Shopping for oscilloscope: need help

[lawrence11]

So I am looking for an oscilloscope.

I need to measure digital signal up to 150 mhz, debug, and "maybe" see signal integrity. 2 or 4 channels will do for me.

I offered a bid on a MSO1104Z, wih has 100 mhz bandwidth and 1 Ghz sample rate, but its RPL-1116 probe for logic analyzer can only handle minimum pulse width of 10ns, wich sucks for 1 ghz sampling.

https://www.ebay.ca/itm/Rigol-MSO1104Z-Mixed-Signal-Oscilloscope/222061926178?hash=item33b3ec1722:g:vdsAAOSwwpdW8ZiV

I have my eyes on the hantek MSO5202, wich is 200 mhz 2 channels, and has logic analyzer abilty, but the question remains, how good is its probe and can it see better than minimal pulse width of
10ns, with 500 mhz sampling rate on the logic analyzer?

https://www.amazon.com/Hantek-MSO5202D-200MHz-Oscilloscope-Analyzer/dp/B00HYYCY5O

Also, on my pcb there is not many test points, there is mostly via with accompanying gnd via for signal via and usually 1mm spacing or less, would it be wise to use a "double needle" type probe  with small spacing in between to maintain small current loop, to get the best measurement. or should I just attach to my only safe secure ground clip. At some point, there will be a usb cable attached to the computer, and the shielding of that cable is direct short between gnd of my pcb and mains earth of scope and computer, so I guess my next revision I will put a 1k resistor seperated segment between usb shield on the pcb and my ground planes. The circuit is 3.3 volt, Do I really risk damaging my board if I accidentally connect ground lead of my scope to a signal pin(via) with my "double needle" type probe.

See what I mean, 15:05 of this video, this is my situation.

[tggzzz]

Oscilloscopes can't observe "digital signals", they can only observe analogue signals that other circuits interpret to be above/below thresholds.

"150mhz" is an unhelpful spec. Firstly, it is MHz. Secondly and importantly, the key parameter is the transition time; if there is repetition period, it is irrelevant.

If you want to observe digital signals, use a logic analyser or printf() statements.

You will also need to understand scope probes and how they interact with your circuit. See references in https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/

There are many threads on this forum discussing scopes, with a lot of useful information therein.

[lawrence11]

My signal is synchronous and has 150 mhz clock, so its quite fast.

I want to see its quality, and debug it.

I want the best for lowest money.

I will use standard probes that come with the product. I may change the physical configuration of the analogc probes for a thin needle type if needed.

[SiliconWizard]

Well, if you want to see a decent version of a 150MHz "digital" signal on a scope, don't expect to get anything useful with 200MHz bandwidth and 1GHz sample rate (that you will get if you use only 1 channel anyway, and one of the smallest memory depths on those Hantek scopes...)

The typical rise time of a 200MHz scope (assuming they really have a 200MHz bandwidth...) is approx 0.35/200e6 ~1.75ns.
A 150MHz digital clock has ~ 3.33ns ON and 3.33ns OFF. Compare that to the above value (which would already be a miracle for those Hantek scopes IMO). Not going to look pretty.

[thinkfat]

If you need to debug a 150MHz digital signal you're going to need at least something in the 500 MHz bandwidth range with a sampling rate to match. Forget about going cheap, decent probes alone are going to cost a small fortune.

And 500 MHz is only going to give you the third harmonic, basically the minimum that allows you to see that its not just a sine wave.

Gesendet von meinem Nokia 6.1 mit Tapatalk

[tggzzz]

My signal is synchronous and has 150 mhz clock, so its quite fast.

Being synchronous is not important, except that you must set up the test equipment so it samples the input at the same time as your circuit's receivers.

If it is 150mhz, then you can debug it with a voltmeter; an analogue voltmeter will be easier than a digital voltmeter.

I want to see its quality, and debug it.

Define what quality you want to see. That's important when it comes to probing.

I want the best for lowest money.

That's a content-free statement. Nobody wants the worst for the most money.

I will use standard probes that come with the product. I may change the physical configuration of the analogc probes for a thin needle type if needed.

So what? You still need to understand probing.

[lawrence11]

I feel destroyed.

What about logic analyzer, what kind do I need to debug this signal?

So even my dream scope is not enough to measure this? I have a hard time believing that...


Here is my dream scope..

https://www.amazon.ca/Rigol-MSO5354-Mixed-Signal-Oscilloscope/dp/B07KMK9724

[Siwastaja]

The point is, if you have a 150MHz digital clock signal, which is kinda square wave, and your scope AND your probes have a 200MHz BW, you see something resembling a sine wave on the screen (which is not how the actual signal is like).

But this is enough if you want to verify the clock frequency, and approximately the amplitude.

If you have a 100MHz probe, even if your scope had 10000 megagigaterasamples per second, you are going to see an attenuated sine-wave-like signal, and you won't have a good idea about the actual amplitude.


If you want to look for voltage overshoot and undershoot, transition time, etc., I'd recommend going for at least 1GHz, preferably 2GHz BW. This is going to be expensive.

Forget signal integrity measurements for 150MHz digital comms at low price (less than a few thousand $). Unless you can score a used high-end unit at a good price.

tggzzz, you are not being very helpful. You don't have to post if you don't feel like it.

[rstofer]

Fourier was all over this problem.

You didn't say but I'll guess your signal is something like a square wave at 150 MHz.  OK, that square wave is made up of all the odd harmonics of 150 MHz from DC to daylight.  The question is, how many harmonics are interesting?  There are simulations out on the web that show graphically how much each harmonic contributes to the signal being "square" (fast rise and fall times).  Or you can look at the coefficients of the various harmonics as they relate to the Fast Fourier Transform.  The harmonics die off quickly in terms of their contribution but they go on forever.

Let's suppose the 9th harmonic is far enough.  No problem, get a scope that can accurately display 9 times 150 MHz and you're pretty much good to go.  Something with a 1.5 GHz bandwidth would probably do a decent job.

It is much easier to find a logic analyzer that can handle 150 MHz but, still, the sampling rate needs to be pretty high if you want to know where the edges are really aligned.  The logic analyzer tells you nothing about signal integrity other than, at some point, there was a transition.  But only in the time window bounded by the sample rate.

[rstofer]

https://www.amazon.ca/Rigol-MSO5354-Mixed-Signal-Oscilloscope/dp/B07KMK9724

That scope is clearly out of my hobby level budget.  What I have done is buy the 100 MHz Rigol DS1054Z for the 4 channels with DSO features.  I also kept my used Tektronix 485 for its two channel 350 MHz bandwidth with no features whatsoever.  Kind of a hybrid approach.  I bought the Tek for a couple of hundred dollars maybe 12 years ago (could be 15, definitely after I retired).

I don't have really good scope probes but I don't need to work with really high frequencies in any event.  If my digital stuff is going fast, it's inside an FPGA and, these days, I can use the factory provided Internal Logic Analyzer (ILA).  Very nice!

[tggzzz]

tggzzz, you are not being very helpful. You don't have to post if you don't feel like it.

There are three key points in using a scope for this kind of thing, and two irrelevant points.

Firstly, the signal's transition time, since that defines the scopes required bandwidth.

Secondly probing technique, because not getting that right means all measurements are untrustworthy.

Thirdly, the sampling rate in a synchronous digital system should only be the same as the clock period, and it should occur at the same instant as the receiver samples the signal.

The signal's clock period is completely irrelevant. For a quick skim over the theory and some measurements, see https://entertaininghacks.wordpress.com/2018/05/08/digital-signal-integrity-and-bandwidth-signals-risetime-is-important-period-is-irrelevant

The scopes sampling rate is irrelevant for signal integrity measurements. Think subsampling, RIS/ETS/etc modes, eye diagrams.

[PCB.Wiz]

My signal is synchronous and has 150 mhz clock, so its quite fast.

I want to see its quality, and debug it.

I want the best for lowest money.

I will use standard probes that come with the product. I may change the physical configuration of the analogc probes for a thin needle type if needed.

If you want to really 'see' 150MHz signals,  a 100 or 200MHz scope is far too slow.  Attaching a probe is also going to disturb the signal.

What exactly is this 150MHz signal ? Is that 150MHz DDR (75MHz clock), or a 150MHz clock.  Is that LVDS ?

If your 150MHz signal is synchronous, you already have a sampling occurring, and the eye diagram of that sample is what actually matters.
If you have a FPGA (seems likely) you can skew the clocks, and so sweep across the signal, to check the eye Tsu/Th.

By using the right tools, and approach, you can  probably get by with an affordable 100~200MHz scope, but do not expect too much of 'see its quality, and debug it'
Get the best sampling rate you can, as even with the highly slewed displays you will have, some phase measurements can be extracted, if you understand what you see is not an exact copy.

You will need to measure the timing margins, to sub-ns precisions, and you do not need a scope to do that. Many FPGAs can do that inbuilt.

[thm_w]

I feel destroyed.
What about logic analyzer, what kind do I need to debug this signal?
So even my dream scope is not enough to measure this? I have a hard time believing that...

Here is my dream scope..

https://www.amazon.ca/Rigol-MSO5354-Mixed-Signal-Oscilloscope/dp/B07KMK9724

What is your budget, you can have your dream scope for $1,200 CAD, home built probe for another $100. But it tops out at about 450MHz I believe.
For logic analzyer, better off getting a USB based one IMO. There are many threads on these topics.

[tautech]

I feel destroyed.

Only because you don't know/understand what could work for you.

If you step up above a 1GSa/s scope to 2GSa/s you'll get much better signal fidelity however passive probe loading of the DUT is always an issue at higher frequencies that particularly with high impedance circuits can destroy the waveform fidelity and give erroneous results.
No way around that without deeper pockets and/or a clear understanding of how measurements can affect circuits and/or their operation.

You may have got better advice if you'd specified MHz instead of mhz as this forum is mainly engineering types that grind their teeth when such basic frequency terminology is wrong.

Anyways, back to a suitable scope.
Let me suggest SDS2202X-E for $619 or if you think BW is king, it's big brother SDS2352X-E.
Only 2 channels but good performers to their rated BW and well beyond.
16ch MSO can be added later if required.
https://www.siglentamerica.com/digital-oscilloscopes/sds2000x-e/

[lawrence11]

I had my eye exactly on that solution from siglent. Clearly THE BEST bang for the buck.

I was feeling destroyed untill I saw that one. Then I just felt hurt.

SDS2352X-E

I also believe I saw ur video?

[tautech]

I had my eye exactly on that solution from siglent. Clearly THE BEST bang for the buck.

I was feeling destroyed untill I saw that one. Then I just felt hurt.

SDS2352X-E

I also believe I saw ur video?

Never shown SDS2000X-E models in a vid only the SDS1000X-E models.

[nctnico]

So I am looking for an oscilloscope.

I need to measure digital signal up to 150 mhz, debug, and "maybe" see signal integrity. 2 or 4 channels will do for me.

Given your requirements you need two oscilloscopes: one cheap 500MHz (Tektronix TDS500/700 series or Lecroy) oscilloscope and a regular more modern bench oscilloscope. For looking at a 150MHz digital signal you'll need at least 500MHz bandwidth. Don't worry too much about probes. It is possible to make a DIY passive divider probe and there are also cheap high frequency active probes made & sold by various people (you should buy one from me  ).

[Sal Ammoniac]

Just a data point: I have a 1 GHz scope (5 GS/sec sampling rate) and 150 MHz digital signals look fine, although I'd probably want a 1.5 GHz or 2 GHz scope if I needed to do precise signal integrity measurements on 150 MHz signals.

Some things to keep in mind that I don't think anyone has mentioned on this thread... The rated bandwidth of a scope (e.g. 350 MHz, 500 MHz, 1 GHz, etc) is the point where the signal is 3 dB down. Most scope makers cheap out on the probes they supply with their higher bandwidth scopes--my 1GHz Agilent scope came with 500 MHz probes.

[SiliconWizard]

Most scope makers cheap out on the probes they supply with their higher bandwidth scopes--my 1GHz Agilent scope came with 500 MHz probes.

Yep!

1GHz probes tend to be pretty expensive too.

Another point for the OP (gonna feel even more destroyed): few of the "affordable" (read: cheap) logic analyzers will allow you to properly deal with 150MHz digital signals.

150MHz digital stuff doesn't look like a big deal these days, but you will already need "pro" equipment to analyze that.
One solution (I think already mentioned) would be to build your own equipment based on some FPGA board. Of course that wouldn't be as complete, as robust and as polyvalent, but it could work for your own specific needs (at least for the logic analyzing part).

[thinkfat]

.
Some things to keep in mind that I don't think anyone has mentioned on this thread... The rated bandwidth of a scope (e.g. 350 MHz, 500 MHz, 1 GHz, etc) is the point where the signal is 3 dB down. Most scope makers cheap out on the probes they supply with their higher bandwidth scopes--my 1GHz Agilent scope came with 500 MHz probes.

500MHz is about the highest bandwidth you can reasonably do with passive probes.


Gesendet von meinem Nokia 6.1 mit Tapatalk

[tggzzz]

.
Some things to keep in mind that I don't think anyone has mentioned on this thread... The rated bandwidth of a scope (e.g. 350 MHz, 500 MHz, 1 GHz, etc) is the point where the signal is 3 dB down. Most scope makers cheap out on the probes they supply with their higher bandwidth scopes--my 1GHz Agilent scope came with 500 MHz probes.

500MHz is about the highest bandwidth you can reasonably do with passive probes.

No; there are many very usable passive probes >1GHz from HP and Tektronix. I frequently use my 1.5GHz HP10020A to look at signals.

As an extreme example, consider the 9GHz P6150, which has an input capacitance <0.15pF. FFI see: http://w140.com/tekwiki/wiki/P6150

[thinkfat]

.
Some things to keep in mind that I don't think anyone has mentioned on this thread... The rated bandwidth of a scope (e.g. 350 MHz, 500 MHz, 1 GHz, etc) is the point where the signal is 3 dB down. Most scope makers cheap out on the probes they supply with their higher bandwidth scopes--my 1GHz Agilent scope came with 500 MHz probes.

500MHz is about the highest bandwidth you can reasonably do with passive probes.

No; there are many very usable passive probes >1GHz from HP and Tektronix. I frequently use my 1.5GHz HP10020A to look at signals.

As an extreme example, consider the 9GHz P6150, which has an input capacitance <0.15pF. FFI see: http://w140.com/tekwiki/wiki/P6150

But you usually don't get those bundled with a scope I'd guess? I imagine it's like the kit lens you'll buy with a camera. A reasonable optic you can start with, but will be upgraded sooner than later.

Gesendet von meinem Nokia 6.1 mit Tapatalk

[tggzzz]

.
Some things to keep in mind that I don't think anyone has mentioned on this thread... The rated bandwidth of a scope (e.g. 350 MHz, 500 MHz, 1 GHz, etc) is the point where the signal is 3 dB down. Most scope makers cheap out on the probes they supply with their higher bandwidth scopes--my 1GHz Agilent scope came with 500 MHz probes.

500MHz is about the highest bandwidth you can reasonably do with passive probes.

No; there are many very usable passive probes >1GHz from HP and Tektronix. I frequently use my 1.5GHz HP10020A to look at signals.

As an extreme example, consider the 9GHz P6150, which has an input capacitance <0.15pF. FFI see: http://w140.com/tekwiki/wiki/P6150

But you usually don't get those bundled with a scope I'd guess? I imagine it's like the kit lens you'll buy with a camera. A reasonable optic you can start with, but will be upgraded sooner than later.

True , but it doesn't change the important point.

Besides, what probes are bundled with a 9GHz scope?

[nigelwright7557]

The problem with sampling oscilloscopes is with a 100MHz scope you still need about 100 samples to get a half decent waveform.
So your 100MHz sampling becomes a 1MHz waveform equivalent scope.
That's why so many good sampling scopes are in the GHz range.

[tggzzz]

The problem with sampling oscilloscopes is with a 100MHz scope you still need about 100 samples to get a half decent waveform.
So your 100MHz sampling becomes a 1MHz waveform equivalent scope.
That's why so many good sampling scopes are in the GHz range.

You don't understand the sampling theorem, nor how sampling is employed in scopes.

Some low-end scopes are 100MHz 1Gs/s, but that is marketing emphasising the highest numbers they can find in the manual.