Question: Cheap pure sine wave genertor 1khz?

[innkeeper]

I'm looking to generate a low thd sine wave to do some THD testing with a DSA.
the internal generator on my DSA is not clean enough for any critical THD testing.

id like it to be 1khz, and will be used for line level input to amplifiers. so 2.1v out max or less. probably down to like 1.2v in that range.

happy building up a circuit or buy one pre built on the cheap.  any ideas?

i could swear i've seen such projects in the past but came up empty in my recent searches.

[Yansi]

Use a decent computer soundcurt plus a suitable software. Easy, done.

You can then easily change test frequency, and even do test like a two tone (IMD).

[RoGeorge]

If you want to build a generator, this type is hard to beat in performance:
https://www.eevblog.com/forum/projects/low-distortion-audio-oscillator-stabilized-via-trigonometric-identity/

There are some linear application notes, too, from the '70s or the '80s, explaining this topology.  Can't find them right now.

[David Hess]

The simplest single operational amplifier Wien Bridge can achieve 0.0025%.  Doing better requires buffering the output and adding common mode suppression which will get you to 0.0003%.  See Linear Technology application note 43.

Doing better than that gets complicated but see Linear Technology application note 67 for an example of what is required to test a modern audio ADC to 18 bits or better.

[Kleinstein]

A good sound card is an easy solution.

For a DIY solution the best bet is likely a Wien bridge oscillator of some kind with really low THD amplifiers. There a quite a few old plans and application notes.  With fixed frequency, there is no real need for the fancy amplitude regulation using instant amplitude measurement, as much less action from the amplitude stabilization is needed. There is an LT appl. Note to get super low THD by combing 3 OPs for each inverting amplifier.

[AG6QR]

If working at a fixed frequency, you can take any oscillator and improve it by suppressing its harmonics with a low-pass filter having a corner frequency somewhere between the fundamental and second harmonic.

Low-pass filters for a fixed frequency aren't that hard or expensive to make. 

[innkeeper]

i think i found a decent thread on this.
https://www.diyaudio.com/forums/equipment-and-tools/205304-low-distortion-audio-range-oscillator.html

most of the solutions are along the lines of what david hess and ag6qr recommended.
theirs a guy there building  decent units but not available at the moment. and a few others that have pcbs for designs etc...the tread is like over 8000 messages long so thats just what i found by skimming it.

certainly no simple solutions,

seems there talking about designs with an output of -140db or better thd+n .. that is along the lines of what i am looking for.

[Fungus]

I'm looking to generate a low thd sine wave

How low? 

There's tons of cheap programmable generators out there if you don't mind adding a microcontroller to set the frequency.

eg. https://www.ebay.com/sch/i.html?_nkw=AD9850

https://www.analog.com/media/en/technical-documentation/data-sheets/AD9850.pdf

Low-pass filters for a fixed frequency aren't that hard or expensive to make.

Probably easier than building something with no distortion.

[Fungus]

There's a whole bunch of other kits/modules on eBay, too:

https://www.ebay.com/sch/i.html?_nkw=DDS+Function+Signal+Generator+Module+DIY

[Gyro]

There are assembled versions available on ebay too, but populating one of these bare boards with genuine components (and maybe lower distortion opamps) ought to get you close to what you need. It uses a J-FET for amplitude control. Jim Williams found that there was some channel resistance modulation when using a J-FET, that could be trimmed out with a little local feedback [Analog Circuit Design - Art, Science and Personalities, Chapter 7 (Max Wein, Mr Hewlett and a Rainy Sunday Afternoon)].

You might get better than the quoted -124dB with some careful component value matching anyway...

https://www.ebay.com/itm/Low-Distortion-Audio-Range-Oscillator-1KHz-Sine-Wave-Signal-Generator-PCB-Board/262844562138

[RoGeorge]

[Analog Circuit Design - Art, Science and Personalities, Chapter 7 (Max Wein, Mr Hewlett and a Rainy Sunday Afternoon)].

That one is a good example, too, available online:
http://www.introni.it/pdf/Williams%2007%20-%20Book%20Chapters.pdf

[Vovk_Z]

The best 1kHz oscillator I know (with best price/thd ratio) is Victor's ultra low distortion oscillator. You can find him at diyaudio.com, he sells sometimes readymade modules via ebay. It is Wien bridge type, and it is cheap. Here it is.
But the answer depends on what TS mean at "pure sine wave"? How much pure?
DDS signal gereators need something like Arduino to program them.

[Grandchuck]

Low distortion is easy but very low distortion is not.  The best one (circuit) here is showing 0.0016% but is probably better since the old HP analyzer is not great with very low distortion measurements.  The circuit is described here:  http://www.redcircuits.com/Page82.htm

The problem with this one is that the dual potentiometer should have very low tracking error and such a part is hard to find.

Having tried several different kinds of very low distortion circuits, my conclusion is they are all 'fiddley".

[innkeeper]

There are assembled versions available on ebay too, but populating one of these bare boards with genuine components (and maybe lower distortion opamps) ought to get you close to what you need. It uses a J-FET for amplitude control. Jim Williams found that there was some channel resistance modulation when using a J-FET, that could be trimmed out with a little local feedback [Analog Circuit Design - Art, Science and Personalities, Chapter 7 (Max Wein, Mr Hewlett and a Rainy Sunday Afternoon)].

You might get better than the quoted -124dB with some careful component value matching anyway...

https://www.ebay.com/itm/Low-Distortion-Audio-Range-Oscillator-1KHz-Sine-Wave-Signal-Generator-PCB-Board/262844562138

i've actually ordered one of these .. figured its a start and maybe i can tweak it.  i was thinking of just replacing the tl071 for something quieter. i didnt know aout that article, i will defiantly try and find it ... i figure it would do until i can make or get somthing better. like vidnic's on diyaudio...

i;ll look into that article if i can find it.

[Gyro]

Low distortion is easy but very low distortion is not.  The best one (circuit) here is showing 0.0016% but is probably better since the old HP analyzer is not great with very low distortion measurements.  The circuit is described here:  http://www.redcircuits.com/Page82.htm

The problem with this one is that the dual potentiometer should have very low tracking error and such a part is hard to find.

Having tried several different kinds of very low distortion circuits, my conclusion is they are all 'fiddley".

This is the first time I've seen independent verification of (the performance of) that circuit, thanks. [Edit: I was planning to build it a few years back but a Topward TAG-4005 came up really cheap on ebay that manages quite a bit better than its specified 0.005%.]

Construction and trimming would be a considerably easier for one (or a few) spot frequencies of course.

Fiddley is the word, that is demonstrated very well in that Jim Williams chapter.

[Gyro]

There are assembled versions available on ebay too, but populating one of these bare boards with genuine components (and maybe lower distortion opamps) ought to get you close to what you need. It uses a J-FET for amplitude control. Jim Williams found that there was some channel resistance modulation when using a J-FET, that could be trimmed out with a little local feedback [Analog Circuit Design - Art, Science and Personalities, Chapter 7 (Max Wein, Mr Hewlett and a Rainy Sunday Afternoon)].

You might get better than the quoted -124dB with some careful component value matching anyway...

https://www.ebay.com/itm/Low-Distortion-Audio-Range-Oscillator-1KHz-Sine-Wave-Signal-Generator-PCB-Board/262844562138

i've actually ordered one of these .. figured its a start and maybe i can tweak it.  i was thinking of just replacing the tl071 for something quieter. i didnt know aout that article, i will defiantly try and find it ... i figure it would do until i can make or get somthing better. like vidnic's on diyaudio...

i;ll look into that article if i can find it.

A good start for sure.

RoGeorge actually found an online copy of that chapter - he linked it above.

[innkeeper]

Jim Williams found that there was some channel resistance modulation when using a J-FET, that could be trimmed out with a little local feedback [Analog Circuit Design - Art, Science and Personalities, Chapter 7 (Max Wein, Mr Hewlett and a Rainy Sunday Afternoon)].

thanks RoGeorge
http://www.introni.it/pdf/Williams%2007%20-%20Book%20Chapters.pdf
also found it here - in the book Gyro pointed to
http://s1.nonlinear.ir/epublish/book/Analog_Circuit_Design_Art_Science_and_Personalities_0750696400.pdf

fascinating article,  Jim Williams is always insightful. I think this is the first thing i've read of his that was not an application note.  i really like how he walks you though the process of what he did and why.

[innkeeper]

Use a decent computer soundcurt plus a suitable software. Easy, done.

You can then easily change test frequency, and even do test like a two tone (IMD).

can you give me an example setup with its thd+n figures?

[Gyro]

...
fascinating article,  Jim Williams is always insightful. I think this is the first thing i've read of his that was not an application note.  i really like how he walks you though the process of what he did and why.

There's another, equally good book, also edited by Jim Williams, in that EDN series called 'The Art and Science of Analog Circuit Design'. I have paperback copies of them both and find them nice to settle down with on the occasional rainy afternoon.

[MasterT]

 Wien-SG_Elektor-1994-12.pdf (131.28 kB - downloaded 276 times.)
I've build one recently, to test my 18-bits differential AD7984 board. I changed OPA to LM4562, and use 7370-18V-0.04mA lamp. Circuits does need "manual" tuning to get the lowest THD, but accounting simplicity it's not an obstacle.  Results thd-2 ~0.0004% (-110dB) and thd-3 ~0.0002 (-114dB) at 4kHz. Not sure about 100kHz, ADC itself has thd-3 ~0.002% at this frequency.
 
My attempts to replace lamp with CdS photocell or thermistor MF52 in the amplitude control loop were really disappointing. Despite  what they say in app note, CdS  is highly non-linear material, and may works only in the <100 mV range, making control loop very limited in the dynamic range sense. 

[Fungus]

You might get better than the quoted -124dB with some careful component value matching anyway...

i've actually ordered one of these .. figured its a start and maybe i can tweak it.

Ok, so -124dB is only "a start"? Got it.   

[innkeeper]

(Attachment Link)
I've build one recently, to test my 18-bits differential AD7984 board. I changed OPA to LM4562, and use 7370-18V-0.04mA lamp. Circuits does need "manual" tuning to get the lowest THD, but accounting simplicity it's not an obstacle.  Results thd-2 ~0.0004% (-110dB) and thd-3 ~0.0002 (-114dB) at 4kHz. Not sure about 100kHz, ADC itself has thd-3 ~0.002% at this frequency.

that is amazing for such a simple circuit, i;ll have to throw that together and put it on the analyzer and check it out. thanks

[innkeeper]

i've hit up victor from the diyaudio forum to see if he has any of his boards available. he is getting better then -150 Db.

[MasterT]

i've hit up victor from the diyaudio forum to see if he has any of his boards available. he is getting better then -150 Db.

What circuits? I've seen link to e-b https://www.ebay.com/itm/Low-Distortion-Audio-Range-Oscillator-1KHz-Sine-Wave-Signal-Generator-PCB-Board/262844562138,
 let's call it JFET-Wein, description says:

Low distortion coefficient of the sinusoidal oscillator circuit board, oscillation frequency: 1KHz, low distortion, distortion factor 0.0064% (- 124dB)

doesn't look right, in my math -120dB is 0.0001% ( or 1ppm), 0.0064% is about -84-85 dB, worse than cheapest sound card has. And NE5532 ? Common, be serious.

Or is it CdS-Wein? http://www.redcircuits.com/Page82.htm
Probably, just simplified version of
https://www.analog.com/en/design-center/evaluation-hardware-and-software/evaluation-boards-kits/dc1858a.html#eb-overview
I trust AD, performance is outstanding, but same with a price and complexity.
Though -120 only:
Table 1. Amplitude and Offset Options for DC1858A (SNR and THD Taken at f OSC = 2kHz)
Range (V) R22 (kΩ) R18 (kΩ) C11 (μF) Sine Out (V P-P ) Offset (V) SNR (dB)* THD (dB)*
0 to 2.5 14 36.5 0.47 2.25 1.25 100 –119
0.5 to 4.5 22.6 12.1 0.47 3.6 2.5 103 –121
0 to 4.096 23.2 17.4 0.47 3.7 2.048 103 –121
0 to 5 28 12.1 0.47 4.5 2.5 104 –118
±2.5 28 ∞ 0.47 4.5 0 104 –119
±5 56.2 ∞ 0.47 9 0 107 –117
±10 118 ∞ 0.1 18 0 100 –107

page 3  https://www.analog.com/media/en/dsp-documentation/evaluation-kit-manuals/dc1858af.pdf

And most resistors precise  0.1% tolerance, they can't get -120dB with CdS even for single Tone generator using common 1% parts, and adjustable frequency doesn't go along with precision parts at all.

[innkeeper]

@masterT

victor sells is an updated version of this:
https://www.diyaudio.com/forums/equipment-and-tools/205304-low-distortion-audio-range-oscillator-21.html#post3116055

update: i think this is the latest schematic
https://www.diyaudio.com/forums/equipment-and-tools/205304-low-distortion-audio-range-oscillator-743.html#post5873578

i am sure there is a more up to date schematic in that thread, but that thread is over 8000 messages long
here is the web page he put up just wiht the specs
https://viccc42.wixsite.com/uld-audio