Sound wise, well, only one channel is done so far so its not really possible to say fo' sho' yet but most obvious thing is theres more gain and in mono it sounds OK! Don't they always though? Bang goes my career in HiFi Journalism eh More news as its made though natch.
Re- Shishdo/LW DC amp Pete, yup, done that and found the same. Mine was very sensitive to what it was fed with, using a buffer helped. I remember the top end being heavily shelved when I fed it with a pot in a box of high resistance value in the workshop which was a short sharp lesson in source impedance and so forth. I was pointed to Miller capacitance, aha!
Just a suggestion for your fettling Dave. This should help.
Seriously though, how do you find the gain with two stages of 5687? Amplification factor is ~16 to 18 according to the data sheet, so two stages even if the cathode resistors aren't bypassed would have a great deal of gain. Is it not too much?
To paraphrase Pink Floyd, "Careful with that 'ammer Lawrence"....
I've only bypassed the 2nd cathode with 150Uf 450v Electrolytic. Gain drives the Sab's in the (smallish) project room to the 'too loud to listen comfortably' level, with just one channel, so 'adequate' I'd say! Non Scientific I'd agree.
I'll do proper testing when the other channel is built.
Actual gain is never the figure you read in the spec.
Using a transconductance of 6.3 s (which is a common actual s, not the s in the data book) and a internal resistance of 2k7 (also not as per data sheet but often found), If load resistor is 5k and cathode fully bypassed gain is 11 if load 10k gain 13. This is for the direct coupled potion. If the input resistance is finite (RC coupled) say 220k losses are not significantly greater i.e. 13.18 for 220k 13.28 with dc.
At each operating point transconductance gain and resistance is different so a more exact analysis can be provided with known quiescent current and voltage.
Overall though not way too much gain, but as Rolls Royce would say when asked the horsepower of their motor: "sufficient".
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Just as Dave says, funny how you forget this stuff when you don't use it for a long time.
I'd guessed at real world gain of about 10 . Assuming two stages of 11x gain gives an overall gain of 121. To drive the 2A3 fully needs a swing of + or - 43.5V (say) on its grid, so the 121x gain means a source needs to swing 43.5/121=0.36V, which is 0.25VRMS (I think?) so pretty sensitive.
Aaaahhh, but, yeah but no but (etc). If you're driving (say) an IT after the 2nd stage the cap on the cathode of the section driving the IT works to extend the LF end of the stage IIRC. There's a relationship betwixt the C and the L, I remember chatting with MrT about this over the weekend looking at a cct of a choke loaded 5687 stage driving a 300B of Nick's from a few years ago.
So if I don't need anymore gain I'm a-leaving the 1st stage unbypassed but bypassing the 2nd with the IT in the anode cct.
I'm still in the process of converting the other channel at the mo, my day job is interfering with fettlin' time, meh!
Yes I wouldn't ever suggest not bypassing an IT stage either.
I only suggested one stage unbypassed because the effect on gain is ginormous and you have thrown so much away that two stages become insufficient.
In the 13 gain example earlier with a 1k cathode resistor gain becomes 1.8.
Compounded across two stages were they both RC a pathetic 3.35.
So for unbypassed cathodes in a cascade only one of the stages unbypassed is appropriate.
I don't know Dave's circuit values so can only throw guesses at the values and give indications from guesses. But if for instance the cathode R is 1k, then sharing the R across two channels doubles the gain to 3.22. By this time if you go RC RC cascade and share one CR across four valves compound gain becomes 27 which may be in the ball park. These are new ideas has been done many times. I had a very nice sounding 112 cascade amplification portion of an amp. It sounded pretty good but was bettered shockingly by the 6em7 VA direct coupled to 6em7 CF as a driver for the 801a and that was probably the best sound I ever created. Non of you guys ever heard it as it never left my upstairs room. But a large part of the sound was due to the capacitorless power supply and the field coil speakers. The amp filled the room so was entirely unportable and the room needed to be used for living space so it is a tearful memory.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
HT 350V via SS reg'd bench supply (aka 'slacker psu').
1st stage 22K 5W in anode, 820R in cathode. O.P. 144V @9mA.
DC coupled
Stage one Gain unbypassed = 3, bypassed would be a healthy 15
Dave the bass wrote:2nd stage, IT in Anode cct, Cathode lifted on 15K 5W, bypassed by 400V 150uF cap. O.P. 187V @ 11mA
IT is basic Hammond 126B 5K 1:1
second stage gain 10.
Compound gain an acceptable 30. Must be a nice amp.
Were more gain than 30 required an interesting concept would be something like an unbypassed 12ay7
With 350v B+ shared equally across valve and load resistor would get you an unbypassed gain of 10 making the compound gain for the amp of 100. Quite likely to sound better than the 5687 with a cathode bypass cap in stage one.
some people's sources and output valves would require a voltage amplification of 100.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Interesting stuff. I would never have thought that an unbypassed cathode resistor would have such a large reduction in gain. I might have guessed at 30% but not 80%! This gives me some interesting things to consider. How do you work out the gain for an unbypassed cathode resistor?
I have some 6EM7s, principally for a Barker-Kaufman one day, but I really should try it as VA and CF sometime, perhaps as a replacement for the 12GN7 driving the GM70s.