Lower voltage on 12AX7 tubes, advantage?

CJVx

AK Subscriber
Subscriber
It seems like Scott and by Dynaco are running their preamp tubes at less than 12 each, maybe other manyfacturers but thise are the only two I have personal experience with. I’m sure this was a carefully chosen arrangement, But why?

*Edit- heater voltages less than 12v. Was a bit ambiguous earlier.
 
Last edited:
Audio Research also did this with some models,such as the SP-3 & SP-6, for example.I seem to recall my Mcintosh C20 does the same.

Studies were done on this topic long ago,when the military and industry were still using vast quantities of firebottles.These studies clearly indicated,with small-signal types operated conservatively,that while initial transconductance was somewhat reduced,it remained much more uniform throughout the tubes lifespan.Also,the MTBF was greatly reduced for all failure modes,with the majority of types carrying on well past their rated hours.A measurable reduction in all noise parameters was also noted. Mebbe a cool tube be a quiet tube;)
 
There is recent research showing that noise in 12AX7 types is reduced at lower heater voltages, but only when anode current is below 1mA.
 
Mac C8/C8/S in the late fifties~ early sixties did the same-with the resultant @ 8 volts DC filament(heaters after voltage rectification w/nearly doubling the 6.3 volt AC with the conversion to DC heaters for all six 12AX7/7025 tubes per preamp when fed from the octal cable connection from a Mac 30, 40, 60, etc. stand alone tube amps connected via the octal umbilical cord..
Or a stand alone PS for said preamps, as there was a MAC optional stand alone PS to provide the necessary power filament/HT for those nice sounding preamps should they be used without Mac amps providing the necessary power for use with other amplifiers with non compatible octal connection, if used..

For the time frame, the lowered Filament voltage, and being DC did result in overall lower system noise floor due to Mac`s configuration, for the engineering mindset at the time..
I owned and used said preamps during the almost mid seventies until I sold them to my B-IL in the mid eighties after building a stand alone very well regulated stand alone SS rectified HT/12 volt Filament PS for their use for him prior to sale..
He refuses to sell them back to me when I have offered generous return/profit on his purchase....
 
There is recent research showing that noise in 12AX7 types is reduced at lower heater voltages, but only when anode current is below 1mA.

Which is an excellent point that is often overlooked. Starved-plate tubes are often used for high-gain preamplifiers, because the noise is lower given low currents which could otherwise be drowned out by the noise.

It's funny how research continually re-invents the wheel, though, and the starved-plate design is continually rediscovered. Back in the 1950s those building instruments to measure very low currents, typically microAmperes, used plate-starved tubes for the low-noise DC amplifier stage. The ultra-low plate current was key. Same thing for the electrometer tubes which could measure picoAmperes.
 
The Acrosound 20-20 is using that feature in its design.
Starved plate current for higher gain, and reduced heater to 4.8VAC for better linearity and noise reduction.
I wasn't aware of this trick, Dave G, has the credit.
 
Sorry , but i don't like running any tubes more than 5% from their nominal voltage. Even Morgan Jones wrote about that in his book, Valve Amplifiers. Some say running them at -5% (6.0v or 12.0v) lowers the noise. i changed my Scotts and Dynas to run at nominal. I think they did it to lengthen tube life.
 
High-reliability uses in the 1950s and 1960s (computers, military, air-traffic control, etc.) typically ran 5% lower to improve life. That's at the low end of the manufacturer's ratings.
 
The Acrosound 20-20 is using that feature in its design.
Starved plate current for higher gain, and reduced heater to 4.8VAC for better linearity and noise reduction.
I wasn't aware of this trick, Dave G, has the credit.

Yes, it is clever. The full circuit and description can be found in "Direct Coupled HiFi Amplifier" by Herbert Keroes, Popular Electronics (1961, Mar).

The starved technique, however, pre-dates Acrosound and goes back to the 1940s, as it was well known then. I read a patent filed in 1946 which references it for starved screens. So it has to be even older than that.
 
Although not related to the small-signal tubes being discussed, here's what I was taught when replacing power tubes in broadcast transmitters: Install new tube, set filament voltage to the proper value (in most cases I dealt with, 7.5V), perform a proper tuneup and let things cook for a couple of days. Then, go back and start reducing the filament voltage til output power starts to drop...then turn it back up a squidge. That would give you the headroom needed to increase filament voltage as the tube aged to maintain a given output power. Wrung a LOT of hours out of tubes that way.
 
Yeah, transmitting tubes certainly are a good bit different than small signal tubes. Finicky beasts, from what I've heard about them. The explanation for the cooking after replacement process apparently arises from the fact that a thoriated-tungsten filament has a considerable burn in period because the cathode is typically not fully activated when the tube is removed from the box. It can take a few hundred hours for the cathode properties to stabilize, as thorium migrates to the surface and reduces, and for the emission to reach its specifications.

The other issue is that the Miller-Larson effect (recrystallization of tungsten) from rapid heating can thin the filament, causing fractures just like a light bulb. So motor-controlled autotransformers were used to gradually bring up the heater current to operating levels. Some tubes were left on for years on end, because this avoided filament fractures. But those filaments are differently constructed from the small signal tubes.
 
Dynaco PAS2/3 ran their filaments in paired series and some of them measured below the
design centers. and they lasted decades, and still test well, after 50 years. they are
"sought after" especially the smooth plates but the ribbed plates sound better
(iff you have music, gear, environment, ears with female vocals).
 
Back
Top Bottom