Wireline emulation for ADSL testing

Good morning :-)

I came to a need where I need to emulate copper wireline in order to test ADSL modem in a lab. I have ADSL DSLAM and ADSL modem which work on a few meters cable but this can't be the equivalent for the real scenario. So, I built (hopefully) equivalent of copper pair emulating L, C and R making it in moduls that can be chained together. Schematic:

formatting link
Values of R, L and C were based on data found in the internet which roughly estimate L to be 600uH per 1km, R = 130 Ohm/1km, C = 50nF/1km After testing the circuit, I found out that ADSL2+ modem synchronizes with a speed of about 1500/800, but only if only the first segment is used (which is R1/R2, L1/L2, C19, L19/L29, R19/R29 - jumpers JP10/JP20 set to 1-3, jumpers JP19/JP29 set to 2-4). The reported line attenuation is about 57dB which is quite a lot. So the question is that why I get that big attenuation if the values should match only 1km of single pair wireline. And the other question is what values should be lowered to match the real environment for ADSL setup. I know that the empirical approach would suit my needs but I'd like to have some estimation instead of soldering random elements ;-)

--
Cheers, 
Bartosz
Reply to
Bartosz
Loading thread data ...

A lumped-constant line is only good for frequencies well below the low-pass limit set by one LC branch in the line. In your case, it is around 27 kHz, which is far too little for any DSL mode.

To make it useful, you have to use much smaller inductances and capacitances, keeping their ratio, and adding more of the LC units to the line.

--

-TV
Reply to
Tauno Voipio

It is just a schematic from Eagle PCB designer ;-)

I've been just crawling around some Spice implementations to find my own one, but I haven't chosen the One yet. I played with LT Spice for a few minutes but I didn't like its editor. But since the important factor is also the size of community that uses it, maybe it would not be that bad solution. Let me try it again.

Reply to
Bartosz

I'll look for it. Many thanks for suggestion.

I don't need a very precise model, but something that will be "similar" and more convinient than a few kilometers of cable on the floor.

No, this is just a pure fake antispam e-mail address. But you can use "kiziuk" username at the same provider that you use as well :-). This will be the right one.

I haven't thought that this would be the very important part for this design. What I was intending to do was to get more familiar with the right "electrical" design and then maybe try to improve PCB design.

Reply to
Bartosz

Please note that an ADSL receiver is actually a wideband radio receiver, and all normal radio design precautions apply.

--

-TV
Reply to
Tauno Voipio

Am 17.03.2015 um 14:13 schrieb Bartosz:

I find the editor quite cool. It uses unusal keys but if you play around for a while, you'll memorize them. But, for example, you can place several parts on your sheet, press F3 for wire and connect them without terminating the wire. Just draw straight over the parts. At first it looks as if you produce a single big short circuit but if you terminate the wire, all short circuits are removed! Also, if you place a part over an existing wire, it will be inserted correctly. I have never seen a schematic program that works like this!

Cheers,

Robert

Reply to
Robert Loos

sent! look for it THAT server likes to put everything into spam and let you figure it ou!

Reply to
RobertMacy

You're right. Absolutely amazing :-) I didn't thought that it is possible, so didn't even try.

Reply to
Bartosz

Good morning !

May be you should read the ITU 992.X standards. As a starting point you should get the normalized Z POTS to feed LTSpice with this model and .. simulate an eye diagram of the transmission. It's a business IMHO !

H.

Reply to
Habib Bouaziz-Viallet

I started to play with LTSpice, but this week, due to lack of time for it, I needed to postpone all of it... Anyway, I'll be digging it more, possibly next week.

Reply to
Bartosz

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.