Hey all,

I've been wanting to up my oscilloscope game. I currently have a modern 100MHz oscilloscope from Siglent, which I'm quite happy with. However, it is not accurate when measuring fast edges, such as rise times for 74 series chips.
So I'm looking at getting a second-hand 1GHz oscilloscope, which should let me accurately show rise times around 2ns. The Siglent is accurate up to 17.5ns. (This is 0.35 / rise time = signal bandwidth, and 5 * signal bandwidth = oscilloscope bandwidth needed)

I'm currently looking at a Lecroy model, in the LC-5x4 series such as this one https://www.ebay.com/itm/123816182341?h ... Sw4mxdQwHR. Most oscilloscopes sold on eBay don't come with probes, so I'll have to add the high bandwidth probe for this model. That's another few hundred.

I'm not familiar with oscilloscopes from that era, nor this brand. Does anyone have some recommendations or gotchas to avoid?

_________________
BB816 Computer YouTube series

No advice for you, but just chiming in seconding your interest in a higher bandwidth scope. The Siglent has a mode that allows you to look at the individual point readings used to make up the scope trace. I have both 100 MHz and 200 MHz versions of the Siglent scope and there aren't very many points making up the rising/fall edge of some signals I was looking at a while back.

I've pulled the trigger on a LeCroy LC-564A https://www.ebay.com/itm/255132438746.
Looks in good shape and has had the CRT replaced by an LCD. According to the seller, there is an internal VGA interface that he was able to hook into to do this mod.

Now I need to find a high bandwidth probe! I'll report back when I get the scope

_________________
BB816 Computer YouTube series

I meant to post earlier but I would love to see the comparison between the same signal with the LeCroy and the Siglent.

I have a Siglent SDS 1204X-E (the 200Mhz version) and after a bit of learning and wrangling I'm very happy with it. I trust it up to about 80Mhz and I think I've learned how to consistently interpret what is actually going on in 100MHz+ signals. If you have the 1104X-E then it's possible to use it at 200Mhz too as the only difference between the 1104 and the 1204 is the version number in the firmware. A bit of googling for 'unlock Siglent SDS1104X-E' should show you how.

This screenshot is 10ns/div in dot mode at 1GSa/s and there's enough detail to understand the ringing I'm seeing on my signal (a 555 timer run through a Schmitt triggered buffer).

However the screenshots don't really do it justice. Turning on screen persistence (the so called Super Phosphor mode) looks way better and the slight variations in sampling each frame give a much better view of the signal. The above photo is all four probes in 500MSa/s dot mode at 5ns/div with something like 30% persistence enabled.

On the subject of probes, the one's that come with the Siglent are ... decent for the price but do seem to limit the signal quality more than the 'scope does. And (depending on who you believe) the probes that come with the 1104 are the same as those that come with the 1204.

Also, I'm glad to see you posting videos again and hope you're fully recovered. Happy scoping!

For reference to future readers, here are some shots from the Siglent SDS 1202X-E and SDS 1104X-U. The 1202X-E is basically one half of the 1204X-E. It has a 200 MHz bandwidth with 1 GSa/s with 1 channel active but drops to 500 MSa/s with 2 channels active (I believe the 1204X-E can do 1 GSa/s with channels 1 and 3 active, but drops to 500 MSa/s adding in channels 2 and 4). The 1104X-U is the budget 4 channel 100 MHz bandwidth scope (w/ 1 channel you get 1GSa/s, w/ 2 you only get 500 MSa/s and w/ 3 or 4 you get 250 MSa/s).

The 1202X-E trace below is similar to the 1204X-E trace in the previous post with the scope in dot mode with continuous trace. This is with a signal of about 500 Hz.

However, if you take a single capture, the effect of bandwidth becomes more apparent. The rising edge is defined by 5 points.



With a 10 MHz signal we have the following at a 5ns scale in dot mode. Note that the lag at about 3.5 volts is due to the circuit. It's not there after I disconnected the oscillator from the circuit in the probe traces at the bottom of this post. Still, it demonstrates scope performance with a not uncommon discontinuity.

With 2 channels active, the 1202X-E drops to 500 MSa/s, considerably dropping the resolution:




The 1104X-U performs similar to the 1202X-E in 1 or 2 channels:
at 500 Hz:



or 10 MHz:



But things aren't so good with 3 or 4 channels active on the 1104X-U:
At 500 Hz:

At 10 MHz:

you only get a sample about every 5 ns.


The probes that come with the two scopes are different. The 1202X-E comes with 200 MHz, PP215 probes. The 1104X-U come with 100 MHz, PP510 probes (I think the 1104X-E comes with these as well). Physically the two look almost the same. The PP215 has a slightly smaller probe tip than the PP510.

I don't have enough experience to judge the probes electrically, but here are some traces of the two, with the 200 MHz probe in yellow and the 100 MHz probe in purple. First both together at 500 MSa/s.

And singly at 1 GSa/s.

Thank you, much appreciated



Yeah, I was able to do this mod on my scope. I did not get updated probes so they are still at 100MHz. This limits the bandwidth of the complete system quite a bit, unfortunately.



I have no doubt that the Siglent can reproduce the ringing, it looks to be of a lower frequency. The issue I'm trying to solve is measuring and displaying the rise time of the signal accurately

_________________
BB816 Computer YouTube series

That is very good data to have. Thanks a lot.

Indeed, at 1GSa/s, you'll only get a sample every nanosecond, which is not a lot of points in this rising edge. I'm sure the scope is applying some averaging over many traces to increase the resolution during that transition for the measurement, but I'm not familiar with how it works and the resulting accuracy you get.
The fastest IC in my system is a DS1035, with a 2ns typical rise time. I would only get 2-3 data points with the Siglent. The LeCroy I got goes to 4GSa/s, meaning 12 data points per edge in this case.

I've been doing some research, and I found the following document talking about the analog frontend bandwidth, added to the probe's, and the resulting accuracy of the analog signal that gets to the ADC.



According to this document, a 100MHz scope paired with a 100Mhz probe has a combined bandwidth of 70Mhz. According to the rule of 3x, this means it can represent with 5% accuracy up to 14,34ns rise times. Faster than that and the error increases.
The LeCroy, on the other hand, would be 1Ghz with a 1Ghz probe. Similarly, that's 700Mhz total bandwidth, which can accurately represent rise times up to 1,43ns.

So this confirms my numbers in the top post. I think I used 0.35 there and they used 0.338 here. This number depends on the scope's internal circuitry and varies by model and manufacturer...

_________________
BB816 Computer YouTube series

And for completeness, a 200MHz scope with a 100MHz probe (my current setup) has a 90Mhz combined bandwidth so with rule of 3x represents an 11,3ns rise time with 5% error, a bit better than not doing the mod!

_________________
BB816 Computer YouTube series

For the Siglents I have, dot mode is showing the actual captured value. In fact, it records them and you can inspect them individually or get a data dump (very useful when you're looking for an event that doesn't happen very often and can't use it as a trigger). Less clear is what's going on in vector mode, but I think it still is operating on the data present at that time as you can switch back and forth between dot and vector modes on a single trace. The manual says
To me the vector trace doesn't look like a simple "connect-the-dots" trace. It seems like some sort of smoothing is taking place.

For completeness, here's what the manual says about dot mode:

For the curious, this "mod" is not available on the budget 1104X-U as it has completely different internals than the X-E variants. Here is an interesting video doing a teardown comparison of the 1104X-U against the 1104X-E.

The "mods" are available on the 1202X-E but not of much use (to me at least) because the scope already has the 200 MHz bandwidth and you need extra items to take advantage of the other mods.

Just saw this on eBay: https://www.ebay.ca/itm/363920774006?hash=item54bb5e5376:g:fLEAAOSwqZhi3o68&autorefresh=true

I wonder if this is legit? That scope would sell for about $25K with the full set of probes new.

Me thinks this is "fell off the truck" special. I know there is still 11 hours to go (at time of posting), but I don't see it getting from ~400GBP to ~$10K in that time. Maybe it's Boris' scope and he's trying to raise cash now that he quit his job.

_________________
Bill

Thanks!



Not a bad deal, but suspicious indeed, account has no reviews. I might put in a low ball offer and see where that gets me, would be nice to have a single modern scope rather than a LeCroy and a Siglent on my bench. Active probe is another 1.5k though

_________________
BB816 Computer YouTube series

Did not get it sold for 1.2k, really good deal I found this other comparable listing going for 21k https://www.ebay.fr/itm/384582328687?ha ... SwI7ViJbl-
As you said though, this was probably not legit, not sure what the buyer is going to get

_________________
BB816 Computer YouTube series

I was watching that - it was at £310 for ages, then went up to £320 then £330 a few minutes before the end. Then in the last second it leaped up to £1271. Obviously a robo-bidder. I wonder what algorithm came up with the £1271 figure.
It all seemed a bit dubious - "no returns on this item". If it was as described, why wouldn't they take it back? To be honest if it was as described and legitimate he would have no trouble selling it to a second hand equipment dealer and probably get more than £1271 for it. All very fishy...

Scope and active probe acquired!

First impressions: that thing is HUGE and NOISY being able to print traces on paper rolls is fun (and useless). The LCD mod done by the seller is of good quality, better than expected.
Nothing more to report yet, I have to learn how to use it.

Attached is a photo of the clock signal from my breadboard computer measured with the active probe. 12ns rise time! No idea how accurate this is, haven't done any calibration.

_________________
BB816 Computer YouTube series