audio-talk

Whats the cathode follower for as its only feeding a 500k load and a grid?

I'm guessing you mean V3/4b, and I'm also guessing there's some kind of 'gap' involving interconnect leads / sockets as after that is the line pre-amp with input selector switch. Which might be in a separate unit .. or not ...?

Nick wrote:Whats the cathode follower for as its only feeding a 500k load and a grid?

I am having a phono out but as to whether I'd use it or whether a CF is really needed ......

Then again why not because with my current chassis I have a spare half valve per channel. Maybe I could do something better with it?

What about a CCS-fed parallel second stage - this would give some degree of consistence with the first stage design. The AN M-10 scheme I have also employs a parallel 6702a second stage (although this does feed an choke-based RIAA and the whole thing has 3 gain stages).

EDIT - If the phono ends up at 2-stage, with the RIAA circuit in there, any advantage if it could be dc-coupled?

Or maybe a 3rd gain stage as this may mean I don't need to bypass the cathode resistors to get ample gain and can get the output signal up towards CD signal levels?

The second stage is currently wired as SRPP so that would make modification significantly easier but personally can't see any benefit of this or a u-follower etc over the proposed CCS fed gain stage.

The AN M10 has a 6072a SRPP phono output stage.

Apparently I am talking out of my bottom re the phono input stage, I didn't know Nick had already sorted out the design of this with a WAD clone + use of CCS in the anode, and minus any cathode bypass caps, so the originally posted diagram of this works just fine as it is.

X65 signal gain & pretty perfect RIAA EQ. So there.

Only proviso CCS current should be 1mA IMO.

Thanks Mike.

Time to get ordering and soldering - that is unless anyone has any comments regarding my paralleld 12AY7/6072a second stage.

Yes can anybody think why it's not a good idea?


What type is the 1k0 resistor in series with the phono output, seems lately every time I try one of those series resistors it ends up sounding awful, although SMD thick film seems quite good. Well in just one instance where I'm using them.

Mike H wrote:After I put 100uF on each I got a Voltage gain of 82.5 after the 0.1uF cap. I expected >=60-ish.

There seems to be seems a split in terms of whether cathode caps should be used, particularly in phono amps. Current proposal is not to use them throughout, essentially in line with the WAD Phono 2 design and I also note there are many other shemes that abstain from their usage such as Andrews.

MikeH seems to be a proponent and MJ does not seem to shy away from their use. One of the WAD mods for the 5751 is to add OSCON cathode caps. Also MJ suggests that we should squeeze every bit of gain out of the first phono stage as possible.

Is the gain benefit by using cathode caps limited when a CCS is employed?

Cathode caps use seem to be more prevalent on line stages so any recommendations as to whether I should consider them and if so, in which stages/positions?

If they are recommended, I presume that they can add colouration. Are the OSCONs one of the better options, (and they are available from Rapid)? Or something like Elnas as I use in my Rocky, or perhaps some other recommened types.

We already established that the input stage as per last schematic that was posted works properly, so I wouldn't fiddle with it any more, just do like that.

little eddy wrote:MikeH seems to be a proponent

Depends, if I want max. Voltage gain possible then yes, if I want local current feedback (cathode feedback) then not. Depends what the output impedance has to be because cathode feedback increases o/p impedance so the ability to drive a following stage could be worse. Unbypassed also increases the peak input signal you can put in which might be a requirement. Swings & roundabouts.

Also MJ suggests that we should squeeze every bit of gain out of the first phono stage as possible.

Or any input stage, on the basis that following stages are amplifying its noise as well. But again in the previously posted diagram, even with unbypassed cathode resistors, the 1st stage has a V-gain of 65 working into the EQ network, which is not what I would call low.

Is the gain benefit by using cathode caps limited when a CCS is employed?

If by benefit you mean bigger signal Voltage, then no, increased.

whether I should consider them and if so, in which stages/positions?

Think it's one of those case of "try it and see", but when I did it in the sim the EQ curve went a bit odd.

If they are recommended, I presume that they can add colouration.

Absolutely.

or perhaps some other recommened types.

Thanks to Nick I discovered the Panasonics

Well it has finally started.

I was confused a little when checking the current amp schematic to find that the SRPP ECC83 stages were not fed with an elevated heater supply. With the 5814a having Vgk max of 90V, thought I'd better do this. It was only later that I seen the Mullard ECC83 has a Vgk max of 180V!

I decided to build a MJ THINGY elevated heater reference and seems to be fine with a CT of 81V.

I followed MJ's implementation as on page 366 where a dummy CT via two 100R resistors. I'm guessing that this is something to do with noise but not quite sure why this would be an issue with a DC heater circuit and if the THINGY is so good. If there is noise on the reference voltage then I presume it will be imposed upon both heater wires and thus cancel out at the valve in terms of any potential impact.

If whole circuit is as last one posted seems like the only one that would have a problem is the last half 5814A; according to datasheet max. cathode to heater is 100V, so an elevated heater for all at 50 - 80V should be fine.

If you're using DC heater supply that is good too, not sure either what benefit having dummy 100R balancing resistors. Depends how good is the regulator I suppose.

little eddy wrote:If there is noise on the reference voltage then I presume it will be imposed upon both heater wires and thus cancel out at the valve in terms of any potential impact.

AFAIK common mode noise is actually the bigger problem on heaters. Differential noise is less of an issue.

Today I made a rig to test my cascaded IXYZ 10m45s CCS.

From the published single regulator charts, for a 5mA current, Rk should be somewhere around 600R. But when I test the cascode arrangement, for the same current, I need more like 530R.

This is quite a difference. Might this be beacuse of my cascade arrangement, variation between chips, or maybe my test rig being operated from only a 9V battery (as recommended by Nick)?

Are these chips quite variable. I wouldn't have thought they would be repeatable but with the above difference, do I need to individually test each pair of cascased chips, or do people think 530R would be fine for each pair?

The "formula" I use for the 10m45s resistor is 3000 divided by the current in ma.

So 3000/5 = 600

But, if you check the Ixys graph that formula is not so accurate.

According to that "formula" 530R would be 5.66ma

little eddy wrote:Are these chips quite variable. I wouldn't have thought they would be repeatable but with the above difference, do I need to individually test each pair of cascased chips, or do people think 530R would be fine for each pair?

Yeah I'm not surprised, yep I think you'll probably have to "tune" each set individually.

Looking at my characteristic curve then if I chose 600R, I would have a current of around 4.5mA compared to the theoretical 5.0mA.

If these were the extremities of the potential range, would I be able to hear the difference in operating points with a 10% difference in Ia and Vg? Or might this be within the 'accuracy' of a normal valve and passive scheme?