New ATA/100 cable design... FINALLY!

[cautery]

Some of you may have seen the article on my design for an ATA/100 Cable from CAT-5E.

While this first prototype was over 5 feet long and easily equaled or slightly exceeded the performance of "standard" 80-conductor ATA/100 cables, it did have a few disadvantages. The connector scheme I worked out was a bit large and clumsy with which to work. The first prototype was also fairly stiff due primarily to its large diameter and solid conductors.

Well, I FINALLY got around to constructing a NEW design!

The new design eliminates both disadvantages. The connector is now simply a single standard 40-pin IDE connector, and I have reduced the cable diameter significantly and drastically improved flexibility and long-term durability. It is still a very difficult cable to manufacture (read: expensive, but less $$ than original design), but it will be a breeze to use!

I am just about to begin testing the new prototype, and if it performs well, I am seriously considering producing them in small numbers for custom applications.... depending on the interest.

If you guys/gals are interested, I will post back with the results...

Oh.... of course the cable is round... and did I mention that the new prototype is 6 feet, 8 inches long!! (length could be significantly extended) AND finally, unlike the first prototype, this design can be constructed much more easily with two drive connectors!

I will post pictures with the performance results, but I don't think I am going to release the construction procedures since I am most likely going to be making them for sale.

Let me know what y'all think.

 

[toft-dk]

That's an awsome job you've done there

 

[Packy]

You're nuts!! I'm extremely impressed

 

[cautery]

Here's a pic of the new cable:

New ATA/100 Cable

I've just finished the preliminary testing on it using HD Tach v2.61 (read only, as I don't have an empty drive til this afternoon). I ran the tests using a 15GB IBM ATA/66, 5400 RPM drive. I'm slightly disappointed in the results, but encouraged that the cable DOES perform as an ATA/66/100 cable. Here is a breakout of the numbers. (I added an ATA/33 cable to make SURE the controller was running at ATA/66 with the new cable) Ignore the absolute values for the tests as this was not an optimal setup. Just look at the relationshio between the different cable test values.

HD Tach v2.61(Random Access, Read Burst Speed, Avg. Reed Speed

Standard ATA/33 Cable:
14.6ms, 20.5mbps, 12372.2kbps

Standard ATA/100 Cable:
14.5ms, 40.1mbps, 12427.8kbps

New ATA/100 Cable:
14.3ms, 30.4mbps, 12400.4kbps

While the Random Access Speed and Average Read Speeds are pretty close, it is obvious that the new design took a hit in the Burst Speed category. I verified that it IS running at ATA/66. My computer shows a warning on boot up for any channel that does not have an "80-conductor" cable (ATA/100).

Why is it taking the Burst Speed hit? Does it really matter? For ordinary use, probably not. But the folks who would be buying this cable will likely be performance geeks (like me ) and will want AT LEAST as good burst speeds as a standard cable delivers. I ran Sandra on the new drive just to see if I could "break" it, and it was fine. It WORKS fine... It's still installed on my boot drive as I type as a matter of fact. I just have to figure out and correct the cause of the burst speed hit.

Some of the possibilities are:

1) Using stranded CAT-5 in this cable as opposed to solid CAT-5E in the original cable. (stranded vs. solid or twist rate differences?)
2) The CAT-5 might have a higher capacitance per foot than the CAT-5E.
3) Deleting the 40 ground wires on each signal wire in favor of simply using adjacent twisted pair.
4) Related to "3" above: Certain signals may not pair well.... notably the DD15:0 (data signal cables).
5) Cable is too long? (signal attenuation or high total capacitance?)

I CAN identify the problem... Unfortunately, it means developing a troubleshooting tree and constructing test cables to rule out/in causes...

Think I'll start by researching the difference in electrical properties of the two different wires I used... Several other things I want to read up on first too... No need in building cables that I don't have to right?

In any event, I will continue to develop the new design as time permits.

 

[J3anyus]

That's insane...