Serious Modifications

[rlyach]

I was wondering how much effort it would be to convert the Stereomour to use 300B tubes and up the output to 8.5 Watts. I am so impressed with the 2A3 version in my bedroom I wanted to replace my downstairs system. I love the parafeed design but I wanted to stick with an integrated system. I plan on purchasing (or building) some 97dB+ speakers to match. I sort of have an idea on the changes necessary, including transformers and chokes. I just don't want mono-blocks or a separate pre-amp. Is this even possible?

 

[Paul Birkeland]

You'd need to change the power transformer, plate chokes, output transformers, parafeed caps (most likely), cathode resistors, cathode bypass capacitors, power supply, chassis plate, ....

You could try the JJ 2A3-40 in the amp and see what PJ says could potentially be a higher power operating point, but I'm not the one to speak regarding the power handling of the iron in that amplifier.

 

[Paul Joppa]

Randy, I've been looking at ways to do - or approximate - that for a couple years. It will likely come about, but not soon - there are several other things ahead of it in the queue, and we're not going to talk about all of them.

 

[rlyach]

Paul,

I would be willing to prototype something on my own. I have already run through many calculations using the Stereomour typology. My initial estimates are to run the 300B a little hot at 80mA with a ~900 ohm cathode resistor and ~700 ohm choke. I think I would use two 6AN4 triodes for the pre-amp stage. I would use between a 2.25K and 2.5K equivalent output impedance. I would also probably run with a 480V B+. AT these conditions you get to 8.3Watts at 7.8% distortion. The 6AN7's are perfect as they have a mu of 70.

Update: If you push it a little more (90mA) and use a cathode resistor equivalent with humpot of 770 ohms you can get 8.8 watts at 6.2% distortion. You would also have to run a B+ of 500 volts.

 

[Paul Joppa]

That would be quite a reasonable design, and should sound quite good , though I haven't heard or measured the 6AN4 so that would be a question to answer.

 

[rlyach]

Paul,

If I use 488 for B+ I can use the same plate choke that comes in the Stereomour kit as long as it can handle 150mA or so. The output transformer is a little more difficult. Does Bottlehead have a 2.5K output transformer suitable for a parafeed configuration? I can use two power transformers, one per channel. Each should have ~370V rms at 150mA for a center tapped transformer. If I use the voltage doubler it will take ~200V rms. Each transformer should also have both 5V and 6.3V filament windings. I can use a C4S board mounted between the 6AN4s to bias them for maximum gain. Once I find a suitable power transformer the rest of the component values can be calculated. I can use your rule of thumb for the parafeed cap (3uF to 6uF) assuming a 15H to 20H output inductance and a 2500 ohm impedance. if I can find the transformers I should be able to lay this out with the power transformers in the center and the chokes/output transformers on the outside rotated 90 degrees from the power transformers  in either the X-Y or X-Z plane.

 

[rlyach]

I just looked at the pre-amp design. The 6AN4 can be biased at 0v and run positive. In fact, this is the preferred condition as it will give the maximum amplification factor. The shunt regulator can be removed and the constant current source can be set from 8mA to 15mA. The input will have to be modified to be capacitively coupled with a 1Meg termination to ground.

 

[Paul Joppa]

Now you pinpoint the issues that make me take my time!

The Stereomour plate choke can handle 50mA, no more. The Paramount choke can handle about 75mADC maximum. Power - and hence size - go as inductance times current squared, so you can easily see why a greater current is expensive! Neither is currently available as a retail product - these are custom designs, optimized for the Paramount.

Bottlehead has only two parafeed output transformers at present, the OT-1 rated 3K ohms and 8 watts at 30Hz, and the OT-2 rated 4K-8K ohms at 4-2 watts respectively, at 25 Hz. Neither is currently available as a retail product.

The PT-4 used in the Paramount has a 200v winding, along with a 6.3v/1.2A and an 8-vCT winding designed for use with the FC-1 filament choke to power 2A3s or 300Bs. The 200v winding can deliver a maximum 80mADC with a voltage doubler. Only the FC-1 is currently available as a retail product.

As you can see, this is a very small corner of the audiophile world, and parts selection is somewhat limited!

 

[Paul Joppa]

I just looked at the pre-amp design. The 6AN4 can be biased at 0v and run positive. In fact, this is the preferred condition as it will give the maximum amplification factor. The shunt regulator can be removed and the constant current source can be set from 8mA to 15mA. The input will have to be modified to be capacitively coupled with a 1Meg termination to ground.

At more than -0.5v or so of grid bias, most tubes will draw grid current which will create nonlinearities (distortion) and rectification (bias shift) unless the source impedance is extremely low. Look up "grid-leak bias" for another aspect of the issue. Don't forget to look at the performance over the full range of grid voltage!

Shunt regulators (there is none in the Stereomour, but it is present in the Paramount) do improve the sound. The C4S has a high source impedance, around 5 megohms, which you would think sufficient to isolate the stage completely from the power supply - but the shunt reg still makes a clearly audible improvement to my ear, and that of almost everyone who has heard it.

And the fewer capacitors, the better. That's why we have no caps at the input of anything, and we use direct coupling where it is possible. Try it yourself!

 

[rlyach]

I thought the LM431 in the C4S board was a shunt regulator. The Stereomour schematic I have shows it on the cathode of the 12AT7. Anyway, here are the plate curves of the 6AN4. It is a very interesting triode. It appears the best place to bias it is 10mA.

 

[rlyach]

Regarding the added caps. I think if you bias the 6AN7 at 1V on the cathode and 140V on the plate (7mA) you can direct drive the 300B stage and eliminate the coupling cap between stages. Therefore there will still only be two caps in the signal path.

 

[Paul Birkeland]

The added voltage and direct coupling of such an amp makes some other problems, such as:

Allowable voltage and power across the C4S
Allowable voltage across available power supply capacitors
Soft-start warmup in the directly coupled DHT circuit

It's worth the investment, it's just not all that easy.

For iron, I would recommend contacting Sowter and asking if they can produce a 2.5K/10 Watt parallel feed transformer (SE05 with no gap?) and appropriate plate loading choke (they have a 50H/100mA choke, sweet!!).

 

[rlyach]

Thanks, I think I may actually attempt this challenge. I really want a parafeed 300B integrated solution. I will contact sowter. There is also electra-print. They make custom transformers as well. Incidentally, I got the voltages reversed in my last post. The cathode of the 300B needs to move up to 140V and the plate of the 6AN7 needs to be set at 70V. This may make things easier in that I don't need as much resistance in the choke, but I will need additional B+, probably around 600V. Actually building this may take me a while. I still have to finish my Tode (1 more day), and then I have to work on the Stereomour for my pastor. I am building a base to match the Hoyt Bedfords I ordered. Thanks for all the help. Paul, Sorry for taking up so much of your time. You guys are awesome and I love your products.

 

[rlyach]

Here is another suggestion. Would it be possible to simplify the paramount enough to fit it in one chassis? I could have a plate custom laser cut and build my own box. This seems more doable and has the added advantage of going with a proven design. Let me know what you think.

 

[Paul Birkeland]

Yes, you could pack two Paramounts into a single chassis.  The design won't need to be simplified any, but I'd suggest using 1/8" thick aluminum for the chassis plate at minimum.  If you wanted to, you could have a single IEC inlet and single power switch.  The parts layout can be mirrored without a whole ton of trouble either.

If you're going completely custom, I would suggest a 10" x 20" chassis to allow for flexibility and easier wiring around parts that would otherwise be on the edge of the chassis plate.

 

[rlyach]

Here is another suggestion. Since I have JJ 40W tubes, I started with the Stereomour, moved the loadline using a 3K output transformer that BH can get, and moved the cathode resistor to 800 ohms. The result is a 6W amp with 4.33% distortion. I have not run the power supply simulation yet but I suspect I am going to be a little short. Is this something that I might be able to do?

 

[Paul Joppa]

You are suggesting taking more current from the power supply. Since heat goes as power which is current squared, that is always a risk! Fortunately in the present case, the PT-6 is able to handle 75mA per channel.

You will get less plate-to-cathode voltage when you draw more current - not a lot less, but maybe 5-10v less(?). You have shown 325 volts, but the design point (at nomimal power line voltage) is 300v, so you can probably count on 290v.

You'll need a new plate choke as well, to handle more than 50mA. The PC-2 that we use in Paramount, and the BH-7 that Magnequest offers, both have exposed terminals with plate voltage on them, and are unsuitable for mounting topside for obvious safety reasons. When looking for an available choke, remember that as a plate choke it must handle a large AC signal as well as the direct current - most power supply chokes are rated for max current near saturation, so look for a rated current about 50% higher than the actual DC current.

 

[rlyach]

In order to compensate for the drop in plat voltage I was planning on raising B+. I ran a simulation and the most I could get was 395V using diodes to maintain channel separation in the power supply in place of the 270 ohm resistors and increasing the supply caps. Based on this data, in order to keep the 2A3 biased at 60V, the most plate current I can get with the current power supply is 60ma and not 75ma. This hardly seems worth it. The Power calculated at this operating point is 5W, which would only increase the volume by 1db. If I had a different power supply I could raise the B+ to 450 and thus run a 325 plate to cathode voltage. This would get the target 6W but too many things will have to change. I did find a 150ma 15H 256 ohm plate choke, but I think the inductance is too low. I did check into custom iron but that gets expensive very quickly. I am currently re-evaluating the Paramount with BeePre.....

 

[Paul Birkeland]

I did check into custom iron but that gets expensive very quickly. I am currently re-evaluating the Paramount with BeePre.....

Yeah, custom iron can get pretty scary in a hurry! 

I know I'll annoy PJ a little bit by saying this, but the Paramount is an outstanding test bed for experiments. 

 

[drewh1]

Hey Randy - There are Paramounts in my future too.  Is there any way to use the Stereomour as a pre-amp, beyond just passive control?