I find all capacitors with low voltage ratings suspect. Normally I won't use anything under 16 VDC unless I have no choice.
-Chris
-Chris
It appears 2.91 is the latest, available from the author's site here:
DL4YHF's Audio Spectrum Analyser
From the text on the page, it appears the latest Windows version it runs on is XP.
Distortion in capacitors is dependent on dielectric material, voltage rating, humidity level and mounting position among other issues.
I showed my bridge method to measure resistors in Linear Audio Volume 1. It has the advantage of using 10 very similar parts and reciprocity to determine distortion. Many folks have tried to simplify this to using reference components and just comparing the devices under test to the reference. That is the quick and dirty method that gets you close but is ultimately limited by the reference.
Testing capacitors show first that when there is a DC voltage on any capacitor the definition of capacitance comes into play. C=V/Q or C dV/dT + V dC/dT = dQ/dT or i. It is the dC/dT term that is often ignored as being very small. There is a baseline of vibration everywhere and this moves the capacitor under test. As the dielectric and electrodes move from this there is a slight variation of capacitance. This is minimized in stacked plate construction by vertically orienting the plates.
The other test I did was on Mylar (Polyester) capacitors which are not among the best for moisture based changes. I used a test set of ten in the 4-1/1-4 bridge and measured the distortion. I then placed them in a desiccation steel container filled with silica gel for a few months. The distortion dropped by a lot. This would indicate that in some locations distortion from capacitors would be lower in winter as opposed to summer. Or doing things such as leaving the power on in gear will warm things up and dry out capacitors. Others have show running signals through capacitors also lowers the distortion.
Now I think it is clear that thicker dielectrics are stressed less under use and yield lower distortion. Of course that dielectric material will also have a significant effect is well known.
On the issue of the National Semiconductor opamps having issues with RF, long ago I mentioned that using a small inductor in series with the input really cleans the issue up. I use a 100 ohm 1/4 watt resistor with 10 turns of fine gauge wire wrapped around it for these inductors. I know of at least one high end piece of audio gear that uses this technique.
Finally I find that quadrature oscillators are most useful for testing systems. My technique as mentioned way back is to use oscillator circuits set just below oscillation and tickling them with a signal to get a very clean output. Using a digitally generated source with well matched tuning elements allows for very precise phase relationships.
I showed my bridge method to measure resistors in Linear Audio Volume 1. It has the advantage of using 10 very similar parts and reciprocity to determine distortion. Many folks have tried to simplify this to using reference components and just comparing the devices under test to the reference. That is the quick and dirty method that gets you close but is ultimately limited by the reference.
Testing capacitors show first that when there is a DC voltage on any capacitor the definition of capacitance comes into play. C=V/Q or C dV/dT + V dC/dT = dQ/dT or i. It is the dC/dT term that is often ignored as being very small. There is a baseline of vibration everywhere and this moves the capacitor under test. As the dielectric and electrodes move from this there is a slight variation of capacitance. This is minimized in stacked plate construction by vertically orienting the plates.
The other test I did was on Mylar (Polyester) capacitors which are not among the best for moisture based changes. I used a test set of ten in the 4-1/1-4 bridge and measured the distortion. I then placed them in a desiccation steel container filled with silica gel for a few months. The distortion dropped by a lot. This would indicate that in some locations distortion from capacitors would be lower in winter as opposed to summer. Or doing things such as leaving the power on in gear will warm things up and dry out capacitors. Others have show running signals through capacitors also lowers the distortion.
Now I think it is clear that thicker dielectrics are stressed less under use and yield lower distortion. Of course that dielectric material will also have a significant effect is well known.
On the issue of the National Semiconductor opamps having issues with RF, long ago I mentioned that using a small inductor in series with the input really cleans the issue up. I use a 100 ohm 1/4 watt resistor with 10 turns of fine gauge wire wrapped around it for these inductors. I know of at least one high end piece of audio gear that uses this technique.
Finally I find that quadrature oscillators are most useful for testing systems. My technique as mentioned way back is to use oscillator circuits set just below oscillation and tickling them with a signal to get a very clean output. Using a digitally generated source with well matched tuning elements allows for very precise phase relationships.
I guess it's not an ongoing issue with National audio parts as they have been discontinued.
Now what?
You method of choking the input for RF on National parts goes way back to the late 70s at least. We shouldn't have to do it.
Now what?
You method of choking the input for RF on National parts goes way back to the late 70s at least. We shouldn't have to do it.
The LME497X0 series is still around, and I believe one of the ostensible "bad behavers" that Ed alludes to. But, yes, most of that series is kaput.
That EOL has since been rescinded. Some of the packages are EOL'd but the 497X0 series lives on (assuredly in SOIC, can't remember which other packages).
Currently still in production. LME49710 | Audio Op Amps | Operational Amplifiers (Op Amps) | Description & parametrics
At $.96/M they are very popular for some folks and adding a few cents for a coil still leaves it as one of the least expensive solutions. (AD797 $5.46/M)
Actually a good choice for a transformer input microphone preamp.
No worries, it was a bit confusing. The National audio line EOL announcement was much better publicized than the revision. More important is that you know that part is still in your armament.
I suspect its not really the same chip. If TI shut down the line then its will be a close cousin on a different line. Scott would be far better able to explain but basically processes don't move easily. Its like trying to get San Francisco sourdough bread anywhere else. its not the same. probably good but different.
The LME49990 was a great part and could address some things that few others could. Despite claims to the contrary TI does not have an equivalent.
Unfortunately polystyrene has both the poorest temperature tolerance and is the most humidity sensitive of the cap materials. Its change with temperature and time is quite good when made well. The cheap parts like the Xicons need to be manually installed and soldered making them but production friendly. I have had them in hermetic packages but thats up there with the audio nutcase caps, but for real reasons. Manufacturing seems to favor 2% tolerances and 1% means you buy all the 2% parts and get the middle of the distribution.
i need to get some current multilayer COG parts and put them on the CLT-1 to see what I measure on the parts distortions. Polystyrene usually measures really low.
Teflon is good but PITA to make and expensive because of it.
I don't see how DA would be a factor in an oscillator. Once its stable the DA artifacts won't show.
Why not?
I don't know what's so bad about adding RF protection. Most chip inputs have nonlinear or PN junctions somewhere near the input that can rectify stray RF input. The National parts may have been more susceptible to this, and/or National actually admitted to this susceptibility and suggested something that can be done to fix it.
If you're making millions of units the parts cost adds up, and without good RF protection you get complaints like "everytime I'm in a taxi and the driver talks on his radio, his voice comes blasting through my earbuds," but this thread's more about quality (lowest distortion oscillator) than quantity, so to me it's a no brainer to add these parts.
I see this for inputs on page 7 of this datasheet (instrumentation amp sold as a balanced studio mic preamp), though the text gives a totally different reason for doing it:
http://www.ti.com/lit/ds/symlink/ina163.pdf
I do not see any difference between the LM4562 and the LM49720. Absolutely the same performance and the same noise problems - same percentage of noisy chips. Seems marketing trick. Also the data sheets are as the copy/paste result.
Last edited:
They're the same chip (the LM4562 and the LME49720). What Demian is alluding to is that National's analog fab was shut down, and that new LME49720/LM4562 chips will be on a new fab process.
The late Bob Pease mentioned the label change was because in one language where lots of stuff is made LM4562 was phonetically close to an obscene term. So it is the same chip with a name change.