How would you build a hybrid LM3xxx amp?

[leadbelly]

I'm just trolling for design ideas on a hybrid amp I am building for a friend. It will drive 4 ohm Klipsch speakers. I want it to be a tube-LM3886 hybrid. For those of you who might want to say "just build one of those VBIGC's on the chip amp forum and leave us alone", I don't like those. The idea of putting a cathode follower in front of the chip just so it looks pretty doesn't appeal to me. I know SY's preamp shows that a cathode follower stage can be a good thing, but the parts count is a little high and pricey for what I want.

So I have two ideas of my own. Feel free to shoot down and recommend something else

1) 2 LM3886's in parallel set up as unity gain inverting current followers. Having 2 in parallel should allow the rails to be as high as +/- 40V. A simple 6SN7 grounded cathode stage in front, probably with a CCS plate load, does all the voltage amplification, and the mu of 20 allows a 2V CD input to swing all the way to the max +/- 40V input voltage. The idea here is to make tubes do what tubes do best, amplify voltage, and let the SS handle the current load.

2) Just a single LM3886, with <30V rails to accomodate the 4 ohm load, but deeply biased into A with a CCS, say 1 or 2 amps worth, and a voltage gain of 10. In front, a simple 6SN7 grounded cathode stage. The idea here is that some feel a class AB output stage is incongruous with tube sound, and so class A will sound better.

Comments?

[xplod1236]

According to the datasheet, the minimum gain for the LM3886 needs to be 10.

[leadbelly]

Originally posted by xplod1236
According to the datasheet, the minimum gain for the LM3886 needs to be 10.

No, only true for non-inverting. http://www.diyaudio.com/forums/showt...061#post112061

[ilimzn]

I should add that rising the noise gain as depicted in that thread is essentially equal to a attenuate and amplify approach. Fortunately, it has an advantage: the noise gain can be rised only at the top frequency end, where the amp actually needs that gain >=10. The LM3886 is sufficiently wide-band so this can be moved almost outside of the audio band. Just use a cap in series with the resistor from the - input to ground. I've used this approach on several chip amps, starting a long time ago with the TDA2030

[leadbelly]

Thanks for the tip ilimzn.

I just thought of a big question: what gain should the tube stage of my option (1) have? People often quote a figure like 1Vpp or 2Vpp for CD input, but would it be better to put in higher gain for certain devices? I know that I am struggling with the level out of an inexpensive portable DVD I bought, it cannot drive normal headphones at an acceptable level.

[arnoldc]

it really depends on your "reference" point. my HDCD player outputs 1V with playing redbook and 2V when playing HDCD encoded discs. since I use a preamp, most of my projects are happy with the 2V, but there are some CD releases that are really low in output.

i'd like to recommend a 2C51 as your driver

[leadbelly]

Hmmm, 2C51, not something I see in a lot of schematics.

No, I'd like to use a tube I already have a lot of, either 6SN7 (mu=20), 6N1P (mu=33), or triode strapped 6EJ7 (mu=55).

[leadbelly]

Well, I'm not sure if anybody's still reading, but after a bit more searching and thinking, this is what I like. It uses both options (1) and (2) of my original post. It's basically just Frank's 6SN7 line stage in front of an inverting, gain of 2 chip amp stage.

[Audio_idiot]

hi,
Interesting idea, 2 questions.
1) is it a must to have gain for the LM3886. how much gain do you racon is sufficient from the tubes?
2) is the 317 a current source and what is the benifit?

Happy New year