I received a comment form people who reportedly well-familiar with using the high-frequency high gain tubes. They suggested me that the resistor between screen and grid (R23) might be eliminated and that would bring some benefits. It very much might be correct, as the location of this resistor (used to prevent high frequency parasitic oscillations) is quite critical. The report was the 6E6P-DR (long life, better quality, extended durability, better constriction) could handle operation with shorted screen and grid.

I personally did not experimented with them without the R23 resistor. Those Milq’s drivers tubes….some of them more wiling to oscillate some less. I never used then without ferrite bids, without R23 and without the greed stoppers. I did come across of quite few occasions when the 6E6P/6E5P were too much “live” and no mater what I do to them, including the shock absorbing tube sockets, helped them.

Still, the Milq’s driver stage is not the fist stage of a phonocorrector and perhaps it is possible to operate it without the R23 resistor (I use currently the Vishay R102 as R23). If anyone has any experience dealing with triode-strapped penthods/tetroids without the R23 resistor then please let me know about your findings.

Rgs,
Romy the Cat

ml8 wrote:

That's true that this resistor have negative impact on sound quality. You can run your 6E5P without it as long as you do not get parazitic oscillation. Just connect screen grid and plate electrodes right on tube socket, using shortest possible cable. By the way, i've tried using 6E5P without any biasing -- 100 V on plate, 0 V on grid. This way 6E5P sounds much better (to say the least) than with any biasing method i was trying (but i didn't tried yours, actually). As long as your source can drive it, you wouldn't get any distortion, but exactly the opposite -- very clean and very live sound.

ml8 wrote:

...but resistor between screen grid and plate, to my hearing, makes sound slightly unnatural, especially in highs. But this may be not true in your case (i didn't try to use this resistor with 6E5P, only with other power tetrodes), so i think i'll suggest you first try it and listen, and then you can make decision about do you need it at all.

Well, I am not thrilled with sound of “opened” 6E5P; I mean with juts fixed bias. It sounded to me too brutal and too aggressive. The “resistor bias”, and partially with a separate “buffered”, soft voltage coursed completely changed the 6E5P sound. Also, the 6E5P is kind off “fast” tube they all have kind of hard sound. The way how we bias the 6E5P make the grid resistor have with miller capacitance a gentile low-pass filter above 100Khz. Still from my point of view the “gas voltage” is the key in the Milq sound. Anyhow, I will try sometimes to get rid the screen lifter or perhaps to put a diode in there… will see what it does.

ml8 wrote:

And about your question... actually, i'm no audiophille by any means, as i have pretty crappy speakers (around ~88 dB, two-way bass reflex) and i'm usibg PC as a source  But still, when i run 6E5P without any biasing (actually there is small bias voltage, which appears on grid leak resistor), sound became much more natural and i thing i can understand now what are you talking about when you're describing the sound of Melquiades.  But i'm using just a single cascade with 6E5P as an amplifier -- no poweramp. Still, it can be pretty loud when i'm want it to be, even with my dumb speakers. So sometimes i just wonder why do you need 6C33C at all, with your high-sensitive speakers.

Hm, you drive 88dB sensitive load with 1.5W buffer? I do not know ml8, it doe not sound credible to me. The fact that it might be “pretty loud” is not really important there are many other things… In fact driving 88dB and 108dB require very different amplification, ever if we disregard power…Anyhow; it is much bigger subject…

http://www.goodsoundclub.com/Forums/ShowPost.aspx?PostID=3688

ml8 wrote:

If you think gas regulators make sound of 6E5P so nice, didn't you try putting gas regulator tube in the grid of the 6C33C, and why?

Sure, I trued. It sounded very bad. I pull out of my butt any imaginable rational with explanations but the core will be the same: I have no idea why the gas buffers did not sound well with 6C33C

ml8 wrote:

You are heavily filtering bias voltage with LCRC filter.  But you're using gas regulators, which are supposed to filter the ripple very well.  So, do you need this additional LCRC filtering at all?

No, it is not necessary to heavy filter the bias before gas tube. As you see the negative voltage is being used to provide bias for output stage and it should very severely “derippled”. Since I generally tend to have an input choke after diode the design of the negative shoulder was obvious. The positive and negative voltages are sourced form the same bridge and to keep the bridge even-loaded I used the same filter entrance for the positive side. Yes, it is not really necessary and did proposed to peoples to simplify the positive supple by taking the voltage for the gas tube form the B+ of the first stages, In fact I had one Milq made exactly like this…

ml8 wrote:

How does sound of your amplifier change when you extract gas regulators from sockets? As far as i understand, there should be no problems (technically speaking) running Melquiades without them, as all voltages should not change...

Well, I’m sure what you intend as is not exactly what you asked. The gas tube has 200V to “light up” in Milq and jut pulling the tube affects the bias of the fist stage. What you most likely are asking is how the first stage sound in that way how it being used in Milq compare with negative voltage is taken juts out of regular, not gas-decupled power supply. The answer is: very much different. I have writhe about is a many places at this forum - you might search. Do not think about the gas tube as a voltage source but rather think about it as a very delicate absorber for the 6E5P’s grid currents…

ml8 wrote:

So sometimes i just wonder why do you need 6C33C at all, with your high-sensitive speakers.

Your point is well taken:

http://www.GoodSoundClub.com/TreeItem.aspx?PostID=4063

and from time to time I ask myself this questions. Dima made it, driving his speakers, so I did it in past in Milq’s headphone versions…. The perfectly see the logical rational behind it but it should be trued and the results should be assessed… It takes time… I am planning to do it …if the upcoming Zaratustra II will turn out to be able sonically compete with the Super Milq’s LF section…

No, I did not try to use grid chokes.  The setting for Milq was flowing: the highest voltage in the gas tube and the highest value of resistor to the grid. Frankly there is a LOT of mystery how it all works out together and how it all inflicts sound. The value of the cap after the gas tube and resistors, the location of the caps, and many other things do affect sound… Evan the type of the gas tube affect sound, not matter how bizarre it is as behind .5MOh resistor it should be absolutely irrelevant:

http://www.GoodSoundClub.com/TreeItem.aspx?PostID=2792

Pretty much how it ended up to be with Milq Bias.: we tried various approaches until we come across something that sounded as we needed and then we locked on it and did not looks at anything furthered.  Here is some history for you:

http://www.GoodSoundClub.com/TreeItem.aspx?PostID=1818

I still would hesitant to short the scene and anode and it would require a very sever control of the tubes. The problem is that it would be necessary to do not only then they new (it not too difficult to put then before a fast scope and to see if they are quiet) but also to check them periodically while they are in use.

Burning the 10 channels of Melquiades for 2 years I do detect that a dying 6E5P/6E6P gains some intermediate noise when it goes south. What is interesting is that 6E5P do it different then other UFH tubes that I used.

For instance the 7788 ultra-high transconductance pentodes (E810F) what they go microphonics did in a gentle way, reacting initially juts to disturbance and then get worse and worse. The WE417 behaved differently: it begin to produce the UHF pings – very unpleasant for tweeters effect…. The 6E5P what got old begin to “boil MF”. I mean from time to time the MF channel begins to pick some “bubbling MF noise” for a few seconds and it is a good indicator that the tube is beginning to go down. Uselessly in my Milqs the 6E5P lives for a little bit less then a year. Ironically but the 6E6P-DR does sooner then the 6E5P….