Jump to content


Photo

How problems manifest w/ too long eSATA cables?


  • You cannot start a new topic
  • Please log in to reply
5 replies to this topic

#1 eshap

eshap

    Member

  • Member
  • 121 posts

Posted 23 June 2010 - 12:52 AM

What sort of problems are expected with an eSATA cable that is too long? No connection at all, random disconnects, or hidden corruption?

By too long I don't necessarily mean out-of-spec length (I suppose the spec is a recommendation in which it's more likely to stay within the required electrical specs), but also factors like cable quality (shielding, impedance, whatever), connectors, and electrical tolerances of the devices at both ends.

I have an external eSATA enclosure that came with a 1m cable which is way too short and a real hindrance. The spec allows for 2m, so I want to get something longer. My assumption is that they included something that short (1m is the max for internal SATA) as a way to minimize problems with all the non-spec hardware floating around; internal SATA connectors used as external, brackets to repurpose internal to external, pre-standard external devices, etc.

Hopefully my hardware is to spec, and not just by looks (the connectors, at least, are true eSATA), but I have no way of knowing.

#2 Mickey

Mickey

    Member

  • Patron
  • 2,281 posts

Posted 23 June 2010 - 09:36 AM

A quality 2-meter cable should exhibit no problems. I've used them in the past in our test lab. Symptoms of a bad cable would be the same whether it's eSATA or SATA: no connection, intermittent connection, or bad throughput.

I suspect the real reason the enclosure came with a 1-meter cable is cost and most users would just set the external unit next to the PC, so a longer cable just gets in the way.

#3 eshap

eshap

    Member

  • Member
  • 121 posts

Posted 24 June 2010 - 12:29 AM

So you think a bad cable would have obvious connection/performance problems but not bad data written (or read)? I have no way to assess cable quality before getting and testing it, nor to know if my controller and enclosure truly adhere to eSATA specs, so I'll just have to try. If hidden corruption is a possibility, 1.5m would be safer.

Re limited cable length, extra 50-100cm of eSATA cable would only be a few more cents. I'm using it next to a PC, a full tower on the floor, and it's barely enough. Combined with a short 1m electricity cable there's no way for me to use it at all without an extension cord. Whatever the reason for skimping on cables, I'm disappointed with Vantec.

#4 mockingbird

mockingbird

    Member

  • Member
  • 173 posts

Posted 24 June 2010 - 03:27 AM

This depends.

What some people do is attach a "SATA to eSata" "adapter" from their motherboard's regular SATA port to a bracket for the back of their case. There is a key difference between SATA and eSata where eSata sends a higher voltage through the cable to allow for the longer length.

So even if you use in spec long eSata cables in this scenario, there is no guarantee.

#5 Mickey

Mickey

    Member

  • Patron
  • 2,281 posts

Posted 24 June 2010 - 08:28 AM

I imagine throughput issues would be due to additional ECC necessary to correct the bad data passing through the cable.

As mentioned, I bet Vantec included a 1-m cable since that's what most folks would need and it's the cheapest type to provide. If enclosure manufacturers are like HDD manufacturers, every penny saved is worth chasing. At least longer eSATA cables are relatively easy to find. I had the misfortune recently of needing 2-m long mini-SAS to 4X SAS cables. I ended up custom ordering those. :(

#6 eshap

eshap

    Member

  • Member
  • 121 posts

Posted 24 June 2010 - 04:05 PM

There is a key difference between SATA and eSata where eSata sends a higher voltage through the cable to allow for the longer length.

I think there's no voltage difference, but there are different tolerances specified for the end devices.

I imagine throughput issues would be due to additional ECC necessary to correct the bad data passing through the cable.

That's fine if you have error correction kicking in. The question is whether there could be undetected errors.



0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users