Richard
The QA400 uses this CODEC: http://www.cirrus.com/en/pubs/proDatasheet/CS4272_F1.pdf If you look at the spec there are two numbers that would encompass the -60 dB info. The JIS spec for dynamic range is based on THD+N at -60 dB, listed at -111dB on the spec sheet. It also lists a -60 dB THD+N of -51 dB (which curiously adds up to -111 db). Most spec sheets will give these numbers. I believe they are weighted somehow.
As to accuracy of harmonic levels below -100 dB, you will find little to give you confidence soine at those levels everything is likely to have some distortion.The -140 dB harmonics from Viktor's oscillators may be the only way to separate distortion sources. Most analyzers will also be questionable at those levels.Even super tweaked you need to figure out validations to confirm the actual residulal of the HP339 or whatever you have in front of the digitizer. Even AP mentioned in some note somewhere adding a passive notch in front to improve the accuracy and resolution of their analyzers.
I think this is interesting but too narrow a pursuit at this level. I'm interested in IM and what can be learned there. Keith Johnson tells me that a tone cluster with a number of closely spaced tones has become his most sensitive indicator. I have one IM analyzer here I have started to play with that has the "Cordell" three tone source. it has too much internal source IM at this stage (.004%) to get to what I want it to be so I have some work ahead.
The QA400 uses this CODEC: http://www.cirrus.com/en/pubs/proDatasheet/CS4272_F1.pdf If you look at the spec there are two numbers that would encompass the -60 dB info. The JIS spec for dynamic range is based on THD+N at -60 dB, listed at -111dB on the spec sheet. It also lists a -60 dB THD+N of -51 dB (which curiously adds up to -111 db). Most spec sheets will give these numbers. I believe they are weighted somehow.
As to accuracy of harmonic levels below -100 dB, you will find little to give you confidence soine at those levels everything is likely to have some distortion.The -140 dB harmonics from Viktor's oscillators may be the only way to separate distortion sources. Most analyzers will also be questionable at those levels.Even super tweaked you need to figure out validations to confirm the actual residulal of the HP339 or whatever you have in front of the digitizer. Even AP mentioned in some note somewhere adding a passive notch in front to improve the accuracy and resolution of their analyzers.
I think this is interesting but too narrow a pursuit at this level. I'm interested in IM and what can be learned there. Keith Johnson tells me that a tone cluster with a number of closely spaced tones has become his most sensitive indicator. I have one IM analyzer here I have started to play with that has the "Cordell" three tone source. it has too much internal source IM at this stage (.004%) to get to what I want it to be so I have some work ahead.
I am now using the best THD instruments that humans have developed for commercial use .... so I use them as secondary standards to transfer their numbers to lesser equipment. Another notch in front of the 725 or the 2722 is an interesting thought, though. But, first, I'll take another look at Victor's with the A-P (looked at it earlier with the 725D).
I am also interested in multi-tone tests and IM tests -- the reason I purchased (at full retail) the IM option for the A-P. That is next on my list of measurements. I don't see any DIY projects for that here however. Or Is there? Improving the 'Cordell's' 3 tone IM test perhaps? Three generators from Victor would be pretty nice.
Thx-RNMarsh
Last edited:
Then with a slight change in input level and you could get this --
View attachment 386330
With a nulled 3H.
THx-RNMarsh
Richard, jou might see reinforcement or partial cancellation, alternatively, at these low levels.
Jan
@1audio
It is not really that curious. Assuming that the distortion is very low, in this case perhaps around 100 dB down, the distortion will "disappear" into the noise with an input level of -60 dBFS. So the THD+N is basically the noise only, since the distortion level is insignificant compared to the noise. And therefore the THD+N is 111 dB below the maximum level. Equal to a dynamic range of 111 dB.
The numbers are not weighted, since that wouldn't normally be applied to a distortion measurement. And the 111 dB is the unweighted dynamic range, so it all fits very well. The A-weighted dynamic range is specified as 114 dB, which seems likely with an unweighted dynamic range of 111 dB.
It is not really that curious. Assuming that the distortion is very low, in this case perhaps around 100 dB down, the distortion will "disappear" into the noise with an input level of -60 dBFS. So the THD+N is basically the noise only, since the distortion level is insignificant compared to the noise. And therefore the THD+N is 111 dB below the maximum level. Equal to a dynamic range of 111 dB.
The numbers are not weighted, since that wouldn't normally be applied to a distortion measurement. And the 111 dB is the unweighted dynamic range, so it all fits very well. The A-weighted dynamic range is specified as 114 dB, which seems likely with an unweighted dynamic range of 111 dB.
yes, I believe that is the case.... for those using sound cards and QA400 etc... they need to take their DUT test data results below -100dB with a 'grain of salt.' Or, just ignore it. I would scale the dB's to only go from 0 to -100dB.
There are many improvements with the more expensive test equipment which do not have this affect until extremely low levels or not detectable at all... just noise contributions to errors.
I also suspect it is related to the notch filter in some way -- because the only comparative difference I see between the A-P and ShibaSoku is at very low levels and only to the 2H.
Thx-RNMarsh
Last edited:
I'll post a scan of the manual for this analyzer I'm playing with. It was built by Chuck Hansen and was a fair amount of work. It needs some tweaks still but I can get it to work to below -120 dB I think. It could be converted to PCB's and the oscillators updated to use current parts instead of the XR-2206 which was discontinued 2 years ago.
If your just in need I may still have one or two of the XR2206.
I can send them to you.
It seems the XR2206 was discontinued some time ago with no alternative parts? I'm looking for a simple low parts count stable oscillator. if it needs more that 6-8 external parts a Wein bridge or state variable oscillator using a TL074 becomes very competitive.
The docs for the IM analyzer are 28 MB. I'll pull the core docs and schematics out. The assembly would not be the best way to duplicate the beast. Give me a day or two.
The docs for the IM analyzer are 28 MB. I'll pull the core docs and schematics out. The assembly would not be the best way to duplicate the beast. Give me a day or two.
Getting parts for the existing units isn't necessary. It works as planned. I think I know how to get another 20 dB out of it, maybe more. Its all wirewrap construction and bolted together in a way that makes access to the PCB's difficult.
Below is a block diagram of the box.
Below is a block diagram of the box.
Attachments
The avatar -
Demian,
Do you intend to modify the one you have or make a new one?
FYI --
The avatar I put up is me greeting a large female deer behind my house. There is a small herd of them passing thru every once in awhile with their new ones. I watched them from afar for months until they got used to seeing me..... Anyway--- after making friends with them (a long story), I tried to greet them the way I saw them greet one another..... they put their nose forward. I did that and she responded in kind. They take a smell of you and it is locked away in their memory after that. From then onward, we were friends and I could walk up to them and pet them or do what ever I wanted. But if others came around they ran away. They can and do discriminate and do not trust unless you earn that trust.
You can communicate with any other animals if you learn their language.... which is mostly non-verbal. They thought I was a female I later learned. One day after being close friends for years and even generations, I acted like a male... I challenged another male who came into our space....and ran him off. The others- all my female deer friends - never let me near them again after that. :-(
Thx-RNMarsh
Demian,
Do you intend to modify the one you have or make a new one?
FYI --
The avatar I put up is me greeting a large female deer behind my house. There is a small herd of them passing thru every once in awhile with their new ones. I watched them from afar for months until they got used to seeing me..... Anyway--- after making friends with them (a long story), I tried to greet them the way I saw them greet one another..... they put their nose forward. I did that and she responded in kind. They take a smell of you and it is locked away in their memory after that. From then onward, we were friends and I could walk up to them and pet them or do what ever I wanted. But if others came around they ran away. They can and do discriminate and do not trust unless you earn that trust.
You can communicate with any other animals if you learn their language.... which is mostly non-verbal. They thought I was a female I later learned. One day after being close friends for years and even generations, I acted like a male... I challenged another male who came into our space....and ran him off. The others- all my female deer friends - never let me near them again after that. :-(
Thx-RNMarsh
The IM box I have is a breadboard. The project looks like it was slated to be an article for Audio Amature or something similar. Chucks documentation is really good.
The limitation I have found is noise, the sources noise swamps its intrinsic IM. Using two external oscillators I can get a residual from the analyzer of -126 dB, but I need a narrow band spectrum analyzer to see it. I'm hoping some tweaks here and there will improve the effective SNR of the system.
I should start a separate thread for this guy. Does anyone have contact info for Chuck (not Charles) Hansen? I don't want to share this if he isn't comfortable with it.
The limitation I have found is noise, the sources noise swamps its intrinsic IM. Using two external oscillators I can get a residual from the analyzer of -126 dB, but I need a narrow band spectrum analyzer to see it. I'm hoping some tweaks here and there will improve the effective SNR of the system.
I should start a separate thread for this guy. Does anyone have contact info for Chuck (not Charles) Hansen? I don't want to share this if he isn't comfortable with it.
Why you still deal with HW limitations, while a good SW wave generator provides such freq. out of the box or even more..
Cheers
Hp