Reduced Filament Voltage

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

toobdood

Well-known member
Joined
Nov 12, 2008
Messages
100
  Here's the question: Does anyone here have experience in running tubes with REDUCED filament voltage?  I know that this is a common practice in some tube mics, but my curiosity is simple: What does it do in simple gain stages? 
 
Try it and find out for yourself.  I mean this in a good way.  It depends on all kind of things cathode oxides, what type of metal the cathode is made of how much plate current...............
 
In addition to the user experiences (which are obviously the most worthwhile), do note that most tube databooks have info about (actually warnings against) significantly different heater voltages. Also lower is not pretty, at least from a technical point of view.

A Philips databook sez: influence on both performance & lifetime. Keep deviation below + & - 15% (design max. value).

Bye,

  Peter
 
Yes - when running the tube at full-power Anode>Cathode, underheating is not a good thing, as it strips the cathode's electron-generating oxide layer.

But if you look at the underheated mic circuits, you will find that they run their tubes waaaay below any ratings.

If you don't need to pull-through too many electrons, there's little reason in making a whole lot of them "available for pull-through" which is what heating does.

Jakob E.
 
If you get the cathode cold enough you no longer have a space charge and are in the "temperature-limited" regime.  The grid hardly works anymore, and about the only thing the tube is good for is producing noise, the predictable amount of it based on the current.

I once tried to reduce the cathode temp in a vidicon in an effort to make the beam landing energetics more uniform, when the thing was used in a slow-scan mode.  It didn't work very well.  An alternative that would have worked better, and someone else actually published results about, was to use a photocathode and illuminate it to produce photoelectrons.  But we weren't in a position to make experimental tubes, and solid-state photosensors were just becoming available and improving rapidly.
 
  I am aware of the bogey specs of keeping a tube within 10% to 15% of the rated voltage.  I also know from my own digging that a Brauner tube mic runs an EF86 at 5 volts: 6.3-5=1.3 this is a 20% DROP in voltage.  I will be performing my own experiments and reporting the results.  Seeing that a NAME tube mic uses this "feature" of tubes I am interested in using this to our sonic advantage. 

  As I understand: dropping the filament voltage to the 20% mark raises the input impedance of the tube caused by the reduction of current from anode to cathode.  With tube mics this feature is a big gain because the situation requires the input impedance of the tube to be higher than the theoretical impedance at normal operation.

  As for the "Thou Shalt Keep All Tubes Within Their Specified Ratings" commandment: I work on tube guitar amplifiers where this commandment is ALWAYS broken.  Yes tubes wear out MUCH faster in guitar amps than what they do in other applications but I have also pulled tubes in 40 year old consumer hi-fi amps that still have plenty of juice.

  I Also think that the Filament Voltage Spec is a little specious at best.  How many here have had experience with OLD tube mics that have this same feature.  Of course lowered filament voltage could easily be applied to a tube mic.

  I am too busy fixing stuff in my shop before I get the chance to mess with the stuff I want to do.  I am throwing this out there for anyone interested and therefore interested in sharing their results.

  Peace
 
  To briefly continue: Not only do we see reduced filament voltage in tube mics, but we also see the tube run waaaay under it's specs.  (Thanks Gyraf)  I am a tubehead and I WANT to get the most out my favorite little glowbugs.  Peace
 
As Gus said already, it's a bit of a suck-it-and-see. Some cathodes won't like it.

Since setting up a valve is a bit of a simultaneous equation, changing the heater voltage considerably means you should pay attention to the operating point of the valve, as you may find that it shifts as you vary the heater voltage.
 
  Thoughts of the old EVH tone on those first few records: an old Marshall on a variac turned down to 89 vac.  Hmmmmm
 
I use 5 to 5.5V on tubes all the time. In terms of lifespan it really is not that much of a difference.
 
I have played with underheating tubes and this is a good case of it depends.
  Some tubes go noisy in tube microphone circuits when underheated.  The best thing to do is read up on oxides and cathode alloys and do a bunch of experiments.  I built different tube circuits and measured and listened.  Sometimes a tube sound better underheated sometimes it is noisy and sometimes I don't hear much of a difference.  It depends on a bunch of stuff
  The VM1 I looked at runs at 5.4VDC.
 
  Okay here's the skinny on a ECC802 (a 12AU7 variant from JJ) I have the tube being run at 9.3vac.  This is  about a 26% reduction in filament voltage and I do hear imaging differences.  What does surprise me is how well everything hangs together.  Plate voltage is 100v, 1k on the cathodes, 56k plate load resistors passing audio quite well into my hacked stereo.  The soundstage sounds forward, but thick at the same time.  I like it.  This sound is not for everything, but I just have Sunrise On The Sufferbus on right now.  I'll try some Nick Drake, Jane Child, Pink Floyd and Mastodon and see if I like what it does to extremes.

  Peace, C
 
  I have an OLD 6SN7 that is the holy grail of fuzzy warmth that I use when I want a sound from my rig that is thick enough to walk on.  The JJ ECC802 in this configuration can duplicate it AND deliver more clarity.  I must sound like an audiophile bonehead, but damn I love this stuff.  Mastodon sounds like its going to kill you with THICKNESS.  This is what I was hoping for: something to get me the results that could only previously get with a RARE nos octal.

  I know that the other factors come into play: The tube is loafing along at 0.1 of a watt 1ma at 100v.  But this is the stink that I want!

  I've got King's X's XV on right now and dUg Pinnick sounds like he's singing from my bench!

  By the way if you saw the PS2, the hacked into triode flea powered EL84 amp culled from an old console stereo a pair of cheap 'n cheerful speakers and the tack-soldered mess that is the working circuit on my bench you would laff yer azz off and say: HOW IS THIS POSSIBLE!?!?

  I know from my experience as a guitar amp tech that so much of what we love of tubes for is their distortion products.  This experiment is no exception: How could I duplicate a sound only ONE tube could give me.  This experiment shows me that it's possible to get tubes to behave in sonically delicious ways that I can have control over and that's what this is about for me: making junk sing.
 
just a side note.
obviously, quite some tubes out there will tolerate underheating to some extent.
some tubes can take extreme underheating, some will die very fast.

in Neumann U47, the VF14 is run at over 35% reduced heater voltage. there are a lot of mics out there with the same old tube and happy.

on the other hand, some specialty miniature tubes like e.g. AC701 (neumann, schoeps mic's) should in no case run below the rated voltage.
a freelancer for neumann once mentioned that earlier batches of M147 or M149 had this very problem, since they were using miniature tubes, but a little underheated.

regards
volker
 
volki said:
some specialty miniature tubes like e.g. AC701 (neumann, schoeps mic's) should in no case run below the rated voltage.
a freelancer for neumann once mentioned that earlier batches of M147 or M149 had this very problem, since they were using miniature tubes, but a little underheated.

Volki, do you know if this also applies to the 5840? In the Royer mod (MXL2001) the heater supply is (loosely) specified at 5-6V. So the tube runs more or less underheated.
 
  Thus is the reason for my question:  What experience do others have?  What collective knowledge do we possess regarding this subject.  The fact that many old tube mics have 40 or 50 year old tubes in them is a fact worth noting.  What materials are these unobtanium tubes made of?  I do mean this in the most literal sense: What are they made of?  More specific what class/type of tubes best lend themselves to this form of abuse?

  More in line with my own immediate needs: Are the classics of the 12A*7 series and similar tubes good candidates for this line of research?

  As for the 5840 in the Royer, and the EF86 in a Brauner the same idea applies: Reduced current flow raises the input impedance of the grids above the theoretical limits at normal operation.  I know from my own experiments that tubes can pass signals at VERY low voltages for either plate or filament.

  If YOU have knowledge or experience please share.  Peace, C
 
_I_ don't have knowledge, but what do you mean with input impedance in this context? Do you mean the amount of grid leakage and hence the highest usable grid circuit resistance?

I my tube mics I have used EF804S (electrically equivalent to EF86 but better) at 1mA and normal 6V filament. I did some experiments with reduced filament voltage and got more noise (well, perhaps 1-2 dB). The operating point of this tube does not change even with very large grid leakage resistors. This might be due to the high quality of the tube in question. One must of course keep in mind that the grid probably collects impurities and consequently the leakage current will increase. I guess keeping cathode colder keeps grid colder and therefore grid emission lower. Is this true? Anyone?

Steve Bench has an excellent article about directly heated triodes with starved filaments. It is a must for tube freaks. Surprising results.

http://greygum.net/sbench/sbench102/dht.html

-Jonte

 
Back
Top