I've got a Wavetek 188 function generator that is versatile and does a nice job but it does have some distortion that I'd like to get rid of. I've got a good service manual and I've gone through the calibration steps but I've still got quite a bit of distortion as you can you see in the attached screenshot. Any thoughts on where I should focus my energies? Many thanks.
Attachments
Function generators are not generally designed with low distortion as a primary goal, if you really need a low distortion source look around for an old Leader or HP sine wave generator.
Thank you guys. Kevin, I'll start looking for a new one.
Jack, what is the squarer?
Sent from my iPhone using Tapatalk
Jack, what is the squarer?
Sent from my iPhone using Tapatalk
Sound card in your computer if capable of doing good measurements may be capable of good performance as a source too. I recommend using an external attenuator with the sound card and running it about 6 - 10dB below FS output or wherever best linearity is achieved.
Kevin, I did that this weekend (everything but the outboard attenuator) when I was adjusting my SH-9010 equalizer. It worked okay but I couldn't get rid of a 60hz ground loop between the soundcard, test leads I was using to measure the SH-9010 output levels and my Fluke 8502a. I need to get my isolation transformer hooked up and figure out where my ground loops are coming from.
I didn't have the ground loops when I used the Wavetek 188 but when I used the soundcard to generate test tones and measure the DUT, hum cropped up. Maybe I'll try two soundcard--one to generate and one to measure. I have a Presonus Audiobox coming in the mail tomorrow.
I didn't have the ground loops when I used the Wavetek 188 but when I used the soundcard to generate test tones and measure the DUT, hum cropped up. Maybe I'll try two soundcard--one to generate and one to measure. I have a Presonus Audiobox coming in the mail tomorrow.
See the diagram towards the end of the manual:
http://sites.fas.harvard.edu/~phys191r/Bench_Notes/A9/Wavetek188.pdf
The 188 uses a "sine shaper" probably a 2nd order low pass filter for the sine wave.
The square wave is generated by funnelling a triangle wave to a comparator which "snaps" when the triangle output hits a certain reference voltage -- but this technique creates a flood of harmonics.
If you wanted to use it as a fixed frequency generator for THD% you could put a tuneable switched cap filter on the output, but the time and expense are not worth the effort. For the record, Audio Amateur had such a design back in the late 1990's.
Hi Kevin,
I'm looking at extending your solution but could use some advice. I've created high quality test tones at wavetones.com I'm outputting those files using my Raspberry Pi and an HRT Microstreamer DAC, which has a great S/N ratio. But, the Microstreamer has a max variable output of 1.4V or a max fixed output of 2.25V. Are there any reasonable priced low distortion headphone amps that can output between 5 and 10V?
The Wavetek, like most function generators, uses a diode network for sine shaping. There may be some trimmers for tweaking it but don't expect better than about .1% or so ever. No RC's since it has to be constant with frequency.
No normal headphone amp will generate 10V. Most headphones are borderline dangerous with 1 volt. You need to buy or make something. Maybe an old preamp would work for you.
No normal headphone amp will generate 10V. Most headphones are borderline dangerous with 1 volt. You need to buy or make something. Maybe an old preamp would work for you.
I won an auction for a leader LAG-120B. Hopefully that will fill my needs.
I'm still building a few headphone amps to explore topologies and see which sounds best. [emoji6]
Sent from my iPhone using Tapatalk
I'm still building a few headphone amps to explore topologies and see which sounds best. [emoji6]
Sent from my iPhone using Tapatalk
If you increase the voltage on the output transistors of Richard Marsh's headphone amplifier it will get you there -- his original schematic in LinearAudio suggested 12V, but I've run it hotter than that.
You could copy the design of the output amplifier of the Boonton 1120/1121
Pretty rare that you will need over 2.83V, however.
Maybe you gentlemen can answer this for me.
There are two traditional methods of creating sine waves right?
1. Triangle wave - then massaging to get into sine shape.
2. Oscillator - already a sine by the act of oscillating, then it is controlling
the oscillations.
From what I understand 2. gets us lower distortion numbers out of the box yes?
And, 1. is useful for various shapes, triangle, mountain, square, sine and change
in duty cycle.
There are two traditional methods of creating sine waves right?
1. Triangle wave - then massaging to get into sine shape.
2. Oscillator - already a sine by the act of oscillating, then it is controlling
the oscillations.
From what I understand 2. gets us lower distortion numbers out of the box yes?
And, 1. is useful for various shapes, triangle, mountain, square, sine and change
in duty cycle.
Its not that simple. Function generators are for multiple waveforms and voltage controlled frequency (and sometimes other elements). They are usually controlled by charging and discharging caps with current sources and comparators for switching. Triagle waves come for free. A good design can get .1% distortion across the entire frequency range (30 MHz to 1 cycle per day or so for the KH2200 on my bench). RC oscillators are not well suited to these tasks. Each has its optimum use. A good ARB with enough bits can almost do both but has other issues to work through.
- Status
- Not open for further replies.