Scratch Build (First for me)

[porcupunctis]

Here is the schematic for the power supply.  I'm getting 414 volts for B+.  The transformer is a 365V center tap given a 120vac input.  My AC runs high at my house.  Usually around 124 to 125vac but that wouldn't seem to account for the much higher B+. 

[Paul Joppa]

It's typical for a cap-input filter to give 10-20% more DC voltage than the AC supplied. A higher-resistance rectifier might bring it down - 5R4? 5Y3?

[Paul Birkeland]

OK, my first real concern (before I do some modding) is that when I checked my voltages, it seems to be running a bit high.  For instance, Pin 5 of the 6SL7 is running just under 310 volts and the data sheet recommends a 300 volt max. 

That's 300V plate to cathode.  The cathode accompanying pin 5 is at about 150V, so you have 150V remaining between plate and cathode.

My overall B+ is 414 volts which is 50 volts over what was listed on the schematic.

No biggie, you can add some resistance in the power supply if you like, but it wouldn't seem to be necessary.

Here is the schematic with voltage measurements at  points of interest:

Nothing too out of the ordinary there.

-PB

[porcupunctis]

I did manage to find two dusty old 5Y3 GT tubes and gave them a try.  The first one sparked internally and sounded like bacon cooking on a really hot stove.  Pulled that one and tried the other.  This one seemed to work fine and as PJ predicted lowered all my voltages.  The B+ came down to 373 (from 414) which is very close to one of the schematics I found online.

While the 5Y3 was in, I performed all the same voltage measurements as before and of course they were all lower.  There was no real change in sound quality and I seemed to notice a bit more hum at idle.  Based on PB's post, I may go back to the 5U4 rectifier.  At least I know they are both safe to try out.

PB - I would like to try the mod you suggested to the 6L6 wiring.  I won't be able to get to that until next week (Spring break) so in the meantime, I will draw up a new schematic so I can verify that I am interpreting the changes properly.  It was actually one of my future question to ask why that grid needed a voltage divider to ground while all other grids were either lifted from ground or dropped from the B+. 

Here is a question that I hope has an easy answer:  I'm running taps out for both 8 ohm and 4 ohm speakers, however the negative feedback is run off the 8 ohm taps.  Will this work as well no matter which speakers are used or do I need to set up a switch to change where the negative feedback comes from?

Thanks again for all the help (everybody), and I hope this thread is somewhat interesting and useful to other members who may be interested in knowing more about what goes on under the hood.  I know my appreciation for what BH does has grown considerably. 

[Paul Birkeland]

Here is a question that I hope has an easy answer:  I'm running taps out for both 8 ohm and 4 ohm speakers, however the negative feedback is run off the 8 ohm taps.  Will this work as well no matter which speakers are used or do I need to set up a switch to change where the negative feedback comes from?

You can leave it on the 8 Ohm tap.

[porcupunctis]

PB - here is how I interpret your proposed changes to the 6L6 wiring.  Seems reasonable but I would like to know more of the "why".

[Paul Birkeland]

There is an assumption in the pentode characteristic curves that the screen grid is bypassed with respect to cathode.  This is taken for granted in fixed bias amplifiers, as the cathode is generally a couple tenths of an ohm from ground to begin with.

Give it a shot both ways and let us know what you find.

-PB

[porcupunctis]

Ok, I had time to make the cathode bypass change tonight.

Started with one channel only so I could A-B the results.  After a while, I could tell the difference and the change was quite positive.  More detail overall, but especially in the low end where the bass was much more life-like and the kick bass more completely separated. 

I also noticed about one-tenth of an amp lower current draw from the wall.  I haven't measured it yet but I'm pretty sure I didn't lose any power at the outputs.  Sounds more like I might have gained some. 

I want to come back to this amp in a few days and plot some load lines for the 6SL7.  That's my next learning target anyway.  In the meantime, I have the following projects to occupy the next few days:

1.  An update to my Stereomour "45" that I have been following.  I understand there was an error in the manual and want to get that corrected.  Really love that amp.

2.  I have a tape path kit for my 1500 that is crying for my attention.  When I do that, I want to run some direct out lines for my Eros tape repro that I recently picked up.  I want to go through the resistance and voltage checks on the Eros before pulling it off the bench.

I have some busy days ahead.

[Paul Birkeland]

If you want to mess with the circuit some more, do some light reading on the purpose of the cathode follower, then Miller capacitance, then Miller capacitance of a pentode.  (This will explain my snickering at the cathode follower)

-PB

[porcupunctis]

That was some interesting reading.  May need to do some more but here is what I'm taking away:

Miller capacitance happens.  More prevalent in triodes and almost negligible in pentodes.  This makes a tone of sense since you essentially have two plates with just a little space between and a voltage difference between them.  That is pretty much textbook definition for a capacitor.  I think of the adjustable capacitors in old radios.

OK, so on to the cathode follower:  Low impedance output and mostly designed as a buffer and/or feeder into a tone circuit.  Well suited for high capacitance follow-on loads which the 6L6 is not.  The cathode follower will generally lower gain to about 70% of the original.  Bright spot is very low distortion without negative feedback.

One previous builder of this circuit cited the low distortion as the main reason for choosing the cathode follower.  But it seems to me that total distortion would be the sum (or maybe product) of the distortion of the two 6SL7 stages together.  If he wanted to eliminate as much distortion as possible, why not skip the second stage altogether and get some extra gain in the process?

I will have to look at previous posts but I remember someone suggesting that I go with just the one stage.  Now that I've done some research I can see the logic in that approach. 

I'm starting to get the joke.

[Paul Birkeland]

Yeah, the funny takeaway is that the first 6SL7 stage (the voltage amp) is actually harder to drive than the 6L6.

-PB

[2wo]

I think one of the reasons for CF is the 6SL7 is a high gain tube but has a high Rp ~44K which will have a harder time driving the next stage. The CF acts as a buffer...John

[Paul Birkeland]

I think one of the reasons for CF is the 6SL7 is a high gain tube but has a high Rp ~44K which will have a harder time driving the next stage. The CF acts as a buffer.

With the 6L6 wired as a pentode, the high output impedance of the previous stage is inconsequential (6L6 capacitance will be around 10pF).  The theoretical negative impact of the high Rp of the 6SL7 driving the 6L6 would be bandwidth limited to several hundred kilohertz. 

The 6SL7's Miller capacitance will be more like 130pF., which is right on the borderline of being so high that a 6SL7 shouldn't drive another 6SL7 in an audio frequency power amp.   

-PB

[porcupunctis]

So, should I try eliminating the second stage of the 6SL7 and take the output from pin 2 on to the 6L6?  It would seem that there is nothing to lose here.

[porcupunctis]

OK, I just went for it.

It now has a single input stage feeding the 6L6 from pin 2.  A small but noticeable increase in gain.  Overall, the amp seems to sound the same however there is more quiescent hum.  My workbench puts me in a very near-field listening position and I'm using fairly efficient 4 ohm speakers.  The hum would probably not be noticeable in other situations. 

Going to do more listening now.