John Curl's Blowtorch preamplifier part II

[buzzforb]

Moving on: (if possible) Let me show what we USUALLY use for the 4Q input stage to cascode it. There are a number of advantages, including more real voltage across the input part to lower and linearize its inherent input capacitance.
Easier part selection, only 3 extra resistors (minimum) added, better high voltage standoff.

To my simple mind, the bias net of the cascode jfets is interesting, as itappears to be floating. What affect does this apparent intentional choice have on performanceor is it unimportant in the schem of things.

 

[Bonsai]

Seriously? I remember >30 years ago integrated amps tended to have an input sensitivity of around 250mV for full output. Then CD came along with the 2V max output standard. I would have thought everything else would have moved in the same direction in the interests of compatibility, maybe taking 700mV or so average output as a target for analogue sources.

(not that I've been paying attention the last couple of decades - I quit buying hifi magazines around the time I started buying baby food, not coincidentally)

Yes, you are correct about amplifier input sensitivity. The last 20 years or so, most power amps settled on c. 1V input sensitivity, with a cluster around 700-800 mV, and other at 1.5 to 2V. If you stick with 1V you will be in good company in general.

 

[bcarso]

What would indicate less than ideal performance from the depletion mode devices? I have had no negative experiences with them as cascodes on some products in production for a while. What have I missed?

Just low voltage for the part below as it were. Which is not a big problem if they are super low pinchoff JFETs.

 

[iaxxaxxai]

It is a shame that great jfets are harder to find, and have gone up in cost, but that is not a problem that I can fix. Actually, most really good bipolars are hard to get, too. It would appear that the most popular bipolars, today, are devices that I used more than 40 years ago. This is progress?

This jumps out at me. I watch these threads and see discussion of circuits that seem very educational, but realize that attempting to even experiment with the concepts would be a major hassle as far as parts sourcing.
It's depressing. Like, why bother. It's easier to get good tubes. How sad is that?

 

[iaxxaxxai]

I can find good quality black plate 6SN7s all day long, but find a good complimentary JFET pair? Forget about it!!!

 

[Steve Eddy]

I can find good quality black plate 6SN7s all day long, but find a good complimentary JFET pair? Forget about it!!!

Funny isn't it how tube circuits can sound absolutely fantastic, yet there are no complimentary devices in the tube world.

se

 

[jan.didden]

Funny isn't it how tube circuits can sound absolutely fantastic, yet there are no complimentary devices in the tube world.

se

Very thoughtful Steve!

 

[Bonsai]

Funny isn't it how tube circuits can sound absolutely fantastic, yet there are no complimentary devices in the tube world.

se

He he. I like that.

 

[wrinkle]

Funny isn't it how tube circuits can sound absolutely fantastic, yet there are no complimentary devices in the tube world.

se

All we need are positron based antimatter triodes...

Mind you there will be a few issues getting the wires to connect.

 

[staccatiss]

Actually, your eq is for gmo, and not gm in general.
In general:

gm = |2 * (IDSS/VP) * [1 - (VGS/VP)]|


and to use your gmo, we have

gm = gmo * sqrt(ID/IDSS)


And do note, that we are still using model equations.

So many words for Gm = 2*Idss/Vp

 

[jacco vermeulen]

Forget about it

If you and a million others were willing to pay $25 for a JFET pair every time, it wouldn't be an issue.
You skipped the most vital criterium, the fact that vacuum valves still is a multi billion dollar business.

(two decades ago, the situation was reversed : folks got out of tube amps and into SS, as it became harder to find valves and/or decent OPT's)

 

[EUVL]

You can still use BF862, 2SK3557, 2SK209, 2SK2145, ......

I do not see any problems.
You just need to start building circlotrons, or Linsley-Hood-type quasi-push-pull's.


Patrick

 

[EUVL]

And progress is iphone, 256GB USB sticks the size of a finger nail, ..... not analog electronics.
Why? Analog does not sell in 3-digit millions a quarter at 3-digit USDs.

Sad, but true.


Patrick

 

[godfrey]

And yet 741 opamps and 2N3055s just won't go away. I guess they're good for industrial stuff.

 

[SY]

If you and a million others were willing to pay $25 for a JFET pair every time, it wouldn't be an issue.

I would IF they were as close to nominal spec as a pair of tubes are. The problem is the three or four to one spread in gm and idss, meaning that I have to buy twenty FETs to get two I can use optimally (if I'm lucky and they're not all clustered over to the wrong side of the spec limits). In many circuits, you can design around this, but not in all- and unfortunately, not in the circuits where I like to use FETs.

 

[jacco vermeulen]

Likely the reason I run a semi museum.

(afaih, there are still folks ready & willing to pay dearly for mil-spec matched 6SN7 pairs)

 

[dadod]

Funny isn't it how tube circuits can sound absolutely fantastic, yet there are no complimentary devices in the tube world.

se

Sorry, but as not native Englesh speaker I would like to know if "complimentary" has same meaning as complemetary in American English?

 

[SY]

No, it's a common writing error. "Complementary" is the right word unless the devices are telling you that you're a handsome and intelligent fellow.

 

[jacco vermeulen]

Buy a valve, get a transistor for free. (complimentary towel of the tube hospitality world)

Speaking of semi's and tubes : hybrid folded cascode (bad boy JC Morrison)

 

[EUVL]

> The problem is the three or four to one spread in gm and idss, meaning that I have to buy twenty FETs to get two I can use optimally

I am sure Jacco can help you out.

But serious, you can still get matched pairs with specific Idss at reasonable price (less than tubes).
And tubes also need matching, not true ?


Patrick