Re: My TDR Evaluation of the SG503 012-0482-00 Cable has been REVISED
All the frequency domain information is present in the TDR result. But it requires a bit of mathematics to do the conversion.
I am waiting on receipt of a pair of 1 MHz square wave <40 ps rise time generators from Leo Bodnar.
Once those arrive, I am going to write a long series on doing vector network analysis in the time domain using a DSO on the EEVblog forum. Aside from the tutorial aspect, I'll be writing software to take data from a DSO and generate magnitude and phase vs frequency. I'll probably leave creating Smith charts to some one else, but I will do the basic time to frequency part as I am very familiar with that. And that's the hard part for most EEs. I'm a geophysicist. So TDR is my natural habitat, though elastic waves at less than 250 Hz rather than EM waves at 10 GHz.
I requested the 1 MHz clock instead of the standard 10 MHz because most DSOs have such horrible front ends. The lower clock rate lets me window off the ringing from the step. It still needs to be accounted for, but it's a lot easier if things don't overlap in time. A 10 ft RG58 delay line should suffice for most low end DSOs.
I have an 8753B/85046A VNA & S parameter set, so I'll be comparing TDR using Leo's unit and a DSO to the HP VNA, and if I can get my Tek 11801 to transfer data to a PC using the 20 ps calibrator and an SD-26 head.
My motivation is pretty simple. 35 years ago while in grad school I tried to build a 40 m DC receiver. It did not work. I later confirmed that the reason was a lack of any test gear other than a 5 MHz recurrent sweep Heathkit IO-18 and a Radio Shack DMM and VOM. My input BP filter was fine except it was about 500 KHz to low.
|
Re: 475A with High +110 rail
On Wed, May 1, 2019 at 02:46 AM, tom jobe wrote:
The online 'parallel resistor calculators' show that your 150k and 100k pair
give you exactly 60k of resistance.
I wonder if it would be okay to leave your pair of resistors in there as a
permanent fix?
Obviously, using 5% tolerance resistors of 100 kOhm and 150 kOhm in parallel will not guarantee to get 60 kOhm. The values of R1486 and R1487 as such aren't critical but their combination basically determines the 110 V: V110 = V50 * (R1486 + R1487) / R1487. I haven't checked the circuits in detail but accuracy worse than 1% for the combination may spoil scope calibration state. You may consider using 3 resistors in a parallel/serial config for R1486 - or even slightly changing R1487 to get the correct ratio but checking with your DVM will be necessary unless you use 1% resistors. Building R1486 from a series connection of a fixed, reasonably stable (metal film) 56 kOhm resistor with a 10 kOhm trimpot in series is another possibility, giving enough resolution in trimming with a range between about 110 V plus and minus 5%, since it allows varying R1486 from about 56 kOhm to about 66 kOhm. The Ohmmeter function on almost any DVM will be accurate enough for trimming. Raymond
|
Re: My TDR Evaluation of the SG503 012-0482-00 Cable has been REVISED
Hi Reginald,
Testing cables with a TDR reveals important characteristics of how a cable will respond when a pulse is applied to it. It is a quick way to check a cable at a glance. But it misses many other characteristics, which depending on the ultimate use of the cable, might be equally important, and in some cases, even more important.
If I had a Vector Network Analyzer I would have tested all three cables that way because it will reveal some of those other characteristics but I ran out of time.
I do have a Scalar Network Analyzer which I thought about using for the test but I had to get back to other things.
Your comment touches on this aspect of the cables characteristics I have often wondered about. No matter how well it is made it is going to have issues such as you have pointed out. This is an unavoidable fact of physics. And it will have an effect on how this cable performs in the frequency domain.
Dennis Tillman W7PF
toggle quoted message
Show quoted text
-----Original Message-----
From: Reginald Beardsley
Sent: Tuesday, April 30, 2019 3:46 PM
Subject: Re: [TekScopes] My TDR Evaluation of the SG503 012-0482-00
Cable has been REVISED
FWIW the reflection in the cable has a pole at 242 MHz for a 36" cable
and 0.678 velocity factor.
An interesting experiment would be to measure the output of the SG503 at
the generator and at the end of the cable. The cable spec may have been
chosen to compensate for roll off in the generator at the upper
frequency limit or to prevent a bump in the response below 250 MHz which
a longer cable would produce. -- Dennis Tillman W7PF TekScopes Moderator
|
Re: 475A with High +110 rail
Hi Micheal,
Thank you for the nice repair summary!
It was good that Dave Hills brought some of his logic, reason and knowledge to your problem.
The online 'parallel resistor calculators' show that your 150k and 100k pair give you exactly 60k of resistance.
I wonder if it would be okay to leave your pair of resistors in there as a permanent fix?
Maybe some kind soul will comment on that question for us amateurs?
Those Tektronix Service Manuals are like a series of college courses on analog electronics!
tom jobe...
toggle quoted message
Show quoted text
On 4/30/2019 4:21 PM, Mlynch001 wrote: On Mon, Apr 29, 2019 at 08:07 AM, Dave Hills wrote:
With 134v on Q1496 collector, you should see +61v at Q1490's base. This will
completely shut down the regulator by reverse biasing Q1494 b-e. That said it
only leaves R1483, Q1496, CR1498, and R1488 as likely suspects. R1483 would
have to have drifted VERY low, (unlikely), or R1488 VERY high, (possible, but
not likely to cause these symptoms), leaving a leaky or shorted Q1496 or
CR1489. Note that an OPEN CR1489 would likely cause Q1496 to subsequently
fail. Also, check that CR1489 has not been installed backward.
Please forgive the caps, not shouting, just no other way to highlight for
emphasis.
Dave
Dave:
Your advice helped me find the issue! Tom Jobe suggested that I study the 110V regulator circuit description and the schematic, which I did as well. Lots of good advice from everyone got me pointed in the right direction.
Here is what I found and how I fixed the issue:
After the aforementioned study and re-reading all the comments, I decided to pull CR1489 off the board and test it. It tested good on my curve tracer, meter and component tester. Reinstalled on the board and there was no change. (of course) I also pulled all transistors, including Q1496 (again) and tested it on my little home built curve tracer and re- ran it through two additional transistor testers, all showed it to be good. (again, no Change) You mentioned that Q1490 should have a base voltage of about 61V with the 110V rail at 135V, I found it to actually be low, at 41.5V, the schematic calls for 50V, and since you said I should have about 61V; I knew that the voltage divider formed by R1486 and R1487 was not right. R1487 tested 49.9K so I moved on to R1486 and it tested "weird", meaning high resistance sometimes and then open at others. This prompted me to remove it completely from the board. After getting it off the board, I found it to be "open", so un-soldering the component probably finished it off.. I took a 68K resistor from the bench and temporarily soldered it into place of R1486, checked the 110V rail and BANG! the 110V rail was now at 119.4V (down from 135V) with no other changes. This told me I had found the issue, now I had to find a proper replacement for R1486, which was originally 60.4K (321-0364-00 - RES.,FXD,FILM: 60.4K OHM, 1%,o.125W). Doing a little math, I found that installing 100K and 150K resistors in parallel would deliver 59K, so VERY close to what I needed. Installed these two resistors parallel in place of R1486, re-tested, this brought the base of Q1490 to 50.06V and gave me exactly 110.65V on the 110V rail. Now all I have to do is find the proper replacement part and my scope in good to go! THANK YOU ALL for your advice!
Sincerely,
|
After accidentally leaving my 1502 on overnight I came back to find
it dead. I eventually traced it down to a HV problem.
This circuit is so simple it's ludicrous but I'm hampered by having
lost my HV probe. In any event, the output across the HV transformer
is a square wave of about 50V; it's supposed to be 500V. If I
disconnect C4328 (connected directly to the transformer), I see the
expected 500V across the transformer. The HV diodes appear to be OK,
but something is loading it down. I've disconnected the CRT and
controls with no change, so it has to be on the board.
I don't see any unusually low ohm readings either.
Any ideas for tracking down the failed component or should I just
shotgun the multiplier and replace everything? One cap is 0.033
at 600V and the other 3 are 0.027 at 1200V.
I do have a 1KV supply so I could check leakage on them, although
I hate to remove good parts from a board.
BTW, instead of those little pin connectors, the CRT cable and
intensity control cable are soldered directly to the board. This looks
like a repair attempt to me.
thanks
--
Paul Amaranth, GCIH | Manchester MI, USA
Aurora Group of Michigan, LLC | Security, Systems & Software
paul@... | Unix & Windows
|
Are you planing to make the book in electronic format? I would like to be able to search for specific words ... etc. Printed book is better than nothing, but I don't have space for all the books I would like to have, so I buy them in electronic format. Thanks Tony Fleming On Tue, Apr 30, 2019 at 7:35 PM Dave Brown <davebr@...> wrote: As mentioned previously, Stan gave the museum authorization to reprint his
book and we do sell it on CD on our eBay store. We are in the process of
reprinting it in softbound. We have our review copy back and are making
final tweaks. I'll update the group when we have them available on our eBay
store. Our first run will be small as we don't know how much demand there
is for this book. When we run out, we will reprint more.
vintageTEK
|
As mentioned previously, Stan gave the museum authorization to reprint his book and we do sell it on CD on our eBay store. We are in the process of reprinting it in softbound. We have our review copy back and are making final tweaks. I'll update the group when we have them available on our eBay store. Our first run will be small as we don't know how much demand there is for this book. When we run out, we will reprint more.
vintageTEK
|
Re: 475A with High +110 rail
On Mon, Apr 29, 2019 at 08:07 AM, Dave Hills wrote:
With 134v on Q1496 collector, you should see +61v at Q1490's base. This will
completely shut down the regulator by reverse biasing Q1494 b-e. That said it
only leaves R1483, Q1496, CR1498, and R1488 as likely suspects. R1483 would
have to have drifted VERY low, (unlikely), or R1488 VERY high, (possible, but
not likely to cause these symptoms), leaving a leaky or shorted Q1496 or
CR1489. Note that an OPEN CR1489 would likely cause Q1496 to subsequently
fail. Also, check that CR1489 has not been installed backward.
Please forgive the caps, not shouting, just no other way to highlight for
emphasis.
Dave
Dave: Your advice helped me find the issue! Tom Jobe suggested that I study the 110V regulator circuit description and the schematic, which I did as well. Lots of good advice from everyone got me pointed in the right direction. Here is what I found and how I fixed the issue: After the aforementioned study and re-reading all the comments, I decided to pull CR1489 off the board and test it. It tested good on my curve tracer, meter and component tester. Reinstalled on the board and there was no change. (of course) I also pulled all transistors, including Q1496 (again) and tested it on my little home built curve tracer and re- ran it through two additional transistor testers, all showed it to be good. (again, no Change) You mentioned that Q1490 should have a base voltage of about 61V with the 110V rail at 135V, I found it to actually be low, at 41.5V, the schematic calls for 50V, and since you said I should have about 61V; I knew that the voltage divider formed by R1486 and R1487 was not right. R1487 tested 49.9K so I moved on to R1486 and it tested "weird", meaning high resistance sometimes and then open at others. This prompted me to remove it completely from the board. After getting it off the board, I found it to be "open", so un-soldering the component probably finished it off.. I took a 68K resistor from the bench and temporarily soldered it into place of R1486, checked the 110V rail and BANG! the 110V rail was now at 119.4V (down from 135V) with no other changes. This told me I had found the issue, now I had to find a proper replacement for R1486, which was originally 60.4K (321-0364-00 - RES.,FXD,FILM: 60.4K OHM, 1%,o.125W). Doing a little math, I found that installing 100K and 150K resistors in parallel would deliver 59K, so VERY close to what I needed. Installed these two resistors parallel in place of R1486, re-tested, this brought the base of Q1490 to 50.06V and gave me exactly 110.65V on the 110V rail. Now all I have to do is find the proper replacement part and my scope in good to go! THANK YOU ALL for your advice! Sincerely, -- Michael Lynch Dardanelle, AR
|
Re: My TDR Evaluation of the SG503 012-0482-00 Cable has been uploaded
Hi John,
If you want to send me one I can test it and compare the results to the two 012-0482-00 cables I have.
Contact me off list at dennis at ridesoft dot com if you are interested.
If I had to guess at what will be the most important factor in duplicating the Tek cable it would be
1) Is the characteristic impedance of the RG-223/U cable really exactly 50.0 ohms?
2) Are the BNC connectors exactly 50.0 ohms and do the attach without any impedance variation?
3) Is the method of attachment so good that it has no reflections and mismatch.
Dennis Tillman W7PF
toggle quoted message
Show quoted text
-----Original Message-----
From: John Ferguson
Sent: Tuesday, April 30, 2019 1:10 PM
Subject: Re: [TekScopes] My TDR Evaluation of the SG503 012-0482-00
Cable has been uploaded
On 4/29/19 1:40 PM, Dennis Tillman W7PF wrote:
This raises another question which none of us may ever have the answer
to: What are the unique qualities the Tek SG503 designer wanted this
cable to have to be the ideal match to the SG503. It may be something as
silly, in retrospect, as the stiffness of the cable. Maybe he/she like
stiff cables. Why is it exactly 36in? Maybe he/she had a scrap piece of
stiff coax exactly that length in his drawer and he/she said "Use this"
to the production people when they asked what kind of cable should ship
with the SG503. Meanwhile we are all guessing at whether using this
cable with an SG503 is really important.
Dennis,Thank you so much for this wonderful informative tests and
discussion of these cables. And it would be interesting to kn ow what
motivated Tek to build this particular cable to this spec, while
wondering what part of the spec, or unwritten assumptions about the
ultimate cable assembly's properties were important.
I blanched at the cost of the real thing, and under advice from one of
our members, bought good BNC (not inexpensive) connectors, the rg223
cable, and built my own. But not having a TDR or any other way to test
the result, I don't know how good mine are.
Thanks again,
john -- Dennis Tillman W7PF TekScopes Moderator
|
Re: My TDR Evaluation of the SG503 012-0482-00 Cable has been REVISED
FWIW the reflection in the cable has a pole at 242 MHz for a 36" cable and 0.678 velocity factor.
An interesting experiment would be to measure the output of the SG503 at the generator and at the end of the cable. The cable spec may have been chosen to compensate for roll off in the generator at the upper frequency limit or to prevent a bump in the response below 250 MHz which a longer cable would produce.
|
Re: My TDR Evaluation of the SG503 012-0482-00 Cable has been REVISED
Hey Denis, thanks for putting this doc together - it's very interesting to me as a TDR n00b to see a case study with images and calculations. As for the correction, I take no credit, as several other people noticed this before me, starting perhaps with Albert Otten, who quipped "Most remarkable is the slow speed of light in vacuum in the USA ;-)". Siggi On Mon, Apr 29, 2019 at 9:20 PM Dennis Tillman W7PF <dennis@...> wrote: Hi Siggi,
YIKES! Thanks for catching my error. That was a very stupid typo I made. I
know the speed of light by heart.
I was getting tired when I was doing that part of the write-up and I
remember I was very annoyed at myself that I was having trouble
concentrating.
I revised my calculations. The new Velocity Factor is 0.678 which is much
more common. The diameter of the 012-0482-00 cable is 0.210 inches and it
is stiffer than other coax due to the double shield. Since RG223/U is
double shielded, has a diameter of 0.213inches, and has a Velocity Factor
of 0.67 this may be the coax that Tek used to make these cables. There is
still another thing that needs to be located to make a replacement cable -
locate a true 50ohm BNC connector. Most BNC connectors are not 50ohms. They
can, and do, vary between 50 to 54ohms to match to the different impedances
of supposedly 50ohm cable (which also can vary between 50 to 54ohms).
I just uploaded a revised copy of my evaluation with the corrected formula
and value and a few other changes to clarify my conclusions.
I deleted the previous document. The revised document is at
/g/TekScopes/files/Evaluation%20of%20the%20Tektronix%20012-0482-00%20cable%20for%20the%20SG503%20REVISED.pdf
The corrected document says REVISED right in the title and in the footnote
at the bottom.
Albert Otten's email seems right on cue. Let us know how your RG223 cable
works out.
Dennis Tillman W7PF
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of
Siggi
Sent: Monday, April 29, 2019 10:51 AM
To: [email protected]
Subject: Re: [TekScopes] My TDR Evaluation of the SG503 012-0482-00
Cable has been uploaded
On Mon, Apr 29, 2019 at 1:40 PM Dennis Tillman W7PF
<dennis@...>
wrote:
Hi Albert,
I think my derivation is valid and I know the VF can't be that high.
Did I leave something out in the way I calculated it?
You misstated the sped of light in vacuum, which is 299792458m/s per
Wikipedia, whereas your doc states "The speed of light in a vacuum is
224844343.5m/Sec = 0.2248m/nSec.", which is a 25% error :).
|
Re: My TDR Evaluation of the SG503 012-0482-00 Cable has been uploaded
I can do a TDR assessment of my DIY RG400 cable with an 8753DD VNA I have on hand.
In terms of just amplitude at 200MHz it has shown 0.01dB loss compared to a Tektronix cable (not a 012-0482-00) I had on hand that had a 0.2dB loss.
I'll need to hunt about for the 012-0482-00.
|
Re: Calibration and full checkout needed - Tek 7000 series
If you end up needing to ship, John Griessen should be able to provide tips. The 7904A I recently purchased from him arrived extremely well packed.
The plugin complement I bought with the 'scope is: 7A26, 7A19, 7B15 (getting ready for the 7104 I want!!), and 7B92A.
|
Re: My TDR Evaluation of the SG503 012-0482-00 Cable has been uploaded
One might also be interested in the practicality and cost of properly calibrating an SG503.
|
Re: My TDR Evaluation of the SG503 012-0482-00 Cable has been uploaded
On 4/29/19 1:40 PM, Dennis Tillman W7PF wrote:
This raises another question which none of us may ever have the answer to: What are the unique qualities the Tek SG503 designer wanted this cable to have to be the ideal match to the SG503. It may be something as silly, in retrospect, as the stiffness of the cable. Maybe he/she like stiff cables. Why is it exactly 36in? Maybe he/she had a scrap piece of stiff coax exactly that length in his drawer and he/she said "Use this" to the production people when they asked what kind of cable should ship with the SG503. Meanwhile we are all guessing at whether using this cable with an SG503 is really important.
Dennis,Thank you so much for this wonderful informative tests and discussion of these cables. And it would be interesting to kn ow what motivated Tek to build this particular cable to this spec, while wondering what part of the spec, or unwritten assumptions about the ultimate cable assembly's properties were important.
I blanched at the cost of the real thing, and under advice from one of our members, bought good BNC (not inexpensive) connectors, the rg223 cable, and built my own. But not having a TDR or any other way to test the result, I don't know how good mine are.
Thanks again,
john
|
Calibration and full checkout needed - Tek 7000 series
Hi All,
I need one, good working scope to get started on repairing the other 30 or so scopes around here. I have a multitude of Tek 7000 frames to choose from but I think the 7904 (or 7904A) is a good place to start. Once that's working + cal'ed I can at least troubleshoot and maybe repair the others.
Where can I get one professionally calibrated in the greater Detroit area? I can drive a fair distance, but I'm scared to ship. I've had two 7934's, a 7834, and a 7704A crushed in separate shipping mishaps.
Also...which plug ins are recommended to cal for building my golden scope? I have a collection...
|
Re: 2465 repair attempt with (horizontal?) issues, help needed [Solved!!]
Thank you for detail information! On Tue, Apr 30, 2019 at 8:08 AM satbeginner <castellcorunas@...> wrote: Hi all,
Here a solution I found for a horizontal problem I had on (one of) my 2465.
The problem was that all horizontal traces, readout, etc. were waaaay too
wide, also the time cursors were not shown.
Also when put in X-Y mode I still had a horizontal line, instead of a dot.
But the good news is: I solved it.
My assumption was that the horizontal shaped line must have been some kind
of deflection??
In a way it was, but not some hum or ripple, it turned out to be the
control of the last set of convergence grids.
On the A9 High Voltage board there is the Edge Focus potmeter (R1864),
that sets the voltage on the vertical convergence grid.
Amplifier U1890B and Q1890 are supposed to invert that voltage as an
inverting voltage follower to be used on the horizontal convergence grid.
After some searching I found that the voltage on the Quad Pole #3 (the
horizontal convergence grid) was always -285V, regardless of the setting of
the R1864 potmeter that did change the voltage on quad pole #4.
It was the €0,03 1MOhm feedback resistor R1891 that went all open :-), so
no feedback so the output went as negative as negative can be.
So, instead of extra focussing the beam it did exactly the opposite, it
spread the beam out completely, hence the line and the way too wide traces!
After replacing that resistor I have about the sharpest trace of all my
scopes, so I am a happy bunny!
Now I will put it all together again and start the calibration.
Un saludo,
Leo
|
Re: Looking for fan impeller fastener for 2465
Parts for the "original" 2465 fan (motor, impeller, collet) are near impossible to find.? Even when found, they are old and made from plastic that does not age well.? Most people make substitutes or convert to muffin fans as used in the later model 24xx scopes.? You can search the group archives with the term "impeller" and spend an afternoon learning all the ways others (including me) have been frustrated with this part.? I ended up using a collet from my collection of Dremel tools to replace the collet.? (Re: Removing a 2465 Squirrel cage fan). A more elegant solution can be found in the pictures archives.? Search for "2465 fan collet repair"Or, if you have access to a machine shop there is a drawing for making one:?? /g/TekScopes/files/Fan%20Collet%20-%20Tek%202465%20O-scope.pdf
On Tuesday, April 30, 2019, 12:55:51 AM CDT, Jonathan.yaeger@... <Jonathan.yaeger@...> wrote:
Does anyone have or sale or a source for the impeller fastener, Tek P/N:? 343-104-001?
Thanks in advance.
Jon
|
Re: 2465 repair attempt with (horizontal?) issues, help needed [Solved!!]
Very good sleuthing Leo, congratulations!
I have been eying that circuit suspiciously on my 2467B as a
possible cause of its beam distortions. I think I will take a
closer look.
-Chuck Harris
satbeginner wrote:
toggle quoted message
Show quoted text
Hi all,
Here a solution I found for a horizontal problem I had on (one of) my 2465.
The problem was that all horizontal traces, readout, etc. were waaaay too wide, also the time cursors were not shown.
Also when put in X-Y mode I still had a horizontal line, instead of a dot.
But the good news is: I solved it.
My assumption was that the horizontal shaped line must have been some kind of deflection??
In a way it was, but not some hum or ripple, it turned out to be the control of the last set of convergence grids.
On the A9 High Voltage board there is the Edge Focus potmeter (R1864), that sets the voltage on the vertical convergence grid.
Amplifier U1890B and Q1890 are supposed to invert that voltage as an inverting voltage follower to be used on the horizontal convergence grid.
After some searching I found that the voltage on the Quad Pole #3 (the horizontal convergence grid) was always -285V, regardless of the setting of the R1864 potmeter that did change the voltage on quad pole #4.
It was the €0,03 1MOhm feedback resistor R1891 that went all open :-), so no feedback so the output went as negative as negative can be.
So, instead of extra focussing the beam it did exactly the opposite, it spread the beam out completely, hence the line and the way too wide traces!
After replacing that resistor I have about the sharpest trace of all my scopes, so I am a happy bunny!
Now I will put it all together again and start the calibration.
Un saludo,
Leo
|
Re: 2465 repair attempt with (horizontal?) issues, help needed [Solved!!]
Hi all,
Here a solution I found for a horizontal problem I had on (one of) my 2465.
The problem was that all horizontal traces, readout, etc. were waaaay too wide, also the time cursors were not shown.
Also when put in X-Y mode I still had a horizontal line, instead of a dot.
But the good news is: I solved it.
My assumption was that the horizontal shaped line must have been some kind of deflection??
In a way it was, but not some hum or ripple, it turned out to be the control of the last set of convergence grids.
On the A9 High Voltage board there is the Edge Focus potmeter (R1864), that sets the voltage on the vertical convergence grid.
Amplifier U1890B and Q1890 are supposed to invert that voltage as an inverting voltage follower to be used on the horizontal convergence grid.
After some searching I found that the voltage on the Quad Pole #3 (the horizontal convergence grid) was always -285V, regardless of the setting of the R1864 potmeter that did change the voltage on quad pole #4.
It was the €0,03 1MOhm feedback resistor R1891 that went all open :-), so no feedback so the output went as negative as negative can be.
So, instead of extra focussing the beam it did exactly the opposite, it spread the beam out completely, hence the line and the way too wide traces!
After replacing that resistor I have about the sharpest trace of all my scopes, so I am a happy bunny!
Now I will put it all together again and start the calibration.
Un saludo,
Leo
|
Raymond Domp Frank
Mlynch001
Ancel
satbeginner