Date   
Re: Questions on impedance matching

Richard Knoppow
 

The attenuator may work but the calibration will not be correct if its not terminated, depends on the type of attenuator. Signal generators have a specified source impedance, output levels will be correct when terminated in the characteristic impedance but the variations will often still be correct as in, for instance, a Hewlett-Packard 606A or B. I think the main thing is that the OP was concerned about voltage not power and with audio not RF. Much simpler problem.

On 2/9/2018 1:19 PM, Michael A. Terrell wrote:
The output level is dependent on proper termination, if there is any type of attenuator at the output stage of the generator.
--
Richard Knoppow
dickburk@...
WB6KBL

Re: Questions on impedance matching

Michael A. Terrell
 

The output level is dependent on proper termination, if there is any type of attenuator at the output stage of the generator.

-----Original Message-----
From: Richard Knoppow <dickburk@...>
Sent: Feb 9, 2018 1:21 PM
To: TekScopes@groups.io
Subject: Re: [TekScopes] Questions on impedance matching

I think you are confusing the conditions for voltage
measurement with those for maximum power transfer. If the signal
generators you write of are audio frequency generators
terminating them in the specified impedance may not be necessary
at all. It depends on whether the output has flat frequency
response if terminated in some other value or left unterminated.
This is usually stated in the specs and in any case is easily
measured with a scope or good voltmeter. Where the generator has
calibrated output level the calibrations usually are for the
terminated condition. In that case a simple non-reactive resistor
is all that is necessary. Modern carbon or metal film resistors
are actually quite non-reactive well into high radio frequencies,
I've measured many using a Boonton RX meter.
Most audio devices like microphones, including the contact
type used on string instruments, are designed to work into a high
impedance. Even low impedance microphones are mostly intended to
work into a high impedance but may need a transformer to increase
the voltage at the amplifier. While they are specified as
impedance transformers they are really not. Terminating a
generator like a microphone, in its source impedance reduces the
voltage by half, not usually desirable (although the BBC made a
practice of this). One generally wants the highest signal to
noise ratio so working a voltage source into a high impedance is
best because is maximizes the voltage available.
Loudspeakers are different because they absorb power. So one
wants a good impedance match to maximize power transfer. Works OK
where the speaker has a constant impedance but low frequency
speakers do not because the acoustical impedance seen by the
speaker is not matched. The purpose of exponential horns and bass
reflex cabinets is to obtain a better acoustical impedance match.
A horn is the acoustical equivalent of a transformer (wide band)
while a bass reflex is the equivalent of a quarter-wave
transmission line at RF, and like it works only over a rather
narrow range of frequencies. When driving a speaker an impedance
match is desirable.
I don't know if this is of much help but I think maybe the
stuff you are reading on impedance is aimed at RF not audio. At
RF something of the same sort exists, for receivers getting the
best signal to noise at the antenna terminals is important and
that may come with a mis-matched condition, for transmitting a
match is necessary for maximum power transfer. However, at RF its
easy to get a good match using reactive components (like a L or
pi network) as well as wide band transformers as in audio. One
can't use reactive networks in audio because they are inherently
very narrow band (actually work perfectly at only one frequency).
For your audio generator all that is necessary is to make
sure its output is flat for whatever condition you use and
measure its output with a meter.
Michael A. Terrell

Re: My Next Scope

Tam Hanna
 

I had a friend at Danaher diddle him, and she apparently found nothing...
---
With best regards
Tam HANNA (emailing on a BlackBerry PRIV)

Enjoy electronics? Join 6500 other followers by visiting the Crazy Electronics Lab at https://www.instagram.com/tam.hanna/

Re: Something you didn't know: What did Tek have to do with Packing Peanuts?

Dave Brown
 

Wasn't me so must be more than one of us (big surprise- where I used to work
-Telecom NZ- there were five of us across the country!!)
DaveB, NZ

-----Original Message-----
From: TekScopes@groups.io [mailto:TekScopes@groups.io] On Behalf Of Dennis
Tillman W7PF
Sent: Saturday, February 10, 2018 09:48
To: TekScopes@groups.io
Subject: [TekScopes] Something you didn't know: What did Tek have to do with
Packing Peanuts?

Everyone is familiar with those little beads of loose fill packing
"popcorn".

Most people don't know where those were invented and under what conditions.



http://vintagetek.org/packing-popcorn/



Thanks to Dave Brown for tracking this information down.



Dennis Tillman W7PF

Something you didn't know: What did Tek have to do with Packing Peanuts?

 

Everyone is familiar with those little beads of loose fill packing
"popcorn".

Most people don't know where those were invented and under what conditions.



http://vintagetek.org/packing-popcorn/



Thanks to Dave Brown for tracking this information down.



Dennis Tillman W7PF

Re: recapped 485 (now we're cooking with grease edition)

Brendan
 

This first picture https://ibb.co/fyoe0c is horizontal display delay Alt Im able to control B intensity, everything seems semi normal. This second picture https://ibb.co/mntbDx is B only. Im getting more confused by the minute. When in B only I am unable to control the intensity with the B intensity control but I am able to control it with A. Im ASSuming the delayed sweep Im seeing in Alt is A alternating with B. Im this close. Grrrr

Re: Questions on impedance matching

Ted Rook
 

One of the reasons to put an external 600 ohm load on audio generators is because the
output is usually calibrated for the condition that the 600 ohms is present. If it is absent the
actual voltage present at the output terminals is twice the size, +6dB greater, than the dial
states.

On 9 Feb 2018 at 7:37, David Berlind wrote:

Hi all,

I have some questions on impedance matching that are borne out of some recent threads
regarding challenges in getting a proper oscilloscope reading off an old signal generator.

I've been spending a disproportionate amount of time learning about impedance matching.
I've learned the math (even developed a spreadsheet to game various scenarios), developed
a pretty good understanding of the Maximum Power Transfer Theorem, and I get the
principles of reflection. But here are my questions.

(1) While it's always best to shoot for minimal reflection (optimal power transfer), I've seen
plenty of references to an acceptable over/under range. In other words, not a direct match.
Are there any thoughts on what that range is?

(2) There are various techniques to impedance matching... resistors and transformers..
applying capacitive reactance... what has worked for you in the past? For example, if you go
the reactance route, you minimize loss of power (vs. resistive), but then you tie yourself to a
narrow range of frequencies. That's good if you have a relatively fixed use case. But what if
you have a range of used cases?

(3) My main use case right now is working in the audio range with guitar amps, working with
signal generators and so on. So, a safe diagnostic range is 80-5000Hz (I know overtones
can run up to 20KHz). But let's say I'm matching the output of an old 600 ohm signal
generator to the 1 megaohm input of my Tek oscilloscopes (which all have the same input
impedance). That's an impedance ratio of 1000000 to 600 which works out to 1666:1. By the
official math, minimizing reflection via transformer (if I chose to do so) would require a 40:1
ratio between the primary and secondary windings (or the equivalent accomplished in
stages). I'm looking for suggestions on this "fixed" case.. but would also like to hear about
any flexible solutions (for example, to accommodate a wide range of output impedances).

Thank you.

Re: Help restoring a 547 scope

felix_cantor@...
 

Hi Albert,

It is going to take me a while answering back because I had to order a set of probes from the USA (Not living there), so I will need to wait like two weeks to get them. But I`ll be back here as soon as possible.


Thank you!

Regards.

F.

Re: Questions on impedance matching

Richard Knoppow
 

I think you are confusing the conditions for voltage measurement with those for maximum power transfer. If the signal generators you write of are audio frequency generators terminating them in the specified impedance may not be necessary at all. It depends on whether the output has flat frequency response if terminated in some other value or left unterminated. This is usually stated in the specs and in any case is easily measured with a scope or good voltmeter. Where the generator has calibrated output level the calibrations usually are for the terminated condition. In that case a simple non-reactive resistor is all that is necessary. Modern carbon or metal film resistors are actually quite non-reactive well into high radio frequencies, I've measured many using a Boonton RX meter.
Most audio devices like microphones, including the contact type used on string instruments, are designed to work into a high impedance. Even low impedance microphones are mostly intended to work into a high impedance but may need a transformer to increase the voltage at the amplifier. While they are specified as impedance transformers they are really not. Terminating a generator like a microphone, in its source impedance reduces the voltage by half, not usually desirable (although the BBC made a practice of this). One generally wants the highest signal to noise ratio so working a voltage source into a high impedance is best because is maximizes the voltage available.
Loudspeakers are different because they absorb power. So one wants a good impedance match to maximize power transfer. Works OK where the speaker has a constant impedance but low frequency speakers do not because the acoustical impedance seen by the speaker is not matched. The purpose of exponential horns and bass reflex cabinets is to obtain a better acoustical impedance match. A horn is the acoustical equivalent of a transformer (wide band) while a bass reflex is the equivalent of a quarter-wave transmission line at RF, and like it works only over a rather narrow range of frequencies. When driving a speaker an impedance match is desirable.
I don't know if this is of much help but I think maybe the stuff you are reading on impedance is aimed at RF not audio. At RF something of the same sort exists, for receivers getting the best signal to noise at the antenna terminals is important and that may come with a mis-matched condition, for transmitting a match is necessary for maximum power transfer. However, at RF its easy to get a good match using reactive components (like a L or pi network) as well as wide band transformers as in audio. One can't use reactive networks in audio because they are inherently very narrow band (actually work perfectly at only one frequency).
For your audio generator all that is necessary is to make sure its output is flat for whatever condition you use and measure its output with a meter.

On 2/9/2018 7:37 AM, David Berlind wrote:
Hi all,
I have some questions on impedance matching that are borne out of some recent threads regarding challenges in getting a proper oscilloscope reading off an old signal generator.
I've been spending a disproportionate amount of time learning about impedance matching. I've learned the math (even developed a spreadsheet to game various scenarios), developed a pretty good understanding of the Maximum Power Transfer Theorem, and I get the principles of reflection. But here are my questions.
(1) While it's always best to shoot for minimal reflection (optimal power transfer), I've seen plenty of references to an acceptable over/under range. In other words, not a direct match. Are there any thoughts on what that range is?
(2) There are various techniques to impedance matching... resistors and transformers.. applying capacitive reactance... what has worked for you in the past? For example, if you go the reactance route, you minimize loss of power (vs. resistive), but then you tie yourself to a narrow range of frequencies. That's good if you have a relatively fixed use case. But what if you have a range of used cases?
(3) My main use case right now is working in the audio range with guitar amps, working with signal generators and so on. So, a safe diagnostic range is 80-5000Hz (I know overtones can run up to 20KHz). But let's say I'm matching the output of an old 600 ohm signal generator to the 1 megaohm input of my Tek oscilloscopes (which all have the same input impedance). That's an impedance ratio of 1000000 to 600 which works out to 1666:1. By the official math, minimizing reflection via transformer (if I chose to do so) would require a 40:1 ratio between the primary and secondary windings (or the equivalent accomplished in stages). I'm looking for suggestions on this "fixed" case.. but would also like to hear about any flexible solutions (for example, to accommodate a wide range of output impedances).
Thank you.
--
Richard Knoppow
dickburk@...
WB6KBL

Re: Questions on impedance matching

Ed Breya
 

Exactly what Kevin said, and I would add that if you need to have say a 600 ohm termination as a load for the signal source, you can easily make one with a 600 ohm resistance and appropriate BNC arrangements, to put on the line. A very simple method is to put a BNC Tee in the line, with an appropriate connector with the termination resistor plugged into the added port. For audio frequencies, it will be 600 ohms for all practical purposes. In reality, it's the resistance paralleled with the cable and scope input capacitances, but for bench type applications with short cables and probes etc, you don't have to worry about complexities until reaching much higher frequencies. You also don't have to worry about the exact physical locations of the various system parts - in a small setup, the elements would be maybe up to tens of nanoseconds apart, which is insignificant at audio.

Ed

Re: Questions on impedance matching

Michael A. Terrell
 

A non inductive resistor load will give the flattest response when you look at that 600 ohm source with your scope. It also allows you to read the actual voltage of the signal. That 40:1 transformer would have bad rolloff at both ends of the bandwidth.

It was common in some industries to use a 620 ohm, half watt carbon comp resistor as a terminator on high impedance inputs, since it was the easiest to purchase. I saw it in many CATV head ends and broadcast studios, but closer values are easy to find, these days.

-----Original Message-----
From: David Berlind <david@...>
Sent: Feb 9, 2018 10:37 AM
To: TekScopes@groups.io
Subject: [TekScopes] Questions on impedance matching

Hi all,

I have some questions on impedance matching that are borne out of some recent threads regarding challenges in getting a proper oscilloscope reading off an old signal generator.

I've been spending a disproportionate amount of time learning about impedance matching. I've learned the math (even developed a spreadsheet to game various scenarios), developed a pretty good understanding of the Maximum Power Transfer Theorem, and I get the principles of reflection. But here are my questions.

(1) While it's always best to shoot for minimal reflection (optimal power transfer), I've seen plenty of references to an acceptable over/under range. In other words, not a direct match. Are there any thoughts on what that range is?

(2) There are various techniques to impedance matching... resistors and transformers.. applying capacitive reactance... what has worked for you in the past? For example, if you go the reactance route, you minimize loss of power (vs. resistive), but then you tie yourself to a narrow range of frequencies. That's good if you have a relatively fixed use case. But what if you have a range of used cases?

(3) My main use case right now is working in the audio range with guitar amps, working with signal generators and so on. So, a safe diagnostic range is 80-5000Hz (I know overtones can run up to 20KHz). But let's say I'm matching the output of an old 600 ohm signal generator to the 1 megaohm input of my Tek oscilloscopes (which all have the same input impedance). That's an impedance ratio of 1000000 to 600 which works out to 1666:1. By the official math, minimizing reflection via transformer (if I chose to do so) would require a 40:1 ratio between the primary and secondary windings (or the equivalent accomplished in stages). I'm looking for suggestions on this "fixed" case.. but would also like to hear about any flexible solutions (for example, to accommodate a wide range of output impedances).
Michael A. Terrell

Re: Questions on impedance matching

Kevin Wood G7BCS
 

Hi David,

The situation at Audio is slightly different because the wavelengths in
cables are very much longer so the problems of strict impedance matching
and reflections occurring where there are mismatches only really occur
when you're talking about the distances used in telecommunications.

With a couple of exceptions, in audio one tends to keep the output
impedances low and the input impedances high (say a ratio of 10:1 or more)
such that effects of loading your source are minimal.

The exceptions would be devices such as moving coil phono cartridges or
microphones where the signal levels are so small that you might use a
transformer to better match the impedances and increase power transfer.

Another area is when passing digital audio in SPDIF / AES3 formats where
reflections could cause problems unless impedance matching is carried out,
but again, this is likely to be experienced only with "longer than
domestic" transmission lines.

The idea with an oscilloscope is for the input impedance to be high so
that connecting it to monitor a circuit has a minimal effect on its
operation. In this case, there simply is no need to achieve maximal power
coupling between your generator and the instrument. Significant power
coupling might change the operation of the circuit you are trying to
diagnose!

73 for now

Kevin

Hi all,

I have some questions on impedance matching that are borne out of some
recent threads regarding challenges in getting a proper oscilloscope
reading off an old signal generator.

I've been spending a disproportionate amount of time learning about
impedance matching. I've learned the math (even developed a spreadsheet to
game various scenarios), developed a pretty good understanding of the
Maximum Power Transfer Theorem, and I get the principles of reflection.
But here are my questions.

(1) While it's always best to shoot for minimal reflection (optimal power
transfer), I've seen plenty of references to an acceptable over/under
range. In other words, not a direct match. Are there any thoughts on what
that range is?

(2) There are various techniques to impedance matching... resistors and
transformers.. applying capacitive reactance... what has worked for you in
the past? For example, if you go the reactance route, you minimize loss of
power (vs. resistive), but then you tie yourself to a narrow range of
frequencies. That's good if you have a relatively fixed use case. But what
if you have a range of used cases?

(3) My main use case right now is working in the audio range with guitar
amps, working with signal generators and so on. So, a safe diagnostic
range is 80-5000Hz (I know overtones can run up to 20KHz). But let's say
I'm matching the output of an old 600 ohm signal generator to the 1
megaohm input of my Tek oscilloscopes (which all have the same input
impedance). That's an impedance ratio of 1000000 to 600 which works out to
1666:1. By the official math, minimizing reflection via transformer (if I
chose to do so) would require a 40:1 ratio between the primary and
secondary windings (or the equivalent accomplished in stages). I'm looking
for suggestions on this "fixed" case.. but would also like to hear about
any flexible solutions (for example, to accommodate a wide range of output
impedances).

Thank you.


Questions on impedance matching

David Berlind
 

Hi all,

I have some questions on impedance matching that are borne out of some recent threads regarding challenges in getting a proper oscilloscope reading off an old signal generator.

I've been spending a disproportionate amount of time learning about impedance matching. I've learned the math (even developed a spreadsheet to game various scenarios), developed a pretty good understanding of the Maximum Power Transfer Theorem, and I get the principles of reflection. But here are my questions.

(1) While it's always best to shoot for minimal reflection (optimal power transfer), I've seen plenty of references to an acceptable over/under range. In other words, not a direct match. Are there any thoughts on what that range is?

(2) There are various techniques to impedance matching... resistors and transformers.. applying capacitive reactance... what has worked for you in the past? For example, if you go the reactance route, you minimize loss of power (vs. resistive), but then you tie yourself to a narrow range of frequencies. That's good if you have a relatively fixed use case. But what if you have a range of used cases?

(3) My main use case right now is working in the audio range with guitar amps, working with signal generators and so on. So, a safe diagnostic range is 80-5000Hz (I know overtones can run up to 20KHz). But let's say I'm matching the output of an old 600 ohm signal generator to the 1 megaohm input of my Tek oscilloscopes (which all have the same input impedance). That's an impedance ratio of 1000000 to 600 which works out to 1666:1. By the official math, minimizing reflection via transformer (if I chose to do so) would require a 40:1 ratio between the primary and secondary windings (or the equivalent accomplished in stages). I'm looking for suggestions on this "fixed" case.. but would also like to hear about any flexible solutions (for example, to accommodate a wide range of output impedances).

Thank you.

Re: recapped 485 (now we're cooking with grease edition)

Brendan
 

I now have everything working... EXCEPT b sweep just gives a dot. When a signal is input it looks like XY.

Re: My Next Scope

Chuck Harris
 

Consider that probably 99% of his customers are looking
for a trouble free turn key scope, and don't really have
the capability to feed the scope with a leveled signal
that can hold its amplitude, to required specifications,
at 50KHz, and 350MHz to 420KHz. And, certainly don't have
the capability of doing a transient response check.

I would bet that most of his customers use their scopes
to play around with audio, and to simply enjoy the
tektronix experience that they longed for back when the
2465B was new. Most of his customers would probably have
been perfectly happy with a 2445B, only they lusted for
the 2465B, because it was top of the line.

This gentleman is reasonably technically competent. His
lab is probably adequate to the task. His only major
problem is a moral one.

He does things like reset the hours counter to zero, and
tells you how the scope is New Old Stock... and explains
that any roughness in appearance is simply shelf wear,
and to be expected of a scope that was made 25 years ago.

And, of course, for a few hundred bucks more, he can give
you a scope that is pristine and perfect in every way.

If you complain that he has given you a faux scope, he will
tell you that he has simply released the 2445B's capabilities,
and how Tektronix is the real scammer, because they prevented
the 2445B from reaching its full capabilities.

And, that if you really want a congenital 2465B, he will
take back your faux model, and for an additional price
give you the real thing...

I have also seen him take an early 2465B that had the CMOS
RAM with an separate lithium backup battery, and directly
replace the CMOS RAM with a Dallas NVRAM, and leave the
nearly spent lithium battery in place... Not to good result.

-Chuck Harris

OBTW, the calibration technician is *supposed* to reset the
ON/OFF counter, and the Hours counter after calibration is
finished. It was never intended to be like a car's odometer.

Malcolm Hunter wrote:

On 8 February 2018 at 23:29, tmillermdems <tmiller11147@...> wrote:

The problem is that he gives a 1 year (I think) warranty and voids it if
the seals are broken. By the time you find out you have a counterfeit
2465B, ebay has timed out on their protection.

​So you can't do a bandwidth check?

Malc​

Re: My Next Scope

Chuck Harris
 

For what it is worth, I have taken a few of his faux
2465B's, added in most of the missing parts, and
succeeded in improving the performance.

The biggest problem, that can't be economically fixed,
is the missing matching network for the delay line.
The 2445 family has a pair of spiral inductors that form
a matching, and frequency limiting network. In order to
get close to 2465B performance, he cuts them out, and
jumpers over them, creating a direct connection to the
delay line. This leaves some artifacts on the first ns
of the leading/trailing edge of fast square waves.

Bandwidth is a hit or miss thing, as tektronix used the
slower vertical hybrids in the 2445 family early on.
Later parts seem to all meet the 2465B specifications.

To do this modification, you have to add a couple of 5K
Bourns trimmer pots, and change a few resistors. Then
the only missing part for calibration is an air wound coil
that is used in the leading edge compensation. It can't
easily be added in... or at least I haven't figured out a
way...

I often can get the faux 2465B scope to meet published
spec's after my modification, and a careful calibration.

The modification only adds about $70 to the cost of a
calibration.

-Chuck Harris

johnhstrauch via Groups.Io wrote:

Yes, it is true in my own personal lotsa $ experience. That man has no conscience and may have no soul.
My 310MHz "2465b" that is a 2445 with a main board mod that sure does not get it to 400MHz 2465b specs.
What did I get? Screwed !
JH

Re: My Next Scope

Jack Reynolds
 

Malcolm,

If you have not already done so you should visit 2465B's eBay store and read his "stuff".  He charges premium prices for his highly touted merchandise with a lot of emphasis on his expertise.  The implication is that if you are willing to pay him enough you can have the worry free scope of your dreams.  He is certainly not trying to appeal to the Tek nuts and bolts guys with a lab full of test equipment who are poised ready to jump on the new scope and test it with everything they've got!

Jack Reynolds

On 2/8/2018 8:21 PM, tmillermdems wrote:
I can but not everyone can.
In fact, I bet most that are on the group can't really do it right.

Regards

----- Original Message ----- From: "Malcolm Hunter" <malcolm.r.hunter@...>
To: <TekScopes@groups.io>
Sent: Thursday, February 08, 2018 7:02 PM
Subject: Re: [TekScopes] My Next Scope


On 8 February 2018 at 23:29, tmillermdems <tmiller11147@...> wrote:

The problem is that he gives a 1 year (I think) warranty and voids it if
the seals are broken. By the time you find out you have a counterfeit
2465B, ebay has timed out on their protection.

​So you can't do a bandwidth check?

Malc

Re: My Next Scope

 

I can but not everyone can.
In fact, I bet most that are on the group can't really do it right.

Regards

----- Original Message -----
From: "Malcolm Hunter" <malcolm.r.hunter@...>
To: <TekScopes@groups.io>
Sent: Thursday, February 08, 2018 7:02 PM
Subject: Re: [TekScopes] My Next Scope


On 8 February 2018 at 23:29, tmillermdems <tmiller11147@...> wrote:

The problem is that he gives a 1 year (I think) warranty and voids it if
the seals are broken. By the time you find out you have a counterfeit
2465B, ebay has timed out on their protection.

​So you can't do a bandwidth check?

Malc-=-=-=-=-=-=-=-=-=-=-=-

Re: Tek 2235 baseline

 

Issue solved.

Found open potmeter on attunator board.
Working fine after replacement!

Hugo

Re: My Next Scope

Malcolm Hunter
 

On 8 February 2018 at 23:29, tmillermdems <tmiller11147@...> wrote:

The problem is that he gives a 1 year (I think) warranty and voids it if
the seals are broken. By the time you find out you have a counterfeit
2465B, ebay has timed out on their protection.

​So you can't do a bandwidth check?

Malc​