Cable reflections can be removed by using 50R termination, and injecting using a equal value splitter, which has 3 arms, each made from 2 parallel 33R resistors, and then output using a 50R termination on the cable by the SA, and the others looped through with only a very short cable to connect them.
My advice to you is get a hp 8903, its digital and self notching and will measure down to about -100 db. That's what I use. Incidentally I went on a long quest to build a low distortion 1 kHz sine wave generator. I eventually achieved something that's better than any of my other generators. Somewhere around 0.001 % distortion. In fact it's so low it was almost at the limit of my measurement equipment
My only comment is that the fractional division method of direct digital synthesis employed by these styles of signal generators yields a range of distortion values as one passes through particularly obnoxious divisor ratios. Modern high speed/bit-depth DAC designs that are commonly available now have greatly minimized that, but it will definitely still show up in a swept measurement.
You can do harmonic distortion measurements with a spectrum analyzer by adding up the dBc values of the harmonic peaks. I have a video on my channel demonstrating this with a couple different analyzers.
Its possible that one, both or neither of those distortion meters are average responding RMS calibrated, rather than RMS responding and calibrated. This may account for differences, when looking at the noise or distortion products left over.
I am not sure any modern instrument that can generate an arbitrary waveform combines harmonics to build waveforms these days. They usually use a high speed DAC and an fpga or a vendor asic.
That applies specifically to digital generators. There are still top tier analog signal generators made. These sometimes do use harmonics to get to the desired frequency range, using harmonic mixers, multipliers, etc. (And some of the digital generators still use harmonics for higher frequencies.) Usually digital ARB generators have a limited bandwidth for their modulation, which the ASICS and FPGAs generate, but that arb signal is then mixed with a carrier from an analog chain to get the desired output.
@@ruhnet Interesting. I'll have to see if Shahriar has a deep drive. I'd assume at the higher end they would still use DDS but to generate a quadrature pair. This can construct pretty much arbitrary modulations for sensible signals. I'd assume the filtering would be the hard part, specially over a wide range of frequencies. I noticed in a recent tear down he did, the filtering was done in all sorts of unexpected place.
@@originalmianos the ARB part is almost always done as you said these days, but generally at a baseband frequency like 10MHz or 50MHz or the like (depending on the desired waveform) and then it is mixed with a carrier to hit the desired output frequency which may be in the GHz range. There are very few FPGAs and digital devices that can directly synthesize an arbitrary signal above a few hundred MHz, and those that can are usually crazy expensive.
@@ruhnet Ah, yes, I would assume anything high would be mixed up. That's not really a 'combining harmonics' thing to me. I was recently looking at high speed DACs. It's much much harder than I thought if you want to have low noise.
Great video! I’m really enjoying your latest video series. 👍🏻
Cable reflections can be removed by using 50R termination, and injecting using a equal value splitter, which has 3 arms, each made from 2 parallel 33R resistors, and then output using a 50R termination on the cable by the SA, and the others looped through with only a very short cable to connect them.
My advice to you is get a hp 8903, its digital and self notching and will measure down to about -100 db. That's what I use. Incidentally I went on a long quest to build a low distortion 1 kHz sine wave generator. I eventually achieved something that's better than any of my other generators. Somewhere around 0.001 % distortion. In fact it's so low it was almost at the limit of my measurement equipment
My only comment is that the fractional division method of direct digital synthesis employed by these styles of signal generators yields a range of distortion values as one passes through particularly obnoxious divisor ratios. Modern high speed/bit-depth DAC designs that are commonly available now have greatly minimized that, but it will definitely still show up in a swept measurement.
It will be good to get an update on your Bitcoin Miner production, now since halving.
You can do harmonic distortion measurements with a spectrum analyzer by adding up the dBc values of the harmonic peaks. I have a video on my channel demonstrating this with a couple different analyzers.
Its possible that one, both or neither of those distortion meters are average responding RMS calibrated, rather than RMS responding and calibrated. This may account for differences, when looking at the noise or distortion products left over.
I am not sure any modern instrument that can generate an arbitrary waveform combines harmonics to build waveforms these days.
They usually use a high speed DAC and an fpga or a vendor asic.
That applies specifically to digital generators. There are still top tier analog signal generators made. These sometimes do use harmonics to get to the desired frequency range, using harmonic mixers, multipliers, etc. (And some of the digital generators still use harmonics for higher frequencies.) Usually digital ARB generators have a limited bandwidth for their modulation, which the ASICS and FPGAs generate, but that arb signal is then mixed with a carrier from an analog chain to get the desired output.
@@ruhnet Interesting. I'll have to see if Shahriar has a deep drive. I'd assume at the higher end they would still use DDS but to generate a quadrature pair. This can construct pretty much arbitrary modulations for sensible signals. I'd assume the filtering would be the hard part, specially over a wide range of frequencies. I noticed in a recent tear down he did, the filtering was done in all sorts of unexpected place.
@@originalmianos the ARB part is almost always done as you said these days, but generally at a baseband frequency like 10MHz or 50MHz or the like (depending on the desired waveform) and then it is mixed with a carrier to hit the desired output frequency which may be in the GHz range. There are very few FPGAs and digital devices that can directly synthesize an arbitrary signal above a few hundred MHz, and those that can are usually crazy expensive.
@@ruhnet Ah, yes, I would assume anything high would be mixed up. That's not really a 'combining harmonics' thing to me. I was recently looking at high speed DACs. It's much much harder than I thought if you want to have low noise.
@@originalmianos True---upconverting isn't exactly the same thing as that I guess. 😀
First!