Hi Romy,

I am planning my own DSET build.  I have based my speakers on your Macondo and am busy turning MDF horns in my spare time.  Once that is over I will move onto amplifiers.  I plan to build a fullrange Melquiades for a friend and then if that works out the DSET 6 or 7 channel version for myself.

Of course I have read the threads and am reasonably familiar with the evolution of DSET Melquiades in the context of your system but there are still a few questions.  The question at the front of my mind is the following hypothetical: 

“How would you proceed with another DSET build knowing what you know now?” 

If you are willing I would like to work through this in stages.  This is a project that will evolve with time as I build the speakers and move them into the new room.  Initially there will be 5 Channels:

Raal Lazy Ribbon - Direct Tube Drive
S2 into 400Hz tractrix
S2 into 250Hz tractrix
Fane 8M into 115Hz tractrix
Scanspeak 25W/8565-00 towers – 8 drivers each

Then, if the results are satisfactory and the final position of the speakers in the room permits I will experiment with:

K15/40 Midbass Horn
Injection Channel

A very familiar system to you indeed, and I hope you don’t mind me using it as my starting point.  I don’t want to spend a decade or two trying to develop a satisfying horn system from scratch so I am using your documented experiences to make a strong start so that with any luck and a measure of persistence my system is matured within a couple of years and I can let the whole diy thing pass for other endeavours. 

So, to start, I am unsure of a few things regarding the operation of your amps.

• Bias adjustment in the first stages.  Looking at the final 6 Channel circuit diagram the bias for Channels A, B and C are all fixed and non-adjustable.  How do you manage the varying gain of the 6E5P tubes?
• Gain structure.  What is the output from your pre-amp?  Mine is 1.5V RMS and 33 Ohm.  I am trying to get an idea about my required size for the 30K input resistor.
•   ULF through the Scanspeak line array.  My calcs are that at 15W, the SPL at the listening chair would be about 98dB at 20Hz, and much higher at 100Hz.  This is loud enough for sure, but in your experience is the 6C33C/Scanspeak combination good near the amplifiers maximum power?  

Regards,

Anthony

anthony wrote:

I plan to build a fullrange Melquiades for a friend and then if that works out the DSET 6 or 7 channel version for myself.

anthony wrote:

If you are willing I would like to work through this in stages.  This is a project that will evolve with time as I build the speakers and move them into the new room.  Initially there will be 5 Channels: Raal Lazy Ribbon - Direct Tube Drive  S2 into 400Hz tractrix S2 into 250Hz tractrix Fane 8M into 115Hz tractrix Scanspeak 25W/8565-00 towers – 8 drivers each

anthony wrote:

Bias adjustment in the first stages.  Looking at the final 6 Channel circuit diagram the bias for Channels A, B and C are all fixed and non-adjustable.  How do you manage the varying gain of the 6E5P tubes?

anthony wrote:

Gain structure.  What is the output from your pre-amp?  Mine is 1.5V RMS and 33 Ohm.  I am trying to get an idea about my required size for the 30K input resistor.

anthony wrote:

ULF through the Scanspeak line array.  My calcs are that at 15W, the SPL at the listening chair would be about 98dB at 20Hz, and much higher at 100Hz.  This is loud enough for sure, but in your experience is the 6C33C/Scanspeak combination good near the amplifiers maximum power?

Romy the Cat wrote:

That is very good approach in my view. The problem is in this approach that you have to have high sensitively speakers that you would be able to use with that test full-range Milq. The behavior of the Full range Milq win this other speaker might be not the same as it would be with Milq.  So, generally I would suggest to have a Milq (or any other amp the you might like) to drive your particular acoustic system with all passive filtration and spend at least 6 month to feel comfortable in this configuration and to figure out your crossover strategy. (Or to develop new listening objectives and consequential frustrations). Then you can slowly to do DSET and line filtration.

Romy the Cat wrote:

My only concern is that I feel that Raal Lazy Ribbon are a bit too wide and too midrangy for use above 12K. The standard Raal Lazy Ribbon drop 10dB at 6K that is not bad but if you made Alex to make the ribbon twice more narrow then it will have steeper drop and you have a bit more “punchier” HF with the lower filter order.  The standard Raal Lazy Ribbon as I understand are meant to be used in two-3 way acoustic system where tweeter is crossed at 2.5-3.5K, so you would need to run 3-rd or 4th order to get the result you want. BTW, I very much might be wrong about it but I do think that the narrower ribbon would be beneficial. Does it worth to pay to Alex extra for a custom job and to be able to get “some” improvement? I do not have the answer to this question.

Romy the Cat wrote:

Yeas, there is no bias adjustment in this driver stage nether I see a need to do it. You need to adjust bias if it going away and in some other occasions but in this amp it is not necessary.  The amp is set to get 200V on the driver stage and it would mean approximately 4.3V in grid. Yes, some of the tubes have a few DB leas or more gain but you do not change “gain” by setting bias. If you have on place 190V or 210V it truly goes not make any difference. Keep neat 200V on place and near 15K load and it would be fine. Saying it I would like to note that nothing shall prevent you to add your own bias adjustment if you feel so. Just be informed that any dealing with bias line is supper affective to sound, in fact ever more then dealing with grid in my mind.  You might gain some questionable benefit by having bias fine-dialed but you would need to put a trimmer in the bias line. Well, I find it is to be at least debatable…

Romy the Cat wrote:

My pream has no gain. My frontend have high output and can swing I believe 3.5V. I think Bidat at 0dB digital can push 5V but this would send the Milq grid to plate currents and you do not want to be there with indirect heating. You can calculate the gain of the amp. The driver has average of 33dB, +- a few DB. The output stage has I believe 3-4times. Then you need to subtract whatever you OPT consumes. You will not have a high gain with this amp as it is two stager. If you have a Macondo-like configuration and very large room (let sat over 800-1000 sq foot) and low output frond end and no active preamp then you will feel that you have not enough volume sometime.  The whole beauty of Milg that you have let say 4V in the input and to have the am on A1 you might run a wide range of the frond end, including high output.  The 30K input resistor is kind of less relevant in Milq DSET as the input impedance of other channels will screw all input impedance picture.

Romy the Cat wrote:

You will be able to drive Scanspeak line array from Milq ONLY in case you have VERY small room, you are very lucky (as I was a few years back where the arrays were at active location of the room) and if it were not a full range but 18-22W LF only amp. Otherwise I would go for a different amp. It need o be said that it is very hard to prognosticate anything,

Romy the Cat wrote:

Make the full-range and then you will see where you are. Still, I feel that for MF channels it is beneficial for do with a good DHT output stage instead of 6C33C and perhaps it is a good idea to use DHT driver as well. So, when you strategy your 5ch DSET do consider that you might want to do non-standard Milq for MF.

Romy the Cat wrote:

In your case, if you have Milq driving the line away of 8 25W/8565-00 towers drivers in the room of near 300 sq and tall ceiling I would predict that you will be in a "dangers zone" and will be running out of gain.

anthony wrote:

My interest is mainly in the 10Y/801A/VT25 tubes on one hand and the 45 on the other hand.  The YO186 also interests me a lot especially after reading your impressions of it in your circuit and system, but it is a very difficult tube to get a hold of and is quite expensive.  The 10Y etc. seem relatively reasonably priced and quite available, however I don't have a ready-made circuit or even knowledge of how they work with Macondo, so I am tempted to just follow your DHT lead with the 6E6P and DHT MF, especially considering that I don't want to be testing/developing amplifiers for much longer than I have to even if I may end up with a better result going the longer route.

anthony wrote:

My primary concern about a DHT driver is the low gain of those tubes (10Y is about 8x compared to 6E5P of 32x) and having to use an input transformer to improve gain and its effect on the crossover for MF.

anthony wrote:

I thought I might be pushing the limit.  Do you think it would be possible to increase gain by using an input transformer on the DSET ULF channel to boost signal level?  I am not sure how this would mess with the low-pass filter. Otherwise I could add a decent SS amp or maybe even push-pull 6C33C.

anthony wrote:

Do you have a “final version” of the schematic anywhere?

anthony wrote:

6E5P/6E6P resistor between plate and grid.  Is this resistor influential to sound?  I notice that at one stage you used a “high-quality” S102 resistor here but later on you are not necessarily using them.

anthony wrote:

Is the resistor in the Fundamentals Channel between the OPT and the S2?  Apart from this resistor you run the S2 without any components between it and the OPT.

anthony wrote:

.LPad for MF S2.  Are you still running the LPad between the OPT and S2?  I assume that the four resistors shown in the schematic at that position signify the LPad and not something else.

anthony wrote:

Yellow circle with cross.  On the schematic at the lower-right side there is a yellow circle with a cross next to a 33K resistor.  I am not familiar with this annotation…could you please tell me what it is?

anthony wrote:

Two sets of gas tubes.  Have you considered more than one set of bias, say one for LF and one for HF?  Do you think it could have any positive or negative effect?  I have no idea.

anthony wrote:

mV Meters on B+.  Is their only purpose to check the plate voltage for the individual output stages?  Would you consider leaving them out?

anthony wrote:

Do you use an hour meter to monitor tube life?  If so I imagine that it would be installed straight after the power switch.  Is this true?

Romy the Cat wrote:

anthony wrote: Is the resistor in the Fundamentals Channel between the OPT and the S2?  Apart from this resistor you run the S2 without any components between it and the OPT. . Actually it is not a resistor connected to OTP, look closer. This is a regular screen lifter, like in your question above, it is juts drown perhaps  not so friendly but as you see the red line goes from plate and directly to OPT.

The resistor between the transducer and OPT is the one that I am not sure of.  Do you still have a resistor before the S2?

anthony wrote:

Sorry Romy, I was not clear enough, but the resistor that I am questioning is in the Fundamentals Channel (not the MF Channel as I believe you referred to in your previous post) between the OPT and the S2.

anthony wrote:

I have secured a couple of pair of YO186 and some 1940's Sylvania 45's to play with.  I wish the YO186 were easier to come by but I think that I now have enough for me to get a very good idea of the capabilities of the tube and circuit, and how they will suit my objectives. Still, I am thinking about the 10Y/801A/VT25/VT62.  Romy, I understand that to bring the 10Y into Melquides I would need 7.5V for the filaments and about 400V for the plate.  Plate resistance is higher in the 10's, 5k compared to about 1.2k-1.6k for YO186/45, and I am not sure how this would affect operation of the Milq circuit particularly interstage coupling.  Reading up I have noticed that some guys prefer much more primary impedance in the OPT than the 10k specified in the 10Y datasheet, up to 16k in fact.  Of course this is in the context of a full-range implementation but I notice that the David Slagle OPT that you use for MF has a 9k primary impedance (if my memory serves me right), which is getting near there already. I have noticed that you have written about your interest in the 10Y, and I am for sure no circuit designer, but given your experience with Milq perhaps you could share some of your thoughts about the viability of putting this tube into the Milq for MF.  Compared to the 45, I am not sure what extra the 10Y would bring as far as sound is concerned (I can only go on "impressions" of other people with other systems using the tubes differently), but seeing as though I will be starting from scratch with my DSET I would be able to add the required voltages for 10Y upfront so long as it does not look too difficult to get the tube to work with 6E6P.

Romy the Cat wrote:

However, if I did it again then I would not do it in context of multi-channel amplifier but rather I would make a maket amplifier able would test my ideas with it.

Romy the Cat wrote:

You see, the problem is that the “best” MF amplifier for you will be NOT the best amplifier but rather the best amp that would feet into the schema of your demands and the capacities of the rest of your channels. It is VERY hard to estimate what would it be, not to mention that your audio expectation might be a moving target and most likely will be recompiled after the first revision of you project will be over.

Romy the Cat wrote:

I would not be advising you about the use of 10Y as I do not have proven experience with this tube. My short and environmentally not-kosher experiments with this tube did show great potential but it was a bit wrong direction to me. The 10Y was super clean tube, kind of aka Type 45 only even “cleaner”. I however was not looking for this cleanneness in my MF but rather I was looking for a lot of distortion but very specific type of distortion. The under-voltaged YO186 gave it to me, but I would never accept that same under-voltaged sound for a good full-range amplifier.

anthony wrote:

I am often drawn to the super-clean sound ….

Romy the Cat wrote:

anthony wrote: I am often drawn to the super-clean sound ….   I do not think that use the term of “super-clean sound” in the same way. Vienna Philharmonic does not have “super-clean sound” from the perspective of most of contemporary orchestras. They have different fingering on clarinets, different reeds on oboes, rotary trumpets, narrow trombones, valveless horns and dozens other differences, some the part of design and some of the part of usability. As the result the sound is very distinctively different very far from “clean” but it has completely deferent meaning when colorations actually do server a very positive purpose and that coloration do fort what we call today Viennese Sound.

Romy the Cat wrote:

It is very difficult to go for this level of sonic creativity with audio but only after one does go there then the use of term of “super-clean sound” with some kind “offence” connotation I find reasonable. I do think that until you experiment with “non- cleanliness injections”  (in one way or other) you truly do not use the term of “super-clean sound” in the same way as I do.

1. 25 turn pot in the filament supply.  For the life of me I can't source a 25 turn pot.  The 10 turn Bourns 3590 is the closest I can find.  Do I need to keep looking for a more sensitive pot or do you think 10 turn is manageable?
2. The fuse.  You show a fuse on this schematic (that traces back to the B+ supply) but omit it on the  6-Channel schematic that I linked to earlier in the thread.  Do you use this fuse or have you removed it?
3. HP Filter 50R.  In the schematic above you use a 50R resistor leading into the 6E6P but omit it on the 6 Channel schematic.  As far as I can tell it is not required for the high-pass RL filter so do you use the 50R or has it been removed?
4. Changing valves.  Could you please explain to me the process of changing valves, say from 45 to YO186?  Please treat me as an imbecile in this regard.  I am very interested in the process and think that it will help me better understand its use.  If I were to guess I would say shut down the amps; remove 45; install YO186; turn on amps; adjust 100R pot to read ???? through the measuring points; press play.  Or maybe it is simpler than that...in some photographs I notice a switch on the "DHT Island" but do not know what it is for.

anthony wrote:

25 turn pot in the filament supply.  For the life of me I can't source a 25 turn pot.  The 10 turn Bourns 3590 is the closest I can find.  Do I need to keep looking for a more sensitive pot or do you think 10 turn is manageable?

anthony wrote:

The fuse.  You show a fuse on this schematic (that traces back to the B+ supply) but omit it on the 6-Channel schematic that I linked to earlier in the thread.  Do you use this fuse or have you removed it?

anthony wrote:

HP Filter 50R.  In the schematic above you use a 50R resistor leading into the 6E6P but omit it on the 6 Channel schematic.  As far as I can tell it is not required for the high-pass RL filter so do you use the 50R or has it been removed?

anthony wrote:

Changing valves.  Could you please explain to me the process of changing valves, say from 45 to YO186?  Please treat me as an imbecile in this regard.  I am very interested in the process and think that it will help me better understand its use.  If I were to guess I would say shut down the amps; remove 45; install YO186; turn on amps; adjust 100R pot to read ???? through the measuring points; press play.  Or maybe it is simpler than that...in some photographs I notice a switch on the "DHT Island" but do not know what it is for.

Romy the Cat wrote:

anthony wrote:  The fuse.  You show a fuse on this schematic (that traces back to the B+ supply) but omit it on the 6-Channel schematic that I linked to earlier in the thread.  Do you use this fuse or have you removed it?   Yes, I use 7.5A slow blow at the very entrance of the 6-Channel schematic with the amp drown in cruse mode in mid of 4A. The slow blow is critical as the amp will start up and begin to saturate the chokes it will draw instantly over 3kW. The normal consumption is 375W if I am not mistaken.

Romy the Cat wrote:

anthony wrote: HP Filter 50R.  In the schematic above you use a 50R resistor leading into the 6E6P but omit it on the 6 Channel schematic.  As far as I can tell it is not required for the high-pass RL filter so do you use the 50R or has it been removed?   Well, the 6E6P is very fast tube and have tendency for microphonic effect, in particularly while it I cold. It is a common practice to put a very small resistor (50-100R) right before the grid entrance to stop any stray noised to come to the grid and arouse the tube. This is call grid stopper. The ferrite bids serve the same purpose. So, the grid stoppers with fast tube are not part of semantics but rather a part of regular design hygiene and they are implied with those type of tubes. On the 6 Channel schematic I just did not have space to depict it but I still would use it in particularly if your wires from the coil are too long as they act as antenna picking up all imaginable dirt in the amp chasses.

Romy the Cat wrote:

Yes, that is exactly how you describe the process. You adjust the 100R in order to kill the noise on the output. What I do is use mili-voltmeter on the terminal of S2 and try to set the voltage in there under 3mV. I think anything above 20mV is auditable, Anything under 10mV is dead silent. Another very important factor when you change the tube and try to compare them is that they all will have slightly different gain. So here is what I do. My tuner has 1kHz calibration tone, it is convenient and I do not need to bring a generator or test CD).  1)    Drive the calibration tone and measure the AV at the speaker (I prefer to do it electrically  not acoustically) 2)    Follow your procedures to replace the tubes 3)    Drive the calibration tone with the new tube and measure the SAME AV at the speaker. If the AV not same then use the post OTP attenuator to make it the same.  There is some “kind” into it. Some tubes you do not mind to drive louder (mostly dark tubes:  PX25, RE604, YO186, 2A3 with single plate for instance) and some you would like to run softer (type 45, 30, 2A3 etc…). You will end up with whatever you prefer eventually. Be advised that the output of Fundamentals, HF and Insertion channels will greatly impact how you would like to hear your MF.