Fun little P-P input stage

[UltimateX86]

HI

you have new diagram ? I would like to use 6SU7GT (6188) in entry differenciel, driver mosfet and bipolar output

thx

 

[Jaap]

In another thread: http://www.diyaudio.com/forums/showthread.php?s=&threadid=137332 I have tried to stir up a discussion about a small amp in the Baby Huey class. I would like to use ecl82 but ecl 86 is also o.k.

I wonder if the here discussed topology with fet at the front could be used with a PP ecl82/86.
Another idea is use a double triode at the front (I have a few ecc40) and than use after that the triode of the ecl82/86 as a cathode-follower.

Perhaps this would be a nice (better ?) variant of the Baby Huey (?)

 

[ErikdeBest]

Hi Jaap

I can't remember where exactly, but I recently asked SY about the input stage shown in this thread. He said that if he had to do it again he would probably opt for a LTP input stage, with well matched anode resistors and a very good CCS - that is, the baby huey input stage.

Erik

 

[Allen Wright]

No, that's the FET's drain (plate), not the source connected to the minus rail; note that the FETs are configured as source followers, where you want the drains to be at signal ground.

Then this really illustrates the STUPIDITY of that fet logo.

With no indication of polarity/in/out, one has to go look up the polarity of the fet from a data sheet to be able to understand the circuit. Dumb, I say.

I draw jfets like bipolars, only with the source and drain leads parallel to each other (unlike the angled leads in bipolars) but with an arrow going in or out on the source lead to indicate polarity - just like NPN or PNP bipolars.

Regards, Allen

 

[Miles Prower]

Then this really illustrates the STUPIDITY of that fet logo.

No, not really. The schemo for the JFET is based on how it's constructed: a bar of N type silicon with the drain and source leads connected to opposite ends. (N-Channel JFET) The gate is a P type diffusion layer to create a reverse biased PN junction. Since the gate is P type, the arrow points from the gate to the channel, P to N.

With no indication of polarity/in/out, one has to go look up the polarity of the fet from a data sheet to be able to understand the circuit. Dumb, I say.

It's helpful to label the gate, source, and drain, but not totally necessary.

I draw jfets like bipolars, only with the source and drain leads parallel to each other (unlike the angled leads in bipolars) but with an arrow going in or out on the source lead to indicate polarity - just like NPN or PNP bipolars.

Regards, Allen

Except that there is no PN junction there, although it's something I've seen before on some schemos.

 

[Wavebourn]

Both FETs and BJTs are reversable. However, since geometry of collectors and emitters are different parameters will be different. But I had some Russian military transistors that measured very close, both for collector - base and emitter - base breakdown voltages, and Beta, when reverse connected. They were good as switches.

 

[Allen Wright]

No, not really. The schemo for the JFET is based on how it's constructed: a bar of N type silicon with the drain and source leads connected to opposite ends. (N-Channel JFET) The gate is a P type diffusion layer to create a reverse biased PN junction. Since the gate is P type, the arrow points from the gate to the channel, P to N<<

Miles, I really do NOT need a primary school tutorial on how fets are made and are normally drawn. I learnt that 40 years ago, and have been using jfets in RF and audio (successfully) ever since.

My point is that drawing them the conventional way leads to confusion when you are try to do a quick understanding of a new schematic. Dave/Planet 10 is a friend and I know he ISN'T a dumb bunny, yet he was confused by SY's schematic, as was I, and I'm sure a bunch of others who didn't want to admit they also didn't understand it.

I will continue to draw jfets my way, it may not be "correct" but they allow a far quicker/better conceptual understanding of the circuit without a lot of head scratching & data book looking up.

Regards, Allen

 

[dsavitsk]

My point is that drawing them the conventional way leads to confusion when you are try to do a quick understanding of a new schematic.

Everybody, but SY it seems, draws these with either the gate at the same height as the source (http://www.nxp.com/documents/data_sheet/PMBFJ108_109_110.pdf) or with a little dot to indicate the source (http://www.amb.org/audio/beta22/b22_sch.png). The only thing confusing here is SY

Miles, I really do NOT need a primary school tutorial on how fets are made and are normally drawn. I learnt that 40 years ago, and have been using jfets in RF and audio (successfully) ever since

Please refrain from making the tone here like it is as AA. We like Miles, and I for one found his post informative.

 

[Zen Mod]

putting S where it belongs is convenient



I must look at datasheet or known schematic each time when I need to draw proper Jfet or mosfet symbol .......

 

[SY]

With no indication of polarity/in/out, one has to go look up the polarity of the fet from a data sheet to be able to understand the circuit. Dumb, I say.

Allen, I really don't get that. The gate arrow points "out." That means "p-fet" in every schematic I've ever seen.

As for the position of the arrow, all I can say is that the National Semi databooks draw it exactly the way I do, and my FETs came from them. My pet peeve is the people who draw tube cathodes as a little circle.

edit: Siliconix, too. See, for example http://www.datasheetcatalog.org/datasheet/vishay/70233.pdf
edit: And Teledyne in their datasheets.

 

[TheGimp]

Both FETs and BJTs are reversable. However, since geometry of collectors and emitters are different parameters will be different. But I had some Russian military transistors that measured very close, both for collector - base and emitter - base breakdown voltages, and Beta, when reverse connected. They were good as switches.

There are (were) some J-Fets that had symmetrical g-s/g-d geometry that allowed them to be used in very interesting circuits.

Somewhere I had a patent on one that took advantage of this for the discharge of a capacitor in a dual slope integrating A/D converter circuit (when I worked for TI).

 

[Tubelab_com]

There are (were) some J-Fets that had symmetrical g-s/g-d geometry that allowed them to be used in very interesting circuits.

I remember some Motorola j-fets from the 70's that stated in the datasheet that they were reversible. That is not universally the case. In fact it is somewhat rare. It may be possible to get some BJT's to function in a reversible manner, but it would be far less than optimum. An early point contact germanium transistor might work if you could find one!

 

[ilimzn]

Quite a number of small JFETs will work with no great difference with S and D reversed at audio frequencies - usually the clue is that Vgs and Vgd max are the same, for FETs intended for amplification. On FETs intended as switches, some unusual combinations are possible in order to drive the channel into deep cut-off for minimum leakage. VFETs are in the oposite cathegory, with well defined S and D, in fact, S-D and D-S will measure different on an ohm-meter. And, of course, |Vgsmax|<|Vdsmax|<|Vgdmax| since pinch-off voltages can be rather high, so Vgd has to sustain |Vgsmax|+|Vdsmax| in cut-off.

I do agree with AW that the symbols are somewhat misleading, also those for MOSFETs. IGBTs make for even more confusion, so do P-SITs, a species of JFET seemingly indiginous to the ex-USSR...

 

[Mr. dB]

It's the bastard child of Dave Berning and Bill Johnson.

Yeah, Murray Zeligman built some Stereo-70 mods that were almost exactly like this.

 

[SY]

Murray's were a bit different, but same basic idea. Two sides of that tube FET stack. On one side, the gate was grounded and the grid fed with the input signal. On the other side, the grid was grounded and the gate fed with the input signal. There was a LOT of HF imbalance in that approach, so it seemed better to make it a bit more symmetrical.

Do you know Murray? Super guy, great ears, and really knew his music. I apprenticed with him back in the Precedent Audio days.

 

[dgta]

SY, this is an interesting circuit indeed. But what's the bottom line? How does the Syclotron compare to a well executed baseline common cathode plus split load, as illustrated by your Red Light front end? Did you do any comparisons? Distortion, output swing, symmetry, PSRR et al?

Inquiring minds want to know what we get for the extra tube and fets

 

[SY]

Here's the bottom line: I don't currently use it in the amps in my living room.

With an input transformer to give it good balance, it works about as well as a cross-coupled inverter using an input transformer or other balanced input. I have a somewhat improved version that I'll pick up and optimize/measure when I have the next two projects off my plate, just to see if I can work a distortion null. Without an input transformer, the balance is, as with the cross-coupled inverter, a function of mu- balance is proportional to mu/(mu + 1).

The advantage over a voltage amp/cathodyne is common mode rejection. It will be interesting to see if I can squeeze out a linearity advantage as well.

 

[Mr. dB]

Murray's were a bit different, but same basic idea. Two sides of that tube FET stack. On one side, the gate was grounded and the grid fed with the input signal. On the other side, the grid was grounded and the gate fed with the input signal. There was a LOT of HF imbalance in that approach, so it seemed better to make it a bit more symmetrical.

Do you know Murray? Super guy, great ears, and really knew his music. I apprenticed with him back in the Precedent Audio days.

Murray was a friend of a friend. I spoke to him on the phone a few times, but we never met. I only know about the contents of his hybrid-70 because I was once left alone in a room with one for an hour or two, and traced out the circuit. It was a bit of a violation of someone's trust, but I had to know.