any references that show that?
Vishay Z-foil claims "< 0.1ppm/V"
so -140 dB would at worst require 1 R/V of output
Bruce Hofer of AP Analog Design presentation seems pretty clear on Bulk Metal Foil having lowest V coefficient - admits most other technology R don't spec it but he is certainly in a position to have measured lots of devices to very high resolution
We were not talking about Voltage coefficient. We were talking about tempco.
The tempco can be 10 ppm/C but if that is limited to 0-100C and is actually of the K*Xexp2 with a limit of 10ppm at 100c then that will have more odd order distortion a than 10ppm tempco that is is 10 ppm at 0C and 100C.
Then there is the issue of thermo electric effects.
But look at the Linear Audio article which shows the distortion of bulk metal and other types.
hard for me to tell what you're talking about much of the time
or where you get your numbers:
60 C seems a OK amp internal temp target, 10 mW is 1.5 C in a 150 C/W R
1.5 C * ±0.05 ppm/ºC is ±0.075 ppm - seems like we can typically manage better than -140 dB from TC if we care to spend for the R, vent the case
thermocouple emf follows the delta_T between junction of differing metals
by soldering both ends of the R into Cu PCB trace of similar weight we have somewhat reduced the differential delta_T
I don't know the exact test conditions - probably documented on Vishay site
but lets just go with the 1.5 C delta, to get: 75 nVpk
so EMF is estimated as a series 75 nVpk following the power envelope curve, themal excursion heavily filtered at audio by the thermal time constants of the R
once again not quite reaching -140 dB in a 1 V signal
or where you get your numbers:
60 C seems a OK amp internal temp target, 10 mW is 1.5 C in a 150 C/W R
1.5 C * ±0.05 ppm/ºC is ±0.075 ppm - seems like we can typically manage better than -140 dB from TC if we care to spend for the R, vent the case
thermocouple emf follows the delta_T between junction of differing metals
by soldering both ends of the R into Cu PCB trace of similar weight we have somewhat reduced the differential delta_T
I don't know the exact test conditions - probably documented on Vishay site
but lets just go with the 1.5 C delta, to get: 75 nVpk
so EMF is estimated as a series 75 nVpk following the power envelope curve, themal excursion heavily filtered at audio by the thermal time constants of the R
once again not quite reaching -140 dB in a 1 V signal
Potential distortions due to passive components, (somewhat distantly) related to what components account for sonics of JC's preamp.
good question, Simon seems to be digging deep under the noise floor to "explain" why amps driving dynamic loudspeakers with dB of power compression sound different
the example is single digit, at worst low 10s of C delta_T in properly selected feedback R with few ppm /C TC
vs sometimes over 100 C delta_T in voice coils with >30000 ppm /C
http://www.klippel.de/uploads/media/Nonlinear_Modeling_of_Heat_Transfer_03.pdf
apparently Golden Ears can "listen through" these loudspeaker distortions of thousands times greater magnitude, even sharing fundamental cause, presumably similar "audio character" to "know" you've changed from "old" Holco to "unapproved" TX2575
the example is single digit, at worst low 10s of C delta_T in properly selected feedback R with few ppm /C TC
vs sometimes over 100 C delta_T in voice coils with >30000 ppm /C
http://www.klippel.de/uploads/media/Nonlinear_Modeling_of_Heat_Transfer_03.pdf
apparently Golden Ears can "listen through" these loudspeaker distortions of thousands times greater magnitude, even sharing fundamental cause, presumably similar "audio character" to "know" you've changed from "old" Holco to "unapproved" TX2575
Last edited:
This is the distortion of a naked Vishay bulk metal film resistor test. An individual resistor would have 6 db more distortion.
Argue with yourself all you want. But it would be more productive if you looked at the material even the stuff on resistor distortion in this thread.
Argue with yourself all you want. But it would be more productive if you looked at the material even the stuff on resistor distortion in this thread.
Attachments
Somewhat?!
I am trying of course to be diplomatic. However, after exhausting the basic topic in part I and part II, and continuing to reassure people who merely want to copy JC's design that he will not provide any further details, things have run fairly far afield --- although I still think, touching on areas of interest in general.
In this latest diversion one of the things lacking is disclosure of the test instrumentation details.
Brad Wood
Which LA article? The first one you did was woefully lacking in details, as pointed out by Samuel Groner.
It did and it was based on an error. (Jan in editing changed a symbol in the schematic which led him solo on a merry trail of errors.) My response corrected that and added a few details.
Read the article, it is quite complete. The technique is being used by two test equipment manufacturers and one resistor maker that I know about.
Next try my article on how to measure distortion in solder joints and audio interconnects.
Yes just re-read them. Perhaps it would be a good idea to reference the letters exchange http://www.linearaudio.net/userfiles/file/letters/Volume_1_LTE_ES.pdf when referencing the article.
>apparently Golden Ears can "listen through"
>these loudspeaker distortions of thousands
>times greater magnitude
It doesn't take "golden ears"
Distortion does not mask distortion.
If a signal is run through 10 distortion producing
stages, (audibly) the differance between no stage
and after stage 1 is near equal to between stage
9 and stage 10 .............
>these loudspeaker distortions of thousands
>times greater magnitude
It doesn't take "golden ears"
Distortion does not mask distortion.
If a signal is run through 10 distortion producing
stages, (audibly) the differance between no stage
and after stage 1 is near equal to between stage
9 and stage 10 .............
- Status
- Not open for further replies.