Network Experts: What's the best hand-held network testor I can get for $500, $1000, $2000? and best software?

[brxndxn]

Our industrial installations are getting more and more computer network dependent - and I want to be able to confirm every cable our electricians run. We're having problems where people are setting a switch on 100mbps FDX and the cable is only really capable of HDX.. Also, fiber runs are beginning to be more and more of a pain in the ass. It especially gets confusing when a different contractor runs all of the cables for our design and they tell us they tested everything...

So.. give me some ideas on the tester I should buy.. It needs to be hand-held, durable, and I need to be able to test 10,100,1000cat5,fiber,HDX,FDX,etc.. I wanna see different price ranges to see if I'm really getting useful features for the higher prices..

Also, if there is any good link testing software, lemme know.. I have a very small tablet pc that I use for configuring switches and it would be nice to have it tell me cable and link diagnostics..

 

[spidey07]

You'll need a cable certifier and fluke is where to get it. They have different models. Oh, and any competent installer will provide you with the actual test results. If they don't, never use them again.

Also duplex isn't really related to the cable. If you set the switch port to 100/full then the other end of the link (NIC) will go to half duplex. It's called a duplex mismatch and it will happen every time when you set one end to 100/full and leave the other at auto.

 

[imagoon]

We use the older version of this:

http://www.flukenetworks.com/f...ingsupportAndDownloads

They are not cheap. I think the base unit is around $7000~$8000 and the full kit of modules bumps the price to about $17000 or so dollars.

They are excellent units. They have a software program called linkware that goes with them that lets you download and view / generate testing reports.

Also like spidey said, the full duplex / half duplex issue is a the device level. As long as the cable is certified, it should not have issues passing data. Even in half duplex the each unit is still using the same pairs to transmit data.

 

[ScottMac]

Originally posted by: spidey07
You'll need a cable certifier and fluke is where to get it. They have different models. Oh, and any competent installer will provide you with the actual test results. If they don't, never use them again.

Also duplex isn't really related to the cable. If you set the switch port to 100/full then the other end of the link (NIC) will go to half duplex. It's called a duplex mismatch and it will happen every time when you set one end to 100/full and leave the other at auto.

Poorly terminated cable can cause the link to downshift to half duplex (or, at the least, perform *better* at half-duplex).

I agree that Fluke makes some great testing equipment. I've had been using JDSU stuff a lot lately too and it's very good, durable, and well-supported.

 

[ScottMac]

Originally posted by: imagoon
We use the older version of this:

http://www.flukenetworks.com/f...ingsupportAndDownloads

They are not cheap. I think the base unit is around $7000~$8000 and the full kit of modules bumps the price to about $17000 or so dollars.

They are excellent units. They have a software program called linkware that goes with them that lets you download and view / generate testing reports.

Also like spidey said, the full duplex / half duplex issue is a the device level. As long as the cable is certified, it should not have issues passing data. Even in half duplex the each unit is still using the same pairs to transmit data.

Yes, true ... but the key is that only one side is transmitting at a time, so the effects of crosstalk (from poor / improper termination) are not as strong. Cabling that doesn't work well in full-duplex mode often works just fine in half-duplex mode. This is also a hallmark symptom of a split-pair on pins 3&6 (extremely high crosstalk).

 

[spidey07]

Originally posted by: ScottMac
Yes, true ... but the key is that only one side is transmitting at a time, so the effects of crosstalk (from poor / improper termination) are not as strong. Cabling that doesn't work well in full-duplex mode often works just fine in half-duplex mode. This is also a hallmark symptom of a split-pair on pins 3&6 (extremely high crosstalk).

You know as I was driving home rethinking my post..."classic split pair and crosstalk" popped into my head.

But I don't know of any 100 Base-T PHY that will change duplex unless the link pulse changes.

I was thinking more about how TOO LONG runs perform better at full duplex than half. I'm slipping.

 

[imagoon]

Originally posted by: ScottMac

Originally posted by: imagoon
We use the older version of this:

http://www.flukenetworks.com/f...ingsupportAndDownloads

They are not cheap. I think the base unit is around $7000~$8000 and the full kit of modules bumps the price to about $17000 or so dollars.

They are excellent units. They have a software program called linkware that goes with them that lets you download and view / generate testing reports.

Also like spidey said, the full duplex / half duplex issue is a the device level. As long as the cable is certified, it should not have issues passing data. Even in half duplex the each unit is still using the same pairs to transmit data.

Yes, true ... but the key is that only one side is transmitting at a time, so the effects of crosstalk (from poor / improper termination) are not as strong. Cabling that doesn't work well in full-duplex mode often works just fine in half-duplex mode. This is also a hallmark symptom of a split-pair on pins 3&6 (extremely high crosstalk).

True but I did qualify it with "if the cable is certifed" when a cable (end to end + jacks etc) tested and passes we call it certified here. I didn't think of the split pair issue though.

 

[RadiclDreamer]

Fluke, I dont trust anything else.

 

[brxndxn]

Originally posted by: spidey07

Originally posted by: ScottMac
Yes, true ... but the key is that only one side is transmitting at a time, so the effects of crosstalk (from poor / improper termination) are not as strong. Cabling that doesn't work well in full-duplex mode often works just fine in half-duplex mode. This is also a hallmark symptom of a split-pair on pins 3&6 (extremely high crosstalk).

You know as I was driving home rethinking my post..."classic split pair and crosstalk" popped into my head.

But I don't know of any 100 Base-T PHY that will change duplex unless the link pulse changes.

I was thinking more about how TOO LONG runs perform better at full duplex than half. I'm slipping.

We often have to use runs that are 'supposed to be' 100mbps FDX - but after years of sitting in an industrial environment will only seem to work at either 100mbps hdx or run slow as hell at 100mpbs FDX.. For us, we care about communications reliability first and foremost.. We don't always get to use new installations - so sometimes we're mixing new and old.. fiber and cat5.. proprietary industrial protocols over ethernet.. etc.. In the end, it all has to work - even if it isn't fast.

For my current project, I've set the switches and the nics all to 'autosense' and things have been working fine so far. Most times, we are recommended to set everything at 100mpbs FDX - but I have a hard time believing that is a best practice - especially when I have seen things slow down because of incorrect FDX settings.

But either way.. any links to the cable tester I can get?