EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: Bebop on December 02, 2020, 11:05:30 am
-
Hi,
I'm looking to get an oscilloscope for ham radio work (it'll be HF and below) and general electronics work.
I've had a look at what's about and price-wise the Rigol DS1202Z-E seems the most tempting, but the Siglent SDS1202X-E is a similar price and spec, I also considered Rigol DS1054Z but I'm wondering if the extra headroom of the 200MHz might be better with ham radio work.
A couple of specific questions having had a look at the spreadsheet.
As I see it the Rigol has a waveform update rate of only up to 30,000 wfms/s compared to the Siglent with 100,000 wfms/s, however the Rigol has a memory depth of 24M compared to only 14M for the Siglent.
Additionally the Rigol has Intensity Grading Display of only 64 to Siglents 256 and shared controls compared to Siglents independent controls.
I'm just looking for advice as to which of the features are 'more important', is it worth the drop in memory depth for the gain in wfms and IGD?
Thanks,
James
-
You're asking all the right questions, except one.
200Mhz - definitely go for that.
I wouldn't worry about wfms/sec or IGD.
The real decider for radio work should be the FFT/frequency analysis. The Siglent definitely wins on that.
-
You're asking all the right questions, except one.
200Mhz - definitely go for that.
I wouldn't worry about wfms or IGT.
The real decider for radio work should be the FFT/frequency analysis. The Siglent definitely wins on that.
And it's 500uV/div sensitivity and 1 Mpts FFT.
-
You haven't said what you want to measure and why.
Frequently a scope is the wrong tool for RF, a spectrum analyser being more appropriate.
When considering a scope, be very aware of the consequences of sensitivity, linearity, and for a digitising scope, the number of bits.
-
Thanks for the replies. So is the memory depth not too big of a deal?
In terms of what I'd be measuring, yes I know that a spectrum analyser in many situations is more helpful but that is well outside by price range and also not so helpful for general electronics debugging, for instance I'm currently working on an oscillator which isn't working, but a scope would be helpful in seeing what's going on. I'm guessing that the FFT spectrum analyser feature of the scope could be very useful.
So the frequency analysis is better in the Siglent? That does seem like an important point. I missed that question as the details weren't on the spreadsheet!
-
Thanks for the replies. So is the memory depth not too big of a deal?
In terms of what I'd be measuring, yes I know that a spectrum analyser in many situations is more helpful but that is well outside by price range and also not so helpful for general electronics debugging, for instance I'm currently working on an oscillator which isn't working, but a scope would be helpful in seeing what's going on. I'm guessing that the FFT spectrum analyser feature of the scope could be very useful.
So the frequency analysis is better in the Siglent? That does seem like an important point. I missed that question as the details weren't on the spreadsheet!
The SDS1202X-E FFT is about as good as it gets for this class of DSO yet until you also study the user manual that will be updated with later FFT functionality that was added in firmware you can't see the full picture of what these can do from just the datasheet.
Unless you're into very long records/captures using both channels when it is shared the lesser mem depth shouldn't be a problem.
-
Thanks for the replies. So is the memory depth not too big of a deal?
It depends on what you're using the scope for.
If 'radio' is a principle use then I'd say the memory's less important then some other things.
-
I am a Ham and I build/study HF electronics. I have an Agilent DSO3024A with 4MB of memory. I also have a dedicated Spectrum analyzer.
My advice is go for bandwidth first as long as quality and usability are on par. I have not be limited by my meager 4MB.
I have found the Scope fft not very useful due to the dynamic range limitations but that is inherent and shared limitation for all scopes unless they have a specific SA front end.
4 channels is incredibly useful if you can afford it.
-
In terms of what I'd be measuring, yes I know that a spectrum analyser in many situations is more helpful but that is well outside by price range and also not so helpful for general electronics debugging, for instance I'm currently working on an oscillator which isn't working, but a scope would be helpful in seeing what's going on. I'm guessing that the FFT spectrum analyser feature of the scope could be very useful.
I suggest you understand the limitations on your measurements due to limited linearity and limited dynamic range and an oscillator sufficient for looking at waveform shape in the time-domain.
For general purpose debugging, any scope is useful. For looking at the phase noise in an oscillator or multiple low-level signals or lock-in transients, less so.
As for cost, I have two general purpose spectrum analysers capable of >18GHz (HP8562A, Tek492). The more expensive one cost me £350. Of course I needed patience before I located those!
For looking frequencies and how they change over time, it is hard to beat a modulation domain analyser. Mine cost me £400, but again I needed patience.
So the frequency analysis is better in the Siglent? That does seem like an important point. I missed that question as the details weren't on the spreadsheet!
With any datasheet, the trick is to read and understand what isn't stated.
-
Thanks for the replies. So is the memory depth not too big of a deal?
In terms of what I'd be measuring, yes I know that a spectrum analyser in many situations is more helpful but that is well outside by price range and also not so helpful for general electronics debugging, for instance I'm currently working on an oscillator which isn't working, but a scope would be helpful in seeing what's going on. I'm guessing that the FFT spectrum analyser feature of the scope could be very useful.
So the frequency analysis is better in the Siglent? That does seem like an important point. I missed that question as the details weren't on the spreadsheet!
If you just want to "investigate" RF things below 1GHz the TinySA is a remarkably capable tiny spectrum analyzer, and cost well under $100. Couple this with a nice DSO like folks are mentioning and you've got sub-1GHz RF things well covered :)
Best,
-
If you just want to "investigate" RF things below 1GHz the TinySA is a remarkably capable tiny spectrum analyzer, and cost well under $100. Couple this with a nice DSO like folks are mentioning and you've got sub-1GHz RF things well covered :)
Yep. I forgot about that one.
For the price of a Siglent you could get a Rigol+TinySA. Definitely worth checking out.
There's plenty of videos of TinySA out there...
-
So the frequency analysis is better in the Siglent? That does seem like an important point. I missed that question as the details weren't on the spreadsheet!
I can't say for the DSOs mentioned here but on the SDS2102X Plus I have the FFT is quite good. If you didn't know where the FFT display came from (a DSO) you would think it was from a dedicated FFT analyzer, it's that good :)
Maybe some of the FFT DNA from the SDS2000X+ has trickled down to the SDS2102X-E SDS1202X-E ::)
Best,
-
My favorite piece of equipment for Ham radio equipment and fixing things is an HP 8935 service monitor.
No real scope in this thing though, the scope it has is audio and pretty useless.
Combines SA and RF gen and AF gen, power meter and Modulation in one big lump
It is also a good HF, VHS, UHF receiver.
-
I can't say for the DSOs mentioned here but on the SDS2102X Plus I have the FFT is quite good.
But it's not going to go up to 1GHz or have the sensitivity of a spectrum analyzer.
-
HP 8935:
Goes to ONE GHz just fine (rated for this) It is really a cell phone tower analyzer so has to be good at cell freq.
Then again from 1.7 GHz to 2 GHz (cell Freq)
No too accurate below 10 MHZ but really OK down to maybe 1 MHz
Not as accurate as power meter, but pretty good
Twin audio generators
When you are testing a radio you can modulate the Mhz input antenna signal with one audio generator and use the other audio generator to inject audio into mic input.
Modulates signal either AM or FM wih big ranges on both modulations
Has about 100 Watt input limitation.
Smallest range is 10 KHz on SA, plenty for any repair job. So you can easily see one KHz side band. Freq counter is down to one Hz, limited by the standard. If I need better I drag out the old two ton Marconi
-
I can't say for the DSOs mentioned here but on the SDS2102X Plus I have the FFT is quite good.
But it's not going to go up to 1GHz or have the sensitivity of a spectrum analyzer.
Never said it goes to 1GHz, that's why I mentioned the TinySA. It's pretty good up to ~100MHz, and quite good below 10MHz.
It performs like a dedicated FFT analyzer at the lower frequencies (of course one needs to set it up properly, which isn't obvious unless you fully understand FFT basics). You would never expect that a DSO 8 bit ADC is behind this. Couple that with the superb low noise, high resolution, wide-band front end and Siglent just got this SDS2000X+ series DSO right ;)
Like mentioned, hopefully some of the 2000X+ DNA has trickled down to the other DSOs.
-
Thanks for the replies. So is the memory depth not too big of a deal?
I don't think so. My main scope has 8M, that was the optional high end "deep memory" at the time it was made, and that's way, way more than I've ever needed. If you capture a really big chunk of waveform all it means is you've got pages and pages and pages of waveform to scroll through and everything just takes longer to process. You're better off learning how to use triggering properly and capture the stuff you want to look at. instead of trying to grab everything and sort it out after the fact.
-
Like mentioned, hopefully some of the 2000X+ DNA has trickled down to the other DSOs.
In reverse, X Plus DNA came from SDS1202X-E the first DSO from Siglent to use the Zynix range of processors.
Their UI evolved from SDS2000 then SDS1000X then SDS2000X, later X-E models and X Plus was following on from SDS5000X with the new UI these 2 model series use today both of which are touch and mouse capable.
-
Thanks for the replies. So is the memory depth not too big of a deal?
I don't think so. My main scope has 8M, that was the optional high end "deep memory" at the time it was made, and that's way, way more than I've ever needed. If you capture a really big chunk of waveform all it means is you've got pages and pages and pages of waveform to scroll through and everything just takes longer to process. You're better off learning how to use triggering properly and capture the stuff you want to look at. instead of trying to grab everything and sort it out after the fact.
I'm afraid that is horribly simplistic way of looking a things.
Some things are simply long processes that you need to sample for long time with maximum bandwidth.
For that, you need long memory. If you're looking at repetitive signals at short timebases, even 10k samples are enough.
To get 500ms of startup transient with 20 MHz bandwidth and 100MSPS/s you need 50MS memory. For higher bandwidths, much more..
There is no replacement for input bandwidth, there is no replacement for long sample memory, if you need any of those.
I think it would be more correct to say that you never did things that needed more memory.
I also don't run out of memory most of the time with KS MSOX3000T, and that one has even les memory than your scope. But occasionally, I have to fire up Picoscope. I do agree that sifting through 100MSPS of data is much easier on 24" monitor on PC though....
-
To get 500ms of startup transient with 20 MHz bandwidth and 100MSPS/s you need 50MS memory. For higher bandwidths, much more..
That's quite a specialist job though, and one that could be handled by a cheap(ish) USB 'scope if the need arises. No need to spend insane amounts of money on an oscilloscope just-in-case.
-
Thanks for the replies. So is the memory depth not too big of a deal?
It depends on what you are working on. Deep memory is very handy to make longer captures and then zoom in on details or move left/right to see more of a signal (but the Siglent scope can't do the latter because it only captures enough data to fill the screen even with the memory length set to a higher number). More memory gives more possibilities and less chance of needing to 'make do'. BTW another option is the GW Instek GDS-1202B. This model has 10Mpts per channel and uses all the memory you tell it to use. Otherwise it is pretty much on par (both in price and the hardware) with the Siglent you selected.
-
Like mentioned, hopefully some of the 2000X+ DNA has trickled down to the other DSOs.
In reverse, X Plus DNA came from SDS1202X-E the first DSO from Siglent to use the Zynix range of processors.
Their UI evolved from SDS2000 then SDS1000X then SDS2000X, later X-E models and X Plus was following on from SDS5000X with the new UI these 2 model series use today both of which are touch and mouse capable.
Well then, that should be a superb FFT implementation regarding the SDS1202X-E ;)
Best,
-
Like mentioned, hopefully some of the 2000X+ DNA has trickled down to the other DSOs.
In reverse, X Plus DNA came from SDS1202X-E the first DSO from Siglent to use the Zynix range of processors.
Their UI evolved from SDS2000 then SDS1000X then SDS2000X, later X-E models and X Plus was following on from SDS5000X with the new UI these 2 model series use today both of which are touch and mouse capable.
Well then, that should be a superb FFT implementation regarding the SDS1202X-E ;)
SDS2000X Plus is better, much better and so it should be at 4x the price. ;)
While neither are a SA the principles of usage are similar however unlike a SA the timebase setting in a scope can influence FFT results. Some scopes with tiny FFT only require a small # of waveform cycles to be displayed while with larger Mpt FFT the more info FFT has to work with (more cycles) the better the result.
Garbage in = garbage out. ;)
-
Thanks for the replies. So is the memory depth not too big of a deal?
I don't think so. My main scope has 8M, that was the optional high end "deep memory" at the time it was made, and that's way, way more than I've ever needed. If you capture a really big chunk of waveform all it means is you've got pages and pages and pages of waveform to scroll through and everything just takes longer to process. You're better off learning how to use triggering properly and capture the stuff you want to look at. instead of trying to grab everything and sort it out after the fact.
Not always. In many cases you don't really know at what timeframe an event takes place but usually the areas of interest are easy to spot (think about the power up sequence example from 2N3055). The advantage of deep memory is that you can capture a lot at a high enough samplerate to capture every tiny detail without needing to re-capture an event ad nauseam. There is also a causal relation between every event in a signal. If you have to capture a signal in little bits and pieces you can never be sure that signal piece C occurs due to / whether piece A or piece B being present in a signal. In the end the trigger abilities of an oscilloscope aren't that sophisticated.
-
There are niche cases where super deep memory can be a benefit, if you anticipate encountering those cases often then it may be worthwhile to put more weight on that feature, but personally I put very little on it. The scope I use most often is a TDS3000 which has by modern standards very shallow memory, 2.5kpps IIRC. Occasionally I find that limiting but most of the time it isn't enough of an issue for me to go drag the boatanchor off the bench if I'm working on something in another room. I came from an analog scope that had no memory at all though.
-
Thanks for the replies. So is the memory depth not too big of a deal?
In terms of what I'd be measuring, yes I know that a spectrum analyser in many situations is more helpful but that is well outside by price range and also not so helpful for general electronics debugging, for instance I'm currently working on an oscillator which isn't working, but a scope would be helpful in seeing what's going on. I'm guessing that the FFT spectrum analyser feature of the scope could be very useful.
So the frequency analysis is better in the Siglent? That does seem like an important point. I missed that question as the details weren't on the spreadsheet!
Here is one FFT with SDS1202X-E.
Yes settings are not optimal at all if example looking 455kHz IF. This image just tell what it can also show.
White dots are average there and whole display used dots mode (because this image purpose was bit different)
Just turn dot mode off resolution etc is same.
(https://siglent.fi/pic/SDS1000X-E/SDS1202X-E__FFT_0dBVrms-455kHz_AM-Mod--first-100Hz--then-120Hz-.png)
Signal is 455kHz carrier modulated using dual tone 100Hz and 120Hz and modulation depth 20% (for noobs 20% is fun because only just this % sideband dBc have same number 20 (-20dBc)
There was used 10MSa/s for keep Nyquist enough high that example possible harmonics etc do not alias so much.
Of course also used capture length and time scale is well over what need and it is here making FFT slow.
With 10MSa/s 1Mpts FFT length give nice freq resolution for 0 - 5MHz band. (of course as can see freq axis is highly zoomed in from full span)
For radio things one important thing is memory length if look one thing. If we are looking (example) AM modulated signals we often want look it in time scale where modulating signal shape is visible and well stable with trig.
Memory lenght is not so important if we use right working peak detect mode for this but then we loose intensity gradation what is also part of information many times. Also then we loose carrier when sampling speed push fNyquist well under carrier. As most old DPO did if we look modulated signal bode.
If we set some imagined numbers. If we want look example 118MHz AM and modulated 100Hz and want we can see also real carrier and also this modulated shape. If think (for easy math) our display have 10 div horizontally. For one cycle 100Hz we need set time scale (aka timebase) to 1ms/div. Now then we have carrier 118MHz not so nice number so round it up to next full hundred 200MHz. Due to Mr Nyquist is here... we need least 400MSa/s ... oh but not... we need also level not only freq... and due to some reality... we need least 400MSa/s/0.8=500MSa/s. Now we have 2.5 sample over one 200MHz carrier cycle. Ok, well made Sinc interpolation can barely give ok looking result but of course more is better.
Now we have some minimum limit. We need 500MSa/s and we need take 10ms acquisition with this speed. 500x0.01=5. We need 5M memory for this if zero overlap.
With 5M memory we can keep 1ms/div (when whole length is 10ms) and 500MSa/s samplerate. With this we have up to 200MHz carrier and we have modulating frequency one cycle on screen when its frequency was 100Hz.
In this case we can use zoom window.. like dual timebase in analog scope... upper window modulated signal shape and bottom window 200MHz carrier sinewave.
This was minimum need for this. As can see memory is important for keep samplerate enough high even with slow time scale.
Of course there is many many other things but this was now looking somehow with radio things approach angle this thing.
In Siglent SDS1202X-E there is 14div in display time axis and max memory 14M so it is well enough for this kind of things. It can do it even with 2 channel on. 7M memory both and 500MSa/s both.
(https://siglent.fi/pub8S/SDS1000X/ESK/SDS1000X-Top-14MHz-AM2kHz-Bottom-14MHz-Carrier.png)
This is what I mean looking modulated shape and looking also carrier itself using zoom.
Do not look other things in this image. It is OLD and not at all from this model. This image was originally part of set for explain freq counter here.
Used carrier was 14MHz. There can see freq counter in right top corner. It need always remember this is not input signal frequency counter, never even designed for it!!! It is trigger freq counter, as can see... it counts in this case rising edges what meet user defined trig setup. Trig threshold is around half way in mod depth. So half is out from count.
But here this image purpose is only show what I mean looking AM shape and same time looking carrier itself.
Now if we have short memory where samplerate is dropped down well under Nyquist.. we get this upper part AM shape also if there is peak detect. But we loose this carrier wave.
Foe this kind of things in practice with 200MHz scope freq range for this purpose we can say we need least 5M memory if want go down to 100Hz Mod and still carrier available.
In this image trigger was done using just rising edge, threshold around half way AM depth and adjusted Trigger Holdoff time for get stable trig. This is simplest method but also very poor method. It is suitable only for around fixed Modulating freq. If modulating freq is changing we need much more clever trig and modern scope have it. One of very good trig function for this is Drop Out trig. Set trig mode Drop Out, Rising edge, trig level half of mod shape top and bottoms, Overtime Type Edge, Time 25us. And thats it.
Now if modulating freq change from around 100Hz up to even 20kHz you do not need adjust trigger at all... whole range it trig rock solid (until carrier or Mod level change) If you use this poor old school Trigger Holdoff adjust you need adjust it frequently when Mod freq is changing.
-
If you just want to "investigate" RF things below 1GHz the TinySA is a remarkably capable tiny spectrum analyzer, and cost well under $100. Couple this with a nice DSO like folks are mentioning and you've got sub-1GHz RF things well covered :)
Yep. I forgot about that one.
For the price of a Siglent you could get a Rigol+TinySA. Definitely worth checking out.
There's plenty of videos of TinySA out there...
I think this is what I’m going to do. Gives me a ‘decent’ (for my price range) scope for both Electronics Ham Radio work and a spectrum analyser for where the scope will start to struggle.