(Quote from your post) "... by building any serious horns you will very fast overgrow the ML2's ability to be flexible."

Yes, this is in fact what I meant by "limitations imposed by of a pair of SET amps such as the ML2s".

Would it be possible to adapt your plans to build a pair of Melquiades with 5 dedicated channels?

Regarding simplification: I would have no interest in simplifying the design if it would not result in an audible improvement.

If you could supply the contacts for the "approved parts", I will order them, and have them shipped to a family member in the U.S., and later (once everything is in one place) shipped on to me as a single shipment.

The chassis and covers could of course be made locally.

Regards,

jd*

This thread is split from the following thread:

http://www.goodsoundclub.com/LatestPosts.aspx?ThreadID=2665

Jessie, I hope you do not mind that I pulled it to the forum explicitly dedicated to the Melquiades Support.

5-channels Melquiades is great idea, I am sorry now that I did not made 4-channels Melquiades as I need now to drive my “Fundamentals Channel” and I do not need one more channel in the same chassis. Still, the 5-channels project would be costly and labor extensive and if I were you I would not dive into it without confirming that a regular single channel Melquiades sound acceptable to you. I would make a prototype with no-chassis and would spend a few months to listen the thing. At that time you will be having you horns and Lamm ML2 and you would be able to observe if Melquiades is valuable to pursue. Who knows, it might be that I’m a fool of it and you did not find the Sound you get from Melquiades worth to continue. A Milq is simple and if you have the Lamm ML2 as a reference (and it is a phenomenal reference!!!) then you would very clearly see if Melquiades is the direction you would like to go. There is another benefit. Since you live on another side of the Globe it is possible that you would like to use some of your local parts and the Milq prototype would be a wonderful opportunity to try some of your choice.

A few words about 5-channels Melquiades that you might to consider. You can pile and many channels as you wish but you have to think in term of packaging everything. Here are some very raw thoughts:

1) You need to decide if you go for single chasses vs. separate chassis for amp and powers supply and if you go for stereo amps vs. monoblocks.  I went for dual chasses monoblocks. If you select the same direction then be advised that you must keep that last largest capacitor of each power supply on the amplifier chassis not on the power supply chassis.

2) You need to decide how you would organize transformers. I went in Super Melquiades for a dedicated transformer for each amps need; however you might combine the positive bias supply with plate of the first stage. I do not see why it might be problem. The filaments would be defiantly a very dedicated transformer. I would make a single huge 20A 6.3V transformer and would drive all tubes from it. There are some benefits to drive 6C33C with parallel 6.3V instead of 12.6V as you would be able to connect one or both heater as you wish.

3) You need to figure out how you would power output stage. Since you do not plan to do lower then 40Hz in your Super Melquiades it does not sound like a complex problem. Make a dedicated 1.5Atransformer with 2-3 power taps and drive all second stages from it. In super my Super Milq I have a single 1A transformer that drives 2 rectifiers and two LCRC chains. One chain drive LF channels and another all the rest. You might do the same or even go for one single supply of the second stage since you do not go for very low frequency that usually “milk” power supplies. From a different perspective… it is “Class A” what kind milk I’m talking about!!!?

4) You need to decide how you will be implementing the input stage. This is the most intricate. Theoretically if you decide to go creasy then you might put a separate pair of gas tubes for each channel. In my Super Milq I decided to use one pair of the gas tubes for all 3 channels. It is a compromise in a way. Take a look at my original plan how I envisioned using it.

It looks confusing but it is not, you will see the very same Dima’s “resistor” bias (groups R2-R3-R4, R5-R6-R7, R8-R9-R10), then you will see the voltage dividers that EQ the deference in sensitivity between the LF and M/F channels (R11, R12, R13, R14), and you would see some filters (R1, C21, C5 and C3). It was exactly how it was built initially but then some changes were made. The resistors were divided to their common dominators and I ended up with the following, this is exactly what I have now running in my amp:

As you see now I drop bias voltage right on the filter resistors and the input look very slick and simple. It took some time for me to find out exactly how I would like to balance the channels (those dummy 6C33C all have different gain) but I have an idea what I would need to crossover them. Then Dima helped me to recalculate the values of the resistor to keep the crossovers in pace and provide for the 6E5P the same bias.  You would need to do the same – start for the fool scale - as you do not know what would be difference in sensitively between your channels. Throatily you would as you will have the horn made at that time but you might observe that it you go for line level filter then you might slightly revise the crossover points (I did). There is a catch in there. You would like to keep the last resistors before grid (R10, R4, R5 and R9) as high value as possible. With supplying 150V into the 3 or more channels the values of them go down…. Perhaps you might consider putting 2 series gas tubes and making then 300V but I do not know how they sound connected in series. With the Milq basing scenario everything is very freaky and whatever comes between the bass tube and the grid has HUGE impact to sound.

5) You would need to think how to cool the thing. Five 6C33C generate some heat. It is possible to use natural air flow (make more holes) to do it but I personally like to keep everything very compact and direct for the sound chain. For instance my coupling caps (including 3 caps by 2uf for bass channels)  sit very close from tubes:

The caps that I use are 85Celica rated; they might be 105Celica with 50% voltage duration. I did not go for 105Celica and used 85 degree. The caps sit on the silicon bath and blow by fan. I have the amp running with no fan and it does fine (the bottom panel of my Super Milq has many ventilation holes that I seal when I use fan). However, I messes temperature after many hours of operation and I do not liked what I have see. It smells “different” and it had once that one of the leg of 6C33C filament got almost unsoldered (!). Nothing failed but I feel that with forced cooling the amp behaves more temperature stable. With fan that operate at 11dB noise it is not really big deal to use fans.

6) You would need to do a lot of drawning in the real-size to do, in order to figure out the size or your parts and how you would layout them. I put together some very initial ideas and you might use them as a starting point:

Anyhow, if you need any help or objections then I would be happy to assist you. Perhaps some other guys who have more experience how to build the things jump to the wagon of your consulting team. I know Jim, Dima, Bud, Chris and some other folk visit this sit and their contribution might be very valuable. Perhaps I would learn something…

Still, I would like to stress this point VERY aggressively you need to male a single amps prototype and listen it. You must have a reference point. Partially you need to do it in order to confirm that each of your channels sound as good as the reference full-range range amp. I wouldn’t make 5-channels amp without knowing what a single channel full-range range amp is capable off…

jessie.dazzle wrote:

There are a few good shops in Paris that have shelves and drawers full of components. And, across the channel ….

http://www.tribute-audio.nl/

However be advised that all that wisdoms about the “good reputation” worth very little. People prefer one of another transformer primary in context of a full range transformer, as any SET transformers is juts a compilation of compromises. However in context of DSET, when you have a dedicated narrow frequency range transformer for each channel then rules of the game are different and you might go away with much simpler transformer and get much better result at the same time. You just do not care about the bandwidth of the transformers anymore.

jessie.dazzle wrote:

Regarding the architecture of the 5-Channel beasts : I should start off by confessing that I am not at all an electrical engineer. I will need to think and ask advice regarding the details. In the prototype stage, though it might be more costly, I have a strong preference for keeping things as direct as possible, dedicating components to only one channel. This approach has the advantage of being easier to get one’s mind around exactly what he might be doing.

Hm, if I were you I would go for horn first. Then you will see what channels you might need. I personally do not feel that you it really make sense to make 5-chanls DSET if you wish do not lower then 40Hz. For instance I run MF and tweeter from the same channel with a transformer of 1.5H inductance. You have to have reasons (crossovers, drivers and so on) to isolate the drivers to desiccated channels and it is not necessary should be 1:1. For instance I will be considering to drive my “fundamental canal” from my current upperbass channels and I will go for a separate amp ONLY if I see/hear the benefits. Perhaps the benefits will manifest themselves (crossover + ability to drive the channels with different current = manage the transients) but how critical it would be? The more you go up in the frequency the less critical it will be. The 3 of your channels will be over 1000Hz and they might be integrated (if the crossovering allow to do it painfully)

jessie.dazzle wrote:

For cooling, I will need to consider the fact that I live with two cats (yesssss), and a fan would likely require an external filter, (which could also serve to attenuate the fan noise). Because I am less interested in keeping things compact, this should not be a problem.

jessie.dazzle wrote:

Could you send the list of “Approved Parts” (and where to get them) for construction of a pair of standard Melquiades.

http://www.goodsoundclub.com/Melquiades.aspx

BTW, Jessie, this morning I opened a Super Melquiades and made some changes in plate current measurements. While I did it I took a shot of the Super Melq inside. The former image was made juts when I finished the amp. The current shot is actual stage of the amp as it has been used for a year. You might or might not find the idea useful but it is what it is…

Red is the signal. Green is the B+

jessie.dazzle wrote:

For the past 10 days I have had to temporarily stop production of the horns and put off researching components for the amps; all this is due to my moving to a new apartment sooner than expected. This move is actually very much a part of the entire project, as there is no way the horns would have fit in the old place. I am looking forward to getting back into it (probably another week before I am up and running again...)

Yes, Jessie, but in case you decide to go for DSET, would it be multi-channel Melquiades or anything else, then your consideration for type and topology of the lower bass would greatly dictate requirements for your DSET. If your lower bass channels run from 40Hz and down then the PS, size of the capacitors and the size of the output transformers for this lower frequency channels would be dominating in your amp. Also, you MUST take under consideration many other factors: efficiency of your LF section, reverberation times in your room at LF, the dimensions of your room (actual and virtual), power handling of your LF drivers, the topology of your LF section and many others.  For instance my LF Super Melquiades channel has somewhere slightly over 20W power and it is perfectly fine with my 6-drivers line array and relatively small room, in fact I have a lot of spare power. However, if my room would be 3 times larger or much less bass efficient then I do not think that the power I get from a single 6C33C would be enough for me.

Since you, looks, like will be dealing with very large room I do not know how it all will play to you. The 6C33C is a fairly not powerful tube and in the way in which it used in Melquiades (fixed bias) it gives better sound but less toleration to work with high AC grid voltages. It is phenomenal rube for bass that does down to 100Ohm on plate and makes the LF transformer very simple to make but it will be still 20W, and in many instances less power if you would like to drive it reliably.

If you find that you need more power then you might go in your LF channels for parallel 6C33C, getting out of it 35W-40W or go for more powerful tube, high voltage tubes for instance. It would be if you would like to stay in the SET world for your LF channel. It is possible that you might decide do not stick to SET for lower LF and go for PP tubes or PP SS amps. It is an open question if properly made SS, hybrids or PP tube amps might be a good complimentary lower LF amplifier to a system where DSET drives the rest of the systems. In fact I will be trying to address this question for myself in context of my system: the LF channel of Melquiades was better then any SS amp that I tried and was better then Zaratrusta Hybrid. However, the new implementation of Zaratrusta is coming (made at very different level, 150W of pure class A, with many other new elements of design) and I really wonder how it will be able to stand up to the sound of Super Melquiades’ LF channel.

In context of this thread: you need to decide much power you would need from your LF channels and then ingrate or do not ingrate this solution into your multi-channel DSET

Rgs,
Romy the caT

Romy,

Sorry for taking so long to get back to you… (this being the summer time in a Latin country, I still have no internet connection at the new place).

Quote from Romy “If your lower bass channels run from 40Hz and down then the PS, size of the capacitors and the size of the output transformers for this lower frequency channels would be dominating in your amp.”

To clarify things, here is the way I plan to handle frequencies under 500Hz :

Upper Bass : Frequency range between 100 and 500Hz – Driver loaded into a 120Hz straight horn, driven by DSET dedicated channel.
 
Mid Bass : Frequency range between 40 and 100Hz – Driver loaded into a 45Hz straight horn, driven by DSET dedicated channel.
 
Lower Bass : Frequency range under 40Hz – via a pair of large direct-radiating woofers in sealed enclosures - driven by a pair of 100 watt hybrid monoblocks (Lamm M1.1s).

Based on the info in the thread “Exceptional loudspeakers drivers: 18-inchers”, I would of course like to try the Aurasound 1808, but I am not at all sure I can find a pair of the original model… In that same thread, you mention the 833A tube could possibly be used to good effect in a DSET channel driving such a woofer… ultimately, I would like to consider this solution, in which case we would be talking about a 6th channel. For now however, I think it would be best to take things one step at a time, and leave this channel out of the DSET, driving it with the M1.1s (I don’t mind building a second version that integrates a lower bass channel, but would rather leave that for later).

Room dimensions : The “room” is actually a sort of loft; the main room plus immediately adjoining areas total to around 400 ft², with ceiling heights varying from 8.25 to 16.5 ft. for an average of about 12 ft, which works out to about 4800 cubic ft. How does this compare with your room?

Quote from Romy : “In context of this thread: you need to decide much power you would need from your LF channels and then ingrate or do not ingrate this solution into your multi-channel DSET” and “If you find that you need more power then you might go in your LF channels for parallel 6C33C, getting out of it 35W-40W…“

In the case of the Upper-Bass channel (100-500Hz) : The driver can handle up to 100 watts, and has an efficiency rating of 101db/w/m. I assume 20 watts would be adequate.

In the case of the Mid-Bass channel (40-100Hz) : The specifications regarding power handling and efficiency of the selected driver (Vitavox AK151) are not clear; the manufacturer says power handling is “determined by the associated horn characteristics”… In any case, 20 watts should not be too much for this driver. You are right, the real question is: Will 20 watts be enough?

I had planned to answer that question by connecting a pair of ML2s to the horns, but am having trouble finding a seller (early version ML2) who will ship to Europe.

Thanks for your thoughts,
Jd*

jessie.dazzle wrote:

Upper Bass : Frequency range between 100 and 500Hz – Driver loaded into a 120Hz straight horn, driven by DSET dedicated channel.   Mid Bass : Frequency range between 40 and 100Hz – Driver loaded into a 45Hz straight horn, driven by DSET dedicated channel.   Lower Bass : Frequency range under 40Hz – via a pair of large direct-radiating woofers in sealed enclosures - driven by a pair of 100 watt hybrid monoblocks (Lamm M1.1s).

jessie.dazzle wrote:

Based on the info in the thread “Exceptional loudspeakers drivers: 18-inchers”, I would of course like to try the Aurasound 1808, but I am not at all sure I can find a pair of the original model… In that same thread, you mention the 833A tube could possibly be used to good effect in a DSET channel driving such a woofer… ultimately, I would like to consider this solution, in which case we would be talking about a 6th channel. For now however, I think it would be best to take things one step at a time, and leave this channel out of the DSET, driving it with the M1.1s (I don’t mind building a second version that integrates a lower bass channel, but would rather leave that for later).

jessie.dazzle wrote:

Room dimensions : The “room” is actually a sort of loft; the main room plus immediately adjoining areas total to around 400 ft², with ceiling heights varying from 8.25 to 16.5 ft. for an average of about 12 ft, which works out to about 4800 cubic ft. How does this compare with your room? In the case of the Upper-Bass channel (100-500Hz) : The driver can handle up to 100 watts, and has an efficiency rating of 101db/w/m. I assume 20 watts would be adequate. In the case of the Mid-Bass channel (40-100Hz) : The specifications regarding power handling and efficiency of the selected driver (Vitavox AK151) are not clear; the manufacturer says power handling is “determined by the associated horn characteristics”… In any case, 20 watts should not be too much for this driver. You are right, the real question is: Will 20 watts be enough? I had planned to answer that question by connecting a pair of ML2s to the horns, but am having trouble finding a seller (early version ML2) who will ship to Europe.

Hm, it is interesting… a few weeks ago I was trying to saddest Jessie that if he end up with 5-chanalls installation then it might not be necessary for to have exactly 5 channels of Melquiades.  However, since I moved myself into 5 channels I begin to ask myself if my position was reasonable.

I do have reasons to see how some of the channels might be combined but after a few weeks exploring of the 5 channels I feel that I would LOVE to have 5 channels independent amplification. Moving to 5 channels make more complicated crossovering, at least in my setting and 5 Milqs would be kind of very handy. It is not that am contemplating the project as I do not really like the building ceremony but if you have a local, persuadable technician, competent but with sense of humor and who would be OK to suffer from a minor abuse, then I would undertake the projects. The 5 channels? Hm, it would take some repackaging…

Hello Romy,

My apologies for not getting back to you sooner… Moving to the new place has turned out to be a huge task, but worth it, as this new home actually has a place to work, not to mention a much better selection of recorded music at the local library. I have also managed to round up most of the materials for construction of the horns. 

While all this has been going on, I have been establishing a sonic reference in the new room by listening a lot to my existing system (which sounds much better on dedicated circuits, isolated ground, and connected directly to 220v. mains… was previously using a large power regenerator set for 110v/60Hz).

Right…

Good to see the success with the new Fundamentals Channel has you thinking… If you don’t have a 5-Channel version running by the time I get around to building mine, it would not be that much more work for me to make a second pair in parallel…

By the way, a big  M E O W  for your work with this new channel (I have some questions on the new horns, but will keep them for the appropriate thread).

In my case, because I will soon have more horns, I am now very much wishing I had started by building a pair of simple amps…

To answer your questions on choice of frequency ranges (sorry if this diverges from the 5-Channel thread)…

First of all, this project is in my case primarily about education. Myself not yet being an  "Experienced Horn Guy", what I need more than anything is to establish a reference. Like most people, I have not yet had the occasion to hear a "properly done" horn system. Since there seem to be very few of these around, I have chosen to educate myself the hard way, and just build one myself.

Due to abilities that come with having a very sordid past, it is relatively easy for me to produce horns having a diameter up to 40 inches. In the interest of establishing a starting point, I am taking my best shot, based on what I am at this point able to understand in matching a driver to a horn, and also of course based on your experiences (thank you very much). I am fully aware that I may have to do some of the work more than once. I will of course make every attempt to insure getting the 45hz horns right the first time.

Looking at your notes in "Over 8,500 Words About My Playback", you spent "a few years" just finding a suitable upper bass driver, and god only knows how long you spent considering drivers for all the other frequencies.

There may be better drivers out there, but with my level of experience, I doubt my abilities to recognize them. So though it may resemble laziness on my part (remember, this is the same person that will be constructing a pair of 45Hz straight horns…), it is smarter from my point of view, to establish this starting point, and get on with the physical work, basing it on existing research ; i.e., if you don’t mind, on your research.  It is best that I do this now, because later I may have already developed a "Horn Ego" sufficiently large to prevent such a shameless approach ! I sincerely hope you are not bothered by this… Please say so if you are.

My choice of drivers, as well as physical horn characteristics are therefore based either on what you have ended up with, or what I have been able to understand you would have preferred.

In the case of the Upper Bass Channel… Why 115Hz horn ?

Simple ; I was lucky enough to convince Fane to make a pair of the discontinued Studio 8Ms (16 Ohm). Again, this driver, and the characteristics of your Upper Bass horn seem like a good place to start.

I was originally not going to start by building a Fundamentals Channel… It turned out that In buying S2s, I ended up with 2 pairs that seem very well matched (I bought 4 drivers thinking I would make a selection, keeping the two best matched). So, I now have what I need to make this channel as well. I will base the Fundamentals horn on your observations, and will likely be starting off with a larger horn (yes, further obscuring the tweeters !), while trying to stick with first order X-overs.

The 45Hz horns with Vitavox AK151s are what I will be using in place of multi-driver line arrays. 45Hz may not be the final number… I have the space to go a bit lower if necessary.

For frequencies below what will be produced by the 45Hz horns, I am not immediately concerned (I have enough work for now), but am seriously considering a pair of McCauley 6174s, direct radiating, in large sealed enclosures. These drivers have Fs of 20Hz, a 94db sensitivity rating, high excursion, and a paper cone…Manufacturer recommends 6ft³ enclosure volume, so not huge. And… a big plus, I can get them in Europe. As previously stated, the DSET amps will not drive this channel (plan to use a pair of 100 watt hybrid monoblocks).

As always, thanks for your comments,

jd*

jessie.dazzle wrote:

There may be better drivers out there, but with my level of experience, I doubt my abilities to recognize them. So though it may resemble laziness on my part (remember, this is the same person that will be constructing a pair of 45Hz straight horns…), it is smarter from my point of view, to establish this starting point, and get on with the physical work, basing it on existing research ; i.e., if you don’t mind, on your research.  It is best that I do this now, because later I may have already developed a "Horn Ego" sufficiently large to prevent such a shameless approach !

jessie.dazzle wrote:

In the case of the Upper Bass Channel… Why 115Hz horn ?

jessie.dazzle wrote:

Simple ; I was lucky enough to convince Fane to make a pair of the discontinued Studio 8Ms (16 Ohm). Again, this driver, and the characteristics of your Upper Bass horn seem like a good place to start.

jessie.dazzle wrote:

I was originally not going to start by building a Fundamentals Channel… It turned out that In buying S2s, I ended up with 2 pairs that seem very well matched (I bought 4 drivers thinking I would make a selection, keeping the two best matched). So, I now have what I need to make this channel as well. I will base the Fundamentals horn on your observations, and will likely be starting off with a larger horn (yes, further obscuring the tweeters !), while trying to stick with first order X-overs.

jessie.dazzle wrote:

The 45Hz horns with Vitavox AK151s are what I will be using in place of multi-driver line arrays. 45Hz may not be the final number… I have the space to go a bit lower if necessary.

jessie.dazzle wrote:

For frequencies below what will be produced by the 45Hz horns, I am not immediately concerned (I have enough work for now), but am seriously considering a pair of McCauley 6174s, direct radiating, in large sealed enclosures. These drivers have Fs of 20Hz, a 94db sensitivity rating, high excursion, and a paper cone…Manufacturer recommends 6ft³ enclosure volume, so not huge. And… a big plus, I can get them in Europe.

Rgs, Romy the Cat

Once again, let make it the last post off the subject of the 5-chals Milq.

I think it make sence to start from 370Hz-400Hz horns for S2, targeting it for 1000Hz. Then the lower bass horn, for instance the 45Hz, however it should have a chamber of mach larger size, embeblming to tunes the driver for 80-90Hz. I know it sounds ridicules to make a huge horn and waste it for a high crossover point. Still you might find it very beneficial for Sound. People usually do not do it (and I understand why) but you might experiment with it and you might discover then with "enough" space under crossover point the upper bass "might" sound very “opened”. You might go then to a lover crossover point by juts masking out the back chamber. It is hard to predict how high the 45Hz horn will be working. It defiantly will not reach the S2 channels and you will need ONE proxy channel. Would it be Fane in own horn or a compression driver it all depends or many factors but I would most defiantly try to take advantage of a copmression drivers. S2 would be wonderful for this task but you can not run it lower then 450-470Hz. It would be VERY interesting to try it in sub-near0200Hz horn crossed at 470Hz with second order. I will be like nothing else!!! Make sure that you have the diaphragm very precisely aligned.

ml8 wrote:

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

Paul S wrote:

So this amp would not have to deal with a crossover?  One R and one L in the signal path, it looks like a headphone amp, maybe a dream circuit for a light piston, HF only, given no cap/coil load at the [ribbon] driver.

Why would not deal with a crossover? The amps will deal with crossover in the way identical to the Super Milq does-combining bias and filter just before the grid of the stage.

It will be hell of a mess to circulate everything considering the necessity to preserver the bias, the input impedances,  use different type of filters (perhaps some subtraction filers) but it is perfectly accomplishable)

It might be way simplified if each channel out of negative gas tube. It would be fine to keep one single negative source but I would more then 150V for 4 channels, perhaps 2 gas tubes but I do not know if the nesting of the gas tube would not affect sound as anything the go to the grid if VERY critical and VERY finicky.

Once again where the get the transformers for the thing? It would be 1K to 15R …..The upperbass and fundamentals channels might be cared by an amorphous version of this (inverted):

http://www.lundahl.se/pdfs/datash/1689.pdf

But how about the HF transformers? I would need something similar with less turns…. Anyhow I do not want to think about it until I verify that one signal 6E5P sound acceptable to drive compression drivers.