Tubelab...Why "sandy" SE amps?

[Richard Ellis]

Where would a MOSFET be incorporated in an SE design & Why.
Iv'e heard of these high voltage 'brick' regulators with some big honking TO-3 case darlingtoned down for B+ up to 335 VDC..
__________________________________________Rick.... .........

[Boris_The_Blade]

They make good followers for A2 output stages. You don't have to worry about h-k ratings, and they don't require a whole lot of work to use. (no socket drilling)

They can also make good regulators for screen supplies, etc.

Some fets also work well in constant current sources and sinks for plate loads and other uses.

[kenpeter]

Because sand keeps the vacuum from escaping?

[leadbelly]

Originally posted by Richard Ellis
Where would a MOSFET be incorporated in an SE design & Why.

I take it that you don't actually read most of the stuff on this forum?

[Miles Prower]

Originally posted by Richard Ellis
Where would a MOSFET be incorporated in an SE design & Why.

Source followers make excellent grid drivers for finals. The A Number One problem with a lot of DHTs, such as the 845, is that they start pulling grid current even before the Vgk actually goes positive. According to the 845 spec sheet, Class A1 grid bias is -155Vdc, but the input swing is limited to 150Vp. By keeping the grid at -5.0Vdc, they hoped to limit that grid current. Back in "the day" they really had no choice.

These days, we do since a source follower can source the required current, due to the fact that a MOSFET has a way lower rd(on) than any VT. You can also take any final to Class A*2 with a source follower for the same reason. Since the MOSFET has a much higher gain, it also has a much lower Zo. Getting good, low distortion Class A*2 requires a very low Zo.

The source follower is also preferred to an emitter follower since the input and output impedances aren't interdependent, as they are with a BJT.

[Richard Ellis]

No I haven't seen any schemos of that Florida guy ripping on the guitar, sorry listening to Scorpions 'Deep & dark.
Kinda like an Emitter followers.....Lemme go surfin to learn.
__________________________________________Rick.... ......

[tubelab.com]

I did a bunch of testing back during the design of the Tubelab SE, the powerdrive and the 845SE and found that a low impedance driver or an interstage transformer is needed to adequately drive a big triode in the event that the grid may draw current. Of the available low impedance drivers I tested the mosfet source follower worked the best. I have since done a lot of research on cathode followers for use as an output stage. It is possible to make a good compound cathode follower circuit that makes an excellent driver stage, but it requires 5 triode sections.

A triode will produce the lowest distortion and highest gain when loaded with a CCS. The simplest CCS can be made with a pentode tube or mosfet. The simple tube CCS's don't work that great, but are adequate in some circuits. There is a single IC that works much better. There are several comples circuits (usually involving some sand) that outperform all simple CCS's.

In these cases a trade off is made between complexity and "sandyness". All "sand" is not necessarilly evil. You just have to be carefull where and when you use it. As you pointed out there are some power supply circuits where "sand" works out well.

I have been tinkering with the idea of building an "ideal" amplifier without "sand" but it would be quite complex and I doubt that many could tell the difference in sound quallity over a well designed tube amp with some "sand" in supporting roles.

Because sand keeps the vacuum from escaping?

What takes the place of the vacuum when it escapes? Over the last 2 weeks I have found out. When the vacuum leaks out the blue glow leaks in! I just got a new toy off of Ebay. For $9.99 and $40 shipping, I now have a 0 to 1000 volt, 0 to 200 mA power supply. Its first use was to "test" one of my "leaky" 6V6's. It lit up blue and purple, sparks were dancing around inside it for 10 minutes. I tried to take video of the action, but I had set the camera on auto focus which resulted in auto blur. I have more leaky tubes!

[bakus11]

Originally posted by tubelab.com

There are several comples circuits (usually involving some sand) that outperform all simple CCS's.

Hey tubelab, what complex CCS circuits do you think are closest to perfection? I have read a lot of the CCS stuff on the boards (again, and again and again) but have never really been able to glean any hierarchy amongst the various designs. Perhaps even a hint as to specific choices regarding topology if you can't think of a design or thread offhand.

[Sheldon]

This gentleman has done of lot of work designing and characterizing CCS's for tube circuits. You will be hard pressed to do better: http://www.pacifier.com/~gpimm/

Sheldon

[tubelab.com]

I was referring to the CCS magic worked by Gary Pimm. The link is in the previous post. He has also characterized some of the more popular "sand state" CCS devices.