Answers to old-fart electronics quiz

[John Woodgate]

I read in sci.electronics.design that Max Hauser <[email protected]>
wrote (in <[email protected]>) about 'Why does TTL run

The analog world tends to lag
the digital world, very broadly speaking of course, in moving to lower
supply voltages.

Yes, because the analogue circuits frequently feed transducers, heaters,
etc. that require significant power. This is often optimally supplied a
quite a high voltage. Digital circuits rarely need to supply much
power. When they do (e.g. PWM motor drives, if you class those as
digital), then they do use higher voltages.

 

[John Woodgate]

I read in sci.electronics.design that N. Thornton <[email protected]>

I'm not very up on microwave stuff. I had a looksee. Just call me
creditless. I did find EER amplifiers tho, and posted under I think
it was Amplifier Classes or something.

What led you to think it was a microwave technique? Certainly not what I
had in mind.

 

[N. Thornton]

I read in sci.electronics.design that N. Thornton wrote:

What led you to think it was a microwave technique?

It is

I'm tempted to say its a mode of operating microwave ovens where the
output is hot pies, but someone somewhere might take it seriously

Certainly not what I had in mind.

Your turn then.


Regards, NT

 

[John Woodgate]

I read in sci.electronics.design that N. Thornton <[email protected]>

It is

I'm tempted to say its a mode of operating microwave ovens where the
output is hot pies, but someone somewhere might take it seriously



Your turn then.

Pi-mode was a particular case of Class AB, with the standing current
equal to 1/pi times the full load current. It had some claimed
theoretical advantages, which I don't now remember. Mullard (Philips UK)
published a design, but AFAIK, it didn't catch on. Maybe it was before
its time; it used germanium power transistors, and the last thing you
want with them is substantial standing power dissipation.

 

[Gerhard v d Berg]

Weren't the precursors to TTL, RTL and DTL, also 5 volts? Perhaps it's
a legacy issue.

Jim

The designers at TI should have the answers.
Legacy compatibility might have been an important concideration.

Another factor was the multi-emiter (NAND/AND) input.
If the one input was high and the other low, the eb junction would
breakdown if the voltage difference was exceeded by much more
than 6 volts (specs normally stated 5.5v max).
A clamp diode on all input prevented large negative input excursions.
Such excursions would have exsasivated the problem.
That's why it was the recommended paractice to wire spare NAND
inputs to Vcc using a 1K or higher resistor to limit breakdown
current to a safe value and spare the multi-emitter emitter-base junction,
especially if supply spikes above 7 volts were present.

Many 74LSxx circuits employed Schottky diodes to realise the NAND
input and could tolerate 7 volt (according to specs) inputs and also 7v Vcc.

(TI's The TTL Data Book for Design Engineers page 5-4 1979 3rd ed
ISBN 0-904047-27-X)

Just my 2 cents

Gerhard van den Berg

 

[N. Thornton]

I read in sci.electronics.design that N. Thornton <[email protected]>
Pi-mode was a particular case of Class AB, with the standing current
equal to 1/pi times the full load current. It had some claimed
theoretical advantages, which I don't now remember. Mullard (Philips UK)
published a design, but AFAIK, it didn't catch on. Maybe it was before
its time; it used germanium power transistors, and the last thing you
want with them is substantial standing power dissipation.

I'd never heard of this one before. I did a quick google, and nothing.
As you say, it sounds like geraniumicide. 'Germanium' and 'power
transistor' just arent a great combination. It also sounds too
inefficient to see much use today.

I could take a guess at a possible advantage: if used for music, the
majority of the time it would operate in class A, along with the low
distortion benefit of A, but when pushed harder it will be running
Class AB, thus allowing a fair output power and efficiency while
maintaining the Class A advantages for the rest of the time. If its
used for music, it might do Class A for most of it and Class AB for
the drum beats.

With todays greater open loop gains at lower prices, it seems to offer
little thats particularly attractive. I cant imagine offhand how
there'd be another advantage to it, but who knows - not me.


This reminds me, I think there are 2 Class Cs. The modern usage is
peak conduction only for rf use, but ISTR Class C was once a common
description for a version of Class B with valves, as I think was QPP.

I like obscure amp classes and am surprised there arent more of them.
Just call me odd


Regards, NT

 

[Jan Panteltje]

I'd never heard of this one before. I did a quick google, and nothing.
As you say, it sounds like geraniumicide. 'Germanium' and 'power
transistor' just arent a great combination. It also sounds too
inefficient to see much use today.

Well, not sure that makes much difference (someone correct me if I am wrong).
I mean germanium or silicon, first we had OC16 power transistors, later
better ones , trying to remember the ge horizontal deflection
ones for TV lemme look
OK now I have this old book (tables) and it is marked G for germanium and S for silicon,
now lets see,
Ic max and Uvce max
ac128 2 A 16 V low power
AD161 3A 20 V NPN 4W
AD162 3A 20 V PNP 6W
I had a Philips audio amp with these...
AU110 10A 160 V 35 W PNP
ASZ15 10 A 60 V 35 W PNP
ASZ18 10A 35 V 30 W PNP
I had some Motorola big one (used for car ignition), can't remember the number,
high voltage high current.
That is all old technology, perhaps si was just easier to make, or perhaps simply cheaper.
The ge ones needed a much lower Vbe to conduct.
JP

 

[Jan Panteltje]

I'd never heard of this one before. I did a quick google, and nothing.
As you say, it sounds like geraniumicide. 'Germanium' and 'power
transistor' just arent a great combination. It also sounds too
inefficient to see much use today.

Well, not sure that makes much difference (someone correct me if I am wrong).
I mean germanium or silicon, first we had OC16 power transistors, later
better ones , trying to remember the ge horizontal deflection
ones for TV lemme look
OK now I have this old book (tables) and it is marked G for germanium and S for silicon,
now lets see,
Ic max and Uvce max
ac128 2 A 16 V low power
AD161 3A 20 V NPN 4W
AD162 3A 20 V PNP 6W
I had a Philips audio amp with these...
AU110 10A 160 V 35 W PNP
ASZ15 10 A 60 V 35 W PNP
ASZ18 10A 35 V 30 W PNP
I had some Motorola big one (used for car ignition), can't remember the number,
high voltage high current.
That is all old technology, perhaps si was just easier to make, or perhaps simply cheaper.
The ge ones needed a much lower Vbe to conduct.
JP

 

[John Woodgate]

I read in sci.electronics.design that N. Thornton <[email protected]>

This reminds me, I think there are 2 Class Cs. The modern usage is peak
conduction only for rf use, but ISTR Class C was once a common
description for a version of Class B with valves,

I never came across that in the old books I used to read in my youth.

as I think was QPP.

QPP (Quiescent Push-Pull) is another name for Class B, specifically
Class B1 in the case of valves/tubes.

 

[N. Thornton]

Well, not sure that makes much difference (someone correct me if I am wrong).
I mean germanium or silicon, first we had OC16 power transistors, later
better ones , trying to remember the ge horizontal deflection

The prime reason 'Germanium' and 'power transistor' are a grotty
combination is that leakage goes up heavily with temp, and a great
many died from thermal runaway.


Regards, NT

 

[Jan Panteltje]

The prime reason 'Germanium' and 'power transistor' are a grotty
combination is that leakage goes up heavily with temp, and a great
many died from thermal runaway.


Regards, NT

I have seen tubes die from termal runaway.... (wel not runaway perhaps, say overdrive).
remember red anodes, sometimes the glass would even melt.
It is a design issue, you can protect for that with a simple sensor,
that was for example even needed in the SEPP audio amps with
si transistors.
MOSFET power decreased current with temp rize.
Would not be a reason not to use ge.

 

[Ian Buckner]

On Sat, 24 Jan 2004 12:06:36 -0800, "Max Hauser"

Even in digital, the standard 5 volt (or even 3.3 volt) rail is dead.
I'm having to do digital designs with a slew of different regulators
for different chips: 5 volts for bus interface and uP, 3.3 for FPGA
i/o and ram chips, 2.5 for one fpga core, 1.8 for another. Keeping
track of logic level compatibilities and supply sequencing is a
nightmare. I have one mixed-signal VME board with eight regulators.

Snarl.

John

I know what you mean, John. I'm having to do just that (multiple
regulators) at the moment.

Here's a pattern that may hint at what comes next:

Feature size: 0.35um 0.25um 0.18um 0.15um 0.13um
Rail voltage: 3.3V 2.5V 1.8V 1.5V 1.2V

Regards
Ian

 

[John Woodgate]

I read in sci.electronics.design that Ian Buckner
com>) about 'Why does TTL run on 5 volts (was Re: Answers to ...)', on
Tue, 27 Jan 2004:

Here's a pattern that may hint at what comes next:

Feature size:0.35um 0.25um 0.18um 0.15um 0.13um

Differences: 0.1 0.07 0.03 0.02

Rail voltage:3.3V 2.5V 1.8V 1.5V 1.2V

Differences: 0.8 0.7 0.3 0.3

4 more steps of 0.02 um/0.3 V gets to:

Feature size: 0.05
Rail voltage: 0.0

Is it permitted to extrapolate to negative values of feature size?

 

[John Larkin]

I know what you mean, John. I'm having to do just that (multiple
regulators) at the moment.

Here's a pattern that may hint at what comes next:

Feature size: 0.35um 0.25um 0.18um 0.15um 0.13um
Rail voltage: 3.3V 2.5V 1.8V 1.5V 1.2V

Regards
Ian

I guess we'll have to put a tiny point-of-use switching regulator (or
two!) next to every big digital chip. The regs will get to be bigger,
and maybe more expensive, than the chips.

What's next? FPGAs with fans on top? It's all that guy Moore's fault.

John

 

[John Woodgate]

I read in sci.electronics.design that Jan Panteltje
01.evisp.enertel.nl>) about 'Answers to old-fart electronics quiz', on
Mon, 26 Jan 2004:

I have seen tubes die from termal runaway.... (wel not runaway perhaps,
say overdrive). remember red anodes, sometimes the glass would even
melt.

My 50 W PA amplifiers with KT88s suffered real thermal runaway on a hot
day until I improved the ventilation. Real thermal runaway is when the
control grid gets hot enough to start emitting electrons - it doesn't
need many to cause the cathode current to skyrocket. Result - burned out
cathode resistors.

 

[Spehro Pefhany]

I guess we'll have to put a tiny point-of-use switching regulator (or
two!) next to every big digital chip. The regs will get to be bigger,
and maybe more expensive, than the chips.

Is there any technical reason they can't be put internally, if chip
designers could just get over their bias against (external, in this
case) inductors? AFAIK, internal regulators are mostly charge pumps or
linear.

What's next? FPGAs with fans on top? It's all that guy Moore's fault.

Gordon, Michael or Benjamin?

Best regards,
Spehro Pefhany

 

[Jim Thompson]

Is there any technical reason they can't be put internally, if chip
designers could just get over their bias against (external, in this
case) inductors? AFAIK, internal regulators are mostly charge pumps or
linear.


Gordon, Michael or Benjamin?

Best regards,
Spehro Pefhany

I've done a 3.3V to 2.5V, at 1A, on-chip linear regulator in a CMOS
process.

Just depends on where you want the heat and the space utilization.

...Jim Thompson

 

[Bob Stephens]

Is there any technical reason they can't be put internally, if chip
designers could just get over their bias against (external, in this
case) inductors? AFAIK, internal regulators are mostly charge pumps or
linear.


Gordon, Michael or Benjamin?

Best regards,
Spehro Pefhany

Dinty

 

[Spehro Pefhany]

On Tue, 27 Jan 2004 10:37:56 -0700, the renowned Jim Thompson
^^^^^^^^^^^
Gretzky?

I've done a 3.3V to 2.5V, at 1A, on-chip linear regulator in a CMOS
process.

Then a current-mode buck switching regulator with external inductor
should be possible, right? Or is there a problem getting enough gate
drive on the "pass" transistor? The diode to ground could be an
internal MOSFET or could be an external Schottky. Or is all this a
stupid suggestion because MOSFET acreage costs much less than VLSI
acreage? Or for some other reason?

Best regards,
Spehro Pefhany

 

[Jim Thompson]

On Tue, 27 Jan 2004 10:37:56 -0700, the renowned Jim Thompson
^^^^^^^^^^^
Gretzky?


Then a current-mode buck switching regulator with external inductor
should be possible, right? Or is there a problem getting enough gate
drive on the "pass" transistor? The diode to ground could be an
internal MOSFET or could be an external Schottky. Or is all this a
stupid suggestion because MOSFET acreage costs much less than VLSI
acreage? Or for some other reason?

Best regards,
Spehro Pefhany

Now-a-days I don't question the client's reasoning. I do what's
asked. Otherwise I've been losing jobs, particularly due to young
bucks trying to kill off the consultant before I show them up :-(
Fudging data, doing CMOS sims without parasitics to make their
solutions look viable, etc. I'm waiting for production time... then
my rates go up ;-)

...Jim Thompson