|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
72.219.92.247
In Reply to: RE: Low noise resistors? posted by andyr on August 15, 2021 at 20:43:06
'Low noise' resistor is a bit of an oxymoron. The thermal (aka Johnson) noise power of a resistor is proportional to the resistance value. So, for example, a 47k generates 470x the noise power of a 100 Ohm resistor (approx 27dB). A 'low noise' resistor would be one that doesn't add anything to the inherent thermal noise. Perhaps, look for white papers or specs for 'excess' noise from manufacturers to see if a particular resistor brand is really best for noise.
The impact of noise from a resistor depends on where they are in the circuit. I assume these are loading resistors, for MC & MM respectively, in which case they are shunted by the cartridge source resistance. The cartridge source resistance will be much lower than the loading resistor. In the MC case a typical cartridge source resistance is 5 Ohms. When that 5 Ohms is loaded by a 100 Ohm resistor the circuit noise actually goes down by 0.2dB but the signal goes down by 0.4dB so, overall, SNR degrades by 0.2dB - not much - and, I expect, the active devices in your phono stage will add more than that anyway.Regards,
13DoW
Edits: 08/16/21 08/16/21Follow Ups:
I'm retired now, but my day job for many years consisted of design and measurement of low phase noise circuity in the region of -180 dBC. This doesn't correlate directly with amplitude noise, but the worst-case equivalent is orders of magnitude better than any vacuum tube. Our products approached the theoretical limits of physics in this regard, and all used thin film resistors. Leaded components aren't usually marked thin or thick film, but resistors like those sold by Vishay with tempcos around 250 ppm/K are undoubtedly thin film. When resistors of this type are used, the noise floor of active devices generally proves to be the limiting factor.Edit: Just to add some relevance to this, most of our products were RF-related. However, we did sell an opamp-based audio amplifier with an input referred noise of <1 nV/√Hz, and our proprietary in-house design using discrete components was very close to 0.5 nV/√Hz. Both products used standard thin film chip resistors.
--------------------------
Buy Chinese. Bury freedom.
Edits: 08/19/21
Thanks, 13DoW ... they are in fact series resistors.
Andy
Are the 100 Ohm in series between the cartridge and the input devices? IIRC, about the best input referred noise for an audio transistor amplifier is approx 1nV/rt.Hz (e.g. MAT-02 and, I think, there is a J-FET can reach that too). Adding a series 100 Ohms is about the same as adding another 1nV/rt.Hz so the system noise doubles.Regards,
13doW.
Edits: 08/16/21
Thank you 13th DoW - exackly! :-)) They are JFET Gate resistors (for stability).
So my problem is ... what are the best reses to use?
I had thought Vishay VSH were good (in terms of low noise) but, as I can't get them any more, maybe I should use:
* Mills ww
* or, say, Holco metal film ... which I use in other places in the circuit?
Or - as I understand higher wattage resistors have less noise - use 1w metal films?
The other place in the circuit where a series res is used is between the Drain of the 1st gain stage JFETs and the passive RIAA circuit - this is 47K.
Andy
Ah, gate-stoppers! I have no experience to know how small a resistance you can get away with. The noise voltage of those resistors will add (~ 1nV.rt.Hz) directly to the equivalent input noise of the rest of the amplifier but, if your input referred noise without the resistors is at least 5nV/rt.Hz then it is not a big hit. The effect of the noise from the series 47k on the output side is lessened by the gain. If the first stage gain is 470x then the noise contributed by the 47k at the output is the same as that contributed by the 100 ohms at the input. A SPICE simulation is a great way to know the dominant noise sources in a circuit.I did a search on the Vishay website for 'noise' and found nothing relating to resistor generated noise. My conclusion is that 'low-noise' resistors are not a thing - it is purely SQRT(4kTR) V/rt.Hz but I also found the article linked below that states some compositions have a 1/f noise region in addition to the thermal noise (1/f noise is common at low frequencies in semiconductors). I suppose if a resistor with a positive temperature coefficient was dissipating power and got hot, its resistance would go up and then so would the noise generated (noise power is proportional to temperature). In that case a higher power rated resistor might not get so hot but you'd have to calculate to see if the effect is significant. Your gate-stopper resistors will dissipate insignificant power so that is not a problem.
Regards,
13DoW
Edits: 08/17/21
That link is a very interesting article, 13DoW. :-))
I cannot afford the Z-foils ... but I will keep away from wirewound.
So I will use 'standard' metal film - but use 1w versions.
Thanks,
Andy
Andy, just use DALE 1% resistors. They will do the job.
And easily available from Mouser, I see. :-))
Andy
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: