Why does this power supply use low-side regulation?

Hint: Power up the input to each section with a ramp. ...Jim Thompson

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

That's a hint regarding the two sides being tied together via R8/R7 (as opposed to all the other items in the quoted text), right? :-)

It's well-behaved and does nothing untoward while powering up. ...Jim Thompson

R7 and R8, which provide the base drive to the series pass transistors, are driven from the opposite side regulator, at effectively higher than rail op voltage, so each rail is dependent on the opposite side as they ramp up on power on.

As you say, quite elegantly subtle, or am I missing something else ?...

Regards,

Chris

Bwahahahahahaha! I missed that before... Wescott used the wrong spelling ;-) ...Jim Thompson

And, r2 and r13 ensure that there is enough base drive so that the o/p's actually start up, otherwise, there would be no base drive on power up...

Regards,

Chris

Some things about this circuit still puzzle me - for example why is the base of Q4 at 8.2 volts? If the cathode of the 6.2 volt zener is at 15 volts, shouldn't that be 8.8 volts?

I believe R8 and R7 are there to prevent a situation where, with this low-side regulated supply, if one side starts regulating before the other then the full filter capacitor voltage would appear on the other output rather than the regulated value.

--- Are you serious?

If you knew anything at all about HP you'd realize that they made their fortune by maximizing clever, minimizing waste, and putting bulletproof instruments into the hands of those who really needed them and could pay for them.

For an example of clever: Once upon a time, parallax was a problem with CRT based oscilloscopes because the graticule wasn't congruent with the plane of the phosphor.

HP solved that problem by depositing the graticule on the inside surface of the CRT's viewscreen, before the phosphor was laid down, much to the chagrin of Tektronix, AIUI.

Another example is their "Transfer Oscillator", a kind of Lissajous scheme enabling microwave frequencies to be counted using a low frequency (45MHz or so) counter.

And, of course, their fillip, the digital frequency counter, the cells of which we now see expressed in devices like the XX190

-- JF

I wondered the same thing. Maybe Zener knees were softer in the '60s -- it was a kinder, gentler time, the era of free love? :-) Contemporary ones such as this:

---Joel

Unfortunately, I learned my spelling from spell checkers. As part of that, I taught myself to catch the homonyms.

Mostly...

(I get principal and principle messed up, too).

If the base of Q4 is actually at 8.2 volts, that implies a drop of 6.2 volts across R6, for a current of 6.8 mA. So if everything is balanced, there'll be a current of about 3.4 mA through R8 - but if the emitter of Q1 is sitting at about 0 volts that means the collector of Q5 is at about 1.2 volts, and the other end of R8 is at -12.6 volts. That would imply a current of about 2.4 mA through R8, so something isn't working out right. The manual specifies all silicon devices so no error there. If Q4 is actually at 8.8 volts the numbers are closer to working out correctly...so am I smoking something, or the HP drafter? :)

I'd vote for the drafter!

I was just funning with you. I also struggle with principal and principle.

I'm beginning to hate spell checkers... my fingers regularly get scrambled and type "fro" when I meant "for" and the spell checker lets it go by :-( ...Jim Thompson

Yup. hp of old was marvelous quality... except I never liked their scope triggering.

And old GenRad stuff (in Walnut cases :-)

In both companies, the newer stuff is horrible.

...Jim Thompson

in

ing

own

e, is

er

round

on... if

it,

rlington

e will

er

the

ollector

), Q3

an no

f Q2's

re

.,

=A0 =A0 ...Jim Thompson

=A0 =A0| =A0 =A0mens =A0 =A0 |

=A0 | =A0 =A0 et =A0 =A0 =A0|

=A0|

=A0 =A0 =A0 |

quoted text -

It's an ICBO startup circuit...has nothing to do with any crosscoupling between regulators and opposite error amps- the zeners keep the error amps locked off all the way. The 6.2V zeners are used because of the zero tempco...nothing at all special about the circuit.. looks like something right out of the Motorola Voltage Regulator Handbook from the 70s.

I was commenting about the metalwork and paint job. I wouldn't economize on the differential pair. The inverted polarity schematic for inverted polarity output is the issue here.

There are many ways to control start-up and loosely track the outputs that don't depend on beta of power transistors operating in their SOA, or cost a second non-optimal power device in the purchasing system. Somebody must have been impressed,though.

Its impressing that the design team managed to push plastic-cased pass transistors into the build in 1968. I still couldn't do that, in some companies I worked for, as late as 1985.

Adjustment pots with single contacts to their wiper (or any kind of pot for fixed outputs..) or 10uF (1968) capacitors in feedback compensation are in my book of 'why you can't do that'-s.

Considering the size of HP in '68, it's hard to believe there wasn't already a better circuit already in the field that could have been re-canned for the specific job of powering a probe. Seeing as it was given it's own part number and could deliver more than 1A, I expect that was actually it's intended purpose. Interesting to know if it did show up elsewhere.

The fact that Agilent still supplies the product's manual, with minimal revisions (or corrections), is an indication that there are many of them still in the field after all this time. This puts the kybosh on any reservations that I might have, 40 years after the fact.

Time to do the chores...

RL

Legg, Could you point out that beta dependence? I'm not seeing that.

...Jim Thompson

Can't you edit your dictionary? I removed a bunch of crap from the default dictionary a few computers ago.