Williamson ring of two current source?

Version 4 SHEET 1 880 680 WIRE 176 -272 0 -272 WIRE 384 -272 176 -272 WIRE 176 -224 176 -272 WIRE 384 -224 384 -272 WIRE 384 -96 384 -144 WIRE 176 -48 176 -160 WIRE 320 -48 176 -48 WIRE 0 96 0 -272 WIRE 176 128 176 -48 WIRE 384 176 384 0 WIRE 384 176 240 176 WIRE 384 256 384 176 WIRE 176 272 176 224 WIRE 0 384 0 176 WIRE 176 384 176 352 WIRE 176 384 0 384 WIRE 384 384 384 320 WIRE 384 384 176 384 WIRE 384 416 384 384 FLAG 384 416 0 SYMBOL npn 240 128 M0 SYMATTR InstName Q1 SYMBOL pnp 320 0 M180 SYMATTR InstName Q2 SYMBOL voltage 0 80 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value 12 SYMBOL res 160 256 R0 SYMATTR InstName R3 SYMATTR Value 1k SYMBOL LED 368 256 R0 SYMATTR InstName D2 SYMATTR Value QTLP690C SYMBOL LED 160 -224 R0 SYMATTR InstName D1 SYMATTR Value QTLP690C SYMBOL res 368 -240 R0 SYMATTR InstName R1 SYMATTR Value 1k TEXT -80 248 Left 0 !;op TEXT 104 472 Left 0 !.tran 0 10m 0m 10n

offers better temperature stability.

-- Bill Sloman, Nijmegen

"Martin Griffith" schrieb im Newsbeitrag news: snipped-for-privacy@4ax.com...

It's something like a SCR. Therefore choose zero Ohm resistors in the emitter. It will be triggered if you inject a small current into the base of one of the two transistors.

Best regards, Helmut

Cheers Bill, keep up the good work :)

martin

Yes, but neither your circuit:

+12v >------+--------+ | | D1 --- .-. \\ /~~> | | R1 ----- | | 1K | '-' | | | | | |------+--------+ | | D1 --- .-. \\ /~~> | | R1 ----- | | 1K | '-' | | | +------. | | | | |

It is the *Peter Williams* 'Ring of Two' current source that is under discussion (not DTN or Reggie Williamson).

The org ref was/is:

P. Williams, ?Constant Current Circuit,? Wireless World (Letters), Sept. 1966, pp. 456.

The shunt R that James Arthur added was mentioned by another correspondent, Rudge, as in:

J. C. Rudge, ?Constant-Current Circuits,? Wireless World (Letters), Dec. 1966, pp. 610.

With this R in place, Rudge noted that the current has a negative characteristic, thus if this R is chosen fortuitously, it can effectively cancel the positive Rout of the xstrs. Making the net Rout >>, a nice trick.

Walt Jung

snip

Wow you must have a good library

martin

You don't recognize the name "Walt Jung"?

Bob

--
== All google group posts are automatically deleted due to spam ==

You mean "god".........

Of course I do, He is second only to Win Hill, a great achievement

martin

We should see if we can get Mr. Jung and Mr. Hill to engage in one of those no-holds-barred pay-per-view octagon arena Ultimate Fights.

How cool would that be?

Bob

--
== All google group posts are automatically deleted due to spam ==

God's don't need money, but it would be available on bittorrent FOC within an hour or two anyway.

Or we could get the mediocre TED talks to do something creative, for a change

martin

Well, the WW issues I have didn't actually include 1966/67, and I actually had to *buy* a copy (ugh)! And I still haven't nailed them all, even yet. Anyone have 1967 WW's out there?

But, more importantly, I made an error yesterday, in commenting on James Arthur's circuit. His is *not* exactly like Rudge's version. Rudge shows the added R between the two xstr *emitters*, not as JA has it. JA's version addresses the startup issue, which is real.

It is instructive to do a quick DC analysis of the currents, and note the negative characteristics with R3 from emitter-emitter, as Rudge had it.

Sorry for the misspeak, but hopefully this clarifies it.

Walt Jung

That *is* cool Walt, and thanks for the history. As you've noted in a later correction R3 isn't quite the same as Rudge's improvement. Specifically, Rudge's emitter-to-emitter resistor doesn't assure startup, and my R3 doesn't improve output impedance.

One might fear placing R3 across Q2 will ruin the circuit's output impedance, but t'ain't so. R3's current comes from robbing Q2's emitter, providing some emitter degeneration, which minimizes R3's impact on Zout.

A quick SPICE shows z(out) 1.1 - 1.8 megohms with R3 = 100k - 10 meg, and 1.8 meg with R3 omitted. (I used 2n3904 / 2n3906 transistors.)

Rudge's resistor, of about 1.9 megohms, increases z(out)

8x to about 15 megohms, as advertised.

Here's a version using both tricks together:

. +12v >------+--------. . | | 2mA current sink/source . D1 --- .-. Zout ~16 megohms (5-20v) . \\ /~~> | | R1 . ----- | | 1K . | '-' . | | . .----- | -------+------. . | | | | . R4 | | |