being limited to two states (0 and 1) it would have 3 (0, 1 and 2). The little pin thingy that reads each switch would have to be a little more accurate, but the tradeoffs in size would be worth it.

Lets say you have 8 bits, each with 2 values. Thats 2^8 possible values, or anywhere between 0 and 255. But using tits, the possible values would be up to 3^8, or anywhere between 0 and 6,560. Thats a huge space increase using the same number of physical switches, simply because each has 3 states rather than 2.

Now back to the original. I said before that a conventional 80GB hard drive would have 687,194,767,360 (80x1024x1024x1024x2) switches or bits. But using tits, this storage space would become
3^687,194,767,360, or <some number the windows calculator won't go up to> possible values. A huge storage space increase!

I believe the same system would work for RAM. A positive electrical charge, a negative, and no charge.

Granted for simple true or false operations this system might be a little slower. But for large operations, say reading a big file, it would be much faster because the mechanical head wouldn't have to read as many switches.

What do you think?
• 05-08-2005
JaWiB
I don't know...is it even possible for a switch to not be magnetised at all?
• 05-08-2005
nickname_changed
What if it was equally north and south?
• 05-08-2005
Thantos
I think there is a physical limitation involving the surrounding bits. Their magnatic value would affect that "neutral" bit in such a way that I don't think you could get a balanced charge.

IIRC from my days looking at RAM we currently don't use a positive charge and a negative charge but a charge or no charge.

Trinary devices are being developed by keeping better control of the voltage and more sensentive readers. So that instead of 0 - ~3 v being off and 5volts being on we'll have something like 0 - 3volts off, 4 - 6 volts on1, 8+ volts on2 (numbers made up for this paragraph)
• 05-08-2005
major_small
you may want to look into quantum computing
• 05-09-2005
SMurf
Quote:

Originally Posted by stovellp
tits

Once again, I feel the moral fibre of this board slipping down the drain... :rolleyes:

Yes, it's possible, but it would require a radical modification of the current drive head mechanisms in use, which would make them rather pricey for a few years. Personally I think it's a miracle that hard drives work in the first place, why we can't focus on more reasonable solid state provision I don't know.
• 05-09-2005
BMJ
Quote:

Originally Posted by major_small
you may want to look into quantum computing

Bingo.
• 05-10-2005
WDT
Well the idea, in principle, is sound and it 's kind of the basis of Quantum computing. (devising a 3rd state to increase the power of computing) but the problems, to summarise, I believe are as follows (i'm just gathering the sum of all posted knowledge before this and putting it here):
1st there is the problem with the residual charge as mentioned by Thantos.

2nd you could get around this with the "3 charge" states but again that would not just require a slight overhaul of the current architecture, (as you'd have to device the mechanism for reading/measuring the charge. This would also add some computational delay) but why stop there? why not carry on till the law of diminishing returns comes into effect?

there's more I have to say but I have to go. But I hope I explained it as I percieve it clearly
• 05-10-2005
Quantum1024
Well you'd need to come up with a new name for a trinary digit, a tribit perhaps?
I think computer memory would still be measured in bits though for compatibility.
• 05-10-2005
jverkoey
There was an ad I saw a while ago, a flash animation, not sure if it was bogus or not but it had a cool music video and everything.

Anyways, it talks about how instead of having the bits laying flat, they're vertically placed, allowing a lot more space to be fit somehow....not sure how that works *shrugs*
• 05-10-2005
anonytmouse
>> Anyways, it talks about how instead of having the bits laying flat, they're vertically placed, allowing a lot more space to be fit somehow....not sure how that works *shrugs* <<

That sounds like 'perpendicular recording'.

base 3 computer???
• 05-10-2005
sand_man
even if stovellp's idea was possible wouldn't that cause major compatibility problems? Besides that, hex numbering system would be useless
• 05-12-2005
Waldo2k2
the whole reason we use base 2 is because that can be translated to electronically on or off, that translates down to "electronic gates" being open or closed on the processor allowing for boolean logic and mathematical operations. A third bit would be of no use, the entire concept of a processor would have to be redone from scratch.

There are much better ways to get increased storage...like moving away from magnetic storage completely and immersing the market in non-volatile memory storage (thumbdrives, etc.), which is already beginning to happen. Think about the speed possible when your hard drive is just another memory simm? No mechanical parts, no cables, no failures, just raw speed. That's the future right there.
• 05-13-2005
SMurf
Well the thing about flash memory at the moment is that you can only write back to it a certain number of times (Usually in the ,000s). I had a look for the price of some powered RAM in the PCMCIA form factor and it was like £80 for 16MB, dunno if that comes with an ATA interface either, but not cheap. :eek:
• 05-13-2005
Waldo2k2
I think the number is much higher than the thousands...and trust me, it will get cheaper (you can buy a 512mb jumpdrive for ~\$45USD, a year and a half ago a 256mb jumpdrive went for ~\$150USD). At that rate we will be in the multiple gigs for under \$100USD in a few years.