I do not think that it needs to be with center-tapped secondary but rather with symmetric secondary. I have good transformer with Faraday Shield that might take 6.3V or 12.6V and give all my DH voltages and with secondary that would be completely decupled from the IDH filament.

http://www.goodsoundclub.com/PDF/6,3_to_multi_voltage_trans.pdf

This is a cool idea to use BUT the problem that I see is that the hum-killing centering pot will be shifting too much when I move from 2.5V to 4V. So, what I need I think would be essentially something like this:

The Cat

Romy the Cat wrote:

floobydust wrote:

I … use a slightly different approach. I use the center-tap as a fixed DC balance (cathode bias resistor solo from c.t. to ground) and use the balance pot with the bypass cap only to ground for adjustable AC balance. Regardless if you try this approach or not, I always recommend padding the hum balance pot to reduce the effective DC resistance that the pot adds into the circuit. As an example, using a 100-ohm balance pot, your effective (series) resistance with the bypass cap is 25 ohms. Using a pair of 12 ohm resistors (wiper to each end) drops this to ~5 ohms. It helps... and there's no such thing as too quiet.

Paul S wrote:

This thread has gotten distressingly DIY, hasn't it?

floobydust wrote:

On a second note, you can also modify this for both adjustable AC and DC balance on seperate potentiometers. Both potentiometers would be in parallel with the filament (and padded) but you use the wipers independantly, i.e., one has the bypass cap attached and the other has the cathode bias resistor attached.

I was thinking about making the adjustable AC. I have bought 50V high quality wirewound very sexy mini- rheostat but other advisers of mine were very skeptical about this idea. In fact I have a number of people who question the validity of extra series resistance on the DHT filaments. I can see their point, even though I personally do not make experiments with validating the filaments series resistance’s impact to sound.  However, the “no resistor” position does make a common sense to me.  It is known that I am not treat my amp as some kind of festivity of electrical design – I know very little about it and many of my actions at this point, prior to listening, comes from exercising a common sense.  So, thinking about keeping or dropping the filament fine adjustment with extra series resistance I inclined to feel what Paul expressed when he suggest that the bar has already been set too high (at least what I hope) and saving a few pennies here or there does not make sense. So, I have ordered the “right” dedicated transformer specifically for the requirements that I have and the 5-pin-connectors that would service another run of a separate cable from PS to amp with AC for DH filaments.

floobydust wrote:

No easy drawing program available this second (traveling for 3 weeks) but I can send you a simple schematic in PDF if you like via email. In any case, to adapt what you have already have to the split biasing, just remove the cathode bias resistor from the hum balance wiper and connect it to the filament center-tap. This results in a fixed DC balance against the DHT filament and the balance pot now provides an adjustable AC balance only. In many cases, I've gotten better results in hum nulling this way, but it still requires good tubes.

In my new configuration the filament center-tap will be on another chassis, in PS side. The way how it organized now in my amp is that the negative of the Obbligato Cap on the B+ side of output transformer, the bottom of cathode resistor and the top of the Cube 950 cap from the cathode resistor and the ground of the OPT’s secondary are the same physical point on my ground buss thermal. I kind of like that the output tube is locked within the singular referent to ground. No matter that the “ground” with this type of cathode bias is kind of virtual category. I do not want the cathode restore to “drive” 3 feet of cable unit it “see” the return path (which will be the center-tap in your proposal). I am not sure that I am right but I do see a login in how I see it.

That is kind of mysteries to me and very none expected. From the people who use DHT I heard many stories that 2.5V filaments driven by AC are less noisy then 5V or 7.5V. That is kind of self-explanatory as with higher AC voltage the more noise. Still what I am witnessing is that that my 4V tubes have less or identical noise with my 2.5V tubes.

I have as now 6 types of 4V tubes:  RE604, KL71403, AC044, YO186, LK4110 , RE614 all of them single plate tubes and I see absolutely no noise increased when I run them despite I it almost twice higher voltage. What I am talking about “noise” I mean the absolute numbers of AC noise measurement not the auditable noise. I was wondering why it is so.

From my Russian sources I learned about the history of YO-186 tube and it was said that the tube it was made for instrumental use to measure low-levers inductance and the tube was designed especially to be driven by AC on it's filaments.  Later on the tube was used during 30s in cinema amplifiers and with AC on it's hears only. Interesting...

In the Yamamoto article:

http://www.jacmusic.com/Yamamoto/html/Amps/A06/description-ready-product.html

…they say: “The 4Volt technique gives best ratio between diameter and length of the filaments”. It does not sound very credible to me, in fact is sound more like marketing BS.

Anyhow, does anyone have any explanations why my 4V tubes have no more AC noise then 2.5V tubes?

DHT4ME wrote:

This is true. It comes down to thermal cycling. I love the #26 tube but it is real hard to get quiet.  It has a 1.5v 1a filiament. The heavy draw causes thermal cycling which is very hard to null out. There is also more ratiated garbage from the current. A tube with 2.5V is far easier to get noise down. Go the opposite direction and it starts to get bad again. The #10 is a Thoriated Tungsten filament 7.5v and it is a bitch. It has a thin wire filament. The #50 has a 7.5v filament but is oxide coated and much thicker. The #50 is less hum prone than the thin wired #10.

Ok, it is might be true but it would presume that all tubes have the same length and type of filament. In reality it is not accurate as there is 5-6 different type of filaments not to mention that there are a few materials the filaments were made from. So, I think the “4Volt technique gives best ratio between diameter and length of the filaments” but it is NOT the only criterion that determines the AC noise. At least it is how I think.

DHT4ME wrote:

I have never bought any euro DHT's but have heard nothing but high praise for them. I one day would like to try the RS241. It is supposedly stunning.

I am very glad that I put the idea of 4V into the Milq channel that was my parley into the world of European DHT tubes. My initial sentiment was to put the YO186 into the game but then all those Telefunken, Klangfilm (which is BS name), Siemens, Masda, Ferranti, Marconi, Triotron, Osram , Valvo, Philips, Tungsram and others begin to pop up.  I always was a fun of German tubes of Telefunken/Siemens and I figured out that it might be a good chance to dive into it.  In the tube of the power level that was interning Germans had RN604 and AD1. The AD1 has own socket but RN604 has the same socket with YO186. So it was no brainer what to do next.  BTW, the RS241 is replaceable with RN604…

BTW, the 10Y/VT25 tube that you mention was the only tube that I would like to try (I got a stunningly good party of them – probably the best tubes that I ever bought) and this is why I initially wanted to keep 6.3V for my MF channel. (It is 15A transformer and it might be driving 10Y with 7V). Then I discovered that 10Y would need very different operation point to do it’s best. So, I kept my 10Y objectives out for now. Still, the preliminary results I am getting from the 2.5V and particularly 4V tubes, the tubes that I already have, suggests me that I will be a very content pussy with this amp.

I started this thread on November 22 last year, or about 6 month ago. 3 years back I went for single stage amps and 6 month ago I begin to conceptualize a move for a higher gain for MF and consequentially the DHT in output stage. Today the new MF channel for Melquiades is complete; the amp is finally closed up.  There was an interesting mystery there. I put a large grid stopper - 1K and that made very unexpected change – the gain went up for 2-3dB and frequency response got extended up. I measured is a few times. The crossover point is there and not changed with or without the grid stopper. The minus 3dB I get at 33kHz without grid stopper and at 63kHz with grid stopper. I have no explanation as I increased the grid impedance and the filter against Miller shall be faster. Anyhow, I do not complain as you understand…

Anyhow, the DHT channel is powered now with own dedicated filament transformer and I actually was able with it to get noise even lover – to 0,2mV. There is absolutely nothing coming from MF channel. Still, I took high a sensitivity woofer and the very-very minor 60Hz hum was there. So, all fear of people who have problems with AC on filaments and consequential modulations are warranted, thankfully it does not affect me and I stay with AC.

I am sure that for a next few weeks I will be talking about the “new” sound from Macondo/Milq and the searched 2dB of MF did changed a lot. I need some time to play with different configurations but as what I see now suggests that the project exceeded all my expectations – it truly sounds very-very nice.

I just played the Verdi’s Requiem with a killer cast of Ludwig, Ghiaurov, Schwarzkopf, Janet Baker, Gedda lead by Carlo Giulini. That was exactly what a doctor ordered. I have million thoughts about the change sound but I will withhold conclusions until I will be in the full control of what Milq/Macondo. So, far it is exceptionally good.

I have to note that despite that the double-stage DH MF channels look like does better in context of Macondo but it does NOT answer the question that I raised in the title of this thread: “Single-stage Melquiades vs. DHT amps”. My direct–coupled DHT is not just most stages but it is a different amp. The double-stager has much more idling output stage, it has completely different output transformer with different core, it has essentially only polypropylene and foil in PS, it has different crossover coil with different core and twice higher Q, it has exposure to different flavor of output tubes. So, I would let settle my excitement but who know what I would say if I was able to get from my single stage a few more dB and at the same time to load it twice lighter.