Siglent SVA1015X and SVA1032X 1.5, 3.2GHz Spectrum & Vector Network Analyzers

[nctnico]

I have a genuine SVA1032x and I have not noticed any part of it warm up, not even a little.  I've had mine open to tighten the connector screws (for some reason they got loose) and I can tell you that aluminum is heavy.  With that in mind, I have 20w transmitters with much less mass in the heat sync that barely gets noticeably warm, so I ask, what inside a RECEIVER needs to get that hot???  I would be concerned enough to find out what is generating the heat.

I wouldn't.

Both you, Spacehen and  Self Bias discuss this "temperature" without stating any temperature measured by thermometer.

"Measuring" temperature by touching it with a hand is highly subjective and two different people will have wildly different perception what "really warm is".
Not to mention that a room all this is happening in can vary by 10 °C and nobody takes that into account.

To many people pain threshold for touching hot objects is mere 50°C, and some will report 40-45°C as very hot. Some will go into 60°C sauna and say "it is warm but it's OK"...

So before getting all crazy about it, a measurement should be done.  If connectors are at 70°C than something is wrong...

Agreed. And using a thermocouple or other sensor physically attached to the object is highly recommended as a thermal camera will only be accurate on matte black surfaces.

[Self Bias]

Grr...So you snatched it from under my nose.

Manchmal hat man Glück. 

Your screenshots mean nothing as sweeps are not seen as Ref levels and Scale are not applicable to the signals used. Turn the measurement table Off so Sweep and Span settings are visible and just use Markers and their values displayed on the screen.

I'm not sure if that's possible when the "Harmonics" measurement is active, but I'll try once I'm back home. That bug did not occur in normal SA mode, only when the measurement function was active. Regular sweeps work fine, both with RBW <30 Hz and when using other detector settings.

Press Preset (check is is set to Factory Default) and show us some aweeps with the trace visible.

Sure, I'll do that.

Then I would like to see the Mode menu if all the SVA options are actually available.
Does the System info display a SN# and model ?

I could take a screenshot, but I already checked that: It has a SN# and all options are permanently active.

Both you, Spacehen and  Self Bias discuss this "temperature" without stating any temperature measured by thermometer.

You're right, and to add to it: It never felt that hot to me. Like, it wasn't hard to touch or anything, just more warm than I would expect. Funnily enough the attached connector/adapter seemed warmer than the outside of the N-Connector on the device too, so I would assume the heat comes from the inner pin.
It never felt "wrong" to me though, so I didn't measure it yet.

[tautech]

Then I would like to see the Mode menu if all the SVA options are actually available.
Does the System info display a SN# and model ?

I could take a screenshot, but I already checked that: It has a SN# and all options are permanently active.

Good. 
Then just a screenshot with the Mode menu showing please.

[Self Bias]

So, I tried a few things. First up, just to make sure, here is what the SA set to default settings look like:

Setting it to a span where I don't have to wait days for the sweep to complete with a RBW of 10 Hz:

All normal. But switching to Harmonics instantly yields the same problem:

I tried to hide the table as tautech suggested, but I couldn't make it disappear. Instead, I noticed a different effect: The graph is only wrong for the harmonics that are being measured. So, combined with

Then just a screenshot with the Mode menu showing please.

, there you go:

[tautech]

So, I tried a few things.
All normal. But switching to Harmonics instantly yields the same problem:

Thanks.

With Harmonics it seems the sweep is mostly above the current Reference level so wind it up some to see the whole trace. Whether that will bring measurements correct IDK.
TBH I've never played with Harmonics.....a task for today maybe.

[Self Bias]

With Harmonics it seems the sweep is mostly above the current Reference level so wind it up some to see the whole trace.

That will not help sadly; the measured value is always above the reference level. Even with the reference set to +30 dBm, it will then read +36 to +38.

One corellation I've noticed: 30 Hz is the lowest frequency the SA will still do in "Sweep" mode. It defaults to "FFT", but "Sweep" can be manually selected. At 10 Hz and below, it's aways "FFT". Maybe the Harmonics measurement doesn't play well with FFT?

[tautech]

Will do some tests with an official SVA1032X, bare with me....lots on...

[tautech]

Now with a chance to do some very quick tests...
Your Mode screenshot looks correct.

Harmonics tests works as expected.
100 MHz Sine, 50-500 MHz sweep
Engage Measure>Harmonics
However turn Harmonics to Off (Measure = Off) and sweep settings have changed  < this is a bug so will report.

[Self Bias]

Harmonics tests works as expected.

So you don't get unreasonably high values in Harmonics measurement when setting the RBW < 30 Hz or the detector to anything but Pos Peak?

[tautech]

Harmonics tests works as expected.

So you don't get unreasonably high values in Harmonics measurement when setting the RBW < 30 Hz or the detector to anything but Pos Peak?

Not tested.
Show us some detailed results displaying your findings.

[Self Bias]

Show us some detailed results displaying your findings.

Can you specify? I'm not sure what else to provide aside from the screenshots in my earlier post, with the RBW set to 10 Hz / the detector set to Neg Peak with the Harmonics measurement producing readings of up to +200 dBm.

So a good test would be to just repeat your measurement from earlier with the 100 MHz Sine, go to Harmonics measurement, then set the RBW to 10 Hz or the Detector to Neg Peak.

[EDIT] And then also switch between Neg Peak and any option other than Pos Peak. With each change, the reading of the fundamental's power should increase.

[Self Bias]

So just in case my initial post was unclear: Here's my full setup. 120 MHz Sine @ - 20 dBm, Harmonics measurement. One photo is with 30 Hz RBW, The other with 10 Hz. Aside from the sporadically appearing notch, the yellow line gets clamped to the top because the readings are all abnormally high when the RBW is < 30 Hz. The same happens if the detector is set to anything but Pos Peak.

[Bad_Driver]

I did a quick check with my liberated SSA. I used my old RF generator with a lot of harmonics @ 100 MHz (-20 dBm) and I can confirm the strange behavior with RBW <30, with 30 and above it works.

[NY2KW]

I just checked this on my SVA1032x, latest firmware but with a Rigol DG4062 and got the same results.  More interesting is try this with no signal and your will see same thing happen going from 30Hz to 10Z.  With no signal, at 30Hz with no signal, fundamental freq power reads less than -100dBm, then change RBW to 10Hz and the power of the fundamental reads near +10 dBm on the table.

[RoV]

Ok, analyzer behavior is a bit weird, but does it make any sense to measure harmonics with 10 Hz RBW? It becomes quite difficult to center the frequency well enough and requires very high frequency stability of both generator and SA.
I suppose in Siglent they didn't even test the harmonics function at such a low RBW! It's a bug, but not one I would care about...

[antenna]

I tried with mine several times, at several different frequency ranges, and although it would often let me enter whatever RBW I wanted, occasionally it would tell me RBW out of range and stop me when inputting anything under 30Hz.  I think the bug is that it don't always stop you from entering it.

[Self Bias]

Thanks for confirming @Bad_Driver and @NY2K. If you have the time, could you also check setting the detector to something other than Pos Peak, e.g. "Neg Peak" or "Sample"?

Ok, analyzer behavior is a bit weird, but does it make any sense to measure harmonics with 10 Hz RBW?

Good question. from my limited understanding, it looked like the Harmonics function either does a small sweep around the harmonics or a zero span for each of them. Sadly there's no documentation on it. If it's zero span, then I agree and it's probably not a useful feature in most cases.

I tried with mine several times, at several different frequency ranges, and although it would often let me enter whatever RBW I wanted, occasionally it would tell me RBW out of range and stop me when inputting anything under 30Hz.  I think the bug is that it don't always stop you from entering it.

Interesting, I had that feeling too. Did you verify that the tracking generator was off? With it turned on, RBW will always be limited to 30 Hz.

All in all it seems like the validity check for the RBW might just not be working all the time. Though I'm still curious about the detector bug.

[antenna]

Did you verify that the tracking generator was off? With it turned on, RBW will always be limited to 30 Hz.

Yes, it was off.  I was using an external sig gen.

[Self Bias]

With no more feedback on those issues, I guess they don't affect too many people in everyday use. Might still be worth fixing them.

@tautech: Could you report those two findings to Siglent?

[antenna]

I'll toss a stupid question out there because I honestly don't understand this thread (but very interested in it for the same reason)...   

From my understanding, the harmonic test simply avoids sweeping all the stuff between the harmonics allowing a narrow RBW setting to cover all the harmonics in a reasonable amount of time instead of sweeping the whole way, but how small of a harmonic are we looking for? 

I get it, with mine in normal SA mode and set to 1Hz, I can see pretty small signals (IIRC down to -165dBm or so), and that's great for looking at a homebrew duplexer, but from a device harmonics perspective, acceptable harmonics are usually -50 to -100dBc, so is it really necessary to see a harmonic that is sitting down at -160dBm?  I wouldn't consider an SVA1032x the go-to tool for repairing something sophisticated enough to need to see a harmonic down in the natural weeds, so whats the point here?  Is 30Hz not enough?

And if its not enough, go to normal SA mode, set it to 1Hz, and sweep each harmonic manually.

[Self Bias]

I wouldn't consider an SVA1032x the go-to tool for repairing something sophisticated enough to need to see a harmonic down in the natural weeds, so whats the point here?

Not a stupid question at all. I was just exploring the device and found those two bugs while doing so. Whether the malfunctioning configurations are actually useful is a different and very debatable question. I would assume Siglent's fix being to lock out RBWs < 30 Hz für Harmonics measurement anyway. Not sure what they'll do about the detector, but maybe also just limit it to pos peak.

[IM3]

I got rid of my SVA1032X.

The reason is that Siglent is not completely honest about the phase noise in its specifications. They state a phase noise of -98dBc at 1 GHz. My device just manages that, but the rest of the spectrum is much worse. Look at the attached screenshot.

I made the phase noise visible for 3 frequencies and plotted them on top of each other. (signalgenerator R&S SMB100)
1GHz is yellow line,
500MHz is red line,
200 MHz is blue line.

You notice that at 200 MHz the phase noise at 10kHz is 13dB worse than at the specified 1GHz. I call that misleading. A renowned manufacturer ensures that the specifications are met across the entire work area.

Many people want to use this analyzer on VHF and UHF and may get disappointed.

So I bought an SSA3032X-R which has a different RF architecture. This one does get the phase noise right over a wide frequency range. See screen dump. Same settings, same colors, but a much better performance on the phasenoise.

[RoV]

The reason is that Siglent is not completely honest about the phase noise in its specifications. They state a phase noise of -98dBc at 1 GHz. My device just manages that, but the rest of the spectrum is much worse. Look at the attached screenshot.

I have tried to replicate your measurements with my SSA3021X+ upgraded to SVA1032. Unfortunately my generator is probably the weak ring of the chain, being an old Marconi 2019A. However, I get better results.

1GHz is yellow line,
200 and 500MHz are red and blue lines, I wouldn't bet on the order  200 and 500 look slightly better than 1 G.

[IM3]

Yes indeed, your phase noise is better.

I am fairly certain that my measurements on the SVA1032X are not unique to my device because I have read some other reports on the internet (and of course EEVBlog) that the phase noise is not that good over the entire spectrum for this series.

Now you have an SSA3021X+ and I'm not sure if it has the same hardware as an SVA1032X.

Bad phase noise is usually caused by the loop filters in the PLL's. It may be that Siglent has already made improvements to the SSA series.

[tautech]

Yes indeed, your phase noise is better.

What is Samp detection in your screenshots ? Never used it 

Now you have an SSA3021X+ and I'm not sure if it has the same hardware as an SVA1032X.

Instead be sure it does.