Remember this tone arm, with mercury contacts in the arm base to avoid torque from the wiring?
Keith Monks Laboratory Manual - Classic Uni-Pivot Tonearm - Vinyl Engine
Did Nordost or Bybee use this line first? Amazing that these folks say these things and maintain any credibility.
The similarity grows...
"When cables are manufactured they do not have any directionality. However, as they break in, they acquire directionality.
Although the cable signal is an alternating current, small impurities in the conductor act as diodes allowing signal flow to be better in one direction over time. This effect is also called quantum tunneling, which has been observed in experiments over 25 years ago. Regardless of the purity of the metal used, there are still diode effects in all conductors. In addition, the insulation material will change when it is subjected to an electrical field."
So you can't get the direction right off the spool, like everyone else says? What a tangled web we weave.
Last edited:
There are several contributors to this, the argument would be better with a picture or two to illustrate. A source follower to the first order bootstraps the input capacitance but in reality there is a dynamic voltage across Cgs needed to make current. This will make displacement current as well as Vin across Crss both need to be provided by the previous stage. Now consider the phase of the load for more complication.
Crss is a strong function of VDS with the worst case at lowest VDS. If you want to swing to say 3V of the supply rails then 1.5nF is absolutely spot on for an IRF240/IRFP240 pair. Ooops, sorry brain fart - the pairing is IRF240/IRF9240 and the P-channel device has lower capacitance. Call it 1.2nF.
Last edited:
Yes, but not with those specific devices. Input capacitance was just what you'd expect from the ratings- there is some nonlinearity, but it's manifest more near the clipping point (again, exactly what you'd expect from the Crss vs Vds curves).
http://www.st.com/web/en/resource/technical/document/datasheet/CD00002071.pdf
See figure 8.
edit: I looked at similar curves for the 15+ year old devices you mentioned. Crss is indeed worse than in modern devices.
Guys a picture what you are discussing would help. PMA a source follower loaded (for instance in the limit) by a current source does not have any contribution from Ciss at the input. Only a portion of Ciss, related to the ratio of gm to the load conductance, causes a displacement current load on the previous stage. At about a gain of .9 for the source follower the Ciss and Crss contribution is comparable (in your case) but in a normal circuit the Ciss is at a lower bias voltage and at a more non-linear point.
Measuring the capacitance of gate to source and drain virtually grounded, does not apply in situ for an amplifier source follower output stage.
Last edited:
I stated my conditions exactly no need to repeat them. You can, and I have, built a fully bootstrapped follower with 100's of femto-farads of input capacitance. In fact with a little care you can null it almost identically to zero. The circuit was in Jan's magazine but I doubt you read anything outside the posse.
And I think it's both.
I agreed with John on 100 V/uS was a good number. More only if the amp is really very high power.
But I also feel we should let the voltage part of the rip at higher than teeny seeny currents they are sometimes ran at. I believe a decent VAS should be biased at 15-20 mA.
- Status
- Not open for further replies.