Well, I see that ES9038 costs $125 CAD right now, and if I could design device that does the same functionality in the audio range < 20 kHz (there is nothing wrong if it goes higher in frequency, isn't it?) and cost < 10 CAD , than I don't see any issue here.
Tween-T filter more close to the topic, since I could break stability barrier some day and add up some LDR to keep magnitude under control, ditching DDS and turning it into autonomous pure analog domain.
Tween-T filter more close to the topic, since I could break stability barrier some day and add up some LDR to keep magnitude under control, ditching DDS and turning it into autonomous pure analog domain.
Mouser prices are more around 20USD in low quantities and the package is somewhat of a pain to solder by hand. But apart from that I agree, using a very good DAC and adding filters as needed just gives a lot more flexibility for very reasonable overall cost.
Pain is BGA, but the package of ES9038Q2M is not a pain, it is very easy to solder it manually.
You need just a hot air soldering gun (which is absolute "must to have" now).
Much higher problem is low pass filter, if is is not for the fixed frequency.
I tried to make electronically tunable LPF, and had some progress, but stopped this project due to the some reasons.
ES9038Q2M's price is less then half of this LPF BOM price, but this was not an issue.
Other big problem is that DAC+LPF cannot help in IMD measurements, especially multi tone.
Last edited:
2 years ago it was $4, so it is not an expensive DAC but actually, almost any DAC is ok for that Low-distortion Audio-range Oscillator.
I tried Victor's osc as well(but not on his original PCB yet), It's a quite noisy osc(due to high RC impedance) without the ability to use in the sweep, and each one requires trimming(I spent 30min at least). I did see there 3rd harmonic at -160db after superslow multiple AVG and anyway, this level floats up to -150db in time.
DAC+LPF a lot simpler to me with unbeatable THD and THD+N, I believe, for everyone DIYer who can make Victor's osc will be able to make DAC+LPF as well.
I tried Victor's osc as well(but not on his original PCB yet), It's a quite noisy osc(due to high RC impedance) without the ability to use in the sweep, and each one requires trimming(I spent 30min at least). I did see there 3rd harmonic at -160db after superslow multiple AVG and anyway, this level floats up to -150db in time.
DAC+LPF a lot simpler to me with unbeatable THD and THD+N, I believe, for everyone DIYer who can make Victor's osc will be able to make DAC+LPF as well.
I suppose there are different takes on what is "very easy to solder manually" or what is a "must have" tool. For me personally the package is not really a challenge, but I don't view myself as an ideal representation of the average DIY-er. For many it is still easier to buy a product off the shelf and add filters after it. Not too long ago my best signal source was a Khadas Toneboard followed by filters straight out of TI's filter design tool. Now I have some better signal sources (including one of Victor's very fine oscillators) but my previous setup wasn't too bad either.
when you gonna solder QFN need to make sure that pads have no oxide and if you are not sure about that better to presolder these pads.
New QFN soldering is very easy and fast, I drop the IC even not perfectly positioned and under the hot-air move the IC to fit into the footprint.
This way I replace ES9038Q2M for 1min.
New QFN soldering is very easy and fast, I drop the IC even not perfectly positioned and under the hot-air move the IC to fit into the footprint.
This way I replace ES9038Q2M for 1min.
I prepared some active notch(twin-T + OPA1612 composite)+LNA(THAT1510) unit called Cosmos APU(about $70 as a PCBA, $100 calibrated-assembled). The notch ratio is -30db@Q about 50 so 2nd and 3rd harmonics stay at 0db(2nd at -.5-.6db, 3rd -.2-.3db), each unit is adjusted to that condition and a simple calibration file makes FFT looks normally, like without a notch.
The sine source is dual ES9038Q2M +LPF, as I mentioned before, IMO, it is the best approach to get an extremely clean sine signal, you can see the result in red from the passive twin-T(2nd -50db, 3rd -54db).
The sine source is dual ES9038Q2M +LPF, as I mentioned before, IMO, it is the best approach to get an extremely clean sine signal, you can see the result in red from the passive twin-T(2nd -50db, 3rd -54db).
What I meant by "tuning" was the THD compensation of the 2nd and 3rd in the ADC.
I assume that the measurement shown in post # 9.752 was done without a notch filter, but with a DAC -> low pass filter -> ADC.
I assume that the measurement shown in post # 9.752 was done without a notch filter, but with a DAC -> low pass filter -> ADC.
I can but with caps switching by relays because no one tech can do that with such precision, only the best passive parts. All MDACs or variable resistors have way higher distortions and noise.