The cheapest way to get those signals is to generate them using audio editing software (freeware), and burn a CD. Play the CD in repeat mode on even just a mid-fi CD/DVD player, and it will be one of the lowest distortion, most stable, most accurate signal sources ever.
Another way is just to use a computer sound card and the same software.
Most modern onboard computer sound cards will at least do 24/96 that is clean up to at least 45 KHz. Same fire drill - generate the test tone and play it in repeat mode.
If you need more voltage than what the CD player or sound card can provide, get a cheap audio mixer like those by Behringer etc. They will go up to about 10V RMS. They will add a little distortion and some modest FR losses, but you said you are not all that critical.
Real close.
Just use one channel and split it.
Better than 0.01% for a cheap audio interface, and better than 0.001% if you spend a little money.
Nothing. That's what the brick wall low pass filter is supposed to make go away and they are generally very good at doing so.
By the way, is there any chance to see the 1 kHz version of this one available again :
http://www.ebay.com/itm/Ultra-low-d...erator-assembled-and-tested-PCB-/321579762273
http://www.diyaudio.com/forums/equi...on-audio-range-oscillator-11.html#post3116055
?
http://www.ebay.com/itm/Ultra-low-d...erator-assembled-and-tested-PCB-/321579762273
http://www.diyaudio.com/forums/equi...on-audio-range-oscillator-11.html#post3116055
?
Hi,
I am still producing small quantities of these boards. Some 1kHz boards would be available approximately on Tuesday next week. For the reservation you can write me PM.
Victor.
That would be great as I'm developing ADC / ClassD circuits and I keep encountering the limits of the Analogue Sig gens of the 2722 & UPD.
The UPD has lover THD then the Sys2722 with harmonics hovering about -130dB, where as my 2722 they are about -125dB...
I've not read over this massive 416 page thread, is your generator variable or single frequency?
I guess I could purchase 2 boards for IMD testing, also, will it support balanced outputs?
The UPD has lover THD then the Sys2722 with harmonics hovering about -130dB, where as my 2722 they are about -125dB...
I've not read over this massive 416 page thread, is your generator variable or single frequency?
I guess I could purchase 2 boards for IMD testing, also, will it support balanced outputs?
Last edited:
Don't worry, complexity does not scare me, I'm currently building and testing a PCB with over 2000 SMD components...
https://dl.dropboxusercontent.com/u/86116171/IMG_7261.JPG The mass of components rats nested above the board is just to allow Bode & current measurements etc. So far I've only discovered x1 design error on a CCS, the problem is I duplicated the CCS x24 across the pcb!!!
I'm lost about you comment on IMD, I need a clean signal so presumed I need to sum (passively sum?) x2 LDO ?
I'm sure you have much experience - but at such low levels of THD, watch the Voltage Coeff of cheaper SMD resistors - I've found the thin film MELF's from Vishay to be very good. Standard thick film 0805 / 0603 have very poor Voltage Coeff.
https://dl.dropboxusercontent.com/u/86116171/IMG_7261.JPG The mass of components rats nested above the board is just to allow Bode & current measurements etc. So far I've only discovered x1 design error on a CCS, the problem is I duplicated the CCS x24 across the pcb!!!
I'm lost about you comment on IMD, I need a clean signal so presumed I need to sum (passively sum?) x2 LDO ?
I'm sure you have much experience - but at such low levels of THD, watch the Voltage Coeff of cheaper SMD resistors - I've found the thin film MELF's from Vishay to be very good. Standard thick film 0805 / 0603 have very poor Voltage Coeff.
Last edited:
That's about the complexity of my oscillator (~1800 parts). I've made sure that all parts are reasonably easy to source and solder.
IMD can only appear when both frequencies are present. If you sum, say, 20 kHz and 21 kHz, you get the harmonics (40k, 42k, 60k, 63k etc.) from the individual sources, but no significant IMD if the summing stage is sufficiently good. The harmonics don't influence a standard IMD measurement significantly (only the frequency range below, say, 30 kHz is looked at), so the individual sources don't need to be particularly good WRT to THD.
Very true, I'm using multiple series/parallel thin film MiniMELFs in the signal path.
Samuel
Sam,
My bad, I'm not so much concerned about measuring IMD (as in 2nd/3rd harmonics) but the effects the multi-tones might induce other spurious components across the ADC / Class D modulators noise floor - which is why I need spectrally (not just harmonically) "Clean" sources.
I'll experiment with the summation stage....
My bad, I'm not so much concerned about measuring IMD (as in 2nd/3rd harmonics) but the effects the multi-tones might induce other spurious components across the ADC / Class D modulators noise floor - which is why I need spectrally (not just harmonically) "Clean" sources.
I'll experiment with the summation stage....
I recently brought a Krohn Hite 4402B but its really spectrally noisy, worst then my UPD:-
https://dl.dropboxusercontent.com/u/86116171/Krohn Hite 4402B 1KHz 2V RMS -123dB THD.jpg
Heres the output of UPD 1KHz @ 2Vrms 50R load, 2nd harmonic is down at -130dB (in fact the FFT plot its the output of a ClassA stage I designed for a new DAC - but its just showing the limits of the UPD's Analogue generator).
https://dl.dropboxusercontent.com/u/86116171/MDAC2 2Vrms 1KHz 50Ohm OPS only FFT.jpg
This -130dB limit is why I'm interested in your Sig Gen
https://dl.dropboxusercontent.com/u/86116171/Krohn Hite 4402B 1KHz 2V RMS -123dB THD.jpg
Heres the output of UPD 1KHz @ 2Vrms 50R load, 2nd harmonic is down at -130dB (in fact the FFT plot its the output of a ClassA stage I designed for a new DAC - but its just showing the limits of the UPD's Analogue generator).
https://dl.dropboxusercontent.com/u/86116171/MDAC2 2Vrms 1KHz 50Ohm OPS only FFT.jpg
This -130dB limit is why I'm interested in your Sig Gen
Sam,
The limitations with the UPD internal Analogue Sig Gen - the analyzer section has much lower THD, well past -145dB (it has the option of a -36dB tracking Notch filter).
Heres an FFT of a 0dB output (120w) of the TAS5015 ClassD digital input modulator (You can clearly see the poles of its Noiseshaper design) - but note the THD harmonics are buried below the -145dB FFT noise floor.
https://dl.dropboxusercontent.com/u/86116171/120W 8R 1kHz.jpg
I had results of better modulator designs but lost the data due to a HDD crash, but you can see the analyser section of the UPD is capable of very low THD.
The limitations with the UPD internal Analogue Sig Gen - the analyzer section has much lower THD, well past -145dB (it has the option of a -36dB tracking Notch filter).
Heres an FFT of a 0dB output (120w) of the TAS5015 ClassD digital input modulator (You can clearly see the poles of its Noiseshaper design) - but note the THD harmonics are buried below the -145dB FFT noise floor.
https://dl.dropboxusercontent.com/u/86116171/120W 8R 1kHz.jpg
I had results of better modulator designs but lost the data due to a HDD crash, but you can see the analyser section of the UPD is capable of very low THD.
Last edited: