Yvan, thank you for your advice... I've buy your COSMOS ADC to measure everything I can put my hand on ! And since my janascard is set to 440Hz, the APU notches won't fit... but I've buy the janascard with its notch too...
But yes, APU and the load-board will probably be good toys for my "lab"...
Keep your really good work Yvan, and thank you again.
Strange, your results look not that good... With 64k FFT, I can not see any harmonic down to the noise floor at 140dB (janascard, not the Viktor osc).
I think it is not about it but Cosmos ADC input impedance may be slightly too heavy for the Victor's osc output at 2.7Vrms.
Yes, I just DIY based on the circuit diagram, and there are still many details that are not done well enough.
I mainly use it for test amps and the performance is good enough for me now.
Victor's design is awesome!
In addition, IVX's ultra-low distortion DAC with LPF is temporarily unavailable in China. If there is one for sale, I should buy one.
I mainly use it for test amps and the performance is good enough for me now.
Victor's design is awesome!
In addition, IVX's ultra-low distortion DAC with LPF is temporarily unavailable in China. If there is one for sale, I should buy one.
hey, make it by yourself, a decent DAC, I believe, you have. The LPF is a lot simpler than an analog osc, you don't need any kind of adjustments
I first used a passive Hall notch filter with a notch depth of 40dB. Then use the OPA1656 to amplify by 20dB.
The above two pictures were not tested on the same day, so the OSC output levels of these two pictures are not consistent. I may have used a higher OSC output level when using the notch filter.
495 pages and ten years later, we are here.
Let me re-state the second post: A Wien bridge oscillator is the way to go. Suggested by a member who's opinion I share. ( I own an old HP 204D oscillator and it still performs well after 40 years)
Has the original post author tried out a solution that yielded satisfactory results?
Let me re-state the second post: A Wien bridge oscillator is the way to go. Suggested by a member who's opinion I share. ( I own an old HP 204D oscillator and it still performs well after 40 years)
Has the original post author tried out a solution that yielded satisfactory results?
zergxia I think you'll never see H2 with 1656 <-140-145db if you'll not match the opamps inputs impedance. You can take RT6863S, a very cheap Chinese opamp(1/3 of opa1612 cost) with very high symmetry of the input stage, and in that case, you'll see <150-160db H2. Also, you can try to select from 5-7pcs of OPA1612 the best one, and you'll see <-150db. I've no idea why but 1612 sample2sample is not consistent regarding H2 performance with unmatched input impedances.
I've got Ok result with the posted "Victor's osc" schematic, but yes, I spent a whole day adjusting 1000.2Hz and tried 10 opamps)) 1kHz H2 -150db(after I add some trimpot to adjust the symmetry), H3 -155db. However, the osc is way too noisy to see even -150db without multiple AVG, hence, as a ref sine source it is a very slow instrument. I needed to calibrate 1000pcs of ES9822Pro ADCs and Wien bridge oscillator with so high RC impedance appeared to be unacceptable noisy. Currently, I using DAC+LPF way to get clean as possible ref sine + digitally generated ani-harmonics bouquet to get H2, H3 at -160-170db.
Thank you IVX, (I am wondering if the original post author has changed his username??)
From what I understand, and I'd have to check the specifications at work tomorrow (we have a gigantic instrumentation document library), the harmonics measure well below what a typical sinewave generator such as the 204, even 333 measure.
I am amazed that digital technology (along with analog LPF, please please tell me they're analog!) manages to kick *** on vintage technology (in regards to harmonics).
I didn't even know it was possible to measure LF spectrum to -160dBc.
Well done, digital.
But we shall meet yet again on the battlefield.
From what I understand, and I'd have to check the specifications at work tomorrow (we have a gigantic instrumentation document library), the harmonics measure well below what a typical sinewave generator such as the 204, even 333 measure.
I am amazed that digital technology (along with analog LPF, please please tell me they're analog!) manages to kick *** on vintage technology (in regards to harmonics).
I didn't even know it was possible to measure LF spectrum to -160dBc.
Well done, digital.
But we shall meet yet again on the battlefield.
Huh, I'm not sure if I'll sell Cosmos DAC just because the commercial interest is not so hot. The same thing with Cosmos APU, I prepared 1000pcs and I hopped it could be popular in the DIY area as an inexpensive PCBA. Not really, almost zero selling of APU PCBA, it seems everyone DIYer already has such a simple device as a Notch filter with <-150db harmonics resolution and THD+N -132db
So, I'm not sure if I'll go to produce Cosmos DAC to sell it 50pcs/year, better if you make it by yourself. There are no secrets, and the idea is trivial wiping out all harmonics and noise >1kHz.
gain wire, you can measure 1kHz sine <<-160db harmonics if you go higher with the output voltage. AP SYS2 uses 2Vrms, IMO, it was a stupid choice even for 1980. 10Vrms that's the real deal. LPF after a DAC of course analog one, and low as possible impedance if you don't need to get the same noisy sine as Victor's osc is.
The performance of the Cosmos ADC is strong enough that only some people will need to use a performance test where thd+n is better than -120dB, so the sales rate of the APU should be significantly slower than that of the ADC.
I think DAC sales should be higher than APU and lower than ADC. If you plan to sell DAC, please remember to list the item on taobao in China.
I think DAC sales should be higher than APU and lower than ADC. If you plan to sell DAC, please remember to list the item on taobao in China.
Here's the thing, at least for me.Huh, I'm not sure if I'll sell Cosmos DAC just because the commercial interest is not so hot. The same thing with Cosmos APU, I prepared 1000pcs and I hopped it could be popular in the DIY area as an inexpensive PCBA. Not really, almost zero selling of APU PCBA, it seems everyone DIYer already has such a simple device as a Notch filter with <-150db harmonics resolution and THD+N -132db
Buying a device like a Cosmos APU (which I did, BTW) built by someone else is like having somebody check your work. You even have referenced double checking performance of your products with an AP system, just as a reference. Same thing. Plus, more time for me to work on other parts of the audio system since I didn't have to develop an ADC or APU myself.
If I wanted to buy an AP system - as well as the software contract - I would, and then I wouldn't even need a Cosmos ADC or APU! The AP isn't perfect, but it certainly is good enough.
But, for us hobbyists, no matter how serious we might be, an AP is probably not a great investment. I once estimated that using an AP would cost way more than $100 per hour based on how much I'd use it as a hobbyist. Personally, I'm willing to spend real money to support products from a reputable company that do the job, so I'm not trying to be cheap. Just sensible. Like not buying a Ferrari for going to the grocery store. If I built audio gear as a business, it would be a different story.
However, I certainly get why selling 50 units a year is not worth your energy. Maybe selling the post DAC low pass filters would be a good substitute. Or, even, just offering the pcb layout you used as a download. Then it could become a de facto standard for everyone to reference like the Victor oscillator has been for so many years.