[ARC5] Solid State 6AL5

[Tom Lee]

Hi Richard,

In the context of detection, nonlinearity is the desired attribute, 
hence the somewhat tortured grammar. But there was a method to my madness.

As to semiconductors, it isn't accurate to say that the forward voltage 
is independent of current. It most certainly depends on current, just as 
a vacuum tube's forward drop does. It's simply that an exponential is so 
much more dramatic that we are fooled into thinking that solid-state 
devices have a fixed forward voltage. In fact, it will increase about 
100mV for every factor of 10 in current (for ordinary diodes; ideally, 
it should be 60mV per decade); that's a 16% increase in forward voltage, 
and that's without including the additional parasitic resistance that 
all real devices possesses. Compared with a vacuum tube, it may seem 
"constant" but one must be careful about quantitative aspects in certain 
contexts.

--Cheers,
Tom

-- 
Prof. Thomas H. Lee
Allen Bldg., CIS-205
420 Via Palou Mall
Stanford University
Stanford, CA 94305-4070
http://www-smirc.stanford.edu
650-725-3383 (public fax; no confidential information, please)

On 2/5/2018 1:35 PM, Richard Knoppow wrote:
>    I am a little confused, here you say that semiconductors are more 
> linear than vacuum tube diodes due to the 3/2 law leading to poor 
> _nonlinearity_ which to me means better linearity. Is this sentence 
> supposed to read _poor linearity_ if so it makes more sense.
>     Also note that semiconductors have a constant voltage drop, 
> something like mercury vapor rectifiers. Ge is about 3/4 volt, Si 
> about 1 volt. It is independent of current. Vacuum tube rectifiers 
> have a voltage drop which is dependent on the current but is not like 
> a linear resistor since it does not vary linearly with the current 
> drawn, perhaps this is where the 3 halves power law comes in. For this 
> reason the output voltage of a semiconductor rectifier is usually 
> greater than a vacuum tube diode and can be corrected for only one 
> value of current.
>    BTW, some early detector circuits using solid state semiconductors 
> had a source of bias voltage, usually a battery, to improve 
> sensitivity. When I was making crystal sets in the dim, distant, past 
> I didn't know about this and never tried it but it shows up in early 
> books on "wireless".
> On 2/5/2018 12:33 PM, Tom Lee wrote:
>> Hi Peter
>>
>> Assuming that impedances are matched, a vacuum tube diode will always 
>> be less sensitive than a semiconductor diode as a detector -- the 
>> 3/2-power law leads to poor nonlinearity. That more-linear 
>> characteristic is one reason there are some audiophiles who insist 
>> that tubes sound better.
>>
>> The 1N34 is a good detector partly because matching impedances to it 
>> is straightforward. Even though a silicon device has a better slope 
>> near the origin, the extremely high impedance there can't be matched 
>> in practice, so that potential lies unrealized. Adding a tiny bias 
>> current helps, but purists dislike the extra bits.
>>
>> Cheers
>> Tom
>>
>