SA bandwidth for precompliance testing

[jmw]

If I'm interested in EMC precompliance testing (with both conducted and near-field probing), are the lower-specced analyzers around 1.5 Ghz capable for the job or is more range needed? Let's say I'm mostly interested in looking at power supplies, which usually switch in the 100s of kHz - 1s of Mhz range.

[tautech]

The SSA3021X has probably been the most popular for EMI work @ $1395 and mainly due to the improvements to 3.2 GHz that can be made to it. Of the few EMI fiddlings I've done with my SSA3032X I've found most of the emissions in the hundreds of MHz and not well up into GHz.

If your other needs are more varied the new SVA1015X with it's additional capabilities might suit you better but AFAIK you won't be able to improve on its BW. It's too new for other improvements to have been found for it yet but you can be sure some members here will be looking.
Both series of these SA's are advertised as suitable for EMI work.

Hunt out member dazz1 threads as he's venturing down this path and has a SSA3021X.....tweaked I think.

[nctnico]

If I'm interested in EMC precompliance testing (with both conducted and near-field probing), are the lower-specced analyzers around 1.5 Ghz capable for the job or is more range needed? Let's say I'm mostly interested in looking at power supplies, which usually switch in the 100s of kHz - 1s of Mhz range.

For conducted emissions the maximum frequency is somewhere around 30MHz to 100MHz. For probing radiated emissions it really depends on the power supply. I've seen simple buck converters emitting noise to up to 300MHz. All in all 1.5GHz range should be plenty for power supply work unless you are working on really high frequency DC-DC converters.

[CopperCone]

since you don't have a preselector you would wanna go 5dB lower then the standard you are trying to meet I think.

[nctnico]

I'd be careful recommending a VNA for use as a spectrum analyser. A spectrum analyer and VNA may look the same at first glance but their purpose and therefore construction are different!

[TK]

I think the Siglent SVA1015X is a spectrum analyzer with the option to do some vector network analysis.  It seems to have less capabilities than the SSA3021X and I am not sure if any vector network analysis is required for pre-compliance EMI testing.

[nctnico]

I think the Siglent SVA1015X is a spectrum analyzer with the option to do some vector network analysis.

If that where true then Siglent would have bolted some SMAs onto their spectrum analyser and be done with it. But they didn't because they couldn't. The requirements are totally different between a spectrum analyser and a network analyser. A network analyser has to measure the phase and amplitide differences between the outgoing and incoming signal very precisely. OTOH a network analyser doesn't have to care so much about harmonics and mixing products.

[TK]

I think the Siglent SVA1015X is a spectrum analyzer with the option to do some vector network analysis.

If that where true then Siglent would have bolted some SMAs onto their spectrum analyser and be done with it. But they didn't because they couldn't. The requirements are totally different between a spectrum analyser and a network analyser. A network analyser has to measure the phase and amplitide differences between the outgoing and incoming signal very precisely. OTOH a network analyser doesn't have to care so much about harmonics and mixing products.

I don't understand what you are saying... can you please clarify?

[tautech]

I think the Siglent SVA1015X is a spectrum analyzer with the option to do some vector network analysis.

Exactly. Plus other stuff as well.

It seems to have less capabilities than the SSA3021X and....

Less only in BW and some lesser accuracy spec than SSA3kX models, but here we're talking of precompliance tests. If the test results are close to max permitted EMI values and instrument accuracy is 'near enough' then obviously a design re-spin/tweak is gunna be prudent rather than have to risk a second $$ compliance test.

....I am not sure if any vector network analysis is required for pre-compliance EMI testing.

Correct, there's no VNA requirement for EMI work unless you're wanting to build and tune your own EMI antennae so for this the new SVA could be very useful.

[nctnico]

I think the Siglent SVA1015X is a spectrum analyzer with the option to do some vector network analysis.

If that where true then Siglent would have bolted some SMAs onto their spectrum analyser and be done with it. But they didn't because they couldn't. The requirements are totally different between a spectrum analyser and a network analyser. A network analyser has to measure the phase and amplitide differences between the outgoing and incoming signal very precisely. OTOH a network analyser doesn't have to care so much about harmonics and mixing products.

I don't understand what you are saying... can you please clarify?

Just know that a spectrum analyser and a VNA are designed and optimised for different purposes. A VNA is typically worse at doing spectrum analysis compared to a spectrum analyser.

[TK]

Just know that a spectrum analyser and a VNA are designed and optimised for different purposes. A VNA is typically worse at doing spectrum analysis compared to a spectrum analyser.

You are right, but there is a market for SA with VNA capabilities, like the JDSU (ex GenComm) 754A, Anritsu MS2711D, Keysight handheld SA, etc.  Maybe they are more Antenna and Cable Fault analyzers and not true VNA, but for the HAM Radio market, the combined function is very useful.

[Smokey]

...

....I am not sure if any vector network analysis is required for pre-compliance EMI testing.

Correct, there's no VNA requirement for EMI work unless you're wanting to build and tune your own EMI antennae so for this the new SVA could be very useful.

(Sorry, I know it's an old thread... )

What about tuning antenna matching network components on a PCB for the DUT?  Having a matching network that is way out of whack could cause issues during testing.  Need a VNA to sort that out.

[tautech]

...

....I am not sure if any vector network analysis is required for pre-compliance EMI testing.

Correct, there's no VNA requirement for EMI work unless you're wanting to build and tune your own EMI antennae so for this the new SVA could be very useful.

(Sorry, I know it's an old thread... )

What about tuning antenna matching network components on a PCB for the DUT?  Having a matching network that is way out of whack could cause issues during testing.  Need a VNA to sort that out.

No worries.
Your only issue is connection and calibration at that point. If it's a experimental design you might use a SMA PCB edge connector and do the Cal just before it.
For that reason we chose a F603FE Female Cal kit so to be able to do a Cal at the end of our test cables so to get as near as possible to the DUT. Small stuff you might use a male PCB edge connector so to fit directly to the port adapter however you will find some antennae particularly susceptible to proximity effects so it's better to have them in free air rather than hard against an instrument.

[Smokey]

I hacked some u.FL connectors into my last design for testing.

For my next designs that have chip antennas, I'm planning on putting unpopulated u.FL connectors footprints at both ends of the transmission line on the PCB (at source, and after the matching network right before the the antenna).  That way I can hook the VNA up without the source and match the antenna, and also have a connection at the other end with the antenna removed for direct testing.  u.FL signal and ground pads pretty much line up with the 50ohm microstrip trace.

[pdenisowski]

If I'm interested in EMC precompliance testing (with both conducted and near-field probing), are the lower-specced analyzers around 1.5 Ghz capable for the job or is more range needed? Let's say I'm mostly interested in looking at power supplies, which usually switch in the 100s of kHz - 1s of Mhz range.

I just presented a module two days ago on EMC precompliance at our Demystifying EMC conference   I think it will be available on-demand at some point, and I'm planing on turning it into a YouTube video / whitepaper shortly thereafter as well.

In the meantime, my short answer would be that 1.5 GHz is probably fine for SMPS precompliance, for both conducted and radiated emissions.

Perhaps more important than upper frequency limit are the noise of analyzer and a good set of near field probes (and potentially a preamplifier, if analyzer noise is high or probe quality is low).  A noisy instrument and/or low quality probes (or poor probing technique) are common issues in precompliance testing.