Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Trying to figure out how ethernet cable tester works

Status
Not open for further replies.

throbscottle

Well-Known Member
So, I have a cheap LAN cable tester for work, and I'm intrigued by it. It has a sender unit, with battery and some LEDs, and a receiver, which just has LEDs. Haven't had chance to have a proper look inside it, but It has what I'm guessing is a micro inside.

So, it pulses through each conductor, and lights the relevant LED on the sender and on the receiver so you can spot crossed or broken connections. There doesn't appear to be a single return path, so I'm guessing it's "any", and the LED brightness doesn't change respective to the number of good conductors or cable length, so I'm guessing it's constant current.

The main question is therefore, assuming the "any" return path idea is correct, how would this be implemented? I think there's only passive components in the receiver. Just can't get my head round it...

Or am I barking up the wrong tree?
 
No microprocessor, just NE555 timer and CD4017

World's laziest circuit diagram attached - 9 LEDs and 9 diodes in remote pod

WIN_20210420_23_43_43_Pro.jpg
 
Lock-down cabin fever is a terrible thing.
Half finished, pathetic, schematic of main unit is attached.
May or may not replace with better version if I can be bothered :(
 

Attachments

  • BadScan.pdf
    81.6 KB · Views: 486
Seems like random schematics that don't answer the question. Or did I miss something?

Mike.
 
I'd think just a medium value resistor or reverse diode across each LED in the "remote", with eg. the cathodes of all LEDs connected to a common point.
And a resistor and LED in series with each output on the main unit, with a reverse diode across each such circuit.

If one wire at a time on the master is taken positive, with all other other wires pulled to ground, only one LED is forward biassed and the return path is through all the other shunt resistors or diodes in the remote and the reverse diodes in the master, in parallel.
 
Seems like random schematics that don't answer the question. Or did I miss something?

Mike.
It's not the best explanation, but the schematics show all that is needed to make it work.

The 8 wires are connected to 8 outputs of the CD4017. That is triggered from the 555 to give a second or so on each output.

When one wire is high, all the other are low.

On the receiving end any wire that is low is connected via a diode to the common wire in the receiver, to provide an earth path.

A major function of the tester is simply to have the LEDs with numbers against them so that it's easy for the user to work out what is happening.

The tester relies on there being at least one other wire working to give a return path. Also if the cable is wired wrong, with each pin wired to the one above, like this:-

1 -> 2
2 -> 3
3 -> 4
4 -> 5
5 -> 6
6 -> 7
7 -> 8
8 -> 1

that won't be detected by the tester as the lights will still illuminate in the correct order. However, wiring an ethernet cable wrong like that is going to need to be done deliberately.

The tester will be really useful for the usual faults of a wire or two not making connection, a couple of connections swapped, or the far end being from a completely different cable.
 
A high is output to each line in turn.
The diodes provide return path on any or all lines
 
Last edited:
Right, now I have a chance to actually get into the thing. Ok in the remote pod there's a bunch of diodes, common A, bunch of LEDs which my (work) meter is too crappy to test, taking the big terminal as K looks like those are commoned with the diodes. The LED for shield doesn't have a diode with it.

So hexreader has 99% the right of it with the pod.

Hmmm. Ok lets have another look in the controller...
LEDs in anti-parallel with diodes, in series with each output from the chip. It's a UUZ 468CH BA32 SP1407AM that's running the show.

The 555+4017 circuit I came across a few times when trying to look this up, but always showing a single return path, which limits it's usefulness.

Well this has been very educational, thanks everyone!
 
It's not the best explanation, but the schematics show all that is needed to make it work.

The 8 wires are connected to 8 outputs of the CD4017. That is triggered from the 555 to give a second or so on each output.

When one wire is high, all the other are low.

On the receiving end any wire that is low is connected via a diode to the common wire in the receiver, to provide an earth path.

A major function of the tester is simply to have the LEDs with numbers against them so that it's easy for the user to work out what is happening.

The tester relies on there being at least one other wire working to give a return path. Also if the cable is wired wrong, with each pin wired to the one above, like this:-

1 -> 2
2 -> 3
3 -> 4
4 -> 5
5 -> 6
6 -> 7
7 -> 8
8 -> 1

that won't be detected by the tester as the lights will still illuminate in the correct order. However, wiring an ethernet cable wrong like that is going to need to be done deliberately.

The tester will be really useful for the usual faults of a wire or two not making connection, a couple of connections swapped, or the far end being from a completely different cable.
If the 4017 is wired with it's reset pin tied low, then it will go through all 10 states. This will produce no lights while it goes through those two extra states, then it will start again at the beginning. But if the pause occurs other than between the last and the first, it would indicate a shift like you describe.
 
If the 4017 is wired with it's reset pin tied low, then it will go through all 10 states. This will produce no lights while it goes through those two extra states, then it will start again at the beginning. But if the pause occurs other than between the last and the first, it would indicate a shift like you describe.
Hexreader's circuit shows the 9 connecting to the reset line. I've just realised that will give 9 states, where I previously thought it was 8, so there will be a pause between the data lines.

With more than 8 states, the unlikely type of miss-wire that I suggested would be detected.

The screen is also tested on Hexreader's circuit. So in theory, there could be a miss-wire that includes the screen that would show as correct, but that would virtually impossible to happen by accident.
 
Start watching this
video at 6:00 minutes.

mine died and with his, my old one and a new one they are all different. His shows real parts and the last one I bought has a COB.
 
My tester has CD4017 pin 11 (output 9) wired to pin 15 (reset), causing only 9 states for 9 LEDs, no pause.
No use to the OP though, as OP's main unit is different to mine.
 
Start watching this video at 6:00 minutes.

mine died and with his, my old one and a new one they are all different. His shows real parts and the last one I bought has a COB.
Very similar - same casing, even down to the triangle screws. Where his has got a slide switch in that funny oval opening, mine has a tactile push button which actually fits the opening. Oscillator is a 3 pin resonator connected to chip. The casing appears designed for a more complicated device as there's an extra pushbutton space (and other spaces) on the board and space for an extra LED.

I was hoping to find a way to repair it as some dead outputs, but it looks like about 4 pins on the presumed-micro have died, so I'm stuffed there I think.
 
I have another cable tester that I like for a different reason. Usually, there are breaks in the cable at the ends and the wire map type of tester won't catch them.

This other tester has bi-color LEDs wired to a resistors. All the remote needs is the bi-color LED. You get a green light if the pair is straight, red if cross and nothing if the PAIR is bad. The result is instantaneous when you wiggle the ends.

I didn't get a remote and I haven't upgraded the remote I made for 100bT or 1000bt, It just consisted of a Jack and bi-color LED w/o a case,
If I make a case, I might add a telco socket at the remote end.

It also has a machine pin socket so, you can access each end of the cable and switches to open each wire. You can then determine which wire is bad with an ohmmeter and all of the switches off. it actually labels.
 
I have one of those little units, but my favourite is one I got from either ebay or aliexpress something like 20 years ago.
It has eight remote units that it can individually identify, for when doing multi-cable installations - and as well as testing to those for continuity and correct connections, it has a TDR (time domain reflectometer) function that will display the length of the overall cable or distance to any open core point.

Note that some places sell the same unit with either no remotes or only one of them.

Ethernet_Tester_800.jpg

Edit - a couple of examples on aliexpress, showing the main two displays:
 
Last edited:
Nice if you're a cabling professional. I just need to eliminate cabling from the list of reasons for stuff not working...
 
I have friends in the business and they usually get a wiring contractor to do the initial work and qualify the installation. He said that usually if he has to install a cable as long as it passes the quick test, it's usually OK. TDR is the next step and the final step is bandwidth. He went on to say that the switches he uses has qualification tests built in.

Before I became more involved, and it was early Ethernet, like for windows 3.1 machines, we got network cards with silver satin cable so no one knew what they were doing. It was 10bT stuff which got ripped out. I brought some of the 4 conductor CAT5 cable home, I managed the pool of IP addresses our department was given and could ask to have the nameserver record changed. We had a class A address assignment. I didn't have access to the switches, but there were four 100bT switches combined to gigabit fiber. After I left there was an upgrade to VoIP phones. I set up PC's and MacIntosh computers to the Internet and to access the printer. The hard was trying to set up windows in Chinese to access the printer. VPN was required to access the network from outside the network except some mchines running Solaris.

I cheated and set-up a man in the middle "attack" to access a server directly. I used Cygwin running an a PC I still had to use the company VPN. It worked for a while until they, IT, fixed it. I still needed credentials which I had. No one, but me, knew about the more direct access

I did buy two 250 feet reels of CAT6 cable (2 colors) for home. I was given like 500-1000 feet, probably 1000 feet of RG6QS copper clad cable.
 
Well, realised I can cut the output pins of the original IC and use it to clock a 4017 (since it has a "heartbeat" LED I can connect it to), so I've just ordered one of those to stuff in the box.
I think it was connecting to a POE cable that damaged it - would adding normal protection diodes protect it against this or do I need something more robust? Or is it easier to just put up with it getting fried occasionally, and fixing it again?
 
Well, realised I can cut the output pins of the original IC and use it to clock a 4017 (since it has a "heartbeat" LED I can connect it to), so I've just ordered one of those to stuff in the box.
I think it was connecting to a POE cable that damaged it - would adding normal protection diodes protect it against this or do I need something more robust? Or is it easier to just put up with it getting fried occasionally, and fixing it again?
I would suggest adding some resistors in series with each terminal. You should also put clamp diodes to stop the voltage getting outside the supply rails.
It would also be a good idea to have a zener diode to stop the supply voltage being driven too large by the POE.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top