807 SE Schematic Critiques

[antiquekid3]

Hey guys,

I drafted this schematic a week or so ago. Please tell me what you think and if any improvements could be made. I will not be implementing any global feedback just yet. Here's my schematic:

(larger picture: http://imgur.com/1kfQC.jpg)

One thing to note: the maximum plate voltage is 400V for an 807. I have seen several circuits that push this rating. The others I've seen also use lower resistance cathode resistors, but I am hoping by using a larger value, I will prolong the life of the tubes. Here's the curve:

(and a larger picture here: http://imgur.com/MBDan.jpg)

Thanks!

Kyle

 

[aardvarkash10]

I haven't looked at any of the numbers, but thoughts straight aways include:

triode may be fine, but you are likely to want more than just u/l to tame the pentoady-ness of the 807 in u/l mode.

Will the SL7 have enough oomph to drive the 807 in triode mode?

Why not stick a ccs under the 807?

 

[quikie22]

try triode mode for the 807

 

[aardvarkash10]

just had a quick look - pick this to bits.

triode mode, set the current using a ccs at 40ma through the 807. 420v B+

set up your driverto provide 37V pp to the 807 g1

you get pretty close to symetrical 300v swing at the 807 anode

300v pp @ 40ma means about 8w output (math check please brighter ones)

STC 807 datasheet - its a goldmine

 

[iko]

Haven't tried it myself, but I've found a number of reports that basically say 807 triode strapped doesn't sound good at all. Don't mean to stop you from your exploration though, it's best to make up your own mind.

 

[antiquekid3]

I have a SPDT switch to switch between UL and triode. If I don't like one, I'll just keep the switch in the other position. I think UL is a good compromise between triode and pentode modes.

The *SL7 should be powerful enough. I've seen other circuits use similar tubes, all with fairly high gain. I need 64V P-P to drive the 807 to full power. I really like the idea of a CCS. I'll be able to adjust it much more easily.

From what I understand, the 807 is a 6L6 in a different package. I'm not sure if people like 6L6 SE amps or not, but my guess is that they'd sound similar. This whole build is going to be an experiment, so we'll see how it goes.

Thanks guys!

Kyle

 

[aardvarkash10]

the -32v bias is going to give LOTS of distortion... Anode swings down to around 185, up to around 560. Thats 215v down, 185v up...

Here's an 807 SE that works. Note the bias point.

Having said that, I'm all or experimentation! Go for it...

 

[revintage]

the -32v bias is going to give LOTS of distortion... Anode swings down to around 185, up to around 560. Thats 215v down, 185v up...

Can not find those figures in the diagram. If we go for 6300ohm load we get 150-375-560 and asymmetry is 22% which correpond to 5% 2H just before clipping. Where is the "LOTS of distortion"? Think it should be called average. Unfortunately output power will be very low. 160-375-550V, 5000ohm, 23% asymmetry, 2H 5%. But when going for 3300ohm it gets worse. We get 175-375-520V, asymmetry 38%, 8% 2H.

But doublecheck my course figures, made this at a short break at work.

Also remember you get distortion-cancellation in a 2-stage amp, so figures will look a lot, lot better IRL.

About the RH807 its better to check this site about Schade to get it right.

 

[Vinylsavor]

Hi!

The 6SL7 is a great tube, but not that well suited as a driver IMHO, I'd prefer something beefier there.

If you want to stick to the 6SL7, don't load it so heavily. It has a rp of about 45k. You load it with a 100k plate resistor and 100k grid resistor in parallel, that is a 50k load! This will leave about half the mu of the 6SL7 as gain!

Increase the plate load resistor to maybe 150k (you can use higher B+). And increase the grid resistor of the 807 to the maximum allowable value. Or use a different driver.

Best regards

Thomas

 

[revintage]

The only "different" dual-triode driver that comes to mind is E88CC albeit with mu on the low side. But maybe Thomas has an ace up his sleeve . About single triodes there are plenty to choose from.

If you run 6SL7, do some loadlines there to.

 

[artosalo]

I think UL is a good compromise between triode and pentode modes.

I think UL is the best choice. You have high output power like pentode and low output impedance and better linearity like triode.
The two other are compromises, not the UL.

If you want to stick to the 6SL7, don't load it so heavily. It has a rp of about 45k. You load it with a 100k plate resistor and 100k grid resistor in parallel, that is a 50k load! This will leave about half the mu of the 6SL7 as gain!

Increase the plate load resistor to maybe 150k (you can use higher B+). And increase the grid resistor of the 807 to the maximum allowable value. Or use a different driver.

I was just going to write almost the same.

Put 470 kohms Rg to 807. Then use full (or some 400 V) supply voltage for 6SL7 and use high (up to 220 kohms) anode resistor. Then the 6SL7 is very linear and distortion is low with the high output levels you will need.

I quess you have some of those free audio analyzer programs to use together with you pc-sound card to optimize the amplifier to low distortion.

 

[Vinylsavor]

Hi!

The only "different" dual-triode driver that comes to mind is E88CC albeit with mu on the low side. But maybe Thomas has an ace up his sleeve . About single triodes there are plenty to choose from.

The resulting gain with the E88CC would not be much different as with the 6SL7 as drawn, since the 6SL7 would loose a lot due to the load on it's plate.

I had a lot of success with the 6N7 as driver. It is a dual triode but with common cathode. Both triodes need to be wired up in parallel. If you are ok to use two tubes for stereo, that would be worth considering, plate resistor of 50k works nicely. Will give a gain of about 30x. B+ 400V or more, if possible.

If you are ok to use two tubes you could parallel two sections of the 6SL7. But I'd still prefer the 6N7 it runs with more current and a higher bias (-6 to -7v on the grid) which gives better headroom.

Best regards

Thomas

 

[DF96]

CCS bias for an SE output stage more or less guarantees the wrong bias for either small signals or large signals - you just have to choose which you want right and which wrong. In this respect it is worse than normal cathode bias via a resistor. It is only popular because people don't think about the difference between quiescent current and average current.

 

[zobsky]

If you're bent upon using the ultralinear tap, .. why not try an E-linear arrangement. Still not a fan of the 6SL7 in that position.

E-linear amplifier

 

[revintage]

But I'd still prefer the 6N7 it runs with more current and a higher bias (-6 to -7v on the grid) which gives better headroom.

Hello Thomas,

I am puzzled, are we talking about the same tube?

Checked the RCA sheet and the curves looks great w. ref. to linearity. But using 50k(parallelled triodes) with a B+ of 400V gives us a realistic workingpoint in the ballpark of -2V/200V/2mA.

Or is it the Russian one we are talking about?

 

[artosalo]

But using 50k(parallelled triodes) with a B+ of 400V gives us a realistic workingpoint in the ballpark of -2V/200V/2mA.

With parallelled triodes: -2V/200V, Ia is 12 mA.



http://tubedata.milbert.com/sheets/093/6/6N7.pdf

 

[Vinylsavor]

Hi,

paralleled halves, about 5-6ma (total), 1k cathode resistor.
Higher B+ would be a bit better. RC coupled I use it with 450V

best regards

Thomas

 

[revintage]

Hey Arto....

What you show is totally unreadable. Where did you get those figures?

Check this one instead to verify my findings. Loadline is for one triode 100k, same as two triodes 50k. This is taken from RCA 6N7 sheet even if the diagram for some reason says 6Q7.

So still the question is if there are more 6N7´s than one?

 

[Vinylsavor]

Hi!

This is taken from RCA 6N7 sheet even if the diagram for some reason says 6Q7.

These are not the plate curves for 6N7, these are for the triode section of the 6Q7 as the diagram says. I meant the american type 6N7, 6N7G, 6N7GT.
Electrically identical: 6A6

best regards

Thomas

 

[artosalo]

What you show is totally unreadable. Where did you get those figures?

The data is from GE. I can open the link with no problem.

I will put the curves of 6N7 below.



[IMGDEAD]http://kuva.termiitti.com/image/14858.gif[/IMGDEAD]