6sn7 - 6bx7 - sv811-3 amp

[Paracelsus]

HI

I'm designing my first tube amp for a while now, and so far I got to this design.

Stage 1 (6sn7) : Ip=8 mA; Vgk=-5V
Gain = 9
spice calculated harmonic distortion=1.568%


Stage 2 (6bx7): Ip=25mA; Vgk=-25V
Gain = 5
spice calculated harmonic distortion=2.138%


Stage 3 (sv811-3):Ip=65mA; Vgk=-160V
spice calculated harmonic distortion=2.139%
(there must be a lot of innaccuracy in the 811-3 spice model i'm using )


I'm planning on using this amp from a CD player so I used a 2Vrms input signal in my calculations.
(Is this value adequate ?)

I'd like to have you oppinions on where to place an attenuator and which to use. Would you place the attenuator at the input or between the first and second stages?

I'm a bit concerned about the harmonic distortion of the first stages. How do you think I can (should) bring it down a bit ?

I'm also very concerned about the PSRR, as it seems bad to me: a 1V variation at the power supply transates into 8V at the anode of the 811-3. Where am I messing this up ? Can it be fixed ?



Thanks in advance for the time reading this, I'll appreciate any comments.
-Miguel

 

[cm]

I have used a similar circuit. I use 6J5 ( it is half a 6SN7 ) running at 4mA, as the input, 6BX7 at 23mA as the driver parafeed ( LC coupled ) into the grid of a 845. The 6BX7 seems to be very low in distortion if it is driven at lower voltages and at least 25ma current. I have tried different bias pts but the music only came alive at around 20mA. I am using 330V B+ for the 6BX7, and the LC coupling is able to give me a 250Vpp swing to get up to 20W output on the 845.

I use SE CAD and it says I got 0% 2nd harmonic and 0% 3rd harmonic at that operating point. 6BL7 give slightly higher distortion.

I think that if the 2nd order harmonic of the driver stage is about the same as the ouput stage, then they tend to cancel each other and give a lower overall distortion. But I think third order or higher distortion is then more prevalent and it sounds bad. Hence my philosophy of minimizing driver distortion. The first stage distortion I have not modeled. What software are you using for the spice modeling?

I am a tube newbie and this amp is my third tube amp.

 

[Majestic]

CM, without wishing to hijack Paracelsus' thread, I'd be interesting in seeing a schematic if you feel comfortable in sharing it.

 

[fdegrove]

Hi,

But I think third order or higher distortion is then more prevalent and it sounds bad.

When 2nd order HD is cancelled out then yes, 3rd order does come on top of the spectrum but the THD is now also less.

However when the amp now goes into clipping and starts to distort it will now appear to be more unpleasant to the ear compared to when THD was higher but with predominantly 2nd order distortion.

Cheers,

 

[cm]

I don't mind posting schematics but I have any any schematic capture software. I have Protel but it does not have good valve libraries. Any good recomendations?

The 6BX7 is rated at max 8 W for one triode and 12 W for two. I am using only one triode, at nearly 7W. Hope the tube life is going to be OK. I wire the 6BX7s with triode one tube 1 on one channel and the other triode 2 on tube2. I plan to swap the tubes between the two channles every year or so to get maximum tube life. Anyway, at that kind of wattage, you need lower voltages to get the current high enough for the tube to sing delightfully. Hence LC coupling is the only way I could get the 250Vpp without exceeding the power rating of the tube.

Sorry to have hijacked your thread Paracelcus. I hope my info will help you. Personally, I think you should use lower voltage for the driver, and use LC coupling. Then the distortion is almost nothing on the driver, and you get predominantly 2nd order harmonics only at the power tube. I think this will sound very good. Better still use interstage. I am asking NickC to wind me a good interstage. He is very helpful and his prices are great.

Frank is spot on on what sounds good, regardless of what the distortion numbers are. Input stage distortion is normally very low and does not affect the sound quality much. The brand of the tube affects the sound more than the topology and bias point. E.g. my 6J5 metals sounds acceptable . 6J5G are really sweet. 6C5P is worst than the 6J5 metal. Same circuit and bias points.