Resistor noise reduction questions

[Planobilly]

Hi Guys,
I am looking for a high level overview concerning resistor noise and how to reduce it. I have read the following general statements.

Carbon comp resistors are rated at 0 db. I read this somewhere.
Vishay metal film resistors MIL-R-10506 are rated at -40 db http://www.vishay.com/docs/31027/cmfmil.pdf

The idea is to lower both the current noise and the Johnson-Nyquist noise in a typical guitar amp.

I assume the following.
1. Use low noise resistors (easy to do)
2. Reduce the resistance value to the minimum that does not cause other issues...most important on preamp stages (not so easy to do)
3. Lower the resistor temperature ( just how to do that and by how much in a practical way I am not sure how to do)...I assume the ambient temperature in a closed chassis to be around 150 F depending on the type and number of tubes. (lowering the temperature could range from a cooling fan to pretty complex other lower temperature methods)

Ideas? Methods used in other equipment?

Thanks,

Billy

 

[Svein]

1. Resistor noise: http://www.eetimes.com/document.asp?doc_id=1278251
2. Low noise preamplifier: http://cds.linear.com/docs/en/datasheet/lt1115fa.pdf
3. I do not think you can get a low temperature anywhere in a tube amplifier.

 

[sophiecentaur]

The Noise Power produced in a (perfect metal) resistor is independent of its value. The noise power in a bandwidth Δf is Pn=kTΔf. How this affects the noise performance of an amplifier depends on the actual value chosen for the resistances that appear in the input and the specific amplifying device. We are usually interested in Signal to Noise ratio and you need to consider the Impedance of the signal source (microphone, antenna etc) in order to maximise the signal Power into the amplifier at the same time as minimising the Johnson Noise generated by the amplifier resistance.
So

Reduce the resistance value to the minimum that does not cause other issues

here the 'issue' will be that the input signal volts could be potted down unacceptably although the Noise volts will also be reduced with no net improvement of SNR. It's a compromise (Noise matching)
This link gives a good chat about the topic of resistor noise.
If you are using a Junction Transistor, there may be an advantage in choosing the appropriate amplifier mode to suit the Impedance of the signal source. (Same may apply to a thermionic valve but I have no experience of common grid valve amplifiers)

 

[Planobilly]

Thanks guys...I will read those links. The EE Times dock was well written and easy to understand.

As far as the heat issue goes, the tubes are outside of a lot of chassis. Thermal insulation would help some in conjunction with a fan I guess. I am not sure how much reduction in temperature that could provide or how much the temperature would need to drop to be really useful. I have never really measured the bottom of a tube socket which I assume would be the hottest point inside the chassis.

Expierments I have tried in the past with fans have caused more issues than they have solved. That has been a while back and I understand some of the newer fans for computer systems are less noisy.
Billy

 

[tech99]

Thanks guys...I will read those links. The EE Times dock was well written and easy to understand.

As far as the heat issue goes, the tubes are outside of a lot of chassis. Thermal insulation would help some in conjunction with a fan I guess. I am not sure how much reduction in temperature that could provide or how much the temperature would need to drop to be really useful. I have never really measured the bottom of a tube socket which I assume would be the hottest point inside the chassis.

Expierments I have tried in the past with fans have caused more issues than they have solved. That has been a while back and I understand some of the newer fans for computer systems are less noisy.
Billy

Reduce or eliminate any DC current flowing in the resistor, to reduce shot or Johnson Noise. Use a type having low shot noise, eg metal film. For a grid leak, no point making it much higher in resistance than the source impedance. If the circuit use an input transformer, then no grid leak is required. Be sure the bias is correct for the stage so that grid current is not flowing. If there is a pre amplifier, then look at that not at the man amp.