EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: TheWaterbug on January 24, 2024, 12:35:57 am
-
¡Hola!
I am trying to look at raw Gigabit / 10 Gigabit Ethernet waveforms on my CSA8200/80E04 (also have a DSA8300), but I have no idea how to trigger it:
I have a pair of RJ45-to-SMA adapter boxes (https://www.pulseresearchlab.com/products/prl-rj45-sma) that allow me to carry a set of Ethernet signals over eight 50 Ohm coax cables, and then I’ve interrupted some of those 50 Ohm coax cables with a set of differential pickoff boxes (https://www.pulseresearchlab.com/products/prl-860q?variant=29205909512) that allow me to probe any pair with 10x attenuation on the scope:
(http://www.kan.org/download/PRL-RJ45-SMA_PRL-860Q-SMA.JPG)
(https://www.kan.org/pictures/EthernetPickoff_PRL-RJ45-SMA_PRL-860Q-SMA.png)
The chain of adapters and pickoffs seems to be ok, because I can successfully carry 10GBaseT traffic over the link. I’ve tested from NIC to NIC through this adapter chain, and I can successfully negotiate a 10GBaseT link (800 MBaud/pair) (https://en.wikipedia.org/wiki/10_Gigabit_Ethernet), and iperf reports throughput of ~9.40 Gbit/sec, which is the same as I get if I remove my gaggle of adapters and just connect the NICs directly with a Cat7/Cat8 cable. So 1000BaseT should be just fine, since it's only 125 MBaud/pair (https://en.wikipedia.org/wiki/Gigabit_Ethernet).
I have two different Tx signals from the pickoffs going into my scope (with 80E04 sampling head), and that's what I'm trying to get a PAM-5 trace out of.
I was able to get an MLT-3 trace (https://en.wikipedia.org/wiki/MLT-3_encoding) by using an eval board for the Microchip KSZ9031RNX Ethernet transceiver chip (http://ww1.microchip.com/downloads/en/DeviceDoc/KSZ9031RNX-EVAL%20Board%20Users%20Guide%20v1.1.pdf), where this board has a test point where I can extract the 125 MHz TX Clock signal. I connected the Ethernet jack of the Microchip board to a PC, via my cabling-and-adapter setup. Then I used the PC to force 100BaseT (https://en.wikipedia.org/wiki/Fast_Ethernet#100BASE-TX), and brought up a link. The purple trace is the 125 MHz clock, and the green is Pin 1 from the 100Base-TX signal, as seen from the Microchip side. The yellow trace is the RX side, e.g. TX from the PC:
If the 100BaseT link negotiates to the Microchip side:
(https://www.kan.org/pictures/CSA8200_100BaseTX_Rx_Tx_1.png)
and if it negotiates to the PC side:
(http://[url=https://www.kan.org/pictures/CSA8200_100BaseTX_Rx_Tx_2.png]https://www.kan.org/pictures/CSA8200_100BaseTX_Rx_Tx_2.png[/url])
I can draw some conclusions about performance at 1000BaseT because it just uses 5 levels instead of 3, at the same baud rate, but dangit, I want to see it!!!
But I can't figure out how to trigger it. If I let the eval board negotiate at 1000BaseT, I just get cloud of dots.
(http://www.kan.org/download/CSA8200Ethernet.png)
If I go into the Trigger menu and choose Pattern Sync, it says “Pattern Sync Module not available,” but when I go to the Mask menu, I can apply a Mask for Gigabit Ethernet.
Do I need to buy some sort of license module in order to trigger on Ethernet? Or am I just doing it wrong?
I want to see PAM-5!!! And eventually I want to do this with 10GBaseT signals.
-
Hi !
I don't understand your question :
you want to see the PAM-5 signal (1Gb)
- to take a picture of it ?
- or to measure jitter/etc ?
Regarding the 10G signal, honestly, 10GbE base-T (=copper) = sh...t (no one use 10GbE thru copper ; a non-sense given the power / signal integrity /etc)
10Gb is thru fiber (single-mode is far better than MM) : that's all. My 2 cents...
-
Hi !
I don't understand your question :
you want to see the PAM-5 signal (1Gb)
- to take a picture of it ?
- or to measure jitter/etc ?
Regarding the 10G signal, honestly, 10GbE base-T (=copper) = sh...t (no one use 10GbE thru copper ; a non-sense given the power / signal integrity /etc)
10Gb is thru fiber (single-mode is far better than MM) : that's all. My 2 cents...
Yes, I want to take a picture of the PAM-5, if that's possible. I'm not sure if it is.
If I google image search for "1000BaseT PAM-5 (https://www.google.com/search?client=safari&sca_esv=601222021&rls=en&sxsrf=ACQVn08XAKNqdVJrkyJIzF4KsFohTlrRcg:1706138939002&q=1000baseT+PAM-5&tbm=isch&source=lnms&prmd=isvhnmbtz&sa=X&ved=2ahUKEwjk5pXrlveDAxWeJEQIHQLBAI0Q0pQJegQIERAB&biw=1688&bih=1024&dpr=1)" I get a bunch of diagrams, and a very few eye diagrams.
Most of the eye diagrams look like simulations, and the one legit scope capture is mine ;D, from when I posted this same question to an Intel forum 2.5 years ago and got zero responses ::).