25W Single Ended Hammond 193V Choke Loaded 2SK180 L'Amp

[Zen Mod]

you can always flip bias polarity

 

[plasnu]

So true. SIT-1 style single power supply method works well (efficiently) because of low voltage bias.

 

[Ben Mah]

Yep, flipping good idea.

I was thinking about that after my post.

 

[Zen Mod]

being flippable is my second nature

 

[Itsmee]

...So beware of the SITs with very low Vgs in choke loaded follower mode. A combination of low Vgs and high choke DCR may limit Iq.

Servo?

 

[Ben Mah]

Resolved - Zen Mod's solution.

 

[nautibuoy]

Care to share? I don't understand this bias flipping but as I'll soon be starting a build it would be good to incorporate any learning...

 

[analog_sa]

If the current through the choke results in a sufficient voltage drop this could be all the bias you desire, or even more. In case of more you have to flip the polarity. It all depends on current and choke dcr. Consider for example that 5A and 0.5ohm mean -2.5v of bias even without an additional bias source.

 

[nautibuoy]

OK, got it thanks.

So, with the Lundahll LL1694 inductors I'm planning to use, the static resistance of each winding of 0.9 ohms, I'll be using them in parallel so that's 0.45ohms. Assuming a current through the choke of 2.5A that gives a potential difference, or bias, of 1.125V.

Mapping that onto the simulation Ben Mah ran in post #325

https://www.diyaudio.com/forums/pas...-choke-loaded-2sk180-lamp-33.html#post6673284

which shows a bias of -2.5V, I'll likely still need to provide a negative bias supply.

 

[analog_sa]

Yes. If you cannot set the current as high as you like, even at zero bias voltage, you will know it's time to flip.

 

[Zen Mod]

yeah

flipping nature of ZM is exact thing which allowed making of SissySIT


step aside, flip everything, see what's happening

 

[nautibuoy]

The world through ZM's eyes:

 

[Zen Mod]

it would be nice that is so simple flipped

 

[Ben Mah]

OK, got it thanks.

So, with the Lundahll LL1694 inductors I'm planning to use, the static resistance of each winding of 0.9 ohms, I'll be using them in parallel so that's 0.45ohms. Assuming a current through the choke of 2.5A that gives a potential difference, or bias, of 1.125V.

Mapping that onto the simulation Ben Mah ran in post #325

https://www.diyaudio.com/forums/pas...-choke-loaded-2sk180-lamp-33.html#post6673284

which shows a bias of -2.5V, I'll likely still need to provide a negative bias supply.

The Tokins in real life have Vgs that vary from sample to sample, so you cannot rely on the Vgs from simulations.

 

[nautibuoy]

Thanks Ben. I appreciate there is some variability, 'tis why I said '...likely to still need... ...a negative bias supply'.

 

[triode_al]

Dunno how audible that regulator is. The final cap in the biasing circuit is certainly audible but it is not an earth shattering difference.

The noise is shunted through the inputcap of 10 uF to a source at the buffer of typically 5-10 ohms. And the noise is "resisted" through a 20 K gate resistor. So 10/20.000 is more than 60 dB attenuation, while the LM7912 already will be better than -60 dB typically.
So . . . why bother with an extra resistor?

You don't even need an extra RC - that would only slow down the ramp-up of the bias voltage.

 

[Zen Mod]

proper biasing circuit, in this context, is inaudible ...... if we are talking about biasing circuit as variable voltage reference

there is always big elco as total shunt for all audio freq

 

[analog_sa]

You don't even need an extra RC - that would only slow down the ramp-up of the bias voltage.

No RC after the regulator, just a capacitor.

 

[peppennino]

Chokes for MoFo/LuFo

 

[triode_al]

I have a question regarding the connection of the 2SK182 (flat version) to the heatsink.
I got mine with some sillycone mat 5 cm wide, it has the weave of something inside.
It is very different from traditional mica's etc.


it looks more like it is fit CPU heat sink use - in my build, I need it to get some > 60 watts out of the way.

• I also read that there are 'new' CPU cooling compounds, like Arctic MX-4 or -5. Specified at 3.8 watt/mK (milliKelvin??)
• does this heat paste Cooler Master HTK-002 look bad at 0.8 watt/mK in comparison - I understand I have to inverse the relationship of watts/kelvin?
  

What is recommended?

I even saw a kind of compound that says you don't need a mat, because it is isolating . . . ("Heatsink compound does not conduct electric current - does not cause short circuit on very sensitive electronic circuits". Great. After 100 N torque on the large houses. Then a big turn-on smoke. 30 amps drain to earth . . )