Is this a mad idea for a low profile HCPC case?

[RichardH]

Idle musings of a tinkerer...
How about splitting the PC into two - one part with PSU, Mobo, etc in a bog standard case, and then a small "break out box" with just DVD drive, an IR receiver, power/reset switches plus LEDs, maybe a USB hub. You could do it in something the width of the DVD drive and twice as high. All you'd need is an IDE cable plus power and switch cables. Wouldn't that sit nicely in an A/V rack?

Rest of the box could be well hidden away to keep the noise down, reduce visual impact etc.

....Just a thought - anyone think of a reason why not?

 

[buns]

If you could get some nice shielded cable..... i imagine any length of the ide cable would pick up a lot of interference.

How about using an external dvd and IR reciever...... it would be neater still

Ad

 

[Sinzer]

IDE cables (dependant on the speed) should be around 10 - 18 inches.

Some places will produce you a cable that is 48 inches. Not really good enough to space the two components apart.

You could do it with a firewire DVD Drive with 4 meter cable.

Why have an IR controller? Just use an RF keyboard + mouse, you then don't even have to be in the same room

 

[buns]

You would want the IR controller because it means you dont need a keyboard and mouse..... you can control it all from pronto like a consumer system. The IR reciever doesnt need to be in the same room as the pc. Doing it this way makes for a more user friendly solution

ad

 

[[email protected]]

If you wanted to split it, why not look at SCSI or Firewire for the drive cage. After all a lot of servers have the drives in a seperate box to the main system.

THis would get round the limitations of IDE over distance and it would be easier to get the cables.

 

[buns]

or you could stick to IDE (which i assume to be cheapest) and build yourself a well shielded cable like stock code 363-733 at http://rswww.com

ad

 

[Lem]

Serial ATA should allow for longer runs too, and the cables would be a lot narrower. Would be a cheaper option to Firewire when the manufacturers get round to knocking them out!

 

[RichardH]

Must admit I'm not currently seriously thinking of doing this, it just occurred to be when tinkering yesterday. I thought it might be a nice solution, even for a discrete desktop PC.
The max length of cables was something I hadn't considered - sounds like waiting for Firewire might be the only way. I'd just thought about using some of the low profile IDE cables some use to maximise air flow.

Of course the enclosure for something like this would need to be custom made too, which adds to the cost. Could simply use an external DVD drive, but I think it would be nice to have power switches etc available easily, so basically you never have to scrabble around for the main box. Mind you, there are lots of microATX bits around for this kind of think, I suppose. Maybe it's just a silly idea.....

 

[buns]

I like the idea to be honest with you! It is very much along the lines of the fancy automated homes that orange and their likes do. You could have a main machine hidden away somewhere, unheard and unseen with all of the noisy fans and stuff, then you could have a small and quiet machine where you need it!

ad

 

[Sinzer]

What you could do, which would be amazing. Would be to house the PSU seperately, then run powerleads to a custom micro ATX mobo, have a low end CPU with minimal fan use. Buy a gig of flash RAM and embed the OS on there with minimal features and have a DVD drive.

The primary cause of noise is usually PSU and Hard-drive. Remove those two and noise would be non-existant. An embedded OS would be excellent, no time to wait for loading and super fast operations. Quiet expensive though and would only be suitable.

Other than that you could make a custom chipset (fortunately for me I have a friend who is a chipset designer and if pushed would most probably be able to create an embedded system... hmm idea).

 

[RichardH]

Presumeably the IDE cable length issue is down to speed of getting dat back and forth, rather than losses?

 

[Stuart Wright]

Originally posted by RichardH
Presumeably the IDE cable length issue is down to speed of getting dat back and forth, rather than losses?

Yes - and because IDE is relatively slow - I think about an inch a second (worse if the data has to flow up hill) - then 48 inches or more will mean a mean time of upto and even over a minute before your data is retrieved.

 

[Lem]

Beaten you it to.. I have a Via Eden C800 Integrated mini-itx board with a Disk-On-Chip system running a cut-down linux client/server. Not quite sure what I plan to do with it yet..!

 

[Sinzer]

Have you got a bunch of programs that run under Linux, which will do the same as Dscaler, Powerstrip etc?

I would be interested in the details, I am thinking of at least designing something for myself.

 

[buns]

Originally posted by Spectre


Yes - and because IDE is relatively slow - I think about an inch a second (worse if the data has to flow up hill) - then 48 inches or more will mean a mean time of upto and even over a minute before your data is retrieved.

I am no one to question this.....but it does confuse me! What is an ide signal comoposed of? How come it takes so long!? What is different about IDE that allows the flow such minimal speed in comparison to other system?

or am i being dopey and taking a non-serious coment, seriously?

Ad

 

[Sinzer]

http://www.a1-electronics.co.uk/PcHardware/IDECables.shtml

This is quite a straight-forward explanation

 

[Stuart Wright]

Originally posted by buns


I am no one to question this.....but it does confuse me! What is an ide signal comoposed of? How come it takes so long!? What is different about IDE that allows the flow such minimal speed in comparison to other system?

or am i being dopey and taking a non-serious coment, seriously?

Ad

I was taking the mickey out of the previous poster since electricity travels at the speed of light, their post about speed was clearly rediculous. Yes it was a joke.

 

[Kramer]

Originally posted by Spectre
since electricity travels at the speed of light

It's no where near the speed of light (but probably faster than me on my moped).

But seriously, better to have all cables (inc. spkr, S-vid, VGA, power etc..) horizontal then - avoid all that "extra resistance" from current flowing up hill.

 

[Sinzer]

The length of the cable does adversely affect speed though, as when the signal is interrupted all the time and data is lost, then data has to be resent. This is the principle for all data that has error correction, TCP/IP etc. So both explanations are true.

Not to the extreme that you mention nor is it affected by gravity, but to take the pee out of a previous poster for a factor, which is true, is a bit lame

 

[Kramer]

Originally posted by Sinzer
The length of the cable does adversely affect speed though, as when the signal is interrupted all the time and data is lost, then data has to be resent.

So why would a slightly longer cable result in the signal being "interrupted" all the time?

"Speed" is of no significance in this regard.

Obviously longer cables can/will result in signal LOSS, DEGRADATION & INTERFERENCE.

Speed losses, to coin a phrase, would be "below a measurable limit".

But gravity, now this is an important factor in all regards.

We can't escape it - it's everywhere man

 

[buns]

looks like i look silly once again!

I'd also like to see electricity travelling at the speed of light! I hazard that if spectre does such a thing, we should ban him very quickly for such universal catasrophe!

What actually is the biggest difficulty with IDE? if firewire and such are a superior solution outside the box, why arent they used internaly?

ad

 

[Sinzer]

Hmmmm, like most PC arguments it is a question of pedantics.

Interruption is meant in the same sense of loss, degradation and interference. In all cases the dataflow is broken, I should have used dataflow instead of signal, as technically the interrupt would normally refer to a break in execution due to a system interrupt.

And yes the physical speed of the signal is not adversely affected to the point of noticing, however I was referring to the speed of the program/data you are executing/accessing. If you are continually resending data, then the program at the user end will appear to be running slower.

Bad definitions on my part.

 

[RichardH]

Time to resurrect this, with the advent of USB2 to IDE converters.

How about this:

A "breakout box" for the desktop, which houses a DVD ROM, a couple of USB ports, power and reset switches, LEDs, and maybe mic and headphone sockets.

Use a USB2 > IDE converter cable
to connect the IDE DVD rom in the breakout box to the main PC. This overcomes the length limitation caused by IDE.

With those front panel USB ports, which run on flyleads from either the motherboard or an add on PCI card, is there a length limit as per IDE for the flyleads? Because if not, then this USB2 front panel jobbie would give me USB2, and I could bring the front panel thing up into the breakout box. Otherwise, a (powered) USB hub mounted in the breakout box would be OK, I guess.

I don't see a problem bringing leads up for LEDs and power switches, not for headphone/mic. No reason not to make a nice wooden case for it all. If we are worried about cooling, a slimline solution would be one of those PCI slot fans, which I'm sure could be bodged into place nicely.

 

[LeadMagnet]

There was a guy over at www.anandtech.com (Ed Downing) forums that made what you discribe out of an "Apex 1500" $50 DVD player and it looks great. I emailed him about building it and his reply is bellow.

Hi,

It wasn’t really to bad.

The motherboard $149

Ram 512mb 333 DDR $70

Hard Drive (when moved to the 80gb WD) $90

PSU $37

DVD player $53 at WalMart



Aprox $400

Ed