GoodSoundClub - Romy the Cat's Audio Site

In the Forum: Melquiades Amplifier
In the Thread: Single-stage Melquiades vs. DHT amps

Post Subject: Some very good points... and... possibly more logical reasoning?Posted by floobydust on: 1/20/2009

Romy,

Some very good points here... if I downloaded the correct schematic from your site for the 6-channel Melq amp (also has the term DSET in the title) then you have created a 6-channel SET amp with built-in crossover points. Your first comment makes sense but only for the SET stages that clip the low frequency response out. The full-range and low frequency SET sections are still succeptable to saturation and field collapse.

******************************

Re: the direct-coupling, seems the three of us are in agreement, nothing else to add.

******************************

Driving current is still a concern with any tube. You can use some basic formulas to calculate required slew rate versus bandwidth. Your input tube is biased around 4V per the schematic referenced above. My post was specific to the output tube being the 45 or 2A3. The same would apply, but of course the figure would be much lower due to the low voltage swing and whatever the total input capacitance of the stage is. Of course, this is fairly insignificant compared to the input cable capacitance. Looking at the 6C33 output tube, it's quite significant.... for sake of argument, I ran thru the calculations using your schematic's operating point and linked tube data.

Required Slew Rate = 2*Pi*f*V where:
Pi= 3.14159
f= frequency bandwidth in Hertz
V=input voltage swing (your bias voltage)

You can determine your total input capacitance as Ct=(Gain+1)*Cgp+Cgk where:
Ct= total input capacitance
Cgp= Grid-to-Plate capacitance
Cgk= Grid-to-Cathode capacitance
Gain= actual gain of stage in operation (less than mu) - let's assume 2.5 in circuit

Required Slew Current = Ct*5*Slew Rate
note: per many smart engineers, multiply Ct by a factor of 5 to fully overcome input capacitance

Using your 6C33 tube and your bias operating points, here the values fro the spec sheet:
Cgp=30 +/- 7pf
Cgk=31 +/- 7pf
V= 80 to 100 volts (shown on your Melq SET amp)

Putting these into the formula:

Slew Rate = 2*3.14159*20,000*90 = 11309733 volts/second = 11.3 volts per microsecond

Total input capacitance = ((2.5+1)*33.5pf)+34.5pf = 151.75pF

Required slew current = 5*Slew Rate*Ct = 5*11.3v/usec*151.75pf = 8.574ma

Your driver stage calculations work out 13.3ma which is fine...

******************************

The 5814A is a ruggedized and improved version of the 12AU7. It has a beefier heater, tighter specs and generally a more robust mechanical structure, i.e., thicker mica supports and in many cases a 3rd top mica disc to enhance rigidity. Beyond this, you design with it like any 12AU7. A very good tube in general. Unfortunately I don't have Russian bottles in my collection... perhaps I should give the 6E5P a try... thanks for the recommendation.

On a side note, looking at the specs for the 6C33, I can't help but think that this tube might have had the western 6336 as a model.... maybe, maybe not. The 6336 is designed in a similar way and can actually replace up to 3 6080 paralleled tubes. Another beefy guy with two independant triode sections.

******************************

I think I'll agree less on this point... or perhaps we're defining things differently. My point is to limit low frequency to the output stage so you don't have issues of saturation or field collapse, not using it as part of a crossover turning point. In any case, using it a crossover turning point, it's only 6dB/octave and you'll probably want it steeper than that for your MF crossover.

******************************

Yea... valid. I've had quite a bit of experience with much larger and much older high sensitivity speakers from JBL and Altec back in the 70's. I've considered getting some 100+db drivers and building something around them, but there's only so much free time in life... I wish there was more. They also are quite large in implementation and while I do have a fairly large main room in the house (12 foot flat ceilings as well) I've not yet decided how much of it will be consumed by audio gear. Still, a valid point as the numbers dont lie. The JBL 2440 is somewhat of an icon... huge Alnico magnet structure, would be cost prohibitive to produce today, but really nice.

The Feastrex drivers are somewhat unique despite much of marketing and such around them. They are probably one of the most ruggedly built drivers I've ever seen. The moving mass is amazingly light considering the size and they can reproduce upper frequencies that are startling. Case in point, the sizzle on well recorded cymbals and percussive instruments in general, all crisp, clean and transparent. They are also unique in their ability to resolve very low-level detail from a recording and they don't fall apart like many fullrange drivers when pushed hard. And yes, another 5-10dB sensitivity on these would be wonderful, but alas, only around 95-96dB.

Regards, KM

Rerurn to Romy the Cat's Site