Siglent1104X-E versus Square Waves

[maddoxhq]

First time poster here.  I've had a 1104X-E for about 2.5 years now.  Haven't used it a whole lot but finally been kicking myself to get back into things.

I started noticing something concerning when I was attempting to measure some low value capacitors and inductors.  My 100MHz Siglent doesn't seem to be getting that bandwidth.  I've been racking my brain trying to blame myself for it with something I'm missing, but honestly I've got nothing left but to point the finger at the scope itself.

When I test the bandwidth with a square wave I reach the -3dB mark around 20MHz, well below the 100MHz.  Eventually I found the whole "hack" your scope thing and got into it.  I at least wanted to check to make sure it's firmware was set up right and asked it PRBD?  It replied with 100M.  So then I went ahead and threw it the string to increase the bandwidth up to 200MHz.  Asked again asked it and it confirmed I was set to 200M.

The problem is (and was) the bandwidth is still sitting at ~20MHz.

Wondering if my cheap-o Koolertron 30MHz signal generator was to blame I programmed my working PIC project to output the internal clock (8MHz) to a CLKR pin.  I then compared it to the Koolertron.  And the results are nearly identical.  (the PIC seems to provide a cleaner signal, but doesn't fair as well at 8MHz (the pins have a capacitance anywhere between 5 - 50pF, compared to the scopes stated 15pF).  The signal generator seems to do better at frequency precision.

So back to the scope.  I've defaulted it before I ran these tests so I don't think it's a settings issue.  I've got four 100MHz probes, and one 60MHz probe, all the same.  Is there anything I might be missing or misunderstanding here? 

[Kean]

Are you using your probe set to 1x?
Try 10x...

[maddoxhq]

Yeah, this was done with the probe set to 1X, I tried 10X like you said and it does seem to help the attenuation.  I maxed the signal generator out at 30 and only dropped 30mV 68mV (13%).  I have to imagine it drops off pretty quick after 30MHz.

I wouldn't say the waveform looks any better though.

[Kean]

If your signal generator maxes out at 30MHz, then it likely cannot generate a nice square wave all the way up to that frequency.
Best to read the specs for your sig gen regarding square waves, and specs for your o'scope probes re bandwidth in 1x and 10x modes.

Maybe spend some time watching this video to help fill in some fundamentals.  It is long, but in particular check the section I've linked directly to that is relevant.
https://youtu.be/8ZKMrzTGxLQ?t=5496

BTW, when you change the probe from 1x to 10x, you need to tell the scope about it - otherwise the displayed voltages will be wrong.

[Kean]

Not sure which Koolertron 30MHz signal generator you have, but I see this in the specs at https://www.koolertron.com/koolertron-dds-signal-generator-counter-2-4in-screen-display-30mhz-high-precision-dual-channel-arbitray-waveform-generator-frequency-meter.html

Sine Wave: 0~30MHz
Square & Triangle Wave: 0~15MHz

I am guessing even 15MHz is a bit ambitious for a clean square wave.

[DavidAlfa]

Getting sharp square waves is tricky, likely you'll need to load the generator output with 50ohms or so.
Don't expect 30MHz square waves to look nice in any random electronic device, normally you'll see lots of armonics unless using lab-grade equipment.
The last picture seems like your generator reached its limits.
Try a fast mcu like stm32, you'll see a much better output than that.

[tautech]

Some fundamental misunderstanding by the OP.

My SN#13 prerelease SDS1104X-E.
BW tests are always done from a 50 Ohm system and as these DSO's don't have a 50 Ohm input we must use some termination which in this case is a BNC feedthrough type.

We look for the -3dB point which is simply 0.707Vp-p remaining for a 1Vp-p sinewave and either read the frequency from the supply or use the scopes counter. Source = SDG6022X



Step response determines how well a DSO/scope displays a square wave and for the next screenshot the affordable 10 MHz Leo Bodnar fast edge source is used, this particular one provided with a sampling scope capture of 29ps rise time, way too fast for the little SDS1104X-E which can only manage some 3.4ns risetime.



Introduce the OEM PP510 probe (10x setting) and not only does it affect the direct connection to the Bodnar pulser, propagation delays due to probe lead length and some ringing becomes apparent also.
A higher BW probe will provide a better result.

[maddoxhq]

Interesting, I definitely plead no contest in fundamental misunderstandings.

The 50ohm feedthrough BNC has been on the way and I was wondering if that might change some things.  I tried to set up an approximate impedance match with a 47ohm resistor across the signal to ground.  But that seems to attenuate my signal about half, I wasn't expecting that.

Also, from what Kean said, I looked deeper at my signal generator, and completely overlooked the square wave being 0-15MHz.  Further on that I see it has a square wave rise time of <=25ns.  I now get how that's going to mess with Vpp beyond that level.

I'm going to wait for my feedthrough to get here and try all this over again.

When it comes to the "upgrade" of the scope to 200MHz, is it going to require a PP215 to even notice this change?

[tautech]

The 50ohm feedthrough BNC has been on the way and I was wondering if that might change some things.

Plenty. 

I tried to set up an approximate impedance match with a 47ohm resistor across the signal to ground.  But that seems to attenuate my signal about half, I wasn't expecting that.

Normal.

Also, from what Kean said, I looked deeper at my signal generator, and completely overlooked the square wave being 0-15MHz.  Further on that I see it has a square wave rise time of <=25ns.  I now get how that's going to mess with Vpp beyond that level.

Should have settings for HiZ and 50 Ohm output however levels do not change, instead the AWG display values change to match the selected load impedance.

When it comes to the "upgrade" of the scope to 200MHz, is it going to require a PP215 to even notice this change?

Not really, tests were done some years back and 0.5-1dB attenuation using PP510 is about normal @200MHz so barely enough to worry about as the majority of our work is at much lower frequencies

[BillyO]

Your Koolertron generator would only have an absolute maximum minimum rise time of abut 11ns on it's square wave and most likely substantially less than that. 30MHz is probably the sine wave spec.  My guess is the rise time on it's square wave would be more like 35ns.  Either 11ns or 35ns are both woefully inadequate for testing the BW of a 100 MHz scope, let alone one that has been pushed to 200MHz.  Putting a 50 ohm load in the mix will likely "improve" things, but not to the extent where you will be able to verify your BW.

You need a square wave generator that will have a ~1ns or better rise time or a sine wave generator that will go out to 300MHz or better.

[TheDefpom]

The scopes bandwidth limit isn't set to 20MHz is it instead of Full Range?

[tautech]

The scopes bandwidth limit isn't set to 20MHz is it instead of Full Range?

No or we could see that in the channel tab as a B. (BW limit engaged)
First thing I checked. 

[maddoxhq]

Woof, if I was confused before I'm especially confused now.

My 50ohm feedthrough arrived and wanted to familiar myself with it.  I also purchased some BNC jumper extensions to connect my signal generator to my scope directly.

With the 50ohm feedthrough I piped a 1MHz signal directly into the scope at 1Vpp, no offset, AC coupled.  I was surprised to see that it attenuated the signal to about half, 500mV.

I then took a second BNC jumper and put the exact same waveform into the scope on channel 2.  And I get 1Vpp.

tautech, when you responded to my question about the 47ohm resistor you said this was normal, but I guess I wasn't expecting this to be true on a device that was strictly meant for impedance matching.

I guess when it comes to looking at testing bandwidth on an oscilloscope and using a feedthrough.  Should I just know in the back of my mind the signal is getting attenuated and adjust the signal generator so the oscilloscope "sees" 1Vpp, then try to find the .707Vpp point?

I know I'll never be able to with this signal generator as it maxes out at 30MHz, but I did see the spec on the PP510's where it said that they're 100MHz on 10X setting, and 8MHz on 1X.  At least I could test that.

[tautech]

Most AWG's output is a 50 Ohm source and you need tell it the load it will see.
If set to a HiZ output and you place a 50 Ohm terminator on it the output will ~halve.

So typically you use 2 impedances, HiZ or 50 Ohm and you need set for either and use an appropriate network/cabling/termination for the signal transmission.

Also AWG output is normally offset by half the signal amplitude so unless you set the max or min value differently and only when you do does scope AC input coupling offer any advantage.

Keep playing around, you'll learn heaps about this sort of stuff. 

[maddoxhq]

Lol, if I don't lose my sanity first.  I honestly don't think it has a HiZ mode.

I just put the output of the AWG out it's normal leads, to the oscilloscope probe, through the 50ohm terminator.  Set the AWG to 1kHz, 1Vpp.

Now the scope is reading 120mVpp, and the measured frequency is bouncing around between 1.14 - 1.23kHz.  However the measure function sees it as 1kHz.  The trigger is having a hell of a rough go at keeping stable though.

[tautech]

Yes something is not stable between the scopes HW counter and the frequency measurement.
Change input coupling to DC and give us another screenshot.....meanwhile I'll get mine out.

[maddoxhq]

Interesting, I found the issue for the frequency, instead of directly connecting the output to the probe, I've been leaving my bread board jumpers on the probe and just grabbing on to them.

The AC vs DC coupling didn't seem to make any difference before or after.

The fact that I'm sitting at 120mV still is a bit odd.  I took off the feedthrough and put the probe directly to the scope and I get my 1.02Vpp like I would expect.  Frequency is a ton more stable now too.

[tggzzz]

Introduce the OEM PP510 probe (10x setting) and not only does it affect the direct connection to the Bodnar pulser, propagation delays due to probe lead length and some ringing becomes apparent also.
A higher BW probe will provide a better result.

Not if you use the long (10cm/15cm) ground lead. The ringing is caused by the inductance of the ground lead resonating with the probe's tip capacitance at around 100MHz. For an example, see https://entertaininghacks.wordpress.com/2016/09/17/scope-probe-accessory-higher-frequency-results/

For the OP, a beginner, the references here might well prove enlightening: https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/

[maddoxhq]

Plus, now armed with the knowledge of my PP510 has a rated 8MHz BW, I cranked it up to about there and I wasn't at the -3dB mark yet.  That was found at the 10.25MHz mark.

Which honestly does make me feel a bit better, that feels like an appropriate result for a known rating.

Still, this is all leaving the 50ohm feed through on the floor in the corner of the room.  Seems like it's worthless honestly.

[tautech]

Oh, are you using a probe in the system between the AWG and the scope ?
Quite suitable for HiZ measurements but not to terminate into on a 50 Ohm system.
At low frequency and sinewaves you'll get away with it however introducing a probe always affects the measurement fidelity in some way.
Study the last screenshot I posted.

More for you.
1 MHz 1Vp-p from SDG6022X, no different result to what I expect from the cheap SDG1032X.
50 Ohm output via BNC cable into the first termination grabbed, placed onto the scope.



Then remove termination but retain all previous settings
Then change source to HiZ which becomes 2Vp-p displayed on the AWG.
Note waveforms are the same amplitude.



Retain HiZ AWG output and reduce to 1Vp-p

[maddoxhq]

I was using a probe in the system when I was measuring the probes bandwidth yes.

When directly connecting the AWG to the scope earlier, no.

I picked the 50ohm terminator off the floor and tried again.  Channel 1 has the terminator and a 1.5ft BNC cable, Channel 2 has only the 1.5ft BNC cable.  And for funzies i set both channels to square wave to see if that made any difference.

And not really, the passthrough terminated side is still half the amplitude of the just cabled side.  The square waves themselves are pretty much the same for the most part.

I would have expected what you saw though.  Without the feedthrough, nearly double the voltage.   The big thing here is I CANNOT change my AWG to HiZ mode.  I grabbed up everything I could from the interwebs, including some dodgy Chinese software to interface with it.  It's 50 ohms output impedance out of the box, or at least it says it is.

Headed to bed for now, the suns making its way to my neck of the woods.

[807]

The Koolatron is rebadged under various names. It's output impedance is always 50Ω, but the displayed output is always HiZ. Unlike my Uni-T UTG962E it doesn't look like you can change the displayed output to represent 50Ω.

This means that if you were to look at the output of this generator on a spectrum analyzer, it would read half the displayed generator output. i.e. 2V p-p  would read 4dBm rather than 10dBm.

See this video @ 5:37

https://youtu.be/b66nnWb3z54?t=337

[BillyO]

A note on using a 50 ohm feed through terminator:  You never use them with a standard scope probe.  You only use them with 50 ohm transmission lines (straight trough 50 ohm cable) and with direct connect devices that require 50 ohm termination.

A note on standard scope probes: They always feed into a HiZ input.  Never 50 ohms.  The cable on a standard scope probe is not a 50 ohm transmission line.  Not eve close.  The generally have high resistance center conductors that will measure 300-500 ohms.  If you use them with a 50 ohm terminator you will create a 1/7 to 1/11 voltage divider.  This explains your 120mv signal from 1V at the generator

[maddoxhq]

Sorry, tried to make this reply yesterday but the site wouldn't load.

The generally have high resistance center conductors that will measure 300-500 ohms.  If you use them with a 50 ohm terminator you will create a 1/7 to 1/11 voltage divider.  This explains your 120mv signal from 1V at the generator

Now that makes sense to me, I wasn't aware there was such high resistance in the oscilloscope lead.  Mine measures 335ohms.  When it comes to the voltage divider are you just taking 50/(probe resistance + feedthrough impedance)?

[BillyO]

  When it comes to the voltage divider are you just taking 50/(probe resistance + feedthrough impedance)?

Correct.