Buying first scope, 1054z vs GDS-1054b which is better suited to my use?

[Celiturbo]

I've searched around and gathered the differences in spec, considered the 1104x-e also but its considerably more expensive and looks to be a lot more time in going from menu to sub menu vs the others. 

My take away is the 1054z I'd have 100mhz vs 50mhz, slightly higher 1ch/2ch channel memory depth, and a larger community base. The GDS-1054b has independent channel controls, more memory in 3ch/4ch modes, a far better FFT function, supports CAN+LIN decoding, and limited data logging over the Rigol, and its 50k wfms/s vs 30k. The GDS-1054b is also $40 less than the Rigol but that isn't a real big deal to me. I'm assuming the GW would be a bit more user friendly with the independent ch controls and the better FFT has me leaning that direction. However I really don't know how big of a deal the higher bandwidth and deeper single channel memory on the Rigol is for my use. I could be totally off on thinking the GW would be more user friendly as well. 

Rigol 1054z          100mhz, 30,000wfms/s, Memory/ch 24/12/6/6Mpoint, FFT 16k points
GW GDS-1054-b     50mhz, 50,000wfms/s, Memory/ch 10/10/10/10Mpoint, FFT 1M point.

Use case would be audio equipment (Amps A/B D, EQ, filters), 80's arcade boards 6502 based mostly, automotive engine management, boost/buck, and general SMPS repair.  My best guess is most signals I'd be looking at would be under 250khz aside from some SPI/I2C/CAN bus or maybe a controller in the 1-25mhz area.   

I've always made do without, mostly because its usually a case of something quit and I can't justify spending hundreds to replace it before being sure its not dead over a few $ in parts, corrosion, or cracked solder joint. That's also not a situation that when it comes up anyone is thinking to buy a $300 tool to maybe repair one thing either. It's since become a bit of a hobby that overlaps with some other things. While it's been interesting seeing what I can manage with a DMM and cheap component tester over the years, I have better things to do with my time than trying to work ways around what I can't see. I have some understanding through research, youtube, ect.  But I've never had access to a scope so as an example something like what sort of bandwidth/wfms do I need in a scope to look at a 10mhz clock signal and know its good enough that a micro or cpu would be happy running from it, is totally outside my experience.   

How much scope do I need for this sort of use, would the difference in bandwidth or wfms/s matter to me or any other features, is the FFT on the Rigol any better below say 100-150khz where I'd be using it for audio / class D filtering, or is it going to be as bad as the examples I've seen at higher frequencies? 

[Fungus]

First of all, thanks for doing some research before posting and for telling people what you need.

Use case would be audio equipment (Amps A/B D, EQ, filters), 80's arcade boards 6502 based mostly, automotive engine management, boost/buck, and general SMPS repair.  My best guess is most signals I'd be looking at would be under 250khz aside from some SPI/I2C/CAN bus or maybe a controller in the 1-25mhz area.   

I'd go for the GW-Instek based on better user interface and FFT.

50MHz is plenty for those jobs.

[RafaPolit]

I just went through a similar process, and really poor FFT shown on several videos for the Rigol is what moved me away from it.  Also, the 1 GSa/s rate seemed that it was bordering the limits.  In the end, I went with a Keysight 1204G still with concerns of the 1Mpoint memory and the "more limited" 64K FFT points, but betting on the 200wfms/s and 2GSa/s.  Another difficult consideration was 4ch vs 2ch, and moving up the scale with 2ch and paying the same for better features.

I hope I made the right decision, I couldn't find any GW locally.  I have been following YouTube's Kiss Analog channel and he uses a GW (bigger brother to the one you posted) to very good results.  That Rigol has been around for so long, that I think that it is starting to show it's age with others surpassing it's features.

Good luck with your quest,
Rafa.

[Fungus]

In the end, I went with a Keysight 1204G still with concerns of the 1Mpoint memory and the "more limited" 64K FFT points, but betting on the 200wfms/s and 2GSa/s.  Another difficult consideration was 4ch vs 2ch, and moving up the scale with 2ch and paying the same for better features.

The "2 channel" Keysight is actually an MSO, two analog and one digital channel.

I hope I made the right decision, I couldn't find any GW locally.

If you live in the USA the GW-Instek can be had for under $300. It's a complete bargain at that price.

[Celiturbo]

I have been following YouTube's Kiss Analog channel and he uses a GW (bigger brother to the one you posted) to very good results.  That Rigol has been around for so long, that I think that it is starting to show it's age with others surpassing it's features.

I just watched one of his videos comparing a bunch of sub $500 scopes, the FFT on the GW's really stood out, even in Dave's video on it, while a bit slow the GW did hold up a reasonable output compared to the much higher priced scopes. The 1054z has been around forever, they still sell tons of them though so I guess they don't really have a motivation to put out a new one yet, kind of wish they would since its been a bit of the defacto benchmark in the price range.  Keysight or similar would be nice, but for that kind of money it would have to generate some income and this is all just personal projects and maybe fix some things for family.

Thanks

[Celiturbo]

I'd go for the GW-Instek based on better user interface and FFT.

50MHz is plenty for those jobs.

I'm thinking that is the way I will go. 

Thanks for confirmation 50mhz is enough, that really stacks the deck. 

[2N3055]

Just be mindful that scope needs to have rise time good enough compared to signals you are measuring. You can easily have signal that has edges faster than 1 ns while repeating once every second, making it 1Hz signal, that you need a 1GHz+ scope to properly measure....
That being said, GDS1054B should serve you well, just need to keep in mind the limits. It is nice little scope, better than DS1000Z in many ways.

[Fungus]

Thanks for confirmation 50mhz is enough, that really stacks the deck.

It will be a bit more than 50 in reality and signals don't suddenly vanish when they go above that, they attenuate like a low pass filter.

[Celiturbo]

Just be mindful that scope needs to have rise time good enough compared to signals you are measuring. You can easily have signal that has edges faster than 1 ns while repeating once every second, making it 1Hz signal, that you need a 1GHz+ scope to properly measure....
That being said, GDS1054B should serve you well, just need to keep in mind the limits. It is nice little scope, better than DS1000Z in many ways.

This was kind of my concern with the 50mhz vs 100mhz, but it seems people think 50mhz is plenty for my use so I assume situations where it is not would be unusual exceptions.  I've managed for a decade+ without so I can always fall back on previous methods if need be, I expect that will often be the case as I make my way through the learning curve.

[nctnico]

The GDS1054B can also be hacked to 100MHz using a key generator. There is a thread about that on this forum: https://www.eevblog.com/forum/testgear/possible-gw-instek-gds-1000b-hack/msg3552831/#msg3552831

[bdunham7]

I haven't had the GW Instek but I have had the Rigol and I don't recommend it.  It was better than nothing and better in some ways than the cheap DSOs that came before it, but now...no.

The GW Instek is dirt cheap for what you appear to get.  The 50MHz will likely be fine for your listed uses--and it can be hacked although I'm not entirely sure you should do that. The 10M per channel memory is enough.  The FFT will be usable, but you probably won't have much use for it.  If you use it on the 6502-based boards or automotive stuff, please post as I'd be interested in seeing what you do with it. 

[Celiturbo]

The GDS1054B can also be hacked to 100MHz using a key generator. There is a thread about that on this forum: https://www.eevblog.com/forum/testgear/possible-gw-instek-gds-1000b-hack/msg3552831/#msg3552831

I was under the impression it would take the license but was otherwise hardware limited in bandwidth. Looking back though I see you mentioned that as well, am I missing something or is there something new that has not been posted in that thread?

https://www.eevblog.com/forum/testgear/possible-gw-instek-gds-1000b-hack/msg3306676/#msg3306676
https://www.eevblog.com/forum/testgear/possible-gw-instek-gds-1000b-hack/msg3306636/#msg3306636

[Celiturbo]

I haven't had the GW Instek but I have had the Rigol and I don't recommend it.  It was better than nothing and better in some ways than the cheap DSOs that came before it, but now...no.

The GW Instek is dirt cheap for what you appear to get.  The 50MHz will likely be fine for your listed uses--and it can be hacked although I'm not entirely sure you should do that. The 10M per channel memory is enough.  The FFT will be usable, but you probably won't have much use for it.  If you use it on the 6502-based boards or automotive stuff, please post as I'd be interested in seeing what you do with it.

I don't believe the hack actually increases the bandwidth on the GW, looks like there are some other features that can be unlocked though. 

The FFT function I mostly had in mind for tuning filters and cross overs for audio circuits, feed in white noise and see what makes it out the other side. Maybe it would be useful for trouble shooting signal from hall or vr crank/cam sensors or tuning a circuit for a knock sensor where what looks like harmonics is actually the signal you are after. I have plans to run a Speeduino oshw ecu on one of my cars so there is some relevance there I doubt applies to anything oem or commercial.

The old 6502 stuff I've dealt with is mostly 1978-85 pinball machine boards. They have marginally regulated and poorly filtered linear supplies and are full of relays and rather high power 20-70v a/c solenoids that cause all sorts of electrical noise, combine that with 1pf-100pf ceramic bypass caps that are about the size of a cr2024 cell that like to test good after turning highly vibration/shock sensitive. I've never had a scope before but my understanding is they show signal or waveforms well but irregular noise is hard to know much about from the standard view other than the signal gets fuzzy, maybe you can determine if its 50/60hz mains or something but random trash I'd think would be better to see the spectrum it spans when trying to figure out a way to get rid of it.

Like I said I've never had access to a scope or SA so I could be totally off on what can be done with FFT but with my current understanding those are the things that came to mind. I guess you could say I'm not looking to use it as a tool to find a signal, I'm looking to use it to identify what I want to get rid of.  I'm really not sure how much use I will get out of it either, while not a hard point in choosing a scope for me its always nice to have additional functionality if its actually functional in a useful capacity which seems like that would be a hard case to make for the FFT in the Rigol. 

[Fungus]

I was under the impression it would take the license but was otherwise hardware limited in bandwidth. Looking back though I see you mentioned that as well, am I missing something or is there something new that has not been posted in that thread?

https://www.eevblog.com/forum/testgear/possible-gw-instek-gds-1000b-hack/msg3306636/#msg3306636

It appears to be limited in hardware these days.

OTOH that post has a graph that shows it can measure a 100MHz signal with only 4db attenuation. 

[Celiturbo]

It appears to be limited in hardware these days.

OTOH that post has a graph that shows it can measure a 100MHz signal with only 4db attenuation. 

I suspected it would have some headroom.

Better yet that graph has it all the way out to 80mhz before its down 3db. Cant complain. Doesn't look like the 1104b had any time base/div the 1054b lacks either

[Celiturbo]

Thanks Everyone!

I've ordered the GDS-1054b

[bdunham7]

The FFT function I mostly had in mind for tuning filters and cross overs for audio circuits, feed in white noise and see what makes it out the other side. Maybe it would be useful for trouble shooting signal from hall or vr crank/cam sensors or tuning a circuit for a knock sensor where what looks like harmonics is actually the signal you are after....

Like I said I've never had access to a scope or SA so I could be totally off on what can be done with FFT but with my current understanding those are the things that came to mind. I guess you could say I'm not looking to use it as a tool to find a signal, I'm looking to use it to identify what I want to get rid of. 

The FFT function probably isn't precise enough for audio crossovers, etc--I've never tried.  The tool for that is the Bode plot, but I don't think the GDS-1054B has that.  I usually wouldn't think of using an FFT on an automotive sensor, but it might actually be interesting to see what it does with a knock sensor.  If you haven't used a scope before, your initial efforts are probably best spend learning how to use all of the the trigger functions.  The ability to trigger is what determines what you can see.

[Celiturbo]

The FFT function probably isn't precise enough for audio crossovers, etc--I've never tried.  The tool for that is the Bode plot, but I don't think the GDS-1054B has that.  I usually wouldn't think of using an FFT on an automotive sensor, but it might actually be interesting to see what it does with a knock sensor.  If you haven't used a scope before, your initial efforts are probably best spend learning how to use all of the the trigger functions.  The ability to trigger is what determines what you can see.

I've been meaning to setup Visual Analyzer www.sillanumsoft.org for some of the audio stuff anyway though working out high pass or low pass on even an old analog scope looks to be easy enough, I thought FFT would be convenient for band pass filters to see what is making it through. Its a little bit round about but its possible to do a bode plot this way



The knock sensor, trying to identify the sound of detonation against the background noise and normal combustion noise is generally done by calculating the bore size to determine a target frequency and then setting it up with a band-pass filter or tuned sensor to only listen during time knock could occur, preferably not when valves are opening or closing or other cylinders are firing if possible, then monitor a baseline without detonation and use that to set a threshold level for that target frequency, anything over that level in that time window is detected as knock. Another way to do it is to target the harmonics which are actually physical pressure waves colliding in the cylinder rather than electrical ringing. Since the bore size determines the fundamental frequency, collisions within that produce harmonics. If harmonics are present at a high enough level then there was knock, none of the mess with timing a window to listen is needed since normal combustion doesn't make much for harmonics.  Thinking about it this is probably well into the realm of a dedicated SA. I Attached a image that shows cylinder pressure in one graph and sensor signal in another, image quality isn't great but you can get the idea.



 

[Fungus]

its possible to do a bode plot this way ...

Feed it some white noise and do an FFT - instant Bode plot!

[2N3055]

The FFT function I mostly had in mind for tuning filters and cross overs for audio circuits, feed in white noise and see what makes it out the other side. Maybe it would be useful for trouble shooting signal from hall or vr crank/cam sensors or tuning a circuit for a knock sensor where what looks like harmonics is actually the signal you are after....

Like I said I've never had access to a scope or SA so I could be totally off on what can be done with FFT but with my current understanding those are the things that came to mind. I guess you could say I'm not looking to use it as a tool to find a signal, I'm looking to use it to identify what I want to get rid of. 

The FFT function probably isn't precise enough for audio crossovers, etc--I've never tried.  The tool for that is the Bode plot, but I don't think the GDS-1054B has that.  I usually wouldn't think of using an FFT on an automotive sensor, but it might actually be interesting to see what it does with a knock sensor.  If you haven't used a scope before, your initial efforts are probably best spend learning how to use all of the the trigger functions.  The ability to trigger is what determines what you can see.

That little thing has 1 Mpoint FFT, that is very fine resolution possible in frequency domain. It is instantaneous dynamic range that is limited by 8 bit converter but that will also be at least 50dB with averaging. What is missing is synchronized  signal gen  and software to do automatic Bode plot. If you just sweep with siggen and use peak hold, you will get quite accurate looking plot. Then use cursors to readout the frequencies and levels.. For tone control and EQ testing quite sufficient for repair type tasks. It's not lab grade 0.001% type of measurement but quite OK to verify tone controls of preamp in guitar amp.

its possible to do a bode plot this way ...

Feed it some white noise and do an FFT - instant Bode plot!

No, Bode plot is amplitude/phase diagram. What you are saying is simple amplitude/frequency response. Much different thing.

The phase part is what makes it Bode plot. It is basically a low frequency VNA.
You can also plot it in polar coordinates and get Nyquist plot. Both are used (together with Nichols plot) in control systems stability analysis..

To make it simple, simple frequency amplitude response shows how loud certain tone will be. 

Bode plot will show that, and also why your amplifier is oscillating and help you find compensation network that will make it stable. Or show why different transistors in amplifier output stage will/won't oscillate, or see how just swapping opamps for "better ones" is not always a good idea...

[ITTSB.Europe]

But I've never had access to a scope so as an example something like what sort of bandwidth/wfms do I need in a scope to look at a 10mhz clock signal and know its good enough that a micro or cpu would be happy running from it, is totally outside my experience.   

I will pass a tip in favor of consumer education, prior of you spending your money, you better register to watch a few webinars from Oscilloscope manufacturers.
The quick answer is that the scope it should be seven times faster than your signal of 10MHz, so you are looking for a scope at 70MHz bandwidth as minimum.
100MHz of bandwidth this is the favorite choice at professional electronic repairs for consumer products.

I got my GW Instek scope (GDS 2000A) at 2013, and this receives regularly firmware updates even up to date.
If you think the scope as long term investment, then it feels good when the scope runs with latest software.
Such a scope will never be considered as outdated.
   

[Celiturbo]

I will pass a tip in favor of consumer education, prior of you spending your money, you better register to watch a few webinars from Oscilloscope manufacturers.
The quick answer is that the scope it should be seven times faster than your signal of 10MHz, so you are looking for a scope at 70MHz bandwidth as minimum.
100MHz of bandwidth this is the favorite choice at professional electronic repairs for consumer products.
 

I could see it if you have something tracking rising/falling edges on a 10mhz triangle wave, or modulated signals.  Anything I'd be looking at over 1mhz would be a 3.3 or 5v square wave. I believe this is also more of an issue of samples per second than outright bandwidth, samples being the level of detail vs time available. That being the case the 1054b is 1gs/s the same as basically all of the 100mhz scopes. Averaging and memory depth can also play a part. For my use say worst case is a 24mhz 3.3v square wave, measured bandwidth on the 1054b is 80mhz or its -4db at 100mhz vs 108mhz for its 100mhz counterpart, Gs/s, wfms, & memory depth are all the same then that puts the 1054b about -1.5db vs the 1104b at 100mhz, ie the 100mhz scope can see a signal about 30% smaller than the 50mhz scope. I'm looking at 3.3-5v stuff though, not too worried about seeing 50-100mv up there.  Even if the 50mhz scope only had a measured bandwidth of 50mhz I don't see where it would be an issue for my use, which is why I was still considering it against 100mhz scopes and the whole reason I came here looking for insight from real world experience.

I believe Keysight mentioned 5x in one of their videos anyway, either way the problem with gospel from on high is not just that it is meant to cover everyone in all use cases but it is also meant to up sell product. Frankly from a manufacture I expect an accurate detailed spec sheet, good documentation, and a warranty. I don't want purchasing advice from someone trying to get their hand in my wallet. Lets be honest about firmware updates, I know sometimes new features get added but that is very rare exceptions, the majority of the time firmware updates are band aids on equipment that was rushed out the door without proper testing / refinement, or its related to some variation of DRM.  Somehow this has become the normal but it never would have been acceptable on todays scale in the 90's. 

[bdunham7]

I could see it if you have something tracking rising/falling edges on a 10mhz triangle wave, or modulated signals.  Anything I'd be looking at over 1mhz would be a 3.3 or 5v square wave. I believe this is also more of an issue of samples per second than outright bandwidth, samples being the level of detail vs time available. That being the case the 1054b is 1gs/s the same as basically all of the 100mhz scopes. Averaging and memory depth can also play a part. For my use say worst case is a 24mhz 3.3v square wave, measured bandwidth on the 1054b is 80mhz or its -4db at 100mhz vs 108mhz for its 100mhz counterpart, Gs/s, wfms, & memory depth are all the same then that puts the 1054b about -1.5db vs the 1104b at 100mhz, ie the 100mhz scope can see a signal about 30% smaller than the 50mhz scope. I'm looking at 3.3-5v stuff though, not too worried about seeing 50-100mv up there.  Even if the 50mhz scope only had a measured bandwidth of 50mhz I don't see where it would be an issue for my use, which is why I was still considering it against 100mhz scopes and the whole reason I came here looking for insight from real world experience.

Just keep in mind that you only have 1GSa/s on one channel.  If you are looking at 4 channels, you want the response to start rolling off well before 100MHz.  Your 24MHz square wave will be rounded and possibly blurry on your scope and will only look marginally better on the 100MHz version.  I wouldn't worry about it at this point.  And don't take that comparison chart you are referring to as gospel, there were some issues with how that was done.

[tv84]

Lets be honest about firmware updates, I know sometimes new features get added but that is very rare exceptions, the majority of the time firmware updates are band aids on equipment that was rushed out the door without proper testing / refinement, or its related to some variation of DRM.  Somehow this has become the normal but it never would have been acceptable on todays scale in the 90's.

Totally agree, software update ratio is far from being an advantage point.

BTW, in the 90's, the web was booting up so the distribution of software updates was in its infancy not because the bugs weren't there.

[Celiturbo]

Just keep in mind that you only have 1GSa/s on one channel.  If you are looking at 4 channels, you want the response to start rolling off well before 100MHz.  Your 24MHz square wave will be rounded and possibly blurry on your scope and will only look marginally better on the 100MHz version.  I wouldn't worry about it at this point.  And don't take that comparison chart you are referring to as gospel, there were some issues with how that was done.

I get that, the point was the sampling rate is the same between the Rigol and GW scopes weather they are 50mhz or 100mhz, this should make the bandwidth difference between them more a function of sensitivity & rise time, as it is individual sample points that the shape of the waveform will be constructed from.

I get the feeling some comments have been more along the lines of assuming the 50mhz scope is half the 100mhz scope but that would only be true if the 50mhz was reduced to 500Ms/s, comparing a 25mhz signal on a 50mhz 500Ms/s scope to a 1Gs/s 100mhz scope would be a much different result than comparing a 1Gs/s 50mhz to a 1Gs/s 100mhz. 

That chart aside I've seen other side by side testing that shows similar so I think while not an absolute calibrated result its pretty reasonable for comparison sake.

Its on its way anyway, I'm sure it will be fine for my use, if I find it lacking I'll be sure to speak up as to why.

On a side note, local dude I contacted months ago just got back to me this morning about a old analog 150mhz 4ch Hitachi V1150 I think I'm going to pick up also, $50 seems like a steel to me. Same as a Gould 3150 I think. I had originally intended to get my feet wet with something cheap used before buying a new DSO but deals seem to go quick and haven't had any luck, now of course since I ordered the GDS last night right..