Agree on that, so what is inside yours?
My noise floor is something like 10 nV for discrete frequencies, so a 1uV RMS sinewave will make a 40 dB pin above the noise floor of the instrument chain.
I'm just using a ymec software spectrum analyzer, and ESI Juli@ soundcard, so anyone can make the same analyser with relatively little cash. A Christer circuit, (maybe with a few minor modifications), will do fine as front end.
That's why I would like to get absolute figures for the two references you tested, not just a dissertation about where 1/f catches on with the white noise.Lars Clausen said:
Keep in mind the relationship between FFT length, sampling rate, FFT window on the one hand and the noise readout on the graph on the other hand. The difference between the integrated noise over 20kHz and where the plot bottoms out depends on that. This is why noise plots should be calibrated in volts per square root of bandwidth, not just in volts. The exact nature of the analyser is not very important, what matters is getting the calibration right and for noise-like signals this requires quite a bit of care. If the manufacturer of the analyser knows their stuff, the manual will indicate the "effective bandwidth" of an FFT bin as a function of FFT window and FFT length.
Lars Clausen said:
Agree on that, so what is inside yours?
My noise floor is something like 10 nV for discrete frequencies, so a 1uV RMS sinewave will make a 40 dB pin above the noise floor of the instrument chain.
I'm just using a ymec software spectrum analyzer, and ESI Juli@ soundcard, so anyone can make the same analyser with relatively little cash. A Christer circuit, (maybe with a few minor modifications), will do fine as front end.
Lars,
I suggest you start using a low noise pre-amplifier, this will complement your equipment and bring the noise floor down, required to do your measurements.
We (Tentlabs, Grimmaudio), use one based on a good old AD797
best
I wonder if this discussion has gone to far... What are you discussing? Audio or tech related issues??
I know some of Lars' older clock's, from the time where he was at LC Audio. The LClock's.
I have experienced good results with these clocks, but for never ever really believed in the necessity of improving on these devices.
I remember LC did some calculations on how much jitter could be accepted in a high resolution audio DAC (Don't remember the figures
but Lars might). As I remember, the jitter of the LClock C2 was lower than what you may require. Around 10-12ps if I'm not wrong. And actually I did not observe any improvement using newer versions. I have them, but in a box in my basement.
My guess is that there are other places, where improvements will give you way better results. The analog stage would be a good place to start. Why keep improving the clock using hundreds of $$, and then let some crappy 20 cent op-amp destroy it all??
I know some of Lars' older clock's, from the time where he was at LC Audio. The LClock's.
I have experienced good results with these clocks, but for never ever really believed in the necessity of improving on these devices.
I remember LC did some calculations on how much jitter could be accepted in a high resolution audio DAC (Don't remember the figures
but Lars might). As I remember, the jitter of the LClock C2 was lower than what you may require. Around 10-12ps if I'm not wrong. And actually I did not observe any improvement using newer versions. I have them, but in a box in my basement.My guess is that there are other places, where improvements will give you way better results. The analog stage would be a good place to start. Why keep improving the clock using hundreds of $$, and then let some crappy 20 cent op-amp destroy it all??
The common analysis is that jitter is no longer an issue when the error remains below 1lsb at 20kHz input. This presumes that digital audio is incapable of coding signal content below 1lsb. Anyone familiar with the concept of dithering will readily see that the audio content doesn't disappear below 1lsb, and neither do jitter modulation products. So as popular as it is, the "LSB" jitter calculation is fallacious.Hurtig said:I remember LC did some calculations on how much jitter could be accepted in a high resolution audio DAC (Don't remember the figures
Hurtig,
>>The analog stage would be a good place to start. Why keep improving the clock using hundreds of $$, and then let some crappy 20 cent op-amp destroy it all??<<
Some of us threw out all the opamps in their players long ago. Vacuum State has done some 500 opamp-less Upgrades across 8 years now worldwide, and I'm sure Lars/LC Audio have also done a lot with their ZAP module.
Now we are focussing on a really superior clock, the TerraFirma ÜberClock. It's in production and user reports are nothing short of amazing. It's based on a theory that I'm not willing to share here, but as soon as we get adequate instrumentation to verify what we calculated (and hear)we will post graphs and other specifications.
The users reports can be found on our website.
Regards, Allen (Vacuum State)
>>The analog stage would be a good place to start. Why keep improving the clock using hundreds of $$, and then let some crappy 20 cent op-amp destroy it all??<<
Some of us threw out all the opamps in their players long ago. Vacuum State has done some 500 opamp-less Upgrades across 8 years now worldwide, and I'm sure Lars/LC Audio have also done a lot with their ZAP module.
Now we are focussing on a really superior clock, the TerraFirma ÜberClock. It's in production and user reports are nothing short of amazing. It's based on a theory that I'm not willing to share here, but as soon as we get adequate instrumentation to verify what we calculated (and hear)we will post graphs and other specifications.
The users reports can be found on our website.
Regards, Allen (Vacuum State)
Vacum State.... I really don't see why
Consider a well designed analog stage using BJT's, and you will be more than surpriced. Just to bad, that this kind of analog stage is rare.
I know the ZAPFilter from LC Audio. I really think that LC has done a lot of great power amps over the years.But the ZAPFilter is not at all in the same class.
It is single ended, and sounds just like that.
I would thinkthat changing the output to PP, would move the ZAPFilter towards the performance of the power amplifiers.
Consider a well designed analog stage using BJT's, and you will be more than surpriced. Just to bad, that this kind of analog stage is rare.
I know the ZAPFilter from LC Audio. I really think that LC has done a lot of great power amps over the years.But the ZAPFilter is not at all in the same class.
It is single ended, and sounds just like that.
I would thinkthat changing the output to PP, would move the ZAPFilter towards the performance of the power amplifiers.
Just because the company name is Vacuum State, and that we mostly focus on tube equipment, doesn't mean we are no competent in solid state design.
Our discrete analog module is a mixture of jfets and bipolars, and can be wired either unbalanced or balanced.
Three bipolars in the actual signal path bypassing up to 11 opamps (SONY SCD-!) makes a huge difference!
Regards, Allen (Vacuum State)
Our discrete analog module is a mixture of jfets and bipolars, and can be wired either unbalanced or balanced.
Three bipolars in the actual signal path bypassing up to 11 opamps (SONY SCD-!) makes a huge difference!
Regards, Allen (Vacuum State)
I have been wondering a bit about these monster clocks! Why spend 200-500USD on a clock??
I have been looking at standard Xtal Oscillators like this: Xtal Oscillator
It has a jitterperformance of 1ps, and cost less than 3USD. What will the expensive "Audio" clocks do better??
Is it simply the well known Audio WooDoo business, that is always looking for new snake oils to make $$ on??
I have been looking at standard Xtal Oscillators like this: Xtal Oscillator
It has a jitterperformance of 1ps, and cost less than 3USD. What will the expensive "Audio" clocks do better??
Is it simply the well known Audio WooDoo business, that is always looking for new snake oils to make $$ on??
Hurtig said:
The oscillator you refer to specs jitter (1ps) without bandwidth. If you carefully compare the figures given further in the spec (phase noise as function of offset) you will notice it is only 33dB worse than mine.
Indeed, no Voodoo, nor snake oil, but plain figures. And at 29 euro my oscillator is just 10dB more expensive.
An externally hosted image should be here but it was not working when we last tested it.
have any one hear about the 16.9344mhz rubidium master clock ?
i just find one japan buy one and use the 74 Chip divide into 44.1khz...
Do you refer perhaps to this item?
11.2896Mhz Rubidium Master clock for CD made by FEI - eBay (item 170352897811 end time Nov-30-09 06:54:00 PST)
11.2896Mhz Rubidium Master clock for CD made by FEI - eBay (item 170352897811 end time Nov-30-09 06:54:00 PST)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.