From Oscilloscope to Wireshark: A UDP Story (2022) | Hacker News Reader

From Oscilloscope to Wireshark: A UDP Story (2022)

mattkeeter.com

134 pointsofrzeta30 comments

University professor here: showing and decoding wire signals has become a staple in my teaching - very effective, eye-opening even for students that find the OSI layers too abstract to grasp on paper. Btw, I’m using this nice demo board (I’m not affiliated with the brand): https://www.batronix.com/shop/demoboards/Batronix-MSO-Demo-B...

…typically we decode I2C.

Oh, I‘ve equipped the lab with all kinds of oscilloscopes. I find a lower mid-tier product (like Rigol DHO 900/1000 series) most suitable: good enough resolution and nice quality-of-life features (many physical buttons / dials and input channels go a long way!), yet not too feature-packed and physically compact to still be approachable.

Interesting.

Do you happen to know if that batronix board useful to have even if you have a Rigol DHO 924S without the logic probe (PLA2216)?

Yes, it’s absolutely useful. You need the logic probe (which I do have, but never use in „standard“ teaching) only if you plan (or need) to monitor more input signals than the analog ones on the scope.

In fact, to demonstrate the very physical foundation of digital computing, it is very useful to use the analog inputs and zoom all the way in to see the slight imperfections in signal transmission.

Thank you!

> I configured the oscilloscope to collect 100M samples at 1 TSPS

Typo? I didn't think we had sample rates anywhere near that high!

A Keysight UXR can do one quarter of that, 256GS/s, but a Tektronix 6 series is limited to 25 GS/s iirc, so you're right.

Looks like max 50GS/s per their site. That also looks reasonable with the screenshot they have in the article showing 1ns / div horiz. But clarity on the data would be lovely. =)

Actually I take it back: For the series 6B spec page... Real-time: 50 GS/s (2 channels), 25 GS/s (4 channels), 12.5 GS/s (> 4 channels) Interpolated: 2.5 TS/s

Yeah, I used to work at Tektronix, this looks like an A unit.

Also, the 50GS/s is actually more like 25GS/s as it has a Nyquist frequency of 12.5 GHz. This is because they take 2 samples at the same time, instead of 180 degrees out of phase. I wish they had made that clearer to customers, it felt misleading to some.

If you are an absolute nutcase, you could characterize a set of line stretchers and a multiplexer on a high end VNA then offset the inputs of the 4 channels on that UXR with them, take a capture and finally rebuild a 1TSamp/s signal out of the 4 results.

You have to have the 240V model of the scope to run all four channels at full rate (110GHz) though.

Take that Nyquist ;)

The older Tektronix TDS540 series did this, but at much lower rates as was common in those days though. Internally there are differential feeds from the very beautiful hybrid ceramic input boards to 4 DACs, with some clever switching so that a single input can be sampled by all 4 DACs with a suitable offset to create 4x the sample rate when running with all 4 inputs.

The calibration procedure on the scope fiddles with the time alignment to get the different DACs correctly offset so that the combined signal is correct.

The hybrid ceramic input boards in their metal cases are a thing of beauty, fragile (don't ask how I know), but beautiful.

Yup, a lot of scopes actually did this internally and some still do. It's part of why some scopes lose half their BW when you go from 2 ports to 4 ports (some go the other direction and run multiple ports on one very fast ADC), they split the digitizers. It's just very very difficult to keep it working external to the box mainly because of line drift.

> Random equivalent-time sampling takes advantage of the nature of a repetitive signal by using samples from several trigger events to digitally reconstruct the waveform. Since sampling occurs on both sides of the trigger point, pretrigger capability is very flexible. Because repetitive signals are being sampled, the bandwidth of an equivalent-time scope can far exceed its sample rate.

https://www.tek.com/en/documents/application-note/real-time-...

Here's a more specific example: PicoScope 9400 series supports just 500Msps per channel, however it's advertising "70ps transition time and 1TS/s (1ps resolution) random equivalent-time sampling", this sort of "equivalent sampling" is presumably where that seemingly crazy spec comes from.

But you can't use equivalent-time sampling for something non-repetitive like network packets.

For measurements like this the SDK will usually include some utilities to send the same data over and over on a port.

Even if it's the same data, the bit stream will be a variety of 0 and 1s. The period of that waveform will then be 1 frame length / data transfer rate (or rather 1/4 frame length / data transfer rate as this is a QSGMII link). I wonder how the scope triggers on that. Trigger criterium would be a bit pattern, say the Ethernet frame preamble of 7 octects (* 10/8) spread across four streams ...

Otoh, at 5Gbps, a sample rate of "just" 10GS/s would be sufficient (barely).

I rather suspect the oscilloscope is capable of 1TS/s equivalent time sampling, but that mode wasn't used.

Not a typo, but you’re correct about the sample rate - with those settings, the scope was doing interpolation between samples.

By definition, you can't interpolate a sample. A sample is a measured value.

What you can do, if and only if you have an exactly repeating signal triggering at the same point within a cycle, is change the delay between the trigger and sample, and repeat. In other words, sample at different times within the same signal (since it's exactly repeating), to build up samples in time, of that waveform, to whatever time resolution you want.

Of course, you're limited to any noise in the trigger, variation in the signal, etc.

This is how you can record light moving through your garage [1]!

[1] https://www.youtube.com/watch?v=o4TdHrMi6do

Not sure if mkeeter's comment has been ninja edited but it says between samples, it doesn't say it is interpolating to generate new samples.

I understand, but that's my point, it's not interpolated!

The number he's referring to is in units of samples per second. It's not doing interpolation between samples, to achieve a high samples per second, because that's not possible, which is my point. Interpolation results in an imagined value, but samples are measured values.

It would be correct to say that the values between samples are interpolated, but the subject of interpolation isn't applicable for anything mentioned in this comment chain.

Ah you are referring to the 'sps' bit. Ok, but I think the extra sentence is enough clarification of what they mean, even if they're wrong about what the device is doing.

The only time these are interpolating is when they are visualizing, there is no point (hah) in storing interpolated data, you can generate that whenever you want.

Not the original reply, but I support the correction here. Regardless of how pedantic/nitpicking it seems, I remember getting confused about this a lot when learning digital signal processing. Simply because its really easy to upsample.. or look at an upsampled result and get confused by that

I think 'upsample' is the root cause here. Technically that is a misnomer.

This reminds me of back in the day (1985) on my final year project, I'd designed a (6502 or 8080 based) SBC, my PCB design, my assembler and on first boot ... nothing happened.

With only a voltmeter in hand, I measured the voltages on the address lines - thankfully my dumb programming error was a very tight loop so it was easy to deduce where the looping was happening !

Some things were easier back in the day.

Could anyone recommend a budget tooling (has most features desired, please explain) to capture on wire like an oscilloscope. Help understand what kind of sample rates needed etc. features desired, cost and any recommended brand/models to get by. Thx!

The QSGMII signals in this blog are very high speed. Beyond the range of what you can measure with budget tooling. Even the probes for such lines are thousands of dollars.

There are some budget sampling oscilloscopes on the market, but budget is still mid four figures and up. That's before probes, cabling, and other things you'd need. Sampling oscilloscopes are only useful for repeating test patterns sent by SDK tools, not for capturing normal data as it goes by

It's possible to look at Ethernet signals with oscilloscopes in the budget range but you would need appropriate fixtures to tap the line. Even at those speeds, touching a regular oscilloscope probe on to a wire disturbs the circuit so much that it might stop communicating.

At the other end of the scale, looking at classic 10M Ethernet signaling is perfectly doable with a sub-$1k scope, and 100M should also be fine with anything 500MS/s or higher. Note that Cat5 is rated for a bandwidth of 100MHz.

I love the photo of the probe, and I bet it's very expensive. I wonder if the "A071626" sticker changes the impedance of the traces enough for you to notice :)

https://www.mattkeeter.com/blog/2022-08-11-udp/probes_full.j...

Reply on news.ycombinator.com