Ping Sy: LED for Cathode Bias

[rdf]

Just to throw a monkey wrench into the discussion, a LED's internal resistance isn't always a bad thing. It results in a bit of cathode feedback and though I don't understand the mechanism it's my experience a small amount of feedback has a much more significant and worthwhile impact on higher-order distortion components than on the dominant 2nd. The benefits are more obvious to a spectral analysis than a straight THD reading. To use an example, in one test with a CCS loaded 6C45 (I think) biased with two green LEDs, bypasssing the cathode with a 1000 uF capacitor raised the third harmonic distortion component 15-20 dB while having almost no effect on the 2nd.

[audiodesign]

My first experience with led on cathode is positive.

GM70 amplifier with Fiat transformers

[Palustris]

Originally Posted by audiodesign

My first experience with led on cathode is positive.

Mine is too. I use a cheap red LED from RS $1.49 for 2. They bias a 6N23P at exactly 2V and about 15mA with 105V on the plate. I use this stage direct coupled to a 2A3. Sound is clean and clear.

[Wavebourn]

Originally Posted by rdf

Just to throw a monkey wrench into the discussion, a LED's internal resistance isn't always a bad thing. It results in a bit of cathode feedback and though I don't understand the mechanism it's my experience a small amount of feedback has a much more significant and worthwhile impact on higher-order distortion components than on the dominant 2nd. The benefits are more obvious to a spectral analysis than a straight THD reading. To use an example, in one test with a CCS loaded 6C45 (I think) biased with two green LEDs, bypasssing the cathode with a 1000 uF capacitor raised the third harmonic distortion component 15-20 dB while having almost no effect on the 2nd.

Higher anode swing no doubt raised 3'rd harmonic, if the stage biased such a way it clips symmetrically.

[audiodesign]

The led cannot be applied to any vacuum tube stage.

The ac impedance of led can create problems because increase the output impedance of the stage so if on anode there is a choke or a transformer this cut the low frequency.

In my GM70 amplififier I have used this solution (led/optocoupled) with good result because the choke is 70H but in my 813 SE with the same input tube but an interstage transformer with 50H if I use the led I have a cut of at 20Hz at -1.5dB instead of 10 at -1dB.

[SY]

How did you set that circuit up? AC impedance of decent red LEDs at 5mA or more is a couple of ohms, not nearly enough to cause the effects you experienced.

[audiodesign]

I am using green led and the vaccum tube is a D3a in triode connection with a voltage of 70x.

In the original circuit there is a RC on cathode 470uF and 100ohm.

Now my problem is find a capacitor with the sound of led.

1) Sanyo OS-CON
2) ELNA Cerafine
3) ELNA Silmic II

There are not very good.

[audiodesign]

I could try to connect the cathode to ground and bias the tube with negative voltage.

In this case I need an input capacitor but I can use a good 0.1uF Auricap or better a silver mica.

[SY]

You might want to try this with a red, or possibly a different green. At (say) 10mA idle current for the D3a, an HLMP6000 (a tiny, inexpensive unit) will cause a rise of about 150R in the plate resistance, well within the tube tolerance. I haven't measured the green from the HLMP series, but you might give that a shot.

[audiodesign]

Now I am using normal 5mm led.

The correct bias for the D3a should be about 20mA.

I have tested other color led (red, blue) but the result is not good.

Also the paralel of more led does not solve the problem.