I'm in the process of getting an oscilloscope - probably a Siglent SDS804X HD - and thought it'd be handy to add a waveform generator.
I'm fine with doing analysis on my Linux laptop, and was looking at the Uni-T UTG-962G ~£147, and the Siglent SDG1032X ~£277.
So what are you using, and how do you like it?
I'm fine with doing analysis on my Linux laptop, and was looking at the Uni-T UTG-962G ~£147, and the Siglent SDG1032X ~£277.
So what are you using, and how do you like it?
Most arb generators have quite high distortion, so not useful for distortion analysis unless it is very gross distortion.
Some siglent scopes, such as my sds2104x plus, can communicate to external siglent arb generators to perform bode plots (frequency response) internally, though the sds2104x has a 50MHz one built in which covers most of my bode needs, just limited for RF.
Some siglent scopes, such as my sds2104x plus, can communicate to external siglent arb generators to perform bode plots (frequency response) internally, though the sds2104x has a 50MHz one built in which covers most of my bode needs, just limited for RF.
@Potentiallyincorrect What do you find the Bode plots useful for?
@TimS The Siglent SDG1032X has -60dB or 0.075% thd on sinewaves. What do you use to look at distortion?
@TimS The Siglent SDG1032X has -60dB or 0.075% thd on sinewaves. What do you use to look at distortion?
Hi,
I also own Siglent scope and an SDG 2042X.
It was mainly the lower THD due to 16Bit resolution that gave the edge over the SDG1000X series.
Also the easy control via the scope comes handy.
Besides the Bode plot function, I like the easy way to import scoped signals to the SDG for later playback and analysis.
jauu
Calvin
I also own Siglent scope and an SDG 2042X.
It was mainly the lower THD due to 16Bit resolution that gave the edge over the SDG1000X series.
Also the easy control via the scope comes handy.
Besides the Bode plot function, I like the easy way to import scoped signals to the SDG for later playback and analysis.
jauu
Calvin
A USB soundcard is often the easiest way to get clean audio signals, though amplitude is limited of course - not ideal for measuring though, I've used one of the Focusrite boxes for this.
I've recently been spoiled with the QA403, which has output upto +18dBV and super low distortion - but that's obviously quite an investment by comparison.
I use a cheapo dual signal generator from eBay for faster signals, like square waves, and that's good to +17dBV and MHz... So cheap it shows peak-to-peak measurements as "AMPL=xxxx" !
Sometimes I want really fast edges for TDR or suchlike and a digital output from a microcontroller often serves, a few ns risetime.
I've recently been spoiled with the QA403, which has output upto +18dBV and super low distortion - but that's obviously quite an investment by comparison.
I use a cheapo dual signal generator from eBay for faster signals, like square waves, and that's good to +17dBV and MHz... So cheap it shows peak-to-peak measurements as "AMPL=xxxx" !
Sometimes I want really fast edges for TDR or suchlike and a digital output from a microcontroller often serves, a few ns risetime.
If you think you might want Bode plot capability (which I use quite a bit) then you might think about getting a compatible Siglent generator unless you want to write your own code to control the scope and a different generator. If you don't care about Bode plots then there are a gazillion options, including the Uni-T you mentioned. Eevblog is a great place to look for user experiences with different models you might consider.
I have an old BK Precision function generator I got off of ebay that can do most of what I need, but since it is not computer controllable it has some limitations. If you just want to do basic testing and don't need computer control, then a simple old-school function generator may fit the bill. You can sometimes find great deals on used ones on ebay. I did have the seller send me photos of what the square-wave response looked like on his scope to make sure it would work for me. Of course you are always gambling a little with ebay purchases, so not everyone is comfortable going that route.
When I want to inject signals not referenced to circuit ground I use a cheap FG-100 that can be powered from a 5V power bank. It was useless out of the box since these tend to come with fake opamps installed, but after replacing them with a couple of opa2992 it works pretty well.
But I mostly use the waveform generator built into my Picoscope 5244B since it is convenient and can do most of what I need. I like that it is super easy to use for Bode plots, including ones that go up to high enough frequencies to check for peaking that could indicate stability issues with amps. But it is limited to +/- 2V output so is not adequate to test clipping on many audio projects (the BK precision and FG-100 both are fine for that). I have never used the arbitrary waveform capabilities of this device.
Like others, I use an audio interface (Scarlett Solo) and REW for distortion measurements. This gets used quite a bit and has added a lot of enjoyment to the hobby.
Not for audio-related, but for signals > 20 MHz (the limit of my Picoscope generator) I have used a TinySA. It actually has a simple serial-over-USB protocol so testing can be automated. I wrote simple Python scripts for this.
jason
I have an old BK Precision function generator I got off of ebay that can do most of what I need, but since it is not computer controllable it has some limitations. If you just want to do basic testing and don't need computer control, then a simple old-school function generator may fit the bill. You can sometimes find great deals on used ones on ebay. I did have the seller send me photos of what the square-wave response looked like on his scope to make sure it would work for me. Of course you are always gambling a little with ebay purchases, so not everyone is comfortable going that route.
When I want to inject signals not referenced to circuit ground I use a cheap FG-100 that can be powered from a 5V power bank. It was useless out of the box since these tend to come with fake opamps installed, but after replacing them with a couple of opa2992 it works pretty well.
But I mostly use the waveform generator built into my Picoscope 5244B since it is convenient and can do most of what I need. I like that it is super easy to use for Bode plots, including ones that go up to high enough frequencies to check for peaking that could indicate stability issues with amps. But it is limited to +/- 2V output so is not adequate to test clipping on many audio projects (the BK precision and FG-100 both are fine for that). I have never used the arbitrary waveform capabilities of this device.
Like others, I use an audio interface (Scarlett Solo) and REW for distortion measurements. This gets used quite a bit and has added a lot of enjoyment to the hobby.
Not for audio-related, but for signals > 20 MHz (the limit of my Picoscope generator) I have used a TinySA. It actually has a simple serial-over-USB protocol so testing can be automated. I wrote simple Python scripts for this.
jason
I have a Motu M2 (similar to the Scarlett, -110dB noise, 192kHz DAC) which I recently used with REW for measuring speakers.
If I were to use that for distortion and Bode plots, I'd still need an external low noise signal generator?
The Picoscope 5244 looks very nice, however it's over my budget, which brings me back to the Siglent SDG2042X used by Calvin. Unless there are suitable signal generators available elsewhere eg on eBay which I can interface with REW?
If I were to use that for distortion and Bode plots, I'd still need an external low noise signal generator?
The Picoscope 5244 looks very nice, however it's over my budget, which brings me back to the Siglent SDG2042X used by Calvin. Unless there are suitable signal generators available elsewhere eg on eBay which I can interface with REW?
The MOTU M2 should be good for distortion measurements - no external signal generator required. And if REW can do bode plots (you should do your own research to verify that) then the M2 should be fine as-is for that as well.
Yeah, the Picoscope is expensive - I bought it used on ebay at about a 75% discount.
jason
Yeah, the Picoscope is expensive - I bought it used on ebay at about a 75% discount.
jason
OK so REW running on the MacBook can generate the signals and measure the components and distortion. It did provide Bode plots for my speaker measurements. Upper frequency limit for measurements will be around 44kHz I think. (I recall its impedance measurements weren't completely consistent from run to run).
So what would I miss if I go this way rather than going Calvin's way with the SDG2042X?
So what would I miss if I go this way rather than going Calvin's way with the SDG2042X?
Giallograle, for starters I suggest you get to know your M2 better and what it can achieve with REW. The M2 will have hugely better S/N than any scope, which in your $ region would be 8-bit, or perhaps 12 bit at a stretch. Your M2 can have an audio bandwidth to 96kHz, and can support 1Mohm scope probes for 10:1 and 100:1 protection for simple probing of electronics. You can prepare a calibration file for each probe setyp you use to effectively get ruler flat response from circa 2Hz to 96kHz. You can connect your M2 to a laptop and avoid ground/earth hum related hassles. REW can tweak your test sinewave to better than the M2 spec level of harmonic distortion, so you have effectively a world class sinewave generator. You can certainly use REW's scope display to augment how you interpret a frequency spectrum response.
I'd suggest some extra care with your impedance measurement jig is needed - as it can provide an excellent RLC meter, or passive device impedance plotter.
I don't have an M2, but something similar, and so many people use these interfaces for high-end benchwork nowadays, but it does require some effort from you to get the most out of a tool, just like any tool.
Picoscope offer a range of scopes, so perhaps don't just look at the price of the 5244. I bought a 4224A as it had 12-bit performance with a 1MHz AWG, but there is a cheaper 8-bit 2206B that similarly does Bode with FRA4 software - which I use a lot for stability assessment of audio amps, as I want to go beyond 96kHz and below 2Hz.
I'd suggest some extra care with your impedance measurement jig is needed - as it can provide an excellent RLC meter, or passive device impedance plotter.
I don't have an M2, but something similar, and so many people use these interfaces for high-end benchwork nowadays, but it does require some effort from you to get the most out of a tool, just like any tool.
Picoscope offer a range of scopes, so perhaps don't just look at the price of the 5244. I bought a 4224A as it had 12-bit performance with a 1MHz AWG, but there is a cheaper 8-bit 2206B that similarly does Bode with FRA4 software - which I use a lot for stability assessment of audio amps, as I want to go beyond 96kHz and below 2Hz.
Thanks for all the advice. I had a look at the thread:
How to distortion measurements with REW which has a good explanation of the process.
It seems that it's straightforward to use sine wave inputs - I couldn't see any references to sweeps, but REW will do this with speakers. The clear recommendation was to use an external signal generator - either playback of a recording of a sine wave via a second DAC, or an external 1kHz sine wave generator - for better results. Square waves seemed to create difficulties.
Most people seem to be using this setup to test distortion of a 1kHz sine wave. I wonder if this matters?
How to distortion measurements with REW which has a good explanation of the process.
It seems that it's straightforward to use sine wave inputs - I couldn't see any references to sweeps, but REW will do this with speakers. The clear recommendation was to use an external signal generator - either playback of a recording of a sine wave via a second DAC, or an external 1kHz sine wave generator - for better results. Square waves seemed to create difficulties.
Most people seem to be using this setup to test distortion of a 1kHz sine wave. I wonder if this matters?
G, REW includes 'sweeps' in its 'Measure' menu, where a sweep not only captures frequency spectrum response, but also harmonic distortion levels during the sweep and other technical aspects - I'd recommend going through the REW help sections to assist your learning curve.
I don't see a 'clear recommendation' to use an external signal generator for better results. Perhaps appreciate that some don't have a good soundcard interface to start with, or don't know how to use the tool they have, or have an external generator already that they like using, or want an independent measurement source to confirm <0.0000x distortion levels, or .....
Perhaps read up some more on bandwidth and squarewaves as a way of appreciating when and why you may need to apply and capture a squarewave response, and when that could be limited by a tool that has a bandwidth limit. If you are trying to excite a speaker system with frequencies higher than 96kHz then yes you would need to assess how to do that, but if not then .... Some want to excite valve amp output transformer resonances that occur well into the 100's of kHz, so for them another tool/method is needed - horses for courses.
Some people use a simple 1kHz tone as a quick test, especially as it has been a common benchmark to compare equipment over decades (before soundcards). There are many historic and recent test methods that have been of use to practitioners - a quick look at the IM distortion options highlights that aspect.
I don't see a 'clear recommendation' to use an external signal generator for better results. Perhaps appreciate that some don't have a good soundcard interface to start with, or don't know how to use the tool they have, or have an external generator already that they like using, or want an independent measurement source to confirm <0.0000x distortion levels, or .....
Perhaps read up some more on bandwidth and squarewaves as a way of appreciating when and why you may need to apply and capture a squarewave response, and when that could be limited by a tool that has a bandwidth limit. If you are trying to excite a speaker system with frequencies higher than 96kHz then yes you would need to assess how to do that, but if not then .... Some want to excite valve amp output transformer resonances that occur well into the 100's of kHz, so for them another tool/method is needed - horses for courses.
Some people use a simple 1kHz tone as a quick test, especially as it has been a common benchmark to compare equipment over decades (before soundcards). There are many historic and recent test methods that have been of use to practitioners - a quick look at the IM distortion options highlights that aspect.
Thanks, I'm familiar with REW's sweep function having used it with my speakers and it's a powerful bit of software.
Anyway this is not for speakers or valve amplifiers. I'm more interested in accurately characterising discrete components than thd of solid state amplifiers, although quite happy to have that too. Is REW's sinewave generator up to 44khz suitable for that?
Unfortunately the QA401 and newer versions are out of my budget. The SDG2042X is affordable and presumably could be used with REW - but no point if it doesn't add anything for my use.
Is there a good alternative to the Scarlett or Motu DACs for this? Maybe an old Picoscope?
Re test signals, @xrk971 in the first post says:
Anyway this is not for speakers or valve amplifiers. I'm more interested in accurately characterising discrete components than thd of solid state amplifiers, although quite happy to have that too. Is REW's sinewave generator up to 44khz suitable for that?
Unfortunately the QA401 and newer versions are out of my budget. The SDG2042X is affordable and presumably could be used with REW - but no point if it doesn't add anything for my use.
Is there a good alternative to the Scarlett or Motu DACs for this? Maybe an old Picoscope?
Re test signals, @xrk971 in the first post says:
I use an Owon portable oscilloscope and it has a built in function generator for sine, triangle, square, stair step, and different digital pulse trains. I use it to generate test function waves and Oscope to measure the result with DUT inserted in path. Here is example of measuring transmission of 5th order Chebyshev filter.
Measurement shows 2.08mV out (yellow) and 508mV in (blue) at 60mHz for sine wave function.
Measurement shows 2.08mV out (yellow) and 508mV in (blue) at 60mHz for sine wave function.