As a first finger exercise in LTSpice I composed this little volume control. For some reason the ECC82 (12AU7, double 6C4) is popular for making distortion, so let's use it. Low output impedance is a necessity, so the CF is king. There is a misunderstanding about the cathode follower giving it a bad reputation. For driving a huge load like a SS poweramp with 10K input impedance (sic) it takes a type that combines high current with high gain (a property called conductance). ECC82 is like 6SN7 a low mu type which makes a lousy CF. Keeping SS out, I opted for the available 6C45n (or 6S45pi but it's all russian to me..) with high gain and high transconductance. A (really) wise old tech once told me to formulate goals before ordering parts, so one doesn't have to lie about its performance when finished. Let's christen her Bourgeois Sandwich and let her do these tricks:
- be a nice girl and behave with little distortion, unless pleasing
- no hysteria, so behaviour should not suffer from oscillations
- naturally, what comes in gets out, so no amplification unless ordered
- turning the knobs should give the desired effect: half volume at 12 'o clock.
A first design.
- be a nice girl and behave with little distortion, unless pleasing
- no hysteria, so behaviour should not suffer from oscillations
- naturally, what comes in gets out, so no amplification unless ordered
- turning the knobs should give the desired effect: half volume at 12 'o clock.
A first design.
Attachments
Last edited:
Are you possibly related to Dirk Wright? It didn't end well with him either.. Personally I think ECC82 is a stinker but what happens when we arrange a marriage to a sputnik? Bear with me, fireworks to be expected..
Both smell funny but they have a different field of application. I've just finished a test rig and the model does not hold: 6C45P promised 20mA standing current where practice shows just over 4mA. Back to the drawing board 🙂
Your operating point is quite extreme, i.e. -Ug is high (-4.3 V) and therefore Gm is low.
At the attached photo it is marked with red dot.
I would use something like the operating point marked with green dot.
Increase R8 to some 5k6 and fine tune (=search) R7 to get some 20 mA anode current.
At the attached photo it is marked with red dot.
I would use something like the operating point marked with green dot.
Increase R8 to some 5k6 and fine tune (=search) R7 to get some 20 mA anode current.
Attachments
Last edited:
Yes,linearity is better, but dynamic Gm (and Zout) is worse.
Mona
Attachments
They shoot over the moon these sputniks; got ten pieces and they all measure differently. After a lot of nitpicking I ended up with 43 ohm for cathode resistance and a beafy 4K 10W load. On the graphic is a sample. I measured Vak=137,5V and Ia 114,5/4035=28,4mA while the grid is at 112,7-114,5=-1,8V. That's half a volt off from the graphic. 7788 better choice?
Oh, and there's a hum (AC out 2 - 3mV) but faintly noticeable between songs. That could be because of the 1 meter B+ cable plus AVO7 in between 🙂 No ground loop, no progress when changing AC to DC glow or grounding the glow. When I short the sputniks grids it's silent. Time to insert anode stoppers too?
Oh, and there's a hum (AC out 2 - 3mV) but faintly noticeable between songs. That could be because of the 1 meter B+ cable plus AVO7 in between 🙂 No ground loop, no progress when changing AC to DC glow or grounding the glow. When I short the sputniks grids it's silent. Time to insert anode stoppers too?
Attachments
Last edited:
20Vpp is to much for a follower with 12V on the cathode, with 5Vpp the situation will be less catastrophic.
The last solution of disco is much better.But the big current draw is probably the cause of the hum.
Mona
True. I was a bit hasty with my simulation.
At some 6 Vpp out the difference is still huge: 0.008% vs. 0.26%.
Hum problem solved; sometimes the clue is in the datasheet 😉 The 6C45 heater (6V AC) is raised 43V by means of the Morgan Jones "Thingy" ladder with two BJTs. Loading this pre down to 10K is no problem. I'm gathering parts for the PSU and timber for a case; will post some pics and building instructions when done.
If you like solid bass this is the one. It does a good job preserving dynamics, ranging from fine details down to voices and contra bass strings plunging. To my (limited) experience some pre amplifiers have a tendency to loose some quality in the process of controlling the output level. Either low end or highs get attenuated when turning the volume control, although it would come down to a silly part of a deciBell, normally not observed.
The only thing worrying might be the decline of transconductance, note the 3000hrs remark in the datasheet.
If you like solid bass this is the one. It does a good job preserving dynamics, ranging from fine details down to voices and contra bass strings plunging. To my (limited) experience some pre amplifiers have a tendency to loose some quality in the process of controlling the output level. Either low end or highs get attenuated when turning the volume control, although it would come down to a silly part of a deciBell, normally not observed.
The only thing worrying might be the decline of transconductance, note the 3000hrs remark in the datasheet.
Attachments
Last edited:
> The only thing worrying might be the decline of transconductance, note the 3000hrs remark in the datasheet.
The rated Gm amounts to 37 Ohms at the cathode. The projected change of Gm means it could be 26 or 52 Ohms in 2 or 3 years (or 4 months at 24/7).
The 1u cap and hi-fi expectations suggest it shouldn't be driving even as low as 10k load.
Note that 1uFd is still 160 Ohms reactance at 1kHz. It dominates output impedance over most of the audio band.
10,000/10,026 -0.023dB
10,000/10,053 -0.046dB
+/- 0.04dB or 0.01dB per month working steady
The voltage-gain stage and passive parts will drift at least this much. The output stage Gm variation vanishes in the haze.
The rated Gm amounts to 37 Ohms at the cathode. The projected change of Gm means it could be 26 or 52 Ohms in 2 or 3 years (or 4 months at 24/7).
The 1u cap and hi-fi expectations suggest it shouldn't be driving even as low as 10k load.
Note that 1uFd is still 160 Ohms reactance at 1kHz. It dominates output impedance over most of the audio band.
10,000/10,026 -0.023dB
10,000/10,053 -0.046dB
+/- 0.04dB or 0.01dB per month working steady
The voltage-gain stage and passive parts will drift at least this much. The output stage Gm variation vanishes in the haze.