Will RIGOL DG1022U be appropriate for audio repair and calibration work?

[vladc77]

Hi to all,

I am trying to obtain a function generator that I mainly going to use to repair audio equipment. It seems like I will be better off with an audio generator but I'd like to get a newer piece of equipment that I will be able to extend to other repair work in the future. The other options I consider are Instek SFG-1003 or Instek SFG-1013. Please advice if these newer function generators will be good companions for audio. Any advice is highly appreciated!

[Gyro]

Most function generators (either digital or analogue ones) have rather poor sine-wave distortion performance. If you are specifically interested in distortion measurements then you would probably be better finding or making a low distortion RC oscilator. For square wave testing they should be fine though.

[vladc77]

Thank you, Gyro! Would you recommend any models?


Sent from my iPhone using Tapatalk

[rstofer]

I'm not in the audio business but the datasheet lays out the THD at < 0.2% for the sine wave in the audio spectrum.  I have no idea whether this is high, medium or low.  Page 3 here:
file:///C:/Users/Richa.DESKTOP-1K83UAF/Downloads/DG1022_datasheet.pdf

I would expect all digitally synthesized sine waves to have a higher level of distortion.  One solution might be a low pass filter on the output to knock down any harmonics.  But harmonics of what?  If the tone is 1 kHz, there is room for a lot of harmonics if the filter is set to 20 kHz or so.

eBay is loaded with audio signal generators of dubious quality and no warranty.  I didn't see any of the old time HPs and that's where I would go if I was interested in quality sign waves.

Then again, is the function generator just being used to inject a signal and chase it down the rathole?  In that case, a small amount of distortion might not matter.  If, OTOH, the generator is being used as a source in the design of low distortion amplifiers, it might not be the right tool.

I actually want to use the arbitrary feature to derive signals representing potholes in the roadway that I can feed to my analog computer simulation of an automobile suspension.  That would be very cool!

[vladc77]

Thank you again. This is exactly my dilemma. I want to avoid buying an old HPs or similar because they more likely will be out of specs. I thought I should buy the newer FG and with possibility to use it for other tasks too. However, as far as I hear they are not great fit for audio. I am still not sure what I should end up getting.

[Lightages]

If your intent is to merely supply audio signals to the input of a device under test, then a sound card output and the appropriate software will be more than good enough. If you want to inject signals into specific circuits then you could use a music player with certain files saved on it in wav or flac format that were generated using free software. If you are interested in this I am happy to provide more details.

[Gyro]

^^ That sounds like a good method of generating low distortion test signals to me. 

You might want to consider breaking possible ground loops with a decent quality audio transformer (Edit: not needed if using a portable player obviously).

[vladc77]

Definitely, I would like to hear about these options Lightages is talking about. The only concern about this method is to avoid damaging computer and it is more cumbersome. Thank you.

[rstofer]

Thank you again. This is exactly my dilemma. I want to avoid buying an old HPs or similar because they more likely will be out of specs. I thought I should buy the newer FG and with possibility to use it for other tasks too. However, as far as I hear they are not great fit for audio. I am still not sure what I should end up getting.

For your purposes, why not just build one:
http://www.bobsdata.com/audio_sine_wave_generator/index.html

Yes, a Function Generator would be cool to have but somehow I have avoided it all these years.  I might buy the 1022A just because I want the 'arbitrary' function but I get that with the Digilent Analog Discovery gadget.

Speaking of the Analog Discovery, it has a headphone jack output from the arbitrary waveform generator and somebody has gone to a lot of trouble to create an audio specific application:

http://www.thestuffmade.com/audioanalyzersuite/publish.htm

http://store.digilentinc.com/analog-discovery-2-100msps-usb-oscilloscope-logic-analyzer-and-variable-power-supply/

I have version 1 and I like it a lot.  There's nothing like a 27" screen for a logic analyzer.  Actually, I have dual 27" monitors on my development computer so I can get great visualization of the scope on one screen and the other instruments on the other screen.  Very nice for old eyes...

[Gyro]

There are many options for self-constructed standalone low distortion (some extremely low) sine-wave oscillators on the web, eg:

http://www.janascard.cz/PDF/An%20ultra%20low%20distortion%20oscillator%20with%20THD%20below%20-140%20dB.pdf

http://www.edn.com/electronics-news/4389675/Low-distortion-oscillator-tests-measurement-circuits

http://sound.westhost.com/articles/sinewave.htm

http://moorepage.net/RC.html

The actual complexity and component count is quite small, particularly if you only need a fixed frequency, say, 1kHz or a few switched ones.

P.S. It's also possible to generate nice sharp square waves with a bit of CMOS logic and an AC coupled attenuator on the output.

[vladc77]

Thank you for the information! The Digilent option is quite interesting. However, I already got Rigol DS 1054Z Digital Oscilloscope. I want to pair it up with a decent low distortion function generator. I may try building my own. Thank you Gyro and rstofer for the links! On the other hand, I am still looking to get the manufactured one if I'll find the good match. As far I understood, the RIGOL DG1022U is out of the option. It is a high distortion and possibly just too much. I thought that the RIGOL DG1022U for the price and versatility could be a good option but not really good for audio. Am I correct?

Does anyone familiar with Kenwood FG 273a? Will this FG suitable? The specs I found state it is less than 1% distortion. It looks like still high.

It seems like the lowest distortion results can be achieved with the computer/software based approach that Lightages will possibly provide more details. It also can be the most affordable. Am I correct here?

Thank you again!

[Gyro]

Unfortunately the Rigol DS 1054Z is only an 8 bit resolution scope, its FFT function isn't going to go low enough for the generator or amplifier distortion to be the limiting factor. It probably won't be able to see the difference between the various generator options.

You may be able to get some useable results out of it if you pair the generator with a sharp notch filter to remove the fundamental, but even then the scope probably won't have the dynamic range to compare the magnitudes of the different distortion harmonics.

Sorry 

[vladc77]

Will Rigol 1104Z work better for this? Can you recommend these sharp notch filters that may help? Thank you!

[Gyro]

I'm not an expert on Rigol stuff, but it looks as if the 1104Z is a 12 bit scope, so yes, 16 times higher resolution.

For reference I use an old Picoscope ADC216, it's only 333ksps, but it's a true 16bit scope with 96dB resolution and noise level to match.

Edit: You would build the notch filter yourself - Google 'Twin T filter'. The circuit is simple, matching the component to achieve best notch depth less so.

[vladc77]

RIGOL DG1022U can be hacked to become the same as 1104. I am still interested in learning what notch filters I may need to have. I need that scope not only for audio but prefer to use it for audio too.

[Gyro]

Sorry, you probably missed my edit re. the notch filter.

[vladc77]

Thank you! I can see it now. Do you think I can be fine without notch filters with 1104z? I just checked specs here: http://www.tequipment.net/Rigol/DS1054Z/Digital-Oscilloscopes/?search=true#tab-specifications

It says Rigol DS1054x is 12 bit high resolution.

[Gyro]

Well 12 bits would give a maximum resolution of 244uV relative to 1V, ie. 0.0244% or -72dB (assuming my maths is correct  ). The other limiting factor is the noise level of the input circuitry of the scope, which tends to be higher with higher bandwidth. I really don't know how close to the theoretical resolution you would get in practice.

Put it this way, it might be enough depending on your needs but notch filtering would certainly help.

[vladc77]

Thank you for the info!

[vladc77]

I still want to find out if Kenwood FG 273 will be suitable for audio? Is anyone familiar with this one? Thank you!

[Gyro]

From what I can see on the web it has a maximum sine distortion of about 1%, fairly typical of function generators.

[vladc77]

Yes, that's true. I need to have something closer to 0.05%. It seems my options are limited.

[Gyro]

You ought to be able to achieve that sort of level fairly easily if you build something. You really want to be looking for 'Audio signal generator' or 'Audio oscillator' as a search term rather than function generator.

[vladc77]

Agree, my idea of scalable approach is not going to work. I wanted to get something newer as Rigol 1022 but it is no way to go. Anyway, it is going to be cheaper but more narrowed approach. Thank you for your support!

[Gyro]

I wouldn't give up on using a PC soundcard (maybe a USB one with USB isolator) for both generating and analysing - it's probably your cheapest and most promising direction at this point. You ought to be able to get to the 16bit level in the audio band.