Has any one measured the -3 dB point for a SDS2104X HD w 100 to 200 MHz Update?

[tomwilkinson]

I have a SDS2104X HD and I measured the -3 dB point at around 180 MHz, one channel in use.  I was wondering if it is worth buying the 100 MHz to 200 MHz upgrade (SDS2000HD-BW1T2)?  What is the -3 dB point after the upgrade?  If it is not 260 MHz or better, I don't think it is worth the $830.  When I upgraded my SDS2104X+ to 200 MHz its -3 dB point changed to 257 MHz, but I wanted to verify that the same thing will happen with the SDS2104X HD.

[tautech]

Data point fwiw:
SDS2104X Plus -3dB BW ~185 MHz < tested with 3 sources that agreed within 5 MHz
SDS2204X Plus ~300 MHz < tested with one of the 3 used above.

As 2000X HD is also a 500 MHz design I would not expect it to be much different.

Corrected.

[voltsandjolts]

As 2000X HD is also a 500 MHz design I would expect it to be much different.

Did you mean..
"As 2000X HD is also a 500 MHz design I would not expect it to be much different."
?

[Martin72]

I have a SDS2104X HD and I measured the -3 dB point at around 180 MHz, one channel in use.  I was wondering if it is worth buying the 100 MHz to 200 MHz upgrade (SDS2000HD-BW1T2)?  What is the -3 dB point after the upgrade?  If it is not 260 MHz or better, I don't think it is worth the $830.  When I upgraded my SDS2104X+ to 200 MHz its -3 dB point changed to 257 MHz, but I wanted to verify that the same thing will happen with the SDS2104X HD.

Hi,
Not so long ago there was a promotion where you could upgrade to 200Mhz for free.
Maybe members of the forums had taken part and could answer your question directly.
Indirectly, I can say that the bandwidth is always a little more "generous" due to the filter.
Here is a picture of the bandwidth of my SDS2504X HD (legal on 500Mhz), the -3dB point was at 576Mhz.
(The blue line shows the bandwidth at 350Mhz, it is then at approx. 430Mhz).

Martin

[jonpaul]

plots reveal poor measurement technique or setup.

Or perhaps poor scope design.

Ripples perhaps from poor VSWR.

Compare to Tektronix scopes with similar BW, both analog and digital scopes  never show such horrible aberrations.

Suggest to get a fast edge gen (Leo Bodnar pulser) and check transient res with precision 50 Ohm term and precision 50 Ohm cable.

j

[tautech]

Hi,
Not so long ago there was a promotion where you could upgrade to 200Mhz for free.
Maybe members of the forums had taken part and could answer your question directly.

Thanks Martin, perfect.

Yes I checked the free 200 MHz BW upgrade and as your FFT confirms it is 300 MHz. (SDS2104X Plus > SDS2204X Plus)

[Performa01]

For those interested in facts, consider the following:

1.   A frequency response of +0 / -1.16 dB is pretty good for a 500 MHz bandwidth DSO.
2.   Visible ripple always comes from less than perfect output impedance, hence poor impedance matching, of the signal generator; a 10 dB inline attenuator directly at the generator output can rectify that.
3.   Because of the unknown frequency response characteristics of not artificially bandwidth limited frontends, a fast pulser cannot be the right tool to determine the scope bandwidth with any reasonable accuracy.

[mawyatt]

Wonder why folks put so much emphasis on the pulse method to determine scope BW since the scope low pass filter charactistics are usually not known, not like the old Tek analog days of first order filter approximations!!

The pulse technique is only an approximation for BW determination, not absolute, and used as a quick evaluation of BW, swept measurements are better for accurate BW assessment as that's exactly what one's measuring!! You normally don't see folks using swept frequency measurements to determine scope pulse response!!

Anyway, one's probes, cables, adapters, and probing technique will likely introduce more measurement aberrations and whatnot than the quality scope/generator itself.

Best, 

[Martin72]

Wonder why folks put so much emphasis on the pulse method to determine scope BW since the scope low pass filter charactistics are usually not known

I did this for years with the Bodnarpulser, for a trivial reason:
I didn't have a source that could "go" that high to capture the bandwidths of my scopes.
Maybe I wasn't the only one...
Now I have an R&S generator with which I can determine it as long as the scope is well below 1Ghz (R&S SML01).
And even then you can't display such a frequency response for all scopes, with the siglents it works with the FFT peak function.
Nevertheless, I still use the Pulser for the risetime determination, it is simply faster.
For "more precise" then the method mentioned above with a sweep.

[jonpaul]

The pulse is NOT to replace BW measurent but rather for those who need accurate measurements of non-sune waves, especiall rise time, jitter, pulse abberatuon critucal in digital media and signal transmission,.

Yes one can INFE BW from RT but it possible to have acceptable BW but terrible transient resp due to phase nonlinearly, etc.

We NEVER tweak BW in our fast (analog) scope CAL, eg 2465B, 2467B TTEK.

Instead the Leo B 40 pS >>50 Ohm inpat (NO cables, adapters!) and then each comp T coil etc is clearly display .

After a fine transient CAL, fast rise min overshoot and aberration, the BW "come out in the was"

Just my experience since 1970s,

Jon

[mawyatt]

Wonder why folks put so much emphasis on the pulse method to determine scope BW since the scope low pass filter charactistics are usually not known

I did this for years with the Bodnarpulser, for a trivial reason:
I didn't have a source that could "go" that high to capture the bandwidths of my scopes.
Maybe I wasn't the only one...
Now I have an R&S generator with which I can determine it as long as the scope is well below 1Ghz (R&S SML01).
And even then you can't display such a frequency response for all scopes, with the siglents it works with the FFT peak function.
Nevertheless, I still use the Pulser for the risetime determination, it is simply faster.
For "more precise" then the method mentioned above with a sweep.

All one needs is a decent sinewave signal generator to measure scope BW, not necessarily accurate in amplitude just uniform across the frequency range of interest. It takes just two measurements, one at a lower frequency and then another when amplitude drops 3dB. In a well behaved scope this takes into account the usual "unknown" scope filter response, whereas the pulse technique does not.

Not going to discount the pulse use tho, as a scope is a Time Domain instrument and usually used to measure pulses more, just as a SA is a Frequency Domain instrument and usually not utilized to measure Time Domain parameters, altho believe some do TDR by inference.

All this begs the question, since SAs are specified over the frequency range of use (does anyone even know the pulse response of their SA?), why not specify scopes with pulse response rather than BW??

So instead of a 350MHz scope one would have a 1ns Scope, or whatever based upon filter characteristics

Anyway, thought this was interesting, of course YMMV!!

BTW seems the latest scopes are starting to move into the Frequency Domain more, things like high resolution and frequency range FFTs, Bode Plots and so on, very versatile instruments indeed

Wait, we have one of these scopes, well it's pretty good but not the latest 12 bitter, we need, must have, require, demand, requisition, ok we'll now start saving our lunch/beer money for one

Best,

[tautech]

All one needs is a decent sinewave signal generator to measure scope BW, not necessarily accurate in amplitude just uniform across the frequency range of interest. It takes just two measurements, one at a lower frequency and then another when amplitude drops 3dB. In a well behaved scope this takes into account the usual "unknown" scope filter response, whereas the pulse technique does not.

Exactly.
One I checked recently:

Data point for the 100 MHz badged SDS814X HD = -3dB BW ~155 MHz

Source SDG6022X terminated into 50 Ohm feedthrough.
Checked @ 10 MHz = 1V p-p

[Martin72]

At that time I had also measured a DHO804 with a 1V sine wave, whose -3dB frequency deviated only a few Mhz from the calculated value from rise time.
That was probably luck/coincidence.

[gf]

You normally don't see folks using swept frequency measurements to determine scope pulse response!!

The problem, of course, is that you would need both magnitude and phase response to do that.
[ I would not rely on it being a minimum phase response in general. ]

[DaneLaw]

Also use crude sweep, as none of the arb signal gens I got, can muster up in the 3 digits MHz, so I tend to use these relatively cheap signal gens from China that go up around 7Ghz (based on a max2870 synth chip).
obviously not close to linear across that ludicrous wide-span but in the lower sub Ghz range, with a sweep, it seems to be relatively flat. (0-800Mhz sweep below)
but there are likely tons of reasons why that is a problematic approach to use a 40-buck mashup RF device from the local Szhenchen flee market to measure BW.??

0 to 800MHz span plot

Dropping hard 6.5GHz, but both tools (RF siggen & SA4) are way out of useful amplitude tolerances at these higher GHz freq.
Also tested the amplitude linearity with one of these self-contained 0 to 6Ghz RFpower meters that also hold a -32dBm selectable digital attenuator.

[mawyatt]

You normally don't see folks using swept frequency measurements to determine scope pulse response!!

The problem, of course, is that you would need both magnitude and phase response to do that.
[ I would not rely on it being a minimum phase response in general. ]

Exactly!!!

Don't think the scope would use Maximally Flat Phase (Bessel) for Anti-Aliasing filter as the amplitude response is weak. More likely some composite type where pulse display artifacts due to amplitude and phase distortion are minimized. Don't know what that filter is tho, maybe someone (hint Performa01) in the know wrt scope front end Anti-Aliasing filters can comment??

Recall way back when we were working with a system that theoretically would show artifacts equally wrt amplitude and phase effects we designed an in-between filter which should equalize the errors from amplitude and phase effects. After initial testing the system we found the simple Butterworth was slightly better, so this was what the system went forward with and into production.

Best,