Help with splitter design for 7591 PP

[avincenty]

I am building my first amp with a pair of 7591 in PP and have hit a slowdown figuring out the preceding splitter stage. My initial thought was to use a 12ax7 amp stage followed by a 12au7 splitter. I am planning to use a -20 bias on the 7591 so I am looking at a max 40v pp to drive each 7591 tube.

The 12ax7 has to much gain. I am assuming an input of 2vpp on the 12ax7.

If I use two 12au7 then there is not enough gain.

Should I lower the gain stage of the 12ax7 with a lot of negative feedback from the secondary of the transformer? I don't know if there is a good sounding tube to replace the 12ax7 but with less gain.

I'd appreciate any suggestion. I already built the chassis and would have 2x 9 pin sockets for the splitter circuits, it will be a stereo amp. I have built the chassis myself, see attached picture. I think it will at least look good... There are a total of three 9 pin sockets, one of them is for an amplifier section before a tone control circuit. The output of the tone control circuit will have aprox 2Vpp.

Alfredo

[tubewade]

7591s are easy to drive. The AX7-AU7 idea would work. One way to get this is to use a 12DW7. The high mu section is half a 12AX7 and the low mu section is half a 12AU7. Set it up as a conventional concertina. Probably better sound wise, though, would be 12AT7 for the input valve and 6FQ7/6CG7 for the concertina splitter but you would lose symmetry in the amplifiers appearance.

Wade

[SHiFTY]

I'm partial to Eli's El-cheapo 12AT7 differential splitter myself, uses a CCS to get perfect AC balance and it sounds very alive- it's a single stage so very little in the signal path. I use a pair of Telefunken ECC801S which I found in an attic to drive 6V6-type tubes wired in triode. There is enough gain and plenty of drive voltage for 6V6.

I'll wait for his more technical explanation...

[avincenty]

Thanks for the quick response. I think I will give the 12at7 a try. I modeled it in LTSPICE with a 12AU7 concertina splitter and it appears to yield what I am looking for.

I revisited the 12ax7 combination and I could get it to work if I lower the Plate voltage to around 200V. I am new to this, but I have heard that the higher the voltage the better linearity I will get. I could run the 12at7 at 300v.

Alfredo

[tubewade]

Alfredo,

As time permits, study the following thread...

up-to-date El Cheapo schematic

There are some great ideas presented here that may be helpful for your project.

Wade

[zigzagflux]

avincenty:

That's a beautiful chassis you have there.

I am giving a try of a similar design with EL34's using a 12AT7 (mu 60) front end and 6n6p (mu 22) LTP. Gain should be enough to add in enough NFB for good damping factor.

[avincenty]

thanks Wade,

Some of the links are not valid but I should be able to find the info.

Alfredo

[avincenty]

zigzagflux,

Thanks for the compliment on the chassis. It took me a while to build, the woodworking was very time consuming.

were do you get a 6n6p? Is it a good sounding tube?

BTW I attached a back view of the chassis.

Alfredo

[sorenj07]

If a CD player puts out 2V peak to peak, that likely means 2VRMS which means about 2.8VAC peak to peak on transients. You will want a gain of 14.2 at the very least in order to drive your outputs to full power.

If you have room for three tubes, my suggestion is to use a 12BH7 grounded cathode feeding into a 12BH7 long tail pair.

The LTP: Set for 10mA with a CCS (300VB+, 20K plate resistors), the LTP can swing a decently clean 116V pk-pk at maximum input. Gain is 6.7 and bias is at -8.5V or so. A grounded cathode 12BH7 stage with 270VB+, a 22K plate resistor, and an unbypassed 1.4K cathode resistor will give you a gain of 6. (5mA plate current)

Doing the math: Input 2.82Vpk-pk, multiply by 6 for 17V pk-pk or so, and by 6.7 again for 113V pk-pk. You have an extra 9dB or so of headroom, some of which you can "spend" on negative feedback.

[zigzagflux]

Originally posted by avincenty
were do you get a 6n6p? Is it a good sounding tube?

That evil auction place called ebay. My only option, really.

Tube is pretty similar to the ECC99, at a fraction of the price. My samples have been nonmicrophonic and exhibit decent balance triode to triode on about 1/3 to 1/2. Good driver (low Rp) tube. Sounds very pleasant.