ml8 wrote:

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

This thread is a derivation of the “5 Channel Version of Melquiades” thread:

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

Following numerous signs and leads I decided to try how one stage Melquiades might behave. I converted my failed headphone amp into one stage Milq. It took less then 2 hours: to toss the 6C19P, put in the Melquiades driver and to reset bias and plate voltages.

I decided to go for 6E6P just because I am lazy and the 6C19P and 6E6P have virtually identical pins layout. I drive 6E6P-DR at ~7.1W (42mA and 167V) with Amorphous Lanhdale LL 1623 transformer of 50mA/60H connected 17:1

The amp after the first re-assembly worked perfectly: both channels have identical current and identical gain. The amp drove 95dB monitors, sure with no juice, no dumping and no volume. I connected the amp for a very short time for Macondo: the volume was more then fine; in fact I would say 6-8dB louder then with Super Melquiades – enough gain! No noise, no freakishness during start and fade, no ferrite bids and no oscillations… Very nice, indeed!

The biggest question would be how good the Sound would be. I did not listen it and decided to let it burn over night and to listen it over weekend. It will be very interesting journey…

Yes, I ma very enthusiastic to see what resume I will get. I would not use a single-stage amplifiers with low sensitively speaker. Low sensitively speaker and high sensitively speakers are different school of thinking in term of amplification. They have different demands for power supply and different amplifier compliance (and I do not mean just dumpling)…. I have a feeling that amplifier for high sensitively is like walking a slippery, covered in ice slope. Any rough move leads to undesirable consequences…

Anyhow, I concluded that I drove the 6E6P-DR too hard and now I am at 30mA, 180V. The B+ is SS, fast-0soft switching, LCRC filtered. It will be fun to listen the thing today.

BTW, I would like to personally express my gratitude to you. Over the years Dima and I were thinking about those experiments but your post was the actual trigger the made me to reconsider it again and try it…

Also, since you drive then “naked” then may I ask you how the 6E6P/6E5P dies if they have no second stage. I mean if I load them to the HF driver with line level filter then do they do HF “pops” when they are ready to go south?

Also, Dima, since I have multiple, sections on the Lanhdale secedes keeps pushing me to try the ultraliner version of a single-stage Milq… I will probably will… if the 6E6P/6E5P will sound itself more or less usable….

I put the Super Milq and the new one-stage Milq today and was playing them together in order to see which would be more interesting. It is imposable to “compare” as is the upper bass and  lover mid channels as the Super Milq has all kind of line level filers built in. Then only channels that very intentionally free from any filtration in the Super Milq is HF channel I was able to play Vitavox S2 in it’s 400Hz horn with 1000Hz filter driven in absolutely identical condition with both amps: HF channels if Super Milq and the one-stage Milq (when I talk in this post about Super Milq I mean the HF channel of Super Milq)

I would preserve my final judgement for later (there are reasons) but the initials impression is very much unexpected. I was predicted everything but not what I got.

So, currently the one-stage Milq has 1.33dB higher gain then Super Milq and equalizing the volumes I learned that the sound was VIRTUALLY IDENTICAL. In fact to my surprise I was VERY impressed (and completely did not expect) with the transparency of the Melquiades output stage. It hard to judge it neutrality as it is over 1kHz sound and it is possible to do only just the “technical listening”.

Anyhow, the preliminary result is that HF Super Milq and the new one-stage Milq sound the same. There is some minor and near negligible difference but to assess this “difference” it I need to do something… I will repost it later…

So fat – a big surprise to me… Hm, if the damn stage around 6C33C is so transparent then it would be a good idea…. to make a line level preamp around it….

Well,

I spent today a number of hours trying to advance my observations regarding abilities of the one-stage Melquiades. It was quite interesting. Thankfully the electivity today was way better during the second part of day. I loaded the 6E6P heavier (it has a “fascinating” reaction to plate loading) and…. eventually I was able to found a configuration where I do feel that one-stage Melquiades performed better then MF and HF channels of the Super Melquiades. I have to note that the DIFFERENCES ARE MANOR but they are notable.

The one-stage Melquiades in comparing to two-stage Melquiades has:

1) Less granularity of sounds
2) Better ability to care stress, it less “confusing” during crescendos and those chorus' mess
3) "Cleaner" highest frequencies
4) Some sense of more advanced refinement and delicacy
5) Less tackiness
6) Less syntactic upper mid range.

Once again: those advantages are VERY-VERY minor but they are there. I DO feel that the amplitude of those advantages, regardless not low they are is a reasonable incentive to pursue a single-stage Melquiades with high sensitivity. I do recommend to the folks who have near 110dB sensitive speakers and who do not have a huge room to pay attention to those observations, I will. I do not know at this point if I go for it but to make a 4-chennal single-stage Melquiades will be remarkable simple. Well, perhaps to have juts a single resistor and a “smart” tube in a signal pass does matter….

Once again, my applauds to the 6E6P/6E5P…

Well, the more I think about it then more I realized that I would like to try it. I’m contemplating to make 5-6 channels on single-stage Milq: 4 channels for each Macondo channel and 1-2 reserved for any perspective lower bass channel or for perspective 6E5P paralleling for bass horns. The more I think about it the more I realizing how easy it would be. Another very motivating for me factor is that is that if I go for it I was thinking to build-in the single-stage Milq inside my speaker frame…. It would be very interesting solution with virtually no speaker wires and direct coupled drivers.

Also, ml8, since for 30mA I have enough core mass in the Amorphous Lundahs to get any gap and any inductance then I am considering what would be the optimum gapfor OPT. Currently I use Lundahl LL1623 in “D” configuration, 50mA at 60H in primary. I was considering keeping it for upper bass channel (with single tube) and for higher frequency channel to go for 40-50mA but for much less core mass and much less turns (inductance). I will not buy transforms for reserved channels of course and will keep the 5th  or perhaps 6th channels only as a placeholder.

The fun part is that the more I think about that 6 channels on single-stage Milq (with no bass channel) the more I realize how simple it might be… It I design a good layout for building it in my Macondo then it might take literally one day to put it all together…

ml8 wrote:

You see, i'm always struggling to find sound without that damn granularity in it, but so far i did not succeeded. Only few of my amps (DHT with kenotron rectifier) were almost free of granularity, but still they all have other flaws and i took them apart for later projects. Still i don't know exactly WHY they was sounding so smooth -- is it DHT, or kenotron, or something else?

Actually granularity was never a subject of my frustration. I do not think that granularity or it’s lock important itself but rather it is important in context of everything else, like harmonic signature of amp and many other aspects… Else, there is granularity and there is graininess…

I very much do not feel that finer granularity and more refine graininess are unconditionally better. Most of the OTL beat anything else in graininess and granularity but most OTL are hardly useable with conventional speakers because other sonic facts of OTLs. Generally fewer capacitors have very positive effect to granularity, and type of PS and…. God know what else…

BTW, admitting that single-stage Milq has better granularity characteristics I do not automatically claim that it is better version then a two-stage Milq. You see a full range Milq was tested musically and was exposed to very ruthless assessment of it musical neutrality and many other factors. The single-stage Milq did not go over it as I used the single-stage Milq so far only as dedicated amp for a number of single band-pass channels. I did invested some time and effort to assure that the given channels driver by full Milq and the single-stage Milq has the identical harmonic structure (30mA at 16R into ~3000R). However, it is not the assurance that the 4 channels, operating full (over 100Hz) range will not have the problems of some others low granularity amplifiers…

I did an interesting experiment today trying to learn if those Melquiades driver tubes have even gain. The Russian tubes are notorious to have very widely spreading specification, so….

I took 14 6E6P. Some of then were new, some used and some violently used. I stack them one by one into the single-stage Melq, drove it with 100Hz loaded to 12R and measured with my db meter the differences between each tube. To my satisfaction only one of the tubes had after 10 min of burning .25dB difference in gain. It is 1/4db among the 14 tubes… VERY NICE.

This is very good sign an after the 6C33C it sounds like a blessing….

Now are the problems. To my great dissatisfaction Lundahl does not has the transformers that I need. I might used some of the amorphous Lundahl transforms but it for HF channels will not be an ultimate solution.  Not to mention that sine I will be building in the single-stage Melquiades INSIDE the Macondo frame the side is matter and for the given project the overly-large Lundahl core is not a virtue. (Lundahl does not do low impedance small transformers)  I will summarize my install draft what I might need, perhaps someone would point me to an interesting direction.

So, the single-stage Melquiades will need the following transformers: (all channels will drive 30mA and all transformers are 12-13 to 1)

1) 100-500Hz channel high-passed at 50Hz: amorphous core 100-120mA (with possible double tubing), with relatively high primary inductance. I migh use Lundahl 1623A.

2) 500-1000Hz channel band-passed at line-level, amorphous, with moderate primary inductance. I migh use Lundahl 1623A with 200-250mA gap but I would prefers to have 50-60mA with less core mass.

3) 1000-1200Hz channel operating full range, amorphous, with low turns on primary and good size core to be able to care full range though the amplifier. I should be gaped for 50-60mA. I have no prospects at this point.

4) 12500Hz and up channel operating high-passed at 12500Hz, amorphous or powder (or air?) very small core, with ultra low turns on primary. I should be gaped for 35-40mA. I have no prospects at this point.

And email from a site’s visitor:….

“And for my full-range Melquiades, should I still try and build the original version or wait to see result of what you are working on now?”

This is a complicated question about 2-stage Melquiades vs. single-stage Milq. Both amps have the same sound. It was said before many time that the sound of Melquiades derive from driver stage not from out stage.  Dima has built in past a few renditions of Milq with different output stages, including GM 70 and SS and he reports the all amps sound in the same way. Still, will a single stage Milq preserve all qualities of the tested 2-stage Melquiades. I do not know the answer at this point. There are some minor advantages in the way how the single-stage Milq produces the sounds but I have no references about Sound of the single-stage Milq. Once again: the Sounds vs. Sound – the very big difference. Since the single-stage Milq will be DSET by nature it’s final sound might be asses only in contexts of Macondo tuned for the capacity of the given DSET (and vise versa). I do not have this experience at this point. I am wiling to try if I find the right transformers for HF channels.

Some other considerations for people who might look at the direction of single-stage Milq.

Look at the driver of your upper bass horn and make sure that you will have enough power, gain, dumping … and juice to drive this channel with single-stage or with single-stage with parallel tubes if necessary. Another thing is power. With single-stage Milq you get whatever you get power-vice and it is it. There is no room for maneuvers of any kind. If you have the same sensitively channels then you can’t for instance decrease plate loading for a selected channel as you will loose gain. With 2-stage Melquiades you can more idle the tube and more open up the input voltage divider but with single-stage Milq it is not the option as the channels amplifier the full grid throttle. Also, if you consider a single-stage Milq then watch the room size. In my room single-stage Milq has enough gain to drive the Macondo loud enough. Should I have 3 times larges room then I do not know that it would be the case, most likely not…

I think if I had problems with room size and if I had a preamp with gain then I would be able to get 6-10 more dB from preamp. However, I do not know if the single-stage Milq would have enough power to care more grid current. If someone is interested I might make some measurement of the single-stage Milq, it would not be difficult…

be wrote:

ml8, could you tell what voltage and anode load you used the 6s45p without bias? What about grid stopper, etc.?

be wrote:

Thanks ml8, do you know how much it could swing before problems, either due to clipping or grid current?

I would be interesting to know how much plate voltage (and power) 6E5P would have with no bias. If ml8 would promise to try the Milq’s biasing then I will promise to try the no-bias operation :-)…

Surely I would not have in the no-bias operation a power necessary for LF channels but for the HF channels might do it…. if it sound in the same way as the LF channels. I wonder how difference is in sonic signature of 6E5P when Milq’s biasing is being use vs. the use of no-biasing? It is very much possible that they will have different sounds, which would make the stages unintegrateable in context of DSET application… The Milq’s biased 6E5P sounds VERY different then a regularly fixed-biased or cathode biased 6E5P. Ml8 suggests that a drastic change on sound character happens with 6E5P with no bias as well…. It would be interring to try it and to see in which direction the tube goes…. BTW, ml8 did you try driving more current thought anode and getting more power while you were in 0-bias mode?

RonyWeissman wrote:

Hi romy, thanks for info. I am going to stick with the 2-stage plan then as it will not be DSET but full range and my mid-bass is somewhat barbaric at this point.  I have found someone to build for me, though we have not discussed fees yet.  I'll probably be using the Tribute transformer as well.

Yes, sure, if I go a full range amp then I would defiantly stick with 2 stages version. The single-stage is in away too radically-specific way to go and I am not convinced at this point that it is a wining way to.  A single stage has very limited scope of applications - a two stage Milq would make way more universal amp. There is NO HUGE difference between in sound of 2 stages and one stage, the difference does exist but it is way less than I would initially expected…

One of the many incentives for me to try a single stage because the single stage has no excessive power. You see, I for instance would like to couple the voice coils of my drivers to the secondary of my amps, implementing all filters at line level. Currently I split filters doing some at speaker level and some at line level. I can not drive with my 6C33C channel directly my compression drivers because the 6C33C has too much power and because the 6C33C might become violent when it dies. With the sub 2W of 6E5P I am planning to couple the transformers to the drivers and it might be interesting. There are some other reasons: the MF high-pass mystery that I never was able to resolve and few other thighs….

Still, doing whatever crazy and Macondo-dedicated things I will be doing with one stage I will have my full-range Milq as a tested reference trying to get result no worse then the one that I got from a “default” Melquiades ….

Paul S wrote:

I think some of the best horns I ever heard were reconstructed "theater" speakers that had their amps wired directly to the voice coils, just as you describe, above.   This was my first notion of the term "dedicated amp", because of this experience.  It may be that the amps in this case were PP, and almost certainly they were 2-stage, although I can't be certain this many years later.  In any case these amp/speakers projected remarkable vitality and immediacy, with very limited resonances for horns of that type, although I have no idea why this would be so, just an observation.

Romy the Cat wrote:

Still, doing whatever crazy and Macondo-dedicated things I will be doing with one stage I will have my full-range Milq as a tested reference trying to get result no worse then the one that I got from a “default” Melquiades …

It was quite serious, rich and meaningful Sound the Sound the ignited within me lethargy to destroy the Super Milq, pulling form it the MF output transformer for the single-stage Melquiades. The Sound was so balanced and so “compiled” that to take it apart would be an act of real barbarianism. Frankly speaking I was so comfortable and so contented that I was wondering if I even what to go for the single-stage Melq… Sure it might be interesting to try to see what would happen  but it would open a Pandora box for many other prospective “issues”.

It is quote nice as it with no extra efforts invested… I do not know… I will not hurry this case…. If someone make for me the necessary uncompromised MF and HF output transformers then I would make the single-stage Melq. If not then I will keep what I have now and will keep bitching that the world is the place filled with mean and non-cooperative people….

Ah, how wonderful would be if I could write a check to someone and that someone would bring the build amp in my room in a few weeks/month…

Now I re-listen all the things the we try to measure. To surmised everything I have to tell that Dima’s way to looks for optimum operation point of the tube (setting balance for voltage vs. current sufficiency) turned out to be very successful as I did not see absolutely no discrepancies between auditable and measurable. As long the tube has the identical and symmetric current and voltage clipping it does sound the best. I have no idea why audio  people create so much “cult” about finding the best operation points for tubes and why any controversy exist – it look so simple and so unambiguous – for one stage of course.

Anyhow, I was not able to listen the Single-stage 6E5P with horns. The tolerable volume level I got driving Vitavox with 0.6-0.7V… VERY far from any dangers regions. So, I took 16R speaker with 90dB sensitivity and drove it with Single-stage 6E5P. First - the 0 bias on the tube – it was absolutely unusable, sorry ml8…

The best sound I got from this tube with ~ 6.5-6.8W dissipated on plate. More power made the  sound more stiff less power made… less power…Now current vs. voltage (with 16R load) keeping the same power. More voltage and less current made Sound “brown”, more compressed and in a way “inverted”.  More current and less voltage made it sound suffocated. In a way deficiently in voltage or current both produced that over- damping effect at upper bass and lower midrange but that over-damping sounded different for voltage and different for current. It is have to explain in words but it is VERY much the same as over-damping of bass when we adjust VTA of cartridge vs. lighting of bass when we change the tracking mass.

Ironically the best subjective sound I got in the best measurable point. I did a few blind tests  with different tubes when I was trying to set subjectively the best sound adjusting bias without looking at miliampermeter of the plate current. In all cases, finding this operations point and then driving the amp into distortions I has the perfect symmetrical pictures of voltage vs. current. It all was at the same current but at different bias settings (from minis 3.1V to minis 3.8V). So, the “magic” number was 175V and 38mA…

Now is the quandary… Do I have to for the Single-stage Milq pre-select the tubes or I need to put bias adjustment in there? Perhaps only for upperbass channel where I “might” be important. Now is the biggest question: now important all of this? All those experiments took place with amp outputting 3V but during it live the Single-stage Milq will do less then 1V. I do not have a full range 109dB system to assess subjectively how will sound. Listening single channels I can say that the deviation form perfect voltage vs. current setting is WAY less (if any) auditable at 0.7V. So, I need to think how important it might be in the real world.

Jim, returning back to the subject of tube testing….

I do not know if you have seen my most recent obsession; the idea of driving my Macondo with 1.8W of a single 6E5P:

http://www.goodsoundclub.com/LatestPosts.aspx?ThreadID=4241

So, what I learned is that all 6E5P, regardless of vintage and lever of being used up have the same 30 times gain. So, it is a drastic improvement over the very uneven gain of 6C33C. Still, all 6E5P have different ability to produce current – the stupid Russian tubes! So, sine I know the exact operation point for this tube and sine I will have the adjustable base for each channels of the single stage Melquiades then what might OBJECTIVELY indicated to me the life span of the 6E5P? The cathode emission? The transconductance? The use of getter? The darkening of the glass? Where would be an objectively correct point when a tube consider “gone”: 10% less of cathode emission, 30% less of cathode emission, 50% less of cathode emission? Is it wary from tube to tube and some of the sound fine with 50% less of cathode emission and with less bias?