Another radio amateur about to buy a DSO ...

[tautech]

Any chance you could put a VNA on one of yours, in order to settle the issue?

No chance, no 2kX-E currently in stock.

[tggzzz]

Any chance you could put a VNA on one of yours, in order to settle the issue?

No chance, no 2kX-E currently in stock.

Shame.

Are there any others with a 1Mohm//*pF and a nominal 50ohm input, but no VSWR spec?
Ditto 50ohm with a VSWR spec?

Either would provide a useful clue.

[Mike99]

Any chance you could put a VNA on one of yours, in order to settle the issue?

No chance, no 2kX-E currently in stock.

Well now I'm leaning towards the 2202X-E! I'm going to the dealer's on Tuesday for a demo of various models and planning to walk out with one, so if I take my NanoVNA with me I'm sure they'll let me look at the VSWR 

Mike

[tggzzz]

Any chance you could put a VNA on one of yours, in order to settle the issue?

No chance, no 2kX-E currently in stock.

Well now I'm leaning towards the 2202X-E! I'm going to the dealer's on Tuesday for a demo of various models and planning to walk out with one, so if I take my NanoVNA with me I'm sure they'll let me look at the VSWR 

Excellent we look forward to hearing your conclusions.

In that position I would want to try some form of quick and dirty experiment at home, to debug my experimental technique without embarrassing myself in public. Me a pessimist? Shurely shome mishtake 

[Mike99]

Any chance you could put a VNA on one of yours, in order to settle the issue?

No chance, no 2kX-E currently in stock.

Well now I'm leaning towards the 2202X-E! I'm going to the dealer's on Tuesday for a demo of various models and planning to walk out with one, so if I take my NanoVNA with me I'm sure they'll let me look at the VSWR 

Excellent we look forward to hearing your conclusions.

In that position I would want to try some form of quick and dirty experiment at home, to debug my experimental technique without embarrassing myself in public. Me a pessimist? Shurely shome mishtake 

It shouldn't be a problem measuring VSWR. I will try it on my Tek 2430A first.

Mike

[RoGeorge]

18pF at 350MHz is only 25\$\Omega\$, while in the specs the impedance is given as 50\$\Omega\$.  No imaginary part and no || 18pF or other value is specified, which makes me guess the 18pF are compensated such that the 50\$\Omega\$ impedance is resistive only.

[tggzzz]

18pF at 350MHz is only 25\$\Omega\$, while in the specs the impedance is given as 50\$\Omega\$.  No imaginary part and no || 18pF or other value is specified, which makes me guess the 18pF are compensated such that the 50\$\Omega\$ impedance is resistive only.

Possibly, but many scopes - including from reputable manufacturers like HP and Tek - don't bother. Hence the desire to have it measured.

One scope that gets it right has two attenuated (one 50ohm, one 1Mohm), and an RF relay to switch between them. That has the advantage that an overload pops the relay back to 1Mohm before the 50ohm circuits are damaged.

[bdunham7]

Well now I'm leaning towards the 2202X-E! I'm going to the dealer's on Tuesday for a demo of various models and planning to walk out with one, so if I take my NanoVNA with me I'm sure they'll let me look at the VSWR 

18pF at 350MHz is only 25\$\Omega\$, while in the specs the impedance is given as 50\$\Omega\$.  No imaginary part and no || 18pF or other value is specified, which makes me guess the 18pF are compensated such that the 50\$\Omega\$ impedance is resistive only.

There were two previous discussions on this:

https://www.eevblog.com/forum/beginners/whats-the-input-capacitance-of-an-oscilloscopes-50-ohm-input/

https://www.eevblog.com/forum/testgear/need-50-ohm-input-on-an-(old)-oscilloscope-that-has-only-1m-ohm-inputs-bnc-tee/msg4484371/#msg4484371

They're not long, and you'll find that I tested a Tek 485, Tek 2465B and a Siglent SDS2354X+ with a NanoVNA.  The 485 is best of all, but the Siglent is essentially perfect to 350MHz, as good as the 2465B and more than likely at least as good as the 2430A since the attenuator design is similar to the 2465B.  The Siglent had a VSWR of 1.43 @ 500MHz, but it only curved up significantly over the 400-500MHz span.  Below that it was low enough to not worry about, IMHO.

Also in that discussion I measured the Siglent with an LCR meter and it seemed to indicate that the input capacitance of 18pF had some 10k ESR.  Some other theories and diagrams were presented, but what seems clear is that the input capacitance specified for the 1M input is not simply the equivalent of an 18pF directly across the input.  Thus statements observing that an 18pF capacitor would have a very low impedance at the upper end of the scopes BW aren't really relevant because the input has additional complexity.  An 18pF capacitor with a 10k resistor in series still looks a lot like 18pF if the rest of the circuit is 1M.  But even you just slap a 50R resistor across all that, the 18pF now looks like about 10k, not 25R, and the phase angle will be nearly zero. 

Ultimately I don't  have accurate enough stuff to try and model these inputs with any real precision, but it seems pretty clear to me that it isn't just 50R with 18pF across it and then they simply fail to mention the capacitance.  The results from that would be really bad, people would notice. 

[Mike99]

18pF at 350MHz is only 25\$\Omega\$, while in the specs the impedance is given as 50\$\Omega\$.  No imaginary part and no || 18pF or other value is specified, which makes me guess the 18pF are compensated such that the 50\$\Omega\$ impedance is resistive only.

Possibly, but many scopes - including from reputable manufacturers like HP and Tek - don't bother. Hence the desire to have it measured.

One scope that gets it right has two attenuated (one 50ohm, one 1Mohm), and an RF relay to switch between them. That has the advantage that an overload pops the relay back to 1Mohm before the 50ohm circuits are damaged.

That's what the 2430A does. It has a 50 ohm termination resistor and thermal sensor. The manual states 1.3:1 VSWR to 150MHz. I'll measure it shortly.

Mike

[tggzzz]

18pF at 350MHz is only 25\$\Omega\$, while in the specs the impedance is given as 50\$\Omega\$.  No imaginary part and no || 18pF or other value is specified, which makes me guess the 18pF are compensated such that the 50\$\Omega\$ impedance is resistive only.

Possibly, but many scopes - including from reputable manufacturers like HP and Tek - don't bother. Hence the desire to have it measured.

One scope that gets it right has two attenuated (one 50ohm, one 1Mohm), and an RF relay to switch between them. That has the advantage that an overload pops the relay back to 1Mohm before the 50ohm circuits are damaged.

That's what the 2430A does. It has a 50 ohm termination resistor and thermal sensor. The manual states 1.3:1 VSWR to 150MHz. I'll measure it shortly.

Mike

Sounds like the earlier HP1740, except that doesn't have the relay. You can guess a common cause of failure!

[G0HZU]

Well now I'm leaning towards the 2202X-E! I'm going to the dealer's on Tuesday for a demo of various models and planning to walk out with one, so if I take my NanoVNA with me I'm sure they'll let me look at the VSWR 

18pF at 350MHz is only 25\$\Omega\$, while in the specs the impedance is given as 50\$\Omega\$.  No imaginary part and no || 18pF or other value is specified, which makes me guess the 18pF are compensated such that the 50\$\Omega\$ impedance is resistive only.

There were two previous discussions on this:

https://www.eevblog.com/forum/beginners/whats-the-input-capacitance-of-an-oscilloscopes-50-ohm-input/

https://www.eevblog.com/forum/testgear/need-50-ohm-input-on-an-(old)-oscilloscope-that-has-only-1m-ohm-inputs-bnc-tee/msg4484371/#msg4484371

They're not long, and you'll find that I tested a Tek 485, Tek 2465B and a Siglent SDS2354X+ with a NanoVNA.  The 485 is best of all, but the Siglent is essentially perfect to 350MHz, as good as the 2465B and more than likely at least as good as the 2430A since the attenuator design is similar to the 2465B.  The Siglent had a VSWR of 1.43 @ 500MHz, but it only curved up significantly over the 400-500MHz span.  Below that it was low enough to not worry about, IMHO.

Also in that discussion I measured the Siglent with an LCR meter and it seemed to indicate that the input capacitance of 18pF had some 10k ESR.  Some other theories and diagrams were presented, but what seems clear is that the input capacitance specified for the 1M input is not simply the equivalent of an 18pF directly across the input.  Thus statements observing that an 18pF capacitor would have a very low impedance at the upper end of the scopes BW aren't really relevant because the input has additional complexity.  An 18pF capacitor with a 10k resistor in series still looks a lot like 18pF if the rest of the circuit is 1M.  But even you just slap a 50R resistor across all that, the 18pF now looks like about 10k, not 25R, and the phase angle will be nearly zero. 

Ultimately I don't  have accurate enough stuff to try and model these inputs with any real precision, but it seems pretty clear to me that it isn't just 50R with 18pF across it and then they simply fail to mention the capacitance.  The results from that would be really bad, people would notice.

As I mentioned in that other thread, by deliberately adding some inductance in series with the 50R termination the circuit will mimic approx 50R in parallel with a negative capacitance and it will do this over a wide bandwidth. The negative parallel capacitance will cancel some of the existing 18pF over a fairly wide bandwidth.

So that simple trick alone will mask some of the input capacitance when the 50R terminator is selected.

The other way to do it is to arrange the 50R termination as a potential divider made up of a series 20R and then switch in a shunt 30R with a relay when in 50R mode. This will mask the 18pF input capacitance yet further and when combined with the inductance trick, the scope could achieve a decent VSWR over a large bandwidth. The penalty will be a higher noise floor due to the attenuation introduced by the 20R and 30R potential divider.

[Mike99]

The 2430A service manual doesn't show the input arrangement in enough detail but I've just measured mine.

Sorry about the layout I can't get the hang of the attachments.

Also ignore the cal description, it is calibrated over the full measurement range.

R+jX for the 1M input.

VSWR for the 50 ohm input

Mike

[Fungus]

Yes I know these can be hacked but I would be reluctant to try until it's out of warranty.

All you're doing is entering a license key as if you'd bought it from Rigol. Inputting a few numbers can't void the warranty.

nb. It can easily be removed, too.

[Performa01]

I don't have an SDS2000X-E, only SDS2000X Plus. I honestly cannot tell whether the little 2000X-E has a true 50 ohms path like the SDS2000X Plus (and higher models) have, but the following demonstration should still be universal to show how an external through termination is only an option at low frequencies. For higher frequencies, a proper internal 50 ohms path is mandatory to get a reasonable VSWR , hence reasonably accurate measurements.

Look at reply #237 here:

https://www.eevblog.com/forum/testgear/rigol-dho804-test-and-compare-thread/msg5114334/#msg5114334

[Mike99]

I don't have an SDS2000X-E, only SDS2000X Plus. I honestly cannot tell whether the little 2000X-E has a true 50 ohms path like the SDS2000X Plus (and higher models) have, but the following demonstration should still be universal to show how an external through termination is only an option at low frequencies. For higher frequencies, a proper internal 50 ohms path is mandatory to get a reasonable VSWR , hence reasonably accurate measurements.

Look at reply #237 here:

https://www.eevblog.com/forum/testgear/rigol-dho804-test-and-compare-thread/msg5114334/#msg5114334

That's very interesting, thank you. It's nudging me towards the 2202X-E which has a selectable 50 ohm input.

Mike

[Performa01]

That's very interesting, thank you. It's nudging me towards the 2202X-E which has a selectable 50 ohm input.

It would be a sensible choice especially for a radio amateur. While the SDS1202X-E really is the bottom of the barrel instrument in the modern Siglent lineup (and has less features than the SDS1004X-E), the SDS2000X-E series gives you not only higher sample rate and bandwidth together with the 50 ohm inputs required for that, but also more software features like Search and Bode Plot.

The latter should be interesting especialy for a ham, since it works with any Siglent AWG and offers up to three channels up to 120 MHz at a high dynamic range >100 dB. So you can not only characterize your newly designed three-way crossover network for your loudspeaker in one go, but also all sorts of (IF) filters. Yes a VNA is faster, but the Bode Plot can be used down to 10 Hz at high accuracy.

EDIT: Yeah, for three channels we need a four channel scope, so the SDS2000X-E is limited to just one channel...

[ebastler]

Bode Plot [...] should be interesting especialy for a ham, since it works with any Siglent AWG and offers up to three channels up to 120 MHz at a high dynamic range >100 dB. So you can not only characterize your newly designed three-way crossover network for your loudspeaker in one go, but also all sorts of (IF) filters. Yes a VNA is faster, but the Bode Plot can be used down to 10 Hz at high accuracy.

But the three-channel plots would only work with a four-channel scope, right? The SDS2000X-E series only comprises dual-channel models, of which one would be used for the input and one for the (single) output of the DUT, in my understanding.

[Mike99]

The deed is done. I bought the Siglent SDS2202X-E and they gave me a discount 

Here is the input VSWR measured on an FA-VA5 VNA (because my NanoVNA has a couple of inband spurious signals).

Mike

[bdunham7]

Here is the input VSWR measured on an FA-VA5 VNA (because my NanoVNA has a couple of inband spurious signals).

Are you using the basic calibration elements that came with the FA-VA5 or the better ones that they sell separately?

[Mike99]

Here is the input VSWR measured on an FA-VA5 VNA (because my NanoVNA has a couple of inband spurious signals).

Are you using the basic calibration elements that came with the FA-VA5 or the better ones that they sell separately?

Shortly after I put that picture up I realised it was nonsense, and yes I do only have the basic cal kit for that VNA.

I fixed the problem with my NanoVNA and tried again, using the SMA cal kit that came with it. Not ideal at higher frequencies but needs must. The result may not be spot on but it's the shape I expected.

Mike

[Performa01]

Well, this is a 200 MHz scope, right?

From the plot I can see a VSWR <1.5 up to 340 MHz, so there is nothing to complain about.

Most likely there is a difference between the direct path up to 100 mV/div and above that, when the first internal attenuator gets switched in, e.g. 200 mV/div. With attenuator, the VSWR usually gets better.

[2N3055]

Well, this is a 200 MHz scope, right?

From the plot I can see a VSWR <1.5 up to 340 MHz, so there is nothing to complain about.

Most likely there is a difference between the direct path up to 100 mV/div and above that, when the first internal attenuator gets switched in, e.g. 200 mV/div. With attenuator, the VSWR usually gets better.

Anyways, it shows 50Ω path is not made by simply putting 50Ω terminator in parallel with 1MΩ/20pF input...

[ebastler]

Anyways, it shows 50Ω path is not made by simply putting 50Ω terminator in parallel with 1MΩ/20pF input...

This would be a nice opportunity to demonstrate the difference.
@Mike99, any chance you could repeat the measurement with the scope's input set to 1 MOhm and an external 50 Ohm pass-through terminator? Thanks!

[Performa01]

This would be a nice opportunity to demonstrate the difference.

Was this not demonstration enough?

https://www.eevblog.com/forum/testgear/another-radio-amateur-about-to-buy-a-dso/msg5164407/#msg5164407

[tggzzz]

Here is the input VSWR measured on an FA-VA5 VNA (because my NanoVNA has a couple of inband spurious signals).

Are you using the basic calibration elements that came with the FA-VA5 or the better ones that they sell separately?

Shortly after I put that picture up I realised it was nonsense, and yes I do only have the basic cal kit for that VNA.

I fixed the problem with my NanoVNA and tried again, using the SMA cal kit that came with it. Not ideal at higher frequencies but needs must. The result may not be spot on but it's the shape I expected.

We all have the occasional brain failure!

Looks like I was right to question the statements in the earlier post shown below

A cursory inspection of the document referred to by Tautech would lead you to think that it has a 50ohm input.
    Channel Flatness (Inner 50 Ω)
    DC - 60% (BW): ± 1 dB
    60% - 100% (BW): + 1 dB/-3 dB
    Max. Input voltage
    1 MΩ: ≤ 400 Vpk (DC + Peak AC <= 10 kHz)
    50 Ω: ≤ 5V rms

A closer inspection reveals what I regard as specmanship...
    Impedance
    DC 1 MΩ: (1 MΩ ± 2%) || (18 pF ± 2 pF)
    DC 50 Ω: 50 Ω ± 2%

Of course only an owner can say whether that is significant in their application.

At a push I could go for the Siglent SDS2202X-E.

A 50 ohm input would be useful but I was planning to use a feedthrough terminator.

You do get more features with 2kX-E models.
1M/50 Ohm input path.

Are they 50ohm inputs or 1Mohm//50ohm//15pF inputs?

What's the VSWR, which is relevant at RF frequencies?

https://int.siglent.com/u_file/document/SDS2000X-E_DataSheet_DS0102E-E02A.pdf

VSWR isn't specified, which often happens where the manufacturer just puts a 50ohm resistor in parallel with the 15pF 18pF.

Any chance you could put a VNA on one of yours, in order to settle the issue?