Something Stirs in the Spare Room.

[Paul Barker]

Just a quicky on gain. The data sheet I found mu is 65 not 70.

Actual gain is load over load plus internal resistance. The data sheet says 4k8 Ri. If we bypass Rk which you have. Im making a judgement youre hammond choke represents 20k

65x20k/24k = A52

Youre cathode bypass cap 1,000 would satisfy a 1 hz - 3db.

No bypass would give you another 13k2 Ri 65x20k/38k = A of 34”48”

Shared cathode resistor = .63 x 65 = A of 41”55”

I like shared cathode resistor and I like no bypass for the feedback sound in the valve stage of an unbypassed resistor.


So you can decide. I’ll leave it to you to see what gain you need and try sound difference.

“Oops youre using a 100 ohm Rk my mistake” so ive put corrected unbypassed gains in parentheses. Youre bypass cap requirement 1700.

[Cressy Snr]

Hi Paul,
I’m using a 200R cathode resistors on the input stage.
The Philips data sheet for the E88CC gives a gain of 70 and a 5.2K internal resistance. But the Telefunken data sheet shows 65 and 4.8K .

I guess 65 would be the safer figure to take then.

Thanks for the gain calcs and the cathode resistors suggestions.

[Paul Barker]

Oh yes, I accidentally first time saw youre choke loaded diagram then casually saw the 10k diagram saw Rk from that was 100 didnt notice I was on the wrong diagram.

Lucky I left the 200 ohm example visible when I adapted for 100 ohm.

Compount curves not provided so were only making an assumption about mu transconductance and resistance as we only have them for the suggested operating point of the particular data sheet we are following.

All you actually have to do it measure the ac on grid of valve, then ac on grid of next valve for Actual gain.

Then you compare the unbypassed gain and if its enough gain, play music and see if like me you prefer the sound of cathode feedback.

[simon]

Shared cathode resistor = .63 x 65 = A of 41

I like shared cathode resistor and I like no bypass for the feedback sound in the valve stage of an unbypassed resistor.

I haven't followed what a shared cathode resistor is in this context. There's only one triode, what would be sharing it?

[Paul Barker]

Shared cathode resistor = .63 x 65 = A of 41

I like shared cathode resistor and I like no bypass for the feedback sound in the valve stage of an unbypassed resistor.

I haven't followed what a shared cathode resistor is in this context. There's only one triode, what would be sharing it?

With the other channel, twice the current half the resistance, half the resistance less effect on (A)ctual gain.

Introduces reactance between the two channels which I like also. When it was first talked of over 20 years ago, some like it better Im one of those.

My most enjoyable personal amp had two stages of 1a four valves shared one cathode resistor unbypassed. But the greatest key to its sublime sound was the shunt valves and no caps in the power LR power supply. Shockingly good but very big and heavy.

[simon]

Ah I see

[Nick]

I would have though a 150H would provide more than 20K at most audio frequencies. I find its only down to 20k below 25Hz. At 1kHz its closer to 1M.

2 * 3.14 * 150 * 1000 = 942000
2 * 3.14 * 150 * 20 = 18840

[Cressy Snr]

A bit more done on refining the project:

6AS7_SET.jpeg (112.81 KiB) Viewed 1336 times

I tried out the sharing of cathodes on the input stage. The presentation after I had done so 'bothered' me. The big soundstage that had previously extended beyond the outer edges of the speaker cabs had become confined between the speakers and instrument placement within the image had become vague. Unfortunately (either bypassed or un-bypassed) shared cathodes were not for me

After a lot of A/B-ing between resistor bias and LED bias, I came down on the side of red LED bias for the input stage. For me, the sound with LEDs was superior in every way and the image was now big, deep, and extended beyond the bounds of the speaker cabs again.

With the LCLC power supply there was a bit of mechanical noise coming from the existing Hammond PSU choke. This could be heard, with no signal through the amp, from the listening chair. As a consequence, I've uprated the choke using a Danbury, 300mA job from the parts bin (see pic above.) Mechanical silence now reigns.

In terms of the audio circuit, I changed the coupling cap value to 0.047uF. I don't want ultra-low bass frequencies saturating the cheap Hammond output transformers and causing LF distortion. Obviously, there is little power on tap, even with the output valves almost flat out at 11.5W dissipation per anode, so the low cap value is a purely practical attempt to eke out the maximum amount of useable music from the project. It seems to work well. I can't complain about the available volume, given the 87dB efficient speakers.

I'm certainly very pleased with the sound of this project. The project makes a big, warm, fast, detailed and thoroughly pleasant sound.

[Cressy Snr]

Schematics:

6AS7_SET.png (44.27 KiB) Viewed 1330 times

Rectifier has been changed to a GZ37, just to help the power valves stay within their dissipation limits.

6AS7_SET_PSU.png (53.79 KiB) Viewed 1330 times

[Nick]

The big soundstage that had previously extended beyond the outer edges of the speaker cabs had become confined between the speakers and instrument placement within the image had become vague.

Yep, consider what the shared cathode is doing to the stereo signal, adding some out of phase left to the right and vice versa. At best I would describe it as an "effect"

[Paul Barker]

No harm done, glad you tried it Steve.

[Paul Barker]

The big soundstage that had previously extended beyond the outer edges of the speaker cabs had become confined between the speakers and instrument placement within the image had become vague.

Yep, consider what the shared cathode is doing to the stereo signal, adding some out of phase left to the right and vice versa. At best I would describe it as an "effect"

Dave Slagle tried to formulate what happened began a senstence or maybe paragraph, then admitted he needed to think about it another time if I recall. Still waiting….20 years and counting…

I vaguely recall some of us liked it… but might only have had other problems it mitigated for us. I’m not bothered.

I’m going to have to revisit led bias just to check I’m not on the wrong path.

[Nick]

Dave Slagle tried to formulate what happened began a senstence or maybe paragraph, then admitted he needed to think about it another time if I recall. Still waiting….20 years and counting…

I will let you stick with your appeal to authority then. But if you consider the simple case of a tone injected into the left channel, then the shared cathode resistor will work a little like a LTP phase splitter, and create an out of phase image of the left onto the right. It will be even more complex (and wrong) with combined left and right signals containing various phase relations between left and right, but the same effect will happen. Mono signals will go through apparently unchanged, signals to the left and right will be made to seem further left or right . Either way its in the opposite direction to accuracy and as inaccurate as the cutting engineers fixing the first pressing of Electric Ladyland.

[Paul Barker]

Yer

[Cressy Snr]

We all know that choke input supplies are good for SE valve amps at any rate, what with the constant current draw from the rectifier and all that, but I have not been able to get them as quiet as cap inputs or regulated supplies. Probably my own inexperience with them is a factor in that problem, but the great sound quality enabled by an LC supply makes them worth the extra bit of noise.

I've been getting a faint 100Hz hum from the speakers (thankfully not a buzz, now I've uprated the current capacity of the first choke) which can be heard depending on where in the room I stand, so I had a look online to see if there was any info on how to get an LC supply quieter.

I came across a some stuff about resonant choke input supplies where a capacitor is used across the choke, to set up a resonant filter at twice the mains frequency. It is, in effect a noise cancelling, parallel tuned LC circuit.

http://www.diyparadise.com/tunedchoke.html

Anyway I worked out a suitable cap value in parallel with my 10H choke, of 0.253uF. I got near enough with a 0.22//with 0.022, just as an experiment (101.89Hz is the freq with those values.) It seems to work well enough.

OK so this tuned filter does absolutely nothing whatsoever about any the harmonics of the 100Hz, which inevitably will pass unimpeded, so I've now got a faint 200Hz hum instead of the 100Hz. BUT it has quietened the supply by another notch.

Also without a simulator I can't tell whether there is any peaking at 100Hz due to the "crossovers" between peak and notch, not quite coinciding (imperfect cancellation) but in subjective terms the bass seems to have stayed even.