Date   

Re: Risetime calculator (in tekwiki)

Roy Thistle
 

On Tue, Jan 21, 2020 at 02:12 AM, Leo Bodnar wrote:


But (and this is where the trick is)...
I had thought it was known that the pots in the Rise Time Calculator were known to be log taper, or some kind of log taper (by a collector who had one, and had measured their tracking). In fact this whole rabbit hole, IMHO, could have been avoided if wherever this thing(s) are someone would have measured the pots, and trimpots. But of course that not the sort of thing one can really request... as we all value our time, and need to respect other's time too.
By "some kind of taper" I mean the following:
I am supposing that many people forget there was once an astounding variety of high quality (and so expensive) specialty tapered pots to do all the algebraic functions, exponentials, and trig functions too. Some were motor driven, or servo/sycro driven. Others were custom tapered for non-closed form functions (or what many considered not in commonly accepted set.) They could be used to do non-real time analog computation... or slow real-time stuff.
We once had a good assortment of them... if I remember, many were army surplus?... and they were great fun to play with. Since many physical phenomena are governed by "square laws" we definitely had quadratic taper pots. We also used to put resistors, or trimpots in parallel, or across the wiper to end lugs, to change the characteristic of the taper. Some audio types still do this with log/linear taper pots to get the kind of "characteristic" sound they want from the audio controls.
Best regards and wishes.
Roy


Damaged 606

Lars Brinkhoff
 

Hello,

I bought a Tektronix 606 off eBay. When I tested it I get this:

https://imgur.com/a/cLOaOVu

There's a bright spot fixed at the center. There's a movable spot, or
short line pointed towards the center, which is to the lower left in the
photo. Between the two is a squiggly line.

I'm a bit out of my depth here. What could cause this? Any advice
would be welcome.


Re: Risetime calculator (in tekwiki)

Roy Thistle
 

On Sun, Jan 19, 2020 at 11:10 AM, Albert Otten wrote:


The easiest way to draw the markings is linear in angle, which only applies
when the pots are linear.
Working from what you said/did, and my own schematic drawing:
if the 4.95K resistor + the trimpot = 5K for 5, which has been the assumption so far?
if the wiper divides the total resistance of the pot into t, and b so that t+b = 5
if the shaft rotation is related to b linearly, which has been the assumption so far?
then for any of pots i in (A,B,C) you can show t_i = (25 -t_i^2) /10 (1)
and you can show for pot F, t_f = [(25 - t_f^2) / 10] + 5 (2)
For a balanced bridge signifying equality
Sum_(a,b,c)[(25 -t_i^2)/10] = [(25 - t_f^2)/10] +5
all the integers vanish, and
t_a^2 +t_b^2 +t_c^2 = t_f^2
so that
f = +sqrt(t_a^2 +t_b^2 +t_c^2)
quod erat demonstrandum... or a little square box, I suppose.
So by the math as you showed, if the pots are linear, and the trimpots are adjusted to make the 4.95K + trimpot = 5K then you get the required function.
Best wishes and best regards.
All the best
Roy


Re: Risetime calculator (in tekwiki)

Roy Thistle
 

On Tue, Jan 21, 2020 at 10:19 AM, Albert Otten wrote:


You probably meant (2,0,5) Roy?
A =3 , B = 4, F = 5 (and C = 0) would easier be recognized, a familiar integer
"Pythagoras" solution.
Sorry... should be (2,3,4), (29).
Yes, any Pythagorean triple would work, if you take it that a pot can go to true 0 linearly?... most don't?... there is a step non-linearity... but I haven't thought this through yet. Three non-zero integers, which leads to 1 non-zero integer, seemed to avoid that problem, if it is one.
Best wishes. All the Best
Roy


Re: Risetime calculator (in tekwiki)

Chuck Harris <cfharris@...>
 

One thing about engineers from the slide rule era, is they
were very, very, good at approximations, or they couldn't
make the slide rule do beans.

The slide rule handled the fractions, the engineer figured
out where the decimal point went, and the exponents.

-Chuck Harris

GerryR wrote:

Never mind, then-came-the-dawn; three different scales  to get from 0 to 20 and I
assume that proportionally you can try to put your values as close to center dial as
possible for greater accuracy by using the different scales. I can be thick, sometimes!
GerryR
KK4GER


Re: Risetime calculator (in tekwiki)

Albert Otten
 

On Tue, Jan 21, 2020 at 06:30 PM, Roy Thistle wrote:
I assume if you program in the numbers (2,3,5) on the Blue scale which gives sqrt(29). ...
You probably meant (2,0,5) Roy?
A =3 , B = 4, F = 5 (and C = 0) would easier be recognized, a familiar integer "Pythagoras" solution.
Albert


Re: Risetime calculator (in tekwiki)

GerryR <totalautomation1@...>
 

Never mind, then-came-the-dawn; three different scales to get from 0 to 20 and I assume that proportionally you can try to put your values as close to center dial as possible for greater accuracy by using the different scales. I can be thick, sometimes!
GerryR
KK4GER

----- Original Message -----
From: "Chuck Harris" <cfharris@erols.com>
To: <TekScopes@groups.io>
Sent: Tuesday, January 21, 2020 10:26 AM
Subject: Re: [TekScopes] Risetime calculator (in tekwiki)


The dial scale values indicate numbers you want to
feed the equation SQRT(A^2 + B^2 + C^2) = f

In this case, if you set them to microseconds, f will
be in microseconds... nanoseconds, f will be in nanoseconds...

Obviously, you will have to interpolate your settings and
results, to gain more than integer accuracy... which you
should be able to do by eye to at least quarters.

-Chuck Harris

GerryR wrote:
Dennis,
What do the dial scales values indicate??
GerryR
KK4GER



----- Original Message ----- From: "Dennis Tillman W7PF" <dennis@ridesoft.com>
To: <TekScopes@groups.io>
Sent: Tuesday, January 21, 2020 12:09 AM
Subject: Re: [TekScopes] Risetime calculator (in tekwiki)


Hi Everyone,
Before getting caught up in the concept of an analog computer take a step back to
consider what an analog is. It is something that is similar to something else; the
two are said to be analogous.

THE RISETIME CALCULATOR IS AN ANALOG COMPUTER BASED ON LOGARITHMS
The risetime calculator is a simple, but clever, analog computer that relies on an
analogy between the angle of the dial on each pot and the resistance of the pot at
that angle. This in turn relies John Napier's discovery of logarithms, which he first
published in 1614. Logarithms are a method to multiply two numbers together by adding
their exponent. Anyone familiar with a slide rule knows the C and D scales on a slide
rule are logarithmic and not linear. The A and B scales increase twice as fast as C
and D because they represent a number times itself or the square of a number. If I
wanted to multiply 7 times 5 I simply had to add together the logarithm of 7 (in base
10) to the logarithm of 5 (in base 10). The log (short for logarithm in base 10) of 7
is 0.845. The log of 5 is 0.699. 7 x 5 = log 7 + log 5 = 0.845 + 0.699 = 1.544. If we
look up the number whose log is 1.544 it is our answer: 35. Commonly available 10"
slide rules are accurate to 2 or 3 decimal places. This is accurate enough for an
enormous variety and number of calculations so they were widely used until the
introduction of mechanical and then electronic calculators.

HOW IT WORKS
If the three pots on the left side of the risetime calculator have a logarithmic
taper and the current coming out of each was added together the sum would = A + B +
C. That's not what we want. We want A*A + B*B + C*C. To do that the taper has to
change logarithmically but twice as fast as an ordinary logarithmic taper. If the
first pot's taper increases twice as fast as an ordinary logarithmic taper the
current that comes out of it will be analogous to A*A. The same taper used for the
second pot will produce a current analogous to B*B, and the third pot will produce a
current analogous to C*C. Now if we sum together the currents coming from the three
pots we will get A*A + B*B + C*C. Our next step is to find the number, using the pot
on the right side, whose current equals the sum of the 3 pots on the left side. The
minimum current from a pot on the left side will be when it is counterclockwise
representing 1. 1*1 = 1, and the log of 1 is 0 so there would be 0 units of current
coming out of the pot. If the pot were set to 10, then 10*10 = 100 and the log of 100
is 2. So there would be 2 units of current coming from the pot. The currents are
summed from all three pots and the total can range from 0 to 6 units of current. With
the meter we have to balance (or null) the current on the left side with an equal
current from the right side knob. There is no need to take the square root of the
left side provided we square the right side instead. So on the right side we can use
one more pot as long as it has the same taper as the other three pots. We can even
use the same scale. One small drawback to doing this is there exists the possibility
that if all three risetimes happen to be greater than 5.773nSec you won't be able to
balance the bridge since the total risetime will be larger than 10nSec.

ANALOG VS DIGITAL OSCILLOSCOPES
High quality oscilloscopes strive to present an accurate analog of the voltages or
currents in a circuit being probed. Until recently all of the oscilloscopes in the
world were analog and there was no need to use the word analog to describe them. I am
absolutely certain no one gave that word a second thought until another way to make
an oscilloscope became possible in the 1970s. Suddenly it was important to
distinguish whether the oscilloscope made every attempt to present you with a wiggly
line on a CRT that was an analog of what was happening in the circuit you were
measuring or if what you saw looked like a connect the dots page from a coloring book.

How closely the analog displayed on the CRT compared to the actual potential in the
circuit being measured was an indication of the quality of the instrument. Enormous
amounts of money were spent over 100+ years to insure the trace on the CRT was an
accurate analogue of the voltage in the circuit. The result was something that could
be within +/-1%. Today there are two kinds of oscilloscopes. One important attribute
distinguishes analog oscilloscopes from digital oscilloscopes.

* The representation on the CRT of an analog oscilloscope is a CONTINUOUS analog of
the signal being measured (except for sampling plugins). One advantage of an analog
oscilloscope, as long as you use it within its specified limits, is the
representation on the CRT has a high likelihood of being a good analog of the signal
being measured.

* The representation on the display of a digital oscilloscope is a SERIES OF DISCRETE
DOT PAIRS that may (or may not) appear to form one or more patterns that humans will
assume, sometimes erroneously, are continuous and then they might conclude it is an
analog of the signal being measured. The interpretation of the pattern of dots on a
digital oscilloscope requires a detailed understanding of how the samples are taken,
how the samples are displayed, the type of sampling used, and the limitations of
sampling itself.

Analog oscilloscopes are desirable for new or unsophisticated users because the
results are less likely to be misunderstood. Digital oscilloscopes have many
limitations and pitfalls that analog oscilloscopes do not. They require that their
results be interpreted carefully. The digitized data from a signal can be further
processed mathematically to extract additional information from it or, for example,
to reduce the noise in the digitized data. There are many other things that can be
done to a digitally sampled signal after it has been captured to further process it.

Dennis Tillman W7PF





Re: Risetime calculator (in tekwiki)

Roy Thistle
 

On Tue, Jan 21, 2020 at 09:08 AM, GerryR wrote:


why the different scales for the rise time
Hi GerryR:
Blue scale is 0 to 5
Black is double Blue so 0 to 10
And Red is double Black so 0 to 20
That's 2^0, 2^1, 2^2, for 1,2,4 multipliers.
If it is like others, of these type of simple analog computers, I've played with:
If you set A = 3, B =4, C=12, (3,4,12), on the Red scale, you should set F to sqrt(3^2+4^2+12^2) = sqrt(169) = 13 ... so you should set the F knob to 13 on the Red scale...push the button... and the bridge should be very close to balance... if the Rise Time Calculator has been properly calibrated.
I assume if you program in the numbers (2,3,5) on the Blue scale which gives sqrt(29), which is greater than 5, you would not be able to balance the bridge, and should reprogram them in on the Black scale.
I don't have one of the Tektronix Rise Time Calculators... so I am not sure that is indeed the way it works. (I'll try to check the math later.)
By the way, for those of us that love math... (3,4,12) = (13) is a solution to the quadratic Diophantine equation X^2 + Y^2 + Z^2 = W^2, such that all of X,Y,Z,W are integers. If you have a Tektronix Rise Time Calculator remembering that might help you.
Best regards and wishes.
Roy


Re: Risetime calculator (in tekwiki)

GerryR <totalautomation1@...>
 

If you look at the dials, each is calibrated 0 thru 10. Then there are numbers in red, 0 -20 and another series 0-5. Just trying to figure out what the different scales mean in the context of rise time. I realize that the numbers can relate to anything you want to take the root-sum-square of, but why the different scales for the rise time?
GerryR

----- Original Message -----
From: "Chuck Harris" <cfharris@erols.com>
To: <TekScopes@groups.io>
Sent: Tuesday, January 21, 2020 10:26 AM
Subject: Re: [TekScopes] Risetime calculator (in tekwiki)


The dial scale values indicate numbers you want to
feed the equation SQRT(A^2 + B^2 + C^2) = f

In this case, if you set them to microseconds, f will
be in microseconds... nanoseconds, f will be in nanoseconds...

Obviously, you will have to interpolate your settings and
results, to gain more than integer accuracy... which you
should be able to do by eye to at least quarters.

-Chuck Harris

GerryR wrote:
Dennis,
What do the dial scales values indicate??
GerryR
KK4GER



----- Original Message ----- From: "Dennis Tillman W7PF" <dennis@ridesoft.com>
To: <TekScopes@groups.io>
Sent: Tuesday, January 21, 2020 12:09 AM
Subject: Re: [TekScopes] Risetime calculator (in tekwiki)


Hi Everyone,
Before getting caught up in the concept of an analog computer take a step back to
consider what an analog is. It is something that is similar to something else; the
two are said to be analogous.

THE RISETIME CALCULATOR IS AN ANALOG COMPUTER BASED ON LOGARITHMS
The risetime calculator is a simple, but clever, analog computer that relies on an
analogy between the angle of the dial on each pot and the resistance of the pot at
that angle. This in turn relies John Napier's discovery of logarithms, which he first
published in 1614. Logarithms are a method to multiply two numbers together by adding
their exponent. Anyone familiar with a slide rule knows the C and D scales on a slide
rule are logarithmic and not linear. The A and B scales increase twice as fast as C
and D because they represent a number times itself or the square of a number. If I
wanted to multiply 7 times 5 I simply had to add together the logarithm of 7 (in base
10) to the logarithm of 5 (in base 10). The log (short for logarithm in base 10) of 7
is 0.845. The log of 5 is 0.699. 7 x 5 = log 7 + log 5 = 0.845 + 0.699 = 1.544. If we
look up the number whose log is 1.544 it is our answer: 35. Commonly available 10"
slide rules are accurate to 2 or 3 decimal places. This is accurate enough for an
enormous variety and number of calculations so they were widely used until the
introduction of mechanical and then electronic calculators.

HOW IT WORKS
If the three pots on the left side of the risetime calculator have a logarithmic
taper and the current coming out of each was added together the sum would = A + B +
C. That's not what we want. We want A*A + B*B + C*C. To do that the taper has to
change logarithmically but twice as fast as an ordinary logarithmic taper. If the
first pot's taper increases twice as fast as an ordinary logarithmic taper the
current that comes out of it will be analogous to A*A. The same taper used for the
second pot will produce a current analogous to B*B, and the third pot will produce a
current analogous to C*C. Now if we sum together the currents coming from the three
pots we will get A*A + B*B + C*C. Our next step is to find the number, using the pot
on the right side, whose current equals the sum of the 3 pots on the left side. The
minimum current from a pot on the left side will be when it is counterclockwise
representing 1. 1*1 = 1, and the log of 1 is 0 so there would be 0 units of current
coming out of the pot. If the pot were set to 10, then 10*10 = 100 and the log of 100
is 2. So there would be 2 units of current coming from the pot. The currents are
summed from all three pots and the total can range from 0 to 6 units of current. With
the meter we have to balance (or null) the current on the left side with an equal
current from the right side knob. There is no need to take the square root of the
left side provided we square the right side instead. So on the right side we can use
one more pot as long as it has the same taper as the other three pots. We can even
use the same scale. One small drawback to doing this is there exists the possibility
that if all three risetimes happen to be greater than 5.773nSec you won't be able to
balance the bridge since the total risetime will be larger than 10nSec.

ANALOG VS DIGITAL OSCILLOSCOPES
High quality oscilloscopes strive to present an accurate analog of the voltages or
currents in a circuit being probed. Until recently all of the oscilloscopes in the
world were analog and there was no need to use the word analog to describe them. I am
absolutely certain no one gave that word a second thought until another way to make
an oscilloscope became possible in the 1970s. Suddenly it was important to
distinguish whether the oscilloscope made every attempt to present you with a wiggly
line on a CRT that was an analog of what was happening in the circuit you were
measuring or if what you saw looked like a connect the dots page from a coloring book.

How closely the analog displayed on the CRT compared to the actual potential in the
circuit being measured was an indication of the quality of the instrument. Enormous
amounts of money were spent over 100+ years to insure the trace on the CRT was an
accurate analogue of the voltage in the circuit. The result was something that could
be within +/-1%. Today there are two kinds of oscilloscopes. One important attribute
distinguishes analog oscilloscopes from digital oscilloscopes.

* The representation on the CRT of an analog oscilloscope is a CONTINUOUS analog of
the signal being measured (except for sampling plugins). One advantage of an analog
oscilloscope, as long as you use it within its specified limits, is the
representation on the CRT has a high likelihood of being a good analog of the signal
being measured.

* The representation on the display of a digital oscilloscope is a SERIES OF DISCRETE
DOT PAIRS that may (or may not) appear to form one or more patterns that humans will
assume, sometimes erroneously, are continuous and then they might conclude it is an
analog of the signal being measured. The interpretation of the pattern of dots on a
digital oscilloscope requires a detailed understanding of how the samples are taken,
how the samples are displayed, the type of sampling used, and the limitations of
sampling itself.

Analog oscilloscopes are desirable for new or unsophisticated users because the
results are less likely to be misunderstood. Digital oscilloscopes have many
limitations and pitfalls that analog oscilloscopes do not. They require that their
results be interpreted carefully. The digitized data from a signal can be further
processed mathematically to extract additional information from it or, for example,
to reduce the noise in the digitized data. There are many other things that can be
done to a digitally sampled signal after it has been captured to further process it.

Dennis Tillman W7PF





Re: What Tektronix means to me

Harvey White
 

Heh, well, not that I know about lazy.....

74181 and 7489's.  Timing generated by a counter that gave 8 decoded time states per instruction, fetch, operate, adjust, etc.

I actually bought stuff from Digikey (think it was the same) before they went honest and started carrying new stuff.  This was when they were giving polypacks a run for their money.

I knew video, I liked video, so I had a video display, color, with a user definable character set for the top 128 characters. Color, too.

I discovered the 2901 bit slice stuff, and thought it quite complicated.  Somewhere around 1989 or so I ended up translating a 2901 bitslice design (was working from flowcharts and only did the software) from that to a 29116/29117 integrated bitslice design, much nicer model.  That was a retrofit for the ARSR3 airport approach radar as part of the 3 level weather upgrade.  (to be specific, it was in the correlator and formatter... all assembly of the "roll your own" variety).  Never did any bit slice until then.

The first scope I had was a heathkit OM-3.

Then the Tektronix 513D and 512.

Then the next I got was the telequipment D75.

Somewhere in there I built a scope with a 3 inch surplus tube from a schematic I got from the heathkit catalog.  Remember when they had tiny little schematics in their catalogs?   Magnifying glasses can be your friends.  Built it on a breadpan for a chassis.

Went to a Kenwood, 4 channel analog with readouts, then a 2430A, and thence into 7000 series, starting with the 7704A.

Went into logic analyzers...

Started with an HP1630D, then 1650, then 1661, then 16500B, and finally ended up with a 16702A, then went to the 16702B.

All I have left (that I want to give new homes) are the 16500B and the 16702A's.  The 16702B is the mainstay of the logic analyzers.

Somewhere in there I got a 1640 serial analyzer (needs new home, I think).  a Tek 308 (cute, but not too useful), and an HP serial analyzer (which I'll keep).

More fun.

The 7904/16702B are the mainstays of the development cycle, adding in a TDS540A for more digital.

Harvey

On 1/21/2020 11:07 AM, Chuck Harris wrote:
I said I was greedy... It didn't serve my purposes to also
mention that I was lazy.

I thought it wasteful to use ripple carry on the S181's, so
I included the S182's with their carry look ahead, or whatever
it was called.

I used 7489's for my register file. I don't recall my plans
for RAM, or disk storage... I think I was waiting for divine
inspiration... I did have a pile of and 1103's on hand.

My design was front panel, and TTY console only. None of this
video stuff. This was 1974 or 5... And my available TTY was
an old baudot Model 15 tty.

I would have used ECL, but it was available only in simple
functions... and not at hamfests.

At some point, I discovered the 2901 bit slice ALU's, and
school got too busy to continue this nonsense.

I found my cache of pencil drawings (on the backs of green
bar line printer paper) two dozen years ago, shook my head,
and dumped them into the trash bin.

Some years after my fixing my 513D, I found one for $5 with
a smashed up CRT. I swapped out the repaired vertical attenuator
switch, and replaced my side fan with the rear fan and filter
from the newer 513D, along with the case. I probably shouldn't
have done that, as it ruins the authenticity of the scope... but
then, so did all my repairs. I am sure that a real 513D never
had all of those yellow CDE mylar caps, and a red oil filled glass
cased unblanking capacitor.

Some day we all will have to talk about the differences between
the old Vollum designed scopes, and the later plugin scopes.

Clearly someone far more cleaver took over the task. The
circuitry went from brute force designs that used the hell out
of the tubes, to inspirational designs, that cherished the tubes...

I wonder who that was? (Calling: Dennis Tillman.)

My second scope, after the 513D, was a plain Jane 545. I had the
scope for only a couple of days before I learned that what I had
really wanted was a 545A, or 545B... All of the interesting
combinations of the dual timebases were left off of the 545...
relying instead on the user connecting cabling between vertical
outputs, gate signals, and the external trigger inputs... bah!

My third scope was a brand new 2465.

It all gets fuzzy after that.

-Chuck Harris

Harvey White wrote:
Hah.  I think you underachieved on your design <grin>.

I have to remember back more time than I care to remember, but the ALU for the jump
computer was 74181's, and I think that the FPU CPU was 74181's, but 4 of them.  I had
hardware multiply, since I didn't know microcode, nor did I have a memory chip (this
was before even 1702's)  I may have had 7489's as memory, and the arithmetic was
BCD.  Don't quite remember how I did that, but I did.  The address computation was a
separate processor and did some odd things to get the addressing done.  Memory was
2102 1K x 1 static memory (two voltages) which were also part of the video display
controller.  It had its own NTSC driver (a separate NTSC sync generator with a
dot/bar/crosshatch generator, also had R-Y and B-Y encoders. (IVC Camera schematic,
homemade PC boards), and that drove the display.  Since the CPU was sharing the video
display memory, you had a bit of an idea what was going on.

Very old design.....

The next scope I got was a Telequipment D75 when Tektronix was closing them out, many
years later and when I actually had a bit of money.  It had its own share of
problems.......

Harvey


Re: 2465B - Weak Readout Intensity

Bruce Atwood
 

When I was servicing TV's (mid 60's!) we carried, in our tube caddy tiny auto-transformers that would, for most BW tubes plug inline with the socket and boost the heater voltage by a volt or so. For Color tubes you would have to do a couple of cuts and splices. It was always the red gun. We also had a " Cathode Rejuvenator " Which would basically draw and arc between G1 and the cathode, the theory being that you could get a new oxide surface on the cathode. When it worked it was great, when it went bad the CRT was dead. There are several rejuvinators on ebay. If you are resigned to buying a new tube it might be worth it. The other trick, which is truly bold, is to whack the neck against your hand. Here the theory is that you might be able to move the cathode to the side, exposing and unused section of the oxide. Definitely wear leather gloves and a full face shield.


good luck

On 1/20/2020 8:51 PM, flanneltuba@gmail.com<mailto:flanneltuba@gmail.com> wrote:

Thanks Chuck.

That was the sort of conclusion I was dreading. I do have a perfectly good 2465 (no suffix) with what I believe is the same CRT. It pains me though to pull a CRT out of an otherwise healthy scope just to prove that the subject unit has a weak CRT. I suppose it's that or roll the dice and buy a CRT off of EBay. There's a seller from Israel offering 154-0850-01 CRTs for $67, shipping included. "Very Good Working Condition." I'll admit to being a little apprehensive about buying parts from overseas.

One other thought: In the good old CRT television days, (were those the good old days?) one could often perk up a sagging CRT by bumping up the filament voltage on the heater. Sounds like something of a fool's errand for a tek scope of this sort, and more effort than it could possibly be worth, but I would be interested in knowing if anyone's ever given it a try.

Thanks,

- Scott



.



--
Bruce Atwood PhD
Department of Astronomy
The Ohio State University
100 West 18th Ave., Room 4055
Columbus, OH 43210

Phone 614.314.0189
FAX 614.292.2928


Re: What Tektronix means to me

Chuck Harris <cfharris@...>
 

I said I was greedy... It didn't serve my purposes to also
mention that I was lazy.

I thought it wasteful to use ripple carry on the S181's, so
I included the S182's with their carry look ahead, or whatever
it was called.

I used 7489's for my register file. I don't recall my plans
for RAM, or disk storage... I think I was waiting for divine
inspiration... I did have a pile of and 1103's on hand.

My design was front panel, and TTY console only. None of this
video stuff. This was 1974 or 5... And my available TTY was
an old baudot Model 15 tty.

I would have used ECL, but it was available only in simple
functions... and not at hamfests.

At some point, I discovered the 2901 bit slice ALU's, and
school got too busy to continue this nonsense.

I found my cache of pencil drawings (on the backs of green
bar line printer paper) two dozen years ago, shook my head,
and dumped them into the trash bin.

Some years after my fixing my 513D, I found one for $5 with
a smashed up CRT. I swapped out the repaired vertical attenuator
switch, and replaced my side fan with the rear fan and filter
from the newer 513D, along with the case. I probably shouldn't
have done that, as it ruins the authenticity of the scope... but
then, so did all my repairs. I am sure that a real 513D never
had all of those yellow CDE mylar caps, and a red oil filled glass
cased unblanking capacitor.

Some day we all will have to talk about the differences between
the old Vollum designed scopes, and the later plugin scopes.

Clearly someone far more cleaver took over the task. The
circuitry went from brute force designs that used the hell out
of the tubes, to inspirational designs, that cherished the tubes...

I wonder who that was? (Calling: Dennis Tillman.)

My second scope, after the 513D, was a plain Jane 545. I had the
scope for only a couple of days before I learned that what I had
really wanted was a 545A, or 545B... All of the interesting
combinations of the dual timebases were left off of the 545...
relying instead on the user connecting cabling between vertical
outputs, gate signals, and the external trigger inputs... bah!

My third scope was a brand new 2465.

It all gets fuzzy after that.

-Chuck Harris

Harvey White wrote:

Hah.  I think you underachieved on your design <grin>.

I have to remember back more time than I care to remember, but the ALU for the jump
computer was 74181's, and I think that the FPU CPU was 74181's, but 4 of them.  I had
hardware multiply, since I didn't know microcode, nor did I have a memory chip (this
was before even 1702's)  I may have had 7489's as memory, and the arithmetic was
BCD.  Don't quite remember how I did that, but I did.  The address computation was a
separate processor and did some odd things to get the addressing done.  Memory was
2102 1K x 1 static memory (two voltages) which were also part of the video display
controller.  It had its own NTSC driver (a separate NTSC sync generator with a
dot/bar/crosshatch generator, also had R-Y and B-Y encoders. (IVC Camera schematic,
homemade PC boards), and that drove the display.  Since the CPU was sharing the video
display memory, you had a bit of an idea what was going on.

Very old design.....

The next scope I got was a Telequipment D75 when Tektronix was closing them out, many
years later and when I actually had a bit of money.  It had its own share of
problems.......

Harvey


Re: Tektronix 532 LV rail shorted

randolphbeebe@...
 

Hi Rajesh,

I replaced the fuse and went to re-test the voltages and now found no voltage whatsoever on transformer pin 16 that supplies the -150V rail. This is where I measured +89V AC before as well as at C640. I was skeptical about my notes as well and re-checked my old notes to confirm that it was AC voltage and no DC. Confirmed +89 to +90V AC only. This of course varied with how much supply voltage I was feeding the scope with my variac, but it was positive AC voltage.

I am convinced that what I have been dealing with is a faulty transformer from the start and now it is totally cooked.

I am keeping the scope (even though I do not need it) though and will keep a lookout for a donor unit or if someone else needs parts.

Thanks again for your help.

Randy


Re: Risetime calculator (in tekwiki)

Chuck Harris <cfharris@...>
 

The dial scale values indicate numbers you want to
feed the equation SQRT(A^2 + B^2 + C^2) = f

In this case, if you set them to microseconds, f will
be in microseconds... nanoseconds, f will be in nanoseconds...

Obviously, you will have to interpolate your settings and
results, to gain more than integer accuracy... which you
should be able to do by eye to at least quarters.

-Chuck Harris

GerryR wrote:

Dennis,
           What do the dial scales values indicate??
GerryR
KK4GER



----- Original Message ----- From: "Dennis Tillman W7PF" <dennis@ridesoft.com>
To: <TekScopes@groups.io>
Sent: Tuesday, January 21, 2020 12:09 AM
Subject: Re: [TekScopes] Risetime calculator (in tekwiki)


Hi Everyone,
Before getting caught up in the concept of an analog computer take a step back to
consider what an analog is. It is something that is similar to something else; the
two are said to be analogous.

THE RISETIME CALCULATOR IS AN ANALOG COMPUTER BASED ON LOGARITHMS
The risetime calculator is a simple, but clever, analog computer that relies on an
analogy between the angle of the dial on each pot and the resistance of the pot at
that angle. This in turn relies John Napier's discovery of logarithms, which he first
published in 1614. Logarithms are a method to multiply two numbers together by adding
their exponent. Anyone familiar with a slide rule knows the C and D scales on a slide
rule are logarithmic and not linear. The A and B scales increase twice as fast as C
and D because they represent a number times itself or the square of a number. If I
wanted to multiply 7 times 5 I simply had to add together the logarithm of 7 (in base
10) to the logarithm of 5 (in base 10). The log (short for logarithm in base 10) of 7
is 0.845. The log of 5 is 0.699. 7 x 5 = log 7 + log 5 = 0.845 + 0.699 = 1.544. If we
look up the number whose log is 1.544 it is our answer: 35. Commonly available 10"
slide rules are accurate to 2 or 3 decimal places. This is accurate enough for an
enormous variety and number of calculations so they were widely used until the
introduction of mechanical and then electronic calculators.

HOW IT WORKS
If the three pots on the left side of the risetime calculator have a logarithmic
taper and the current coming out of each was added together the sum would = A + B +
C. That's not what we want. We want A*A + B*B + C*C. To do that the taper has to
change logarithmically but twice as fast as an ordinary logarithmic taper. If the
first pot's taper increases twice as fast as an ordinary logarithmic taper the
current that comes out of it will be analogous to A*A. The same taper used for the
second pot will produce a current analogous to B*B, and the third pot will produce a
current analogous to C*C. Now if we sum together the currents coming from the three
pots we will get A*A + B*B + C*C. Our next step is to find the number, using the pot
on the right side, whose current equals the sum of the 3 pots on the left side. The
minimum current from a pot on the left side will be when it is counterclockwise
representing 1. 1*1 = 1, and the log of 1 is 0 so there would be 0 units of current
coming out of the pot. If the pot were set to 10, then 10*10 = 100 and the log of 100
is 2. So there would be 2 units of current coming from the pot.  The currents are
summed from all three pots and the total can range from 0 to 6 units of current. With
the meter we have to balance (or null) the current on the left side with an equal
current from the right side knob. There is no need to take the square root of the
left side provided we square the right side instead. So on the right side we can use
one more pot as long as it has the same taper as the other three pots. We can even
use the same scale. One small drawback to doing this is there exists the possibility
that if all three risetimes happen to be greater than 5.773nSec you won't be able to
balance the bridge since the total risetime will be larger than 10nSec.

ANALOG VS DIGITAL OSCILLOSCOPES
High quality oscilloscopes strive to present an accurate analog of the voltages or
currents in a circuit being probed. Until recently all of the oscilloscopes in the
world were analog and there was no need to use the word analog to describe them. I am
absolutely certain no one gave that word a second thought until another way to make
an oscilloscope became possible in the 1970s. Suddenly it was important to
distinguish whether the oscilloscope made every attempt to present you with a wiggly
line on a CRT that was an analog of what was happening in the circuit you were
measuring or if what you saw looked like a connect the dots page from a coloring book.

How closely the analog displayed on the CRT compared to the actual potential in the
circuit being measured was an indication of the quality of the instrument. Enormous
amounts of money were spent over 100+ years to insure the trace on the CRT was an
accurate analogue of the voltage in the circuit. The result was something that could
be within +/-1%. Today there are two kinds of oscilloscopes. One important attribute
distinguishes analog oscilloscopes from digital oscilloscopes.

* The representation on the CRT of an analog oscilloscope is a CONTINUOUS analog of
the signal being measured (except for sampling plugins). One advantage of an analog
oscilloscope, as long as you use it within its specified limits, is the
representation on the CRT has a high likelihood of being a good analog of the signal
being measured.

* The representation on the display of a digital oscilloscope is a SERIES OF DISCRETE
DOT PAIRS that may (or may not) appear to form one or more patterns that humans will
assume, sometimes erroneously, are continuous and then they might conclude it is an
analog of the signal being measured. The interpretation of the pattern of dots on a
digital oscilloscope requires a detailed understanding of how the samples are taken,
how the samples are displayed, the type of sampling used, and the limitations of
sampling itself.

Analog oscilloscopes are desirable for new or unsophisticated users because the
results are less likely to be misunderstood. Digital oscilloscopes have many
limitations and pitfalls that analog oscilloscopes do not. They require that their
results be interpreted carefully. The digitized data from a signal can be further
processed mathematically to extract additional information from it or, for example,
to reduce the noise in the digitized data. There are many other things that can be
done to a digitally sampled signal after it has been captured to further process it.

Dennis Tillman W7PF





Re: Risetime calculator (in tekwiki)

Chuck Harris <cfharris@...>
 

Hi Dennis,

You are trying to force the math.

Remember the whole point behind logarithms is they
turn difficult to do multiplication, and division
into easier to do addition and subtraction, following
these rules:

LOG(A) + LOG(B) + LOG(C) = LOG(A*B*C)

And,

2*LOG(A) = LOG(A^2)

So,

2*LOG(A) + 2*LOG(B) + 2*LOG(C) = LOG(A^2) + LOG(B^2) + LOG(C^2)

Which is,

LOG(A^2 * B^2 * C^2)

Not,

LOG(A^2 + B^2 + C^2) as you presume.

For it to work the way you want, you would have to do the
antilog of each pot's value before you sum, which is more
difficult.

The real method used is devilishly cleaver. See
Albert's explanation in an earlier posting.

-Chuck Harris (I hope I got this right!)

Dennis Tillman W7PF wrote:
Hi Everyone,
Before getting caught up in the concept of an analog computer take a step back to consider what an analog is. It is something that is similar to something else; the two are said to be analogous.

THE RISETIME CALCULATOR IS AN ANALOG COMPUTER BASED ON LOGARITHMS
The risetime calculator is a simple, but clever, analog computer that relies on an analogy between the angle of the dial on each pot and the resistance of the pot at that angle. This in turn relies John Napier's discovery of logarithms, which he first published in 1614. Logarithms are a method to multiply two numbers together by adding their exponent. Anyone familiar with a slide rule knows the C and D scales on a slide rule are logarithmic and not linear. The A and B scales increase twice as fast as C and D because they represent a number times itself or the square of a number. If I wanted to multiply 7 times 5 I simply had to add together the logarithm of 7 (in base 10) to the logarithm of 5 (in base 10). The log (short for logarithm in base 10) of 7 is 0.845. The log of 5 is 0.699. 7 x 5 = log 7 + log 5 = 0.845 + 0.699 = 1.544. If we look up the number whose log is 1.544 it is our answer: 35. Commonly available 10" slide rules are accurate to 2 or 3 decimal places. This is accurate enough for an enormous variety and number of calculations so they were widely used until the introduction of mechanical and then electronic calculators.

HOW IT WORKS
If the three pots on the left side of the risetime calculator have a logarithmic taper and the current coming out of each was added together the sum would = A + B + C. That's not what we want. We want A*A + B*B + C*C. To do that the taper has to change logarithmically but twice as fast as an ordinary logarithmic taper. If the first pot's taper increases twice as fast as an ordinary logarithmic taper the current that comes out of it will be analogous to A*A. The same taper used for the second pot will produce a current analogous to B*B, and the third pot will produce a current analogous to C*C. Now if we sum together the currents coming from the three pots we will get A*A + B*B + C*C. Our next step is to find the number, using the pot on the right side, whose current equals the sum of the 3 pots on the left side. The minimum current from a pot on the left side will be when it is counterclockwise representing 1. 1*1 = 1, and the log of 1 is 0 so there would be 0 units of current coming out of the pot. If the pot were set to 10, then 10*10 = 100 and the log of 100 is 2. So there would be 2 units of current coming from the pot. The currents are summed from all three pots and the total can range from 0 to 6 units of current. With the meter we have to balance (or null) the current on the left side with an equal current from the right side knob. There is no need to take the square root of the left side provided we square the right side instead. So on the right side we can use one more pot as long as it has the same taper as the other three pots. We can even use the same scale. One small drawback to doing this is there exists the possibility that if all three risetimes happen to be greater than 5.773nSec you won't be able to balance the bridge since the total risetime will be larger than 10nSec.
...
Dennis Tillman W7PF





Re: 2465B - Weak Readout Intensity

Chuck Harris <cfharris@...>
 

If you do decide to go with an ebay seller, pay
particular attention that all of the wires, plugs,
and cables come with the CRT. Most scrappers think
the wires and cables are unimportant, and snip them
off close to the bottle.

You don't want a CRT like that.

Dinos in Greece sells an honest CRT. He knows what
they are for, and tests them before sending them out
to you. Several of the Israel sellers do as well.

You are probably going to spend closer to $150 all told.

The CRTs are slightly different between the 2465 and
the 2465A/B. They will interchange, but the changes
were made for a reason.

-Chuck Harris

flanneltuba@gmail.com wrote:

Thanks Chuck.

That was the sort of conclusion I was dreading. I do have a perfectly good 2465 (no suffix) with what I believe is the same CRT. It pains me though to pull a CRT out of an otherwise healthy scope just to prove that the subject unit has a weak CRT. I suppose it's that or roll the dice and buy a CRT off of EBay. There's a seller from Israel offering 154-0850-01 CRTs for $67, shipping included. "Very Good Working Condition." I'll admit to being a little apprehensive about buying parts from overseas.

One other thought: In the good old CRT television days, (were those the good old days?) one could often perk up a sagging CRT by bumping up the filament voltage on the heater. Sounds like something of a fool's errand for a tek scope of this sort, and more effort than it could possibly be worth, but I would be interested in knowing if anyone's ever given it a try.

Thanks,

- Scott




Re: Risetime calculator (in tekwiki)

Leo Bodnar
 

This is one heck of a rabbit-hole!
I wonder what would the calibration process be?
There are nine adjustable parameters - assuming knobs have grub screws.
Leo

On Tue, Jan 21, 2020 at 12:38 PM, Steve Hendrix wrote:
Very interesting! I've spent more time than I could afford this morning,


Re: Risetime calculator (in tekwiki)

Steve Hendrix
 

At 2020-01-21 05:12 AM, Leo Bodnar wrote:


But (and this is where the trick is) if the shunting resistance is equal
to the nominal pot value then linear component is cancelled out and the
result is purely quadratic function of wiper position (if the track is
linear taper.)
Very interesting! I've spent more time than I could afford this morning, trying to reconcile your math with mine. I finally figured out that the "linear component" just  moves the origin along the X axis. The way I calculated it over the weekend, you'd just block the wiper at the midpoint, using only half the pot range. If the pot value is P, and X is the wiper position ranging from 0 at the stop at the center to 1 at the end, with the two ends tied together the total resistance R = P/2 * (1-X^2). Effectively this is your drawing, with a=0, and the change in how the wiper position is used. With the parallel resistance being exactly the pot value, we can use the full range of the pot and it's nicely squared just as you showed. Apparently the old boys considered that being able to use the full range of the pot instead of needing a mechanical stop, was worth the trouble of having a well-matched resistor across the pot. Labor being relatively cheaper in those days, it wouldn't surprise me to find that they manually selected a resistor to perfectly match each pot.

Steve Hendrix


Re: Risetime calculator (in tekwiki)

GerryR <totalautomation1@...>
 

Dennis,
What do the dial scales values indicate??
GerryR
KK4GER

----- Original Message -----
From: "Dennis Tillman W7PF" <dennis@ridesoft.com>
To: <TekScopes@groups.io>
Sent: Tuesday, January 21, 2020 12:09 AM
Subject: Re: [TekScopes] Risetime calculator (in tekwiki)


Hi Everyone,
Before getting caught up in the concept of an analog computer take a step back to consider what an analog is. It is something that is similar to something else; the two are said to be analogous.

THE RISETIME CALCULATOR IS AN ANALOG COMPUTER BASED ON LOGARITHMS
The risetime calculator is a simple, but clever, analog computer that relies on an analogy between the angle of the dial on each pot and the resistance of the pot at that angle. This in turn relies John Napier's discovery of logarithms, which he first published in 1614. Logarithms are a method to multiply two numbers together by adding their exponent. Anyone familiar with a slide rule knows the C and D scales on a slide rule are logarithmic and not linear. The A and B scales increase twice as fast as C and D because they represent a number times itself or the square of a number. If I wanted to multiply 7 times 5 I simply had to add together the logarithm of 7 (in base 10) to the logarithm of 5 (in base 10). The log (short for logarithm in base 10) of 7 is 0.845. The log of 5 is 0.699. 7 x 5 = log 7 + log 5 = 0.845 + 0.699 = 1.544. If we look up the number whose log is 1.544 it is our answer: 35. Commonly available 10" slide rules are accurate to 2 or 3 decimal places. This is accurate enough for an enormous variety and number of calculations so they were widely used until the introduction of mechanical and then electronic calculators.

HOW IT WORKS
If the three pots on the left side of the risetime calculator have a logarithmic taper and the current coming out of each was added together the sum would = A + B + C. That's not what we want. We want A*A + B*B + C*C. To do that the taper has to change logarithmically but twice as fast as an ordinary logarithmic taper. If the first pot's taper increases twice as fast as an ordinary logarithmic taper the current that comes out of it will be analogous to A*A. The same taper used for the second pot will produce a current analogous to B*B, and the third pot will produce a current analogous to C*C. Now if we sum together the currents coming from the three pots we will get A*A + B*B + C*C. Our next step is to find the number, using the pot on the right side, whose current equals the sum of the 3 pots on the left side. The minimum current from a pot on the left side will be when it is counterclockwise representing 1. 1*1 = 1, and the log of 1 is 0 so there would be 0 units of current coming out of the pot. If the pot were set to 10, then 10*10 = 100 and the log of 100 is 2. So there would be 2 units of current coming from the pot. The currents are summed from all three pots and the total can range from 0 to 6 units of current. With the meter we have to balance (or null) the current on the left side with an equal current from the right side knob. There is no need to take the square root of the left side provided we square the right side instead. So on the right side we can use one more pot as long as it has the same taper as the other three pots. We can even use the same scale. One small drawback to doing this is there exists the possibility that if all three risetimes happen to be greater than 5.773nSec you won't be able to balance the bridge since the total risetime will be larger than 10nSec.

ANALOG VS DIGITAL OSCILLOSCOPES
High quality oscilloscopes strive to present an accurate analog of the voltages or currents in a circuit being probed. Until recently all of the oscilloscopes in the world were analog and there was no need to use the word analog to describe them. I am absolutely certain no one gave that word a second thought until another way to make an oscilloscope became possible in the 1970s. Suddenly it was important to distinguish whether the oscilloscope made every attempt to present you with a wiggly line on a CRT that was an analog of what was happening in the circuit you were measuring or if what you saw looked like a connect the dots page from a coloring book.

How closely the analog displayed on the CRT compared to the actual potential in the circuit being measured was an indication of the quality of the instrument. Enormous amounts of money were spent over 100+ years to insure the trace on the CRT was an accurate analogue of the voltage in the circuit. The result was something that could be within +/-1%. Today there are two kinds of oscilloscopes. One important attribute distinguishes analog oscilloscopes from digital oscilloscopes.

* The representation on the CRT of an analog oscilloscope is a CONTINUOUS analog of the signal being measured (except for sampling plugins). One advantage of an analog oscilloscope, as long as you use it within its specified limits, is the representation on the CRT has a high likelihood of being a good analog of the signal being measured.

* The representation on the display of a digital oscilloscope is a SERIES OF DISCRETE DOT PAIRS that may (or may not) appear to form one or more patterns that humans will assume, sometimes erroneously, are continuous and then they might conclude it is an analog of the signal being measured. The interpretation of the pattern of dots on a digital oscilloscope requires a detailed understanding of how the samples are taken, how the samples are displayed, the type of sampling used, and the limitations of sampling itself.

Analog oscilloscopes are desirable for new or unsophisticated users because the results are less likely to be misunderstood. Digital oscilloscopes have many limitations and pitfalls that analog oscilloscopes do not. They require that their results be interpreted carefully. The digitized data from a signal can be further processed mathematically to extract additional information from it or, for example, to reduce the noise in the digitized data. There are many other things that can be done to a digitally sampled signal after it has been captured to further process it.

Dennis Tillman W7PF





--
Dennis Tillman W7PF
TekScopes Moderator


Re: 7A26 transient response 5 nS all atten

Jean-Paul
 

Nenad and Albert I will first try 99% Iso Alch from Amazon. Found a giant plastic disposable syringe and long needles. Just exercising the V/DIV improved it somewhat but some positions seem mechanically flaky, also AC/DC/GND switches
All plugins need the treatment, .


Nedad: Fine to hear from Belgrad, Do you know the Nikola Tesla Museum?
A good friend from Belgrad is a fine photographer and cinematographer.
I was there in 1982, also to Dubrovnik and Split, I recall street signs in Cyrilic and old railroad trains...and a fine fortress.


Jon in Paris

18921 - 18940 of 182448