I have seen people saying their tube preamps or power amps were dead quiet, but exactly what is considered as dead quiet? I have to admit that my DIY experience is short and lack of any EE background, but I have been able to scratch build from someone else's complete or partial design successful. While I managed to build some that are hum free, none of them is hiss free. Some hiss louder than the others. Due to this inperfection, I haven't built one "dead quiet" tube amp of any kind ....... So, what is your definition of "dead quiet"? Also, it is probably impossible to eliminate hiss, but what are the techniques to reduce them?
"Quiet" for me means that at 1M or more from my speakers, I cannot discern any noise. "Dead quiet" for me means that with my ears within an inch or two of the speakers, I cannot discern any noise.
There are a number of things that determine noise in a circuit: the thermal noise from the tubes, thermal and excess noise from plate load resistors, and power supply noise and ripple are all very important factors to try to minimize.
There are a number of things that determine noise in a circuit: the thermal noise from the tubes, thermal and excess noise from plate load resistors, and power supply noise and ripple are all very important factors to try to minimize.
LOL, my lame way to reduce noise is to use inefficient speakers. The most efficient that I have is at best 89dB 
Seriously, those were inherited from sand amp era. I am going to build a pair based on FE207E just to get a taste of FR. That's the reason I posted the thread. Relatively quiet from a 89dB pair can be a totally different animal with the 95dB fostex!!
SY, what is the typical technique to deal with the thermal tube and load resistor noise? Better resistors, or choke load?
Thanks!!
Seriously, those were inherited from sand amp era. I am going to build a pair based on FE207E just to get a taste of FR. That's the reason I posted the thread. Relatively quiet from a 89dB pair can be a totally different animal with the 95dB fostex!!SY, what is the typical technique to deal with the thermal tube and load resistor noise? Better resistors, or choke load?
Thanks!!
Your method of noise reduction by using poor efficiency speakers is common for car audio. But they have a noise infested ground to deal with.
Reduce hiss:
-Don't use over sized grid stoppers.
-Done use over sized grid pull-down resistors
-Use the lowest Rp tubes that can still get the job done
-Don't use plate resistors larger than needed to get the job done.
-Use resistors with a power rating quite a bit larger than needed.
-Use Metal Film and Non-inductive wire wound resistors.
-Use higher Gm tubes when the option is available.
-DC heaters and/or referencing heaters above ground potential can help hiss and hum depending upon the circumstance.
Reduce hum:
-Proper grounding techniques (this takes research + trial and error to master).
-Have enough power supply filtering
-Proper physical layout techniques (don't place input stages next to the power transformer, don't route signal next to heater, ect)
- When possible, twist the conductors that have equally opposite current flow.
- When routing/wiring power supply, go from the transformer to the rectifier to the filtering caps to the load. Do not ground or get power for the load at any physical point in between.
-DC heaters and/or referencing heaters above ground potential can help hiss and hum depending upon the circumstance.
Reduce hiss:
-Don't use over sized grid stoppers.
-Done use over sized grid pull-down resistors
-Use the lowest Rp tubes that can still get the job done
-Don't use plate resistors larger than needed to get the job done.
-Use resistors with a power rating quite a bit larger than needed.
-Use Metal Film and Non-inductive wire wound resistors.
-Use higher Gm tubes when the option is available.
-DC heaters and/or referencing heaters above ground potential can help hiss and hum depending upon the circumstance.
Reduce hum:
-Proper grounding techniques (this takes research + trial and error to master).
-Have enough power supply filtering
-Proper physical layout techniques (don't place input stages next to the power transformer, don't route signal next to heater, ect)
- When possible, twist the conductors that have equally opposite current flow.
- When routing/wiring power supply, go from the transformer to the rectifier to the filtering caps to the load. Do not ground or get power for the load at any physical point in between.
-DC heaters and/or referencing heaters above ground potential can help hiss and hum depending upon the circumstance.
pchw, that's a difficult question to answer generally. Much of it has to do with choices in basic circuit design. But... wirewound resistors have the lowest excess noise (though the same Johnson noise) of all resistor types. Tube noise usually goes as the inverse of transconductance. Good power supply rejection minimizes noise from that source. And grounding/layout are critical.
(Broken record time) If you don't have a copy of "Valve Amplifiers," run right out and get it. There is a detailed discussion of noise in tube amps and how to design properly to minimize it.
(Broken record time) If you don't have a copy of "Valve Amplifiers," run right out and get it. There is a detailed discussion of noise in tube amps and how to design properly to minimize it.
Yes, I have both MJ's books - hard copies, not printed from Google
They are indeed very informative. The amps that I built after I owned the books definitely have lower noise the the early ones. There is a wealth of advise here. I will reread MJ's books, in particular, those points mentioned here.
Thanks!!
They are indeed very informative. The amps that I built after I owned the books definitely have lower noise the the early ones. There is a wealth of advise here. I will reread MJ's books, in particular, those points mentioned here.
Thanks!!
ray_moth said:
A more general rule is to look at the overall gain of the circuit and use the right amount of gain that will get the job done. It is better to have two gain stages than a single stage operating at max gain with no headroom.
Excessive gain = Excessive noise
For example, when listening to the Cary 805 (6SL7, 300B,845) amp using my 106dB horns, I have to subsitute a 6SN7 to lower the gain and bring down the hiss.
-- josé k.
I've never heard anyone claim this, nor experienced this myself. However, I have head that even though carbon comps generate the most noise of any resistor, they have a different noise spectrum. Carbon comps do change value with voltage which can sometimes create a desired affect when used as high voltage swing plate loads.
I'm an open minded person though, so if you have anything that supports carbon comps as low noise resistors, I'm interested to know about it. Maybe in specific applications perhaps?
Well, I know that some carbon comps can produce crackles and pops for a short time whilst thermally stabilising at start up. Though this does stop, I replace any such all the same and hence I wrote "good" carbon comps.
This is what I suspect is the talked about issue as under normal circumstances I don't find them to be particularly noisy.
The "warmth" or voltage distortion is a separate issue IMO.
So that leaves what MJ said about phono stages (I assume it's OK to reproduce as it isn't patented information) words to the effect that all resistors have thermal noise and film resistors also have excess noise and that the excess noise dominates the thermal noise.
This is what I suspect is the talked about issue as under normal circumstances I don't find them to be particularly noisy.
The "warmth" or voltage distortion is a separate issue IMO.
So that leaves what MJ said about phono stages (I assume it's OK to reproduce as it isn't patented information) words to the effect that all resistors have thermal noise and film resistors also have excess noise and that the excess noise dominates the thermal noise.
Had a look, he states that there is thermal noise with metal film. Earlier in the book he talks about resistors and goes in to detail on metal film and wire wound only. He states that film resistors develop excess noise proportional to the DC voltage drop across them and that excessive noise in wire wound resistors is virtually non-existent. He mentions that for values above 10K, wirewound inductance is negligible.
On page 204 he talks about carbon:
"Carbon film resistors are an anachronism and will not be considered, as their tolerance and noise specifications are so very poor, although their reduced inductance makes them useful as grid stoppers."
On page 204 he talks about carbon:
"Carbon film resistors are an anachronism and will not be considered, as their tolerance and noise specifications are so very poor, although their reduced inductance makes them useful as grid stoppers."
OK, so there is no equivocal answer from Morgan Jones on carbon comps...or maybe he just didn't wish to discuss them?
I've discovered the following about carbon composition resistors. They are subject to these sources of noise:
Thermal noise
Contact noise
Shot noise and
Flicker noise
Related to their fabrication, the crumb structure of the composition is subject to issues such as:
Vibration
Voltage stress
Moisture ingress and
Thermal strain.
Carbon comps appear to be the noisiest resistors.
It looks as if the quietest are foil resistors due to their wide area offering integrity. These are followed by wirewounds. (note that foil are different to thick film which are noisier)
I've discovered the following about carbon composition resistors. They are subject to these sources of noise:
Thermal noise
Contact noise
Shot noise and
Flicker noise
Related to their fabrication, the crumb structure of the composition is subject to issues such as:
Vibration
Voltage stress
Moisture ingress and
Thermal strain.
Carbon comps appear to be the noisiest resistors.
It looks as if the quietest are foil resistors due to their wide area offering integrity. These are followed by wirewounds. (note that foil are different to thick film which are noisier)
His comments there related to a comparison between metal film and wirewound. There's nothing whatever in that section about carbon comps, so I wouldn't leap to the conclusion that MJ is claiming that they're quieter than metal films. They aren't.
His recommendation (which I heartily endorse) is wirewound. No excess noise, lowest voltage coefficient. What's not to like?
His recommendation (which I heartily endorse) is wirewound. No excess noise, lowest voltage coefficient. What's not to like?
jnb said:
What's unclear about this statement, then:
On page 204 he talks about carbon:
"Carbon film resistors are an anachronism and will not be considered, as their tolerance and noise specifications are so very poor, although their reduced inductance makes them useful as grid stoppers."
Jan Didden
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.