Siglent SDS2000X HD, 12bit, 2GSa/s high definition oscilloscope serie.

[dietert1]

Does it really matter whether the bandwidth is 579.70 or 579.75 MHz?
The estimated -3 dB result is 579,691 MHz by linear interpolation.

Regards, Dieter

[KungFuJosh]

I would guess it's due to "limits" in the resolution.
I mean, isn't a difference of 0.027dBm pretty damn small to begin with?

The limitation is because we can only move markers based on the X axis. If we set the marker to 1MHz, it will show us whatever the exact Y value is.

The correct resolution is to add selection based on Y axis to show X axis value.

[KungFuJosh]

Does it really matter whether the bandwidth is 579.70 or 579.75 MHz?
The estimated -3 dB result is 579,691 MHz by linear interpolation.

Regards, Dieter

Of course not!  None of this matters at all, but that's entirely besides the point. 😉

Shouldn't the estimated -3dB result take into account the -0.018dBm measurement at 1MHz, and so be higher than 579.7MHz? Where's my extra 0.013dB?? 😉😉

[2N3055]

Of course there will be discrete steps. Vertically and horizontally.

Vertically because vertical resolution. Horizontally because FFT. It sorts frequencies in discrete bins.
And while some sort of interpolation for the readout might be made, it is useless because all of that detail is already in the error cloud.

So no, I don't see that happening.

[KungFuJosh]

Of course there will be discrete steps. Vertically and horizontally.

Vertically because vertical resolution. Horizontally because FFT. It sorts frequencies in discrete bins.
And while some sort of interpolation for the readout might be made, it is useless because all of that detail is already in the error cloud.

So no, I don't see that happening.

Selecting based on Y axis could still be plenty useful. Maybe not as much in this case, but there are certainly cases where it would be more helpful.

[2N3055]

Of course there will be discrete steps. Vertically and horizontally.

Vertically because vertical resolution. Horizontally because FFT. It sorts frequencies in discrete bins.
And while some sort of interpolation for the readout might be made, it is useless because all of that detail is already in the error cloud.

So no, I don't see that happening.

Selecting based on Y axis could still be plenty useful. Maybe not as much in this case, but there are certainly cases where it would be more helpful.

I was talking about "unlimited resolution" comments. As for inverting how tracking cursors work I have no opinion at this moment. I pretty much use it as it is and is sufficient for me for what cursors are.
Cursors are more of a "tongue at a right angle" estimations to me. 

[dietert1]

Does it really matter whether the bandwidth is 579.70 or 579.75 MHz?
The estimated -3 dB result is 579,691 MHz by linear interpolation.

Regards, Dieter

Of course not!  None of this matters at all, but that's entirely besides the point. 😉

Shouldn't the estimated -3dB result take into account the -0.018dBm measurement at 1MHz, and so be higher than 579.7MHz? Where's my extra 0.013dB?? 😉😉

Is this a 350 MHz SDS2354X HD? Did you apply some kung fu?

Regards, Dieter

[KungFuJosh]

Is this a 350 MHz SDS2354X HD? Did you apply some kung fu?

No kung fu involved, Siglent upgraded it.

Thanks,
Josh

[Sorama]

but is upgraded to 350MHz ?

[KungFuJosh]

No, it's a 350 model upgraded to 500. If not 500MHz, we wouldn't see the bw that high in the screenshots.

[Sorama]

No, it's a 350 model upgraded to 500. If not 500MHz, we wouldn't see the bw that high in the screenshots.

that would have been my next question :-)

tnx.

[KungFuJosh]

This was...interesting. I had a cable "professionally made" directly from Messi & Paoloni. It looks nice. It's the Stealthflex 7 series with their N connector on one side, and BNC on the other.

I compared that against the short Hyperflex 5 cable I made with the same connectors (but for the 5 series).

[dietert1]

Your FFT screen shots demonstrate the correct behavior of a scope frontend. Between 500 MHz and 1 GHz i read about 13.5 dB, that is about the 12 dB expected for a 2nd order low pass. As far as i remember BNC connectors introduce losses already at 1 GHz and shouldn't be used for measurements above 2 GHz.
The wiggles introduced by your cables are a fraction of 1 dB. With a 12 bit scope the differences may be visible in the pulse shape.

Regards, Dieter

[KungFuJosh]

What about the differences in the wiggles? The amplitude seems similar, but does one look better than the other?

Thanks,
Josh

[Performa01]

What about the differences in the wiggles? The amplitude seems similar, but does one look better than the other?

The wiggles are affected by signal generator output mismatch and cable length.

To optimize the output impedance, you can connect a high quality inline attenuator of at least 6 dB (more is better) directly at the generator output.

If such an attenuator is not available, you can try to lower the signal level until you hear a relay acting inside the signal generator.

Even with all these measures, there might still be wiggles left, e.g. just because of the characteristic impedance of the cable, which should nominally be 50 ohms, but still is subject to tolerances.

[KungFuJosh]

...a high quality inline attenuator of at least 6 dB (more is better) directly at the generator output.

How's this?: https://www.pasternack.com/12db-fixed-n-male-n-female-2-watts-attenuator-pe7047-12-p.aspx

I assume this is not good enough: https://www.amazon.com/dp/B0DKFDD3KT/

Thanks,
Josh

[mawyatt]

Mini Circuits is a good source for attenuators (and cables).

https://www.minicircuits.com/WebStore/Attenuators.html#product_detail_one

https://www.minicircuits.com/WebStore/Cables.html#product_detail_one



Best

[KungFuJosh]

https://www.minicircuits.com/WebStore/Attenuators.html#product_detail_one

Is that Pasternack attenuator that costs half as much as the minicircuits one not good enough?

[mawyatt]

https://www.minicircuits.com/WebStore/Attenuators.html#product_detail_one

Is that Pasternack attenuator that costs half as much as the minicircuits one not good enough?

Don't know about Pasternack, but the Mini Circuits UNAT-10 we have is OK.

Best

[Martin72]

...a high quality inline attenuator of at least 6 dB (more is better) directly at the generator output.

How's this?: https://www.pasternack.com/12db-fixed-n-male-n-female-2-watts-attenuator-pe7047-12-p.aspx

I assume this is not good enough: https://www.amazon.com/dp/B0DKFDD3KT/

Thanks,
Josh

You can also buy one from Huber+Suhner, which will set you back just under €200....
Well, what might the criteria be...
I'm no HF expert, so I can only guess.
SWR will probably be important, although all the attenuators mentioned are similar in this respect.
Then probably bandwidth, tolerance, and workmanship.
I would be interested to know, because I have attenuators from H+S and Mini-Circuits, but they are all BNC types, and I don't have an N type.

[KungFuJosh]

Don't know about Pasternack, but the Mini Circuits UNAT-10 we have is OK.

Perfect, thanks! That one didn't come up searching on Mouser because they don't have the N connection listed for it. 🙄

That one is only $30, and that works for me. 😉


Thanks,
Josh

[mawyatt]

Don't know about Pasternack, but the Mini Circuits UNAT-10 we have is OK.

Perfect, thanks! That one didn't come up searching on Mouser because they don't have the N connection listed for it. 🙄

That one is only $30, and that works for me. 😉


Thanks,
Josh

They have some nice cables that are also ~$30. We don't have any (yet?), but will probably order some next time we order from them (or Mouser).

Best

[KungFuJosh]

They have some nice cables that are also ~$30. We don't have any (yet?), but will probably order some next time we order from them (or Mouser).

I may try one at some point for comparison, but I just blew a bunch of money on the M&P stuff, which I'm happy with anyway.

Having them custom make me that Stealthflex 7 cable was about $34 after conversion, so not bad at all really. That price included the fee for the attached report I was also curious about.

Thanks,
Josh

[Performa01]

...a high quality inline attenuator of at least 6 dB (more is better) directly at the generator output.

How's this?: https://www.pasternack.com/12db-fixed-n-male-n-female-2-watts-attenuator-pe7047-12-p.aspx

This one looks very good. It will allow you to explot the full bandwidth of your generator.

I assume this is not good enough: https://www.amazon.com/dp/B0DKFDD3KT/

It doesn't look that bad, but with a rated bandwidth of 3 GHz, it might not perform optimally at frequencies above 1 GHz. On the other hand, cables have a frequency dependant attenuation also, hence this concern might not apply.

With regard to the cable, the Hyperflex 5 is about the best with ~5 mm diameter we can get up to 2 GHz. You would have a hard time finding a better one, regardless the price.

Above 2 GHz, there might be better options, although vastly more expensive.

[mawyatt]

They have some nice cables that are also ~$30. We don't have any (yet?), but will probably order some next time we order from them (or Mouser).

I may try one at some point for comparison, but I just blew a bunch of money on the M&P stuff, which I'm happy with anyway.

Having them custom make me that Stealthflex 7 cable was about $34 after conversion, so not bad at all really. That price included the fee for the attached report I was also curious about.

Thanks,
Josh

You should do a test of the cable to compare with OEM Data sheet. Nice they provide the measurement and for a custom quality cable with N and BNC type connectors at that price!! The M&P cables/connectors look like a very good value and quality product.

Since these newer DSOs are venturing into the upper UHF region, maybe start a thread on cables, connectors (N/SMA/BNC) & Adapters with measured results for comparisons.  Most DSOs utilize BNC inputs, so info should emphasize at least 1 BNC connector/adapter/cable type.

This might prove useful for folks and save some $ as we've all been down the cheap cable/connector route, ourselves more times than we care to admit

Best