Best Brand Name for an Oscilloscope?

[mawyatt]

...
IIRC SDS2504X Plus BW is nearly 600 MHz and it's reported somewhere in the 2kX Plus thread.

First SDS2104X Plus we got hands on is what I tested and as reported in the link to post above and I flat out didn't believe it was ~185 MHz but tested with 3 yes 3 sinewave sources confirmed it was. 
I'd never seen the likes of rated BW exceeded by 85% in any instrument.
...

Recall our SDS2104X Plus after "enhancement" with the kind help of folks here, returned something like 618MHz BW 

Best,

[gf]

And then again, Aliexpress and Ebay are full of low cost HF generator boards that are suitable up to hundreds of MHz with a reasonably constant amplitude.

Yes, but these $$ SI5351, ADF435x, etc. PLL synthesizer boards usually don't output sine wave, which complicates things. However, the amplitude of the fundamental frequency component can still be measured on the scope via FFT if the test point frequencies are chosen such that none of the harmonics alias exactly to the fundamental frequency in the 1st Nyquist band, enabling the separation of fundamental and harmonics via FFT, despite aliasing. If the test point frequencies are furthermore selected such that the fundamental falls onto a FFT bin frequency, then scalopping loss can be avoided. Reference measurements for the fundamental frequency amplitude of each test point can be done with a (tiny)SA if frequency-invariant amplitude is not granted.

EDIT: Btw, isn't the tinySA able to output sine wave when used as signal generator? At least the Ultra?

[nctnico]

And then again, Aliexpress and Ebay are full of low cost HF generator boards that are suitable up to hundreds of MHz with a reasonably constant amplitude.

Yes, but these $$ SI5351, ADF435x, etc. PLL synthesizer boards usually don't output sine wave, which complicates things.

No it doesn't. Near the maximum bandwidth you'll see a sine wave anyway because the 2nd harmonic is typically not very strong while the 3rd harmonic is 3 times the fundamental frequency.

[MrAl]

Are you using the right probes too?  I use 100MHz rated probes.

Direct connection.  The FG is a Siglent SDG1062X connected by 18" of coax.  The fast risetime oscillator connects directly to the scope input.

Oh ok that sounds good, but i wonder if that 18 inches of coax is affecting the readings.  As long as it is good coax i guess you're ok there.

I think the FAST family TTL is around 2ns rise time.  There are ways around that too by doing a little calculation that includes the rise time of the input wave.  If you could have someone with a better scope measure that, i could provide you with a formula to compensate.

Interesting little gadget though.  What main chip is that based on?  That really tells all.

[MrAl]

Are you making these Billy ?

I may, once I have them where I want them and there is enough demand for them.  Otherwise I can just make the files available.

Before you get all wound up about risetime, be aware that risetime tells you nothing about the bandwidth on an oscilloscope. When using risetime the assumption is that the 0.35 number is constant for all oscilloscopes. Unfortunately it isn't. The only proper way to measure the bandwidth of an oscilloscope is to sweep the input using a levelled RF generator.

There is a way to compensate for the rise time of the input wave if that is what you are talking about.
Also, when they quote the 0.35 constant what levels are they touting the rise time is measured at? 
Is it the 10 percent to 90 percent thresholds, or something else?
There is also the settling time which is rarely mentioned, which i see affects some digital scopes.

[MrAl]

I would only put Owon's VDS1022/VDS1022I on the category of good price/performance ratio. Others reported several issues with their offers, including some videos from Dave himself.

Hi,

Would you happen to have a link to one of Dave's reviews?
I searched but didnt find one yet, except with other members reviews.

[nctnico]

Are you making these Billy ?

I may, once I have them where I want them and there is enough demand for them.  Otherwise I can just make the files available.

Before you get all wound up about risetime, be aware that risetime tells you nothing about the bandwidth on an oscilloscope. When using risetime the assumption is that the 0.35 number is constant for all oscilloscopes. Unfortunately it isn't. The only proper way to measure the bandwidth of an oscilloscope is to sweep the input using a levelled RF generator.

There is a way to compensate for the rise time of the input wave if that is what you are talking about.
Also, when they quote the 0.35 constant what levels are they touting the rise time is measured at? 
Is it the 10 percent to 90 percent thresholds, or something else?

No, this is not the threshold but it is a factor that says how steep the signal amplitude drops with increasing frequency. The number 0.35 only applies when an oscilloscope has a 1st order (RC) roll-off. However, in order to maximise bandwidth and minimise aliasing, a lot of DSOs have much steeper (higher order) filters for which the number 0.35 no longer applies. Since this number is typically not specified (although Keysight does for some of their oscilloscopes IIRC), you can't use a steep edge to determine bandwidth accurately. To make matters worse, the bandwidth can also depend on the V/div setting. For low V/div settings the signal needs to be amplified first instead of attenuated and the amplifier may have a different bandwidth than the attenuator. You see some claiming bandwidths that far exceed what the oscilloscope specifies but it can be completely different at low V/div settings. In the end the manufacturer should specify what the oscilloscope is capable off at all V/div settings. Again, testing with a function / frequency generator at various levels gives you a far more detailed picture of what is going on.

This is what I measured on my GW Instek GDS-2204E:

There is also the settling time which is rarely mentioned, which i see affects some digital scopes.

You mean overload recovery time?

[rsjsouza]

I would only put Owon's VDS1022/VDS1022I on the category of good price/performance ratio. Others reported several issues with their offers, including some videos from Dave himself.

Hi,

Would you happen to have a link to one of Dave's reviews?
I searched but didnt find one yet, except with other members reviews.

Duckduckgo'ing for "eevblog owon oscilloscope" yields:
https://www.eevblog.com/forum/testgear/are-owon-oscilloscopes-worth-it/



And the Owon VDS1022/VDS1022I megathread, including even a much improved version of the software modified by the community:
https://www.eevblog.com/forum/testgear/owon-vds1022i-quick-teardown-(versus-the-hantek-6022be)/

[BillyO]

Oh ok that sounds good, but i wonder if that 18 inches of coax is affecting the readings.  As long as it is good coax i guess you're ok there.

  Properly terminated coax is a much better bet than any passive probe.  That 18" bit of coax will out perform my 500MHz Tek 6139 probes.

Interesting little gadget though.  What main chip is that based on?  That really tells all.

It's a 74AC14.  It should be able to get sub 2ns T_r, but just.  Maybe 1.7ns.   That's what I'm after.

[BillyO]

The bigger problem with Leo Bodnar's (at least the one fitted with the BNC) is that it's out of stock.  I don't know how often he does a build run, but I speculate that pandemic shortages have had an effect and that when he does get stock it goes pretty fast?

It looks like they are in stock today.  My order is in..

[mwb1100]

It looks like they are in stock today.  My order is in..

Thanks!  My order is in too.  I honestly have no business monkeying around with this stuff (I'm really a software guy).  But for whatever reason I find it fascinating, so paying a bit less than $100 is easily justified as on-going education.

I still look forward to any info you post about your device (from the looks of it, I think I might actually be able to successfully assemble it in spite of my lack of soldering skill).

[BillyO]

I still look forward to any info you post about your device (from the looks of it, I think I might actually be able to successfully assemble it in spite of my lack of soldering skill).

The next one is going to have SMD resistors.  They will be pretty big ones .. 1206.  Those are fairly easy to solder.  Almost as easy as thru-hole.

[mwb1100]

The next one is going to have SMD resistors.  They will be pretty big ones .. 1206.  Those are fairly easy to solder.  Almost as easy as thru-hole.

The bigger the better!  I'm currently working a "project" that consists of adding eight 0 Ohm resistors to a board to enable some signals to get to a connector.  The pads are for 0402 SMD resistors and my workmanship is pathetic.  I can manage to get one on and working in a single sit-down, but I find it so frustratingly difficult and ends up so cosmetically awful (probably awful more than just cosmetically, too) that I can't bring myself to try the next one for another week or so.

I can't believe all the posts I see where people say they can do 0402 work without magnification - those things are not much larger than a grain of sand!

Anyway - my sad story is really more suitable for the Beginner's forum.  I should stop drifting this thread off-topic anymore.

[tautech]

The next one is going to have SMD resistors.  They will be pretty big ones .. 1206.  Those are fairly easy to solder.  Almost as easy as thru-hole.

The bigger the better!  I'm currently working a "project" that consists of adding eight 0 Ohm resistors to a board to enable some signals to get to a connector.  The pads are for 0402 SMD resistors and my workmanship is pathetic.  I can manage to get one on and working in a single sit-down, but I find it so frustratingly difficult and ends up so cosmetically awful (probably awful more than just cosmetically, too) that I can't bring myself to try the next one for another week or so.

I can't believe all the posts I see where people say they can do 0402 work without magnification - those things are not much larger than a grain of sand!

With good eyesight and mine's long gone 1mm x 0.5mm (0402) component soldering is quite achievable.
Iron tips not need be small but you need a corner or point of the tip to get to the pad and heaps of flux and a reasonably fine leaded solder to make a tidy job.
Technique is everything with SMD.

[pcprogrammer]

I still look forward to any info you post about your device (from the looks of it, I think I might actually be able to successfully assemble it in spite of my lack of soldering skill).

The next one is going to have SMD resistors.  They will be pretty big ones .. 1206.  Those are fairly easy to solder.  Almost as easy as thru-hole.

Easier then thru hole, because there is no clipping of the leads. 0805 is also still doable for older eyes, but it is easier with a microscope.

[nctnico]

I can't believe all the posts I see where people say they can do 0402 work without magnification - those things are not much larger than a grain of sand!

Soldering is definitely doable if you have good eyesight but for a final check of the solder joints, you'll need to use a magnifier. Also having really good lighting is a must. I have about 400 Lux >95% CRI at my workbench surface level.

[MrAl]

Oh ok that sounds good, but i wonder if that 18 inches of coax is affecting the readings.  As long as it is good coax i guess you're ok there.

  Properly terminated coax is a much better bet than any passive probe.  That 18" bit of coax will out perform my 500MHz Tek 6139 probes.

Interesting little gadget though.  What main chip is that based on?  That really tells all.

It's a 74AC14.  It should be able to get sub 2ns T_r, but just.  Maybe 1.7ns.   That's what I'm after.

That looks reasonable then.  I was thinking of the FAST version or FASTr version of TTL.

[MrAl]

I would only put Owon's VDS1022/VDS1022I on the category of good price/performance ratio. Others reported several issues with their offers, including some videos from Dave himself.

Hi,

Would you happen to have a link to one of Dave's reviews?
I searched but didnt find one yet, except with other members reviews.

Duckduckgo'ing for "eevblog owon oscilloscope" yields:
https://www.eevblog.com/forum/testgear/are-owon-oscilloscopes-worth-it/



And the Owon VDS1022/VDS1022I megathread, including even a much improved version of the software modified by the community:
https://www.eevblog.com/forum/testgear/owon-vds1022i-quick-teardown-(versus-the-hantek-6022be)/

Those look interesting too, i was hoping for reviews on the 200 series which i think are the 240, 270, etc.  These are 40MHz and 70MHz rated units.  Not sure which ones are newer, the 200 series or the ones you show here i guess that is the 1000 series.

[MrAl]

Are you making these Billy ?

I may, once I have them where I want them and there is enough demand for them.  Otherwise I can just make the files available.

Before you get all wound up about risetime, be aware that risetime tells you nothing about the bandwidth on an oscilloscope. When using risetime the assumption is that the 0.35 number is constant for all oscilloscopes. Unfortunately it isn't. The only proper way to measure the bandwidth of an oscilloscope is to sweep the input using a levelled RF generator.

There is a way to compensate for the rise time of the input wave if that is what you are talking about.
Also, when they quote the 0.35 constant what levels are they touting the rise time is measured at? 
Is it the 10 percent to 90 percent thresholds, or something else?

No, this is not the threshold but it is a factor that says how steep the signal amplitude drops with increasing frequency. The number 0.35 only applies when an oscilloscope has a 1st order (RC) roll-off. However, in order to maximise bandwidth and minimise aliasing, a lot of DSOs have much steeper (higher order) filters for which the number 0.35 no longer applies. Since this number is typically not specified (although Keysight does for some of their oscilloscopes IIRC), you can't use a steep edge to determine bandwidth accurately. To make matters worse, the bandwidth can also depend on the V/div setting. For low V/div settings the signal needs to be amplified first instead of attenuated and the amplifier may have a different bandwidth than the attenuator. You see some claiming bandwidths that far exceed what the oscilloscope specifies but it can be completely different at low V/div settings. In the end the manufacturer should specify what the oscilloscope is capable off at all V/div settings. Again, testing with a function / frequency generator at various levels gives you a far more detailed picture of what is going on.

This is what I measured on my GW Instek GDS-2204E:

There is also the settling time which is rarely mentioned, which i see affects some digital scopes.

You mean overload recovery time?

Oh so i guess we see some differences AGAIN between CRT and DSO scopes, that's a pain in the neck.

The settling time is the time to settle to within 99 percent of maximum, and that is an important spec of any ADC.  I think the percentage used can be different though, such as 95 percent or something.  So if you shoot a 1v high square wave into the front end and the scope display starts to rise, it's the time it takes to get to 99 percent of the full max which in this case would be 0.99 volts.  This could be a lot longer than the rise time.  This was a big concern for me when i was designing a 100MS/s scope way back in the early 1990's.

[gf]

A suitable ADC driver can charge the S&H capacitor to settling within the sample interval.
At the end, the settling time of the overall system is still predominantly determined by the impulse/step response of the analog frontent.

[pdenisowski]

Lots of really great discussion and comments here about measuring scope bandwidth!

We get so many questions about this topic, I actually made a video explaining the basic concepts of "bandwidth" in oscilloscopes.  Probably a bit too introductory for most of the people posting here, but it does cover a few things discussed in this thread such as calculating BW from rise time, the different frequency responses (Gaussian vs. flat), etc. 

One thing I didn't cover is how to measure the bandwidth of an oscilloscope (i.e. procedures, techniques), but it seems like this might be a good topic for a future video.  I've already learned a few things from this thread myself

[2N3055]

Oh so i guess we see some differences AGAIN between CRT and DSO scopes, that's a pain in the neck.

The settling time is the time to settle to within 99 percent of maximum, and that is an important spec of any ADC.  I think the percentage used can be different though, such as 95 percent or something.  So if you shoot a 1v high square wave into the front end and the scope display starts to rise, it's the time it takes to get to 99 percent of the full max which in this case would be 0.99 volts.  This could be a lot longer than the rise time.  This was a big concern for me when i was designing a 100MS/s scope way back in the early 1990's.

I will politely comment on that..
In my opinion it is not "between CRT and DSO scope"  but how industry was making and looking at scope usage (and design that stems from that) 30 years ago as opposed to today. And also technology gap.

Perfect impulse response (that was industry standard many years ago) proved to be suboptimal for many other things people do with scopes.
For example my Keysight MSO3104T has practically flat BW until it's 1GHz BW. If I'm looking at signal in frequency domain (FFT) or verifying amplitude of 868 Mhz oscillator it will show excellent amplitude accuracy. All while having "only" 450ps risetime.  If I wanted to do the same on scope with Gaussian response (that would yield perfect impulse response) i would need at least 2GHz scope for same amplitude flatness up to 1 GHz.

There are digital scopes that have Gaussian response. They are just not common anymore.

[nctnico]

Oh so i guess we see some differences AGAIN between CRT and DSO scopes, that's a pain in the neck.

The settling time is the time to settle to within 99 percent of maximum, and that is an important spec of any ADC.  I think the percentage used can be different though, such as 95 percent or something.  So if you shoot a 1v high square wave into the front end and the scope display starts to rise, it's the time it takes to get to 99 percent of the full max which in this case would be 0.99 volts.  This could be a lot longer than the rise time.  This was a big concern for me when i was designing a 100MS/s scope way back in the early 1990's.

I will politely comment on that..
In my opinion it is not "between CRT and DSO scope"  but how industry was making and looking at scope usage (and design that stems from that) 30 years ago as opposed to today. And also technology gap.

Perfect impulse response (that was industry standard many years ago) proved to be suboptimal for many other things people do with scopes.
For example my Keysight MSO3104T has practically flat BW until it's 1GHz BW. If I'm looking at signal in frequency domain (FFT) or verifying amplitude of 868 Mhz oscillator it will show excellent amplitude accuracy. All while having "only" 450ps risetime.  If I wanted to do the same on scope with Gaussian response (that would yield perfect impulse response) i would need at least 2GHz scope for same amplitude flatness up to 1 GHz.

There are digital scopes that have Gaussian response. They are just not common anymore.

Actually you can make the oscilloscope Gaussian again by turning the bandwidth limit on. Higher frequency oscilloscopes typically have various bandwidth limits (with 1st order roll off) so you can make the tradeoff yourself.

[2N3055]

Oh so i guess we see some differences AGAIN between CRT and DSO scopes, that's a pain in the neck.

The settling time is the time to settle to within 99 percent of maximum, and that is an important spec of any ADC.  I think the percentage used can be different though, such as 95 percent or something.  So if you shoot a 1v high square wave into the front end and the scope display starts to rise, it's the time it takes to get to 99 percent of the full max which in this case would be 0.99 volts.  This could be a lot longer than the rise time.  This was a big concern for me when i was designing a 100MS/s scope way back in the early 1990's.

I will politely comment on that..
In my opinion it is not "between CRT and DSO scope"  but how industry was making and looking at scope usage (and design that stems from that) 30 years ago as opposed to today. And also technology gap.

Perfect impulse response (that was industry standard many years ago) proved to be suboptimal for many other things people do with scopes.
For example my Keysight MSO3104T has practically flat BW until it's 1GHz BW. If I'm looking at signal in frequency domain (FFT) or verifying amplitude of 868 Mhz oscillator it will show excellent amplitude accuracy. All while having "only" 450ps risetime.  If I wanted to do the same on scope with Gaussian response (that would yield perfect impulse response) i would need at least 2GHz scope for same amplitude flatness up to 1 GHz.

There are digital scopes that have Gaussian response. They are just not common anymore.

Actually you can make the oscilloscope Gaussian again by turning the bandwidth limit on. Higher frequency oscilloscopes typically have various bandwidth limits (with 1st order roll off) so you can make the tradeoff yourself.

That is very true, it converts 1GHz scope to 200 or 500MHz scope and user have control...
I actually use 200Mhz BW limit on Siglents most of the time I work with passive probes.  Or whenever I don't need full BW.
I wish MSOX3104T had more BW limits in addition to 20Mhz.

[MrAl]

A suitable ADC driver can charge the S&H capacitor to settling within the sample interval.
At the end, the settling time of the overall system is still predominantly determined by the impulse/step response of the analog frontent.

Yes i am sorry i misspoke, of course it was the front end i was talking about the the ADC 'system'.
The front end settling time is the thing i looked at most when designing the front end.