Date   

Re: Tek 576 Curve Tracer USB port to a cpu ... ?

 

Hi Ke-Fong Lin,

Digitizing the deflection voltages is a great idea because of its simplicity. It would be easy to capture the deflection voltage for both axes with A/D converters and convert it into X-Y data. A few issues that may make it a bit more difficult:
* The deflection signals are differential so the X axis would require two A/Ds sampling the two vertical deflection plates and two A/Ds sampling the two horizontal plates. Both of the vertical A/Ds would need to sample simultaneously. The same is true for the two horizontal A/Ds.
* There is no information about what you are looking at. For a simple example of measuring a bipolar transistor the data you capture has no information about the Volts/Div of the horizontal axis or the Amps/Div of the vertical axis, or what current step caused the trace to move vertically and by how much.
* You have no idea when to start and stop capturing the data. This might be easy to fix if there is a start sync pulse you could use to trigger the A/Ds. I'm not aware that there is a Stop pulse. It may be possible to use a Z-axis output signal or un-blanking signal, if one is available somewhere inside the curve tracer for this. The A/Ds would start when the un-blanking signal went true and they wouldn't stop until the un-blanking signal went false.

Assuming there is a way to use the un-blanking signal or equivalent this might work.

The way this could work is by pressing a momentary push button to tell your software to take a sample of the next set of curves. It would then wait for the un-blanking signal to go true. The software would automatically write the data streaming from the four A/Ds to local memory. When the un-blanking goes false, indicating the curve tracer is done displaying a complete set of curves the A/Ds would be stopped and the software would automatically send the sampled data to a display or back to the curve tracer's CRT so the user could decide if they wanted to save it permanently or do this over again. If you decide to save the data it would then ask you to enter the settings of the curve tracer knobs or maybe what would be better would be that you have the software generate a pattern on the CRt that could be as simple as a diagonal line for example. By adjusting the line so it starts in the lower left corner of the screen and ends in the upper right corner of the screen you are providing a calibrated range that can then be used by entering the settings of the vertical A/Div knob and the horizontal V/Div knob. Finally you would include the base current steps and how many steps there were.

This actually might work!!!

Dennis Tillman W7PF

-----Original Message-----
From: TekScopes@groups.io [mailto:TekScopes@groups.io] On Behalf Of Ke-Fong Lin
Sent: Wednesday, January 08, 2020 2:00 PM
To: TekScopes@groups.io
Subject: Re: [TekScopes] Tek 576 Curve Tracer USB port to a cpu ... ?

Would it not be a simple matter to somehow convert the 576 deflection
plate voltage signals to X-Y Data? Just thinking out loud here.
As mentioned already, the refresh is at 60Hz so a pen plotter cannot go that fast.
However. it should be possible to have the display "static" and have the plotter slowly do it's work.
But then you would have to design a mechanism to "follow" the "beam" at the slow pen rate.
That would be quite complicated.

In contrast, the spectrum analyzer is a single f(x) function to plot.
And in the case of my HP3580A which is a low frequency SA, the sweeping speed is very slow (can be a few minutes).

Indeed, what looks easy for an Arduino, may not have been in the 70s!







--
Dennis Tillman W7PF
TekScopes Moderator


Re: 'Solder Rot'

KB6NAX
 

Copper tracks on aqueous cleaned boards can form between traces. I once found one that was over an eighth inch long. -Arden


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

John Griessen
 

On 1/8/20 4:50 PM, Miguel Work wrote:
I think that the first step is to make single ended outputs, x y z. BNC outputs in the rear panel for example, I have some drafts, schematics designed to connect to deflection plates.
Yes! That's how to divide and conquer. This seems to have some interest. Maybe there are volunteers for gruuuuunt work needed...

I have examples of volt divider layout that works for 2 kV to record traces with micropython on a STM32f401CEU6, various FETs, diodes, R's, C's. If the pickoff point is higher, then it needs more work, but I think there has to be some under 2kV analog to digitize in there somewhere. (I have not even looked at schematic.)


Re: 454 HV regulation off

KB6NAX
 

You have a bunch of 3 megohm resistors in series in the feedback path from the CRT cathode to the oscillator control Q1414. Check the resistor values and sub the .001 cap. Yep, ceramics do go bad. -Arden


454 HV regulation off

Jack Ohme
 

I created a thread a few days ago about my 454, and believe I have isolated
the issue, in its -1960v supply. The HV reg is totally off, which is
causing my grid voltage to be far too high, and as an effect, my screen
brightness as well. The LV regulation is spot-on, and cranking the HV
regulation control pot, R1401, mentioned on page 6-11 of the manual, to the
max and then dropping it to the minimum causes my screen and trace to drop
to an acceptably low level for a moment before returning to far too bright.
Any idea as to what could be causing this? Something in the actual HV
section, or something in the LV reg that controls the HV? Oh and to put a
number on my -1960v test point, its hovering around -1830 currently.


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

ebayatessnh
 

I did a project for a military sub contractor a few years ago (~2008) that turned an XY display into a raster image using a PIC micro controller with an LCD drive. The PIC sampled the XY signal using the internal AD converters, and updated the memory array that drove the LCD. I don't recall the update rate of the display, but it was way faster than 60 Hz. If you are just looking for a large display, and not data capture this might be a reasonable approach.

If are looking for is a display and data capture, then a Xilinx-Zync processor would be ideal. It has built in AD converters, two fast ARM processors, and a large FPGA all in one chip. In fact, I have a R&S EMI receiver with a bad display that I was contemplating rasterizing with a Zync demo board from Digilent. Digilent has example FPGA code that can drive and HDMI display from the FPGA.

Dave


For Sale: Tektronix 7623A Multi-mode Storage Oscilloscope with 12 Plugins & 5 Manuals near Seattle, WA, USA

maxfrister@...
 

I'm selling a cool 'scope for a friend.

Asking $250, local pickup near Seattle.

Send me an email and I can send pictures.



Tektronix 7623A 100MHz Multi-mode Oscilloscope

Scope and most plugs are tested for basic function and are working. Three plugins will require repairs (see details below).

With instruction and service manuals

100MHz Bandwidth

Storage Modes: variable persistence, fast variable persistence, bistable, fast bistable, non-store.


12 7000 series Plugs

7A26 Dual trace 200MHz Vertical + Manual
7A12 Dual trace 120MHz vertical amplifier
7A12 Dual trace 120MHz vertical amplifier (one switch does not latch)
7A15A Single channel 85Mhz vertical amplifier
7A15AN (same as 7A15) (Unstable trace)
AM-6565U (Same as 7A15). Single Channel 85MHz vertical amplifier


7B71 Delaying timebase 200Mhz Horizontal (needs repair)
7B53A Dual timebase 100Mhz Horizontal + Manual
7B85 Delaying timebase 400MHz

7A22 Differential Amplifier + Manual
7D15 Frequency counter 225MHz
7D13A Digital Multimeter


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

Miguel Work
 

I think that the first step is to make single ended outputs, x y z. BNC outputs in the rear panel for example, I have some drafts, schematics designed to connect to deflection plates.


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

Jim Ford
 

Thanks, John.I do have a project in mind.  Will PM you this evening from home.JimSent from my Verizon, Samsung Galaxy smartphone

-------- Original message --------From: JJ <jajustin@...> Date: 1/7/20 4:37 PM (GMT-08:00) To: TekScopes@groups.io Subject: Re: [TekScopes] Tek 576 Curve Tracer USB port to a cpu ... ? Hi, Jim. I used many resources and tutorials - YouTube has many goodtutorials for beginners. There is Python and there is PYQT5. PYQT5 allowsyou to build a Graphical User Interface (GUI) easily. - it uses Python.It's a long road and takes time to become proficient - a few months atleast. I programmed in Java, C++, Javascript, and many other proprietarylanguages over the past 50 years - so picking up a new language isn't allthat difficult for me.I have also been an analog/digital circuit designer over that period aswell - so I'm with you. Using your hardware skills together with an Arduinoand coupled with Python allows you to bring the analog outside world intoyour PC for data analysis and control.I can send you a few books in PDF format that I used if you like.But please note that you need to love this stuff - it takes time andpatience! But, once you get it, you can do virtually anything you want. Thebest way to learn a new programming language is to jump in and practice.Don't just read a book - you need to practice! Otherwise, it gets tooboring and overwhelming! It's good to have a project in mind - it's verymotivational.Best,John Justin


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

Dave Voorhis
 

On 8 Jan 2020, at 22:04, Tam Hanna <tamhan@...> wrote:

Hi,
Our product names come from a name generator from a previous employer. I stand by it for one reason - confidentiality.

Vendors can, and have, leaked data between customers a lot in the past. In my case, the situation is easy as the names mean nothing.

Tam
---
With best regards
Tam HANNA (emailing on a BlackBerry PRIV)

Enjoy electronics? Join 14k other followers by visiting the Crazy Electronics Lab at https://www.instagram.com/tam.hanna/
Are you responding to my question about referring to 575/576/577 curve tracers as Danaher rather than Tektronix?

If so, I’m sorry, but I can’t make sense of your response.

If it’s a reference to “Stinkely”, then it makes a bit more sense, but doesn’t answer my question about Danaher...


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

Tam Hanna
 

Hi,
Our product names come from a name generator from a previous employer. I stand by it for one reason - confidentiality.

Vendors can, and have, leaked data between customers a lot in the past. In my case, the situation is easy as the names mean nothing.

Tam
---
With best regards
Tam HANNA (emailing on a BlackBerry PRIV)

Enjoy electronics? Join 14k other followers by visiting the Crazy Electronics Lab at https://www.instagram.com/tam.hanna/


Re: 2235 2mV vertical range noise and surprising fixes

Bert Haskins
 

On 1/7/2020 4:00 PM, walter shawlee wrote:
The 2mV vertical range on the 2235 is a triumph of marketing over technology, as it really pushes the performance of the scope to a place where it doesn't willingly want to go. virtually all 2235's exhibit increasing noise on this range as they age, and sometimes the displayed noise is huge (when the psu caps deteriorate in the secondary of the switching supply).

I have had to fix many of these, and generally follow this plan:
----------------------------------------------------------------------------------
1. replace all LV psu secondary caps with modern low ESR versions (820uf becomes 1000uf).
check for ripple after the change at the psu test points, and replace any bad switching rectifiers if needed.
2. tighten all ground hardware all over the scope, especially around the vertical assembly. make sure all case ground screws (2) are there too.

Recently I had one stubborn unit that just wouldn't behave even after re-capping (which it desperately needed), and I traced the fault to a very unusual area. all of the switching supplies are slaved the the adjustment of the -8.6V level. my unit measured -8.42V, which I initially ignored as it seemed to be within spec. BUT, the manual specifies a tighter adjustment range, and incredibly, as I raised the voltage to just over -8.6V with the rear edge trimpot on the motherboard, the displayed ripple almost vanished. I have never seen this interaction before, so I thought I would pass it on, as it may prove useful to anybody re-capping one of these nice little 100MHz scopes.

the difference between good and bad ripple display is a very narrow -8.6V adjustment range, about 150mV, which I found amazing, as I cannot see how this has such a dramatic effect, but it does. no adjustment or parts replacement can totally remove ALL ripple from the 2mV range display, but you can certainly improve it down to about one minor division. bad scopes with poor caps can be 5-10 minor divisions, making it unusable on the 2, 5 and 10mV ranges.

I hope this proves useful to others working on 2235's.
all the best,
walter (@walter2)
sphere research corp.
https://www.sphere.bc.ca/test/index.html

Thanks for this Walter, I'll be sure to check this out on my 2232(s) also.

Thanks again,
  Bert


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

Ke-Fong Lin
 

Would it not be a simple matter to somehow convert the 576 deflection plate
voltage signals to X-Y Data? Just thinking out loud here.
As mentioned already, the refresh is at 60Hz so a pen plotter cannot go that fast.
However. it should be possible to have the display "static" and have the plotter slowly do it's work.
But then you would have to design a mechanism to "follow" the "beam" at the slow pen rate.
That would be quite complicated.

In contrast, the spectrum analyzer is a single f(x) function to plot.
And in the case of my HP3580A which is a low frequency SA, the sweeping speed is very slow (can be a few minutes).

Indeed, what looks easy for an Arduino, may not have been in the 70s!


Re: Push-push switch repair (need some theory of operation)

Ke-Fong Lin
 

Hi everyone,

So I've disassembled the 2 offending switches. BJT/FET is 6PDT and the x1000 is 4PDT.
I've made some photos: https://groups.io/g/TekScopes/album?id=237923

Indeed the "Repairing_tek_made_board-mounted_push-button_switches.pdf " is for a different kind of (more modern?) switches. However, the 5CT1N's switches are much simpler. In fact, by looking at the pictures, the "theory of operation" is pretty obvious.

I'm pretty confident that I can fix them by careful cleaning and re-assembly. Maybe with some extra tweaking of the metal "bends" for more "tension" and hence better contact.
I've ordered and just received isopropyl alcool and an MG chemical "super contact cleaner" pen.

The extra soldered cables on the photo is for power supply to the plug-in. I don't have an extender but needed to probe around with an oscilloscope.

Best regards,


Re: GPIB videos wanted?

Mlynch001
 

I have been playing around with the GPIB on my TDS series scopes. I have a PROLOGIX GPIB to USB interface and use Realterm for communication. So far, I have used the TEKTRONIX programmers manual and followed those basic instructions. Just brushing the surface of GPIB, but wanting to learn more as time goes by.

--
Michael Lynch
Dardanelle, AR


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

garp66
 

hi Ed,

Referring to your response : 12:19pm #162991

I agree with you, 1st step would be
to get at the 576 CRT display XYZ info & possibly a sync pulse out to the back panel to go to a different X-Y scope, monitor or a CRT replacement. The CRT & HV xfmr deaths are the primary headaches !

If you have further details or insight, please continue to bring it up !

thank you,
rick


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

Eric
 

This was along the lines I was thinking, having the video out VGA would let you use any older computer monitor. It would also let us view on a 19 inch or larger screen. The “zoom” in the larger screen would jump up the measurement ability of the unit from just having the larger display. I do like the ability to digitize and store 2 waveforms so that you could overlay the 2 for transistor matching. Replacing the tube seems to me will correct the 2 main terminal issues of the 576 The HV crapping out and the tubes being week.

On 1/8/2020 3:19 PM, Ed Breya via Groups.Io wrote:
Regarding converting the old 570 series CTs for "modern" display, readout, and control, there are degrees of complexity involved, ranging from fairly simple, like providing XYZ output in a unit with CRT or HV beyond hope, to very elaborate changes to full-blown, digital display and programmable functions. I'd recommend against getting too carried away with fancy modernization of these old beauties - it's doable, with a lot of effort, but not very practical. The most common, serious issue is the need to restore operation and usability to a unit with failed display. This can be done with internal modifications to provide XYZ access, to at least put the display on an external XY scope or monitor, or possibly digitize the signals and put in an LCD display, fitted to replace the CRT face inside. Providing readout of the myriad function settings adds quite a bit of complexity. Adding remote control borders on ridiculous. Again, all of this is doable, but not advisable. If you want all that, it's best to start from scratch and build or acquire appropriate SMUs and make a fully programmable system. The ATE industry figured this all out a long time ago.

I can speak readily to the 576 issues since I'm familiar with it. The biggest issue is CRT display failure, due to HV problems, or CRT wearout. As we can see from recent discussions, HV problems can be addressed with rewound transformers or even new designs. CRTs going bad have some chance of extra usable life by rejuvenation, but ultimately they will deteriorate beyond hope. There is no direct replacement for the 576 CRT except for the existing ones that may occasionally show up NOS, or salvaged from junkers. Once your CRT goes, that's it, and finding a replacement becomes ever more difficult over time.

What would be nice is the availability of a modern, LCD replacement, like those made for certain HP gear like the 8566/85660, which suffers the same sort of problem with CRT life. In this case, the XYZ info is readily accessible, and it's already digital, so interfacing is fairly straightforward. This is not the case with the 576.

I think the most essential, and simple "upgrade" for the 576 is to provide the same display info on a fitted LCD, as on the original CRT. The trick would be in digitizing and processing the XYZ info, and presenting VGA or whatever form is appropriate for the LCD. Getting the XYZ and other info from the 576 is fairly straightforward, but of course needs some internal modifications.

So, for anyone contemplating designing and selling a 576 "upgrade," the most important thing is to replicate the original XY display in real time, as would be shown on the CRT, while ideally also storing enough info to eliminate the flicker that results from multiple stepped-curve displays. Properly handling multiple curves may be the trickiest part. Other bells and whistles can be useful, but are secondary. Assuming the basic display can be made, the next most useful function I can picture, offhand, would be to replace the graticule intensity pot (no longer needed) with a button or rotary switch to activate digital storage features - imagine storing a curve of one device, compared to another that's live (or also stored), and so on.

I looked a bit at the 576 circuits, and it appears easy to get the XYZ info, especially if the CRT is eliminated. Unfortunately, many of the control settings are implemented within the X and Y deflection amplifiers, so there are no single-ended, ground-referenced tapping points available that can capture all the required info. The most straightforward access to all the original info is to take the deflection amplifier outputs, and convert them from differential to single-ended, and eliminate the common-mode signal originally set up for the CRT deflection plates. This is all easy to do, considering the low bandwidth requirement, and that the signals are quite strong to begin with. I didn't look at the Z-axis yet, but presume it's even easier to pick off. Synchronization and setting signals for digitizing, like from line and the step generator would likely be required, and should be readily accessible. I can elaborate on all this, if anyone is interested.

BTW my main 576's display recently crapped out. I haven't looked into it much yet, but it appears to be a HV failure, even though it is of the later type with the "good" transformer. Hopefully, it will be something minor.

Ed


Re: 577 Curve Tracer question

David Berlind
 

Hi Dennis.

I have my 577 and my Eico 667 Tube tester ready for the marriage according
to your paper. It's in my project queue and I'm looking forward to the
undertaking. First, I'm going to bring the tube tester up to snuff. Do you
have a shortlist of things to do in order to get the storage into a good
place on the 577 (rather than reading the confusing manual)?

Thanks.

David

On Wed, Jan 8, 2020 at 1:18 PM Dennis Tillman W7PF <@Dennis_Tillman_W7pF>
wrote:

Hi Jason,
Congratulations. You've got yourself a great curve tracer. If you ever get
the need to test vacuum tubes I designed an inexpensive adapter that will
let you do that on a 577. I wrote a paper on how I did it.

In my experience the only thing you will probably need to adjust is the
storage. My impression of all Tek storage tubes is that they are
particularly sensitive to their voltages being properly set. There is a
procedure in the Service Manual for doing this. I found it so confusing
(many of the adjustments seem to interact with each other) that I
eventually gave up and just settled for something that I could live with.
After a few minutes the scope has warmed up it the adjustments I made seem
to make the storage work OK. I'm sure that someone else could do a better
job than I did.

What part of the circuit are U520 and U530 used for? It seems odd that
they were missing. Do you have any idea why?

Dennis Tillman W7PF

-----Original Message-----
From: TekScopes@groups.io [mailto:TekScopes@groups.io] On Behalf Of Jason
A. via Groups.Io
Sent: Wednesday, January 08, 2020 8:24 AM
To: TekScopes@groups.io
Subject: [TekScopes] 577 Curve Tracer question

Hello all! I've recently acquired a 577D1 curve tracer that was missing a
few components. I managed to find everything online that it was missing
(at least at first blush - time will tell on my observation skills and
glasses prescription). Replacement parts 156-0200-00 for U520 and U530
arrived yesterday. Before I plug in two somewhat hard to find parts, is
there anything I should be checking first?




--
Dennis Tillman W7PF
TekScopes Moderator




Re: 577 Curve Tracer question

Jason A.
 

Hi Dennis,

I am a very happy user of your adapter on a 7CT1N! :-) I got the 577 for $125+shipping (still can't believe it, but there may be a reason...) mainly for the additional voltage and current ranges the 577 offers.

The U520 and U530 are on the main board for the 577 as Op Amps showing in the -Horizontal circuit, it appears as drivers to the display board. Their rails are listed to be +/-12v.

I also have a K436, Q1392 and Q310 on the way. Those were the components that were obviously missing. I'm hoping this doesn't turn into a months long repair ordeal, but who knows. It should be worth it in the end if my skills match my ambition.

Thank you and best regards,

Jason


Re: Tek 576 Curve Tracer USB port to a cpu ... ?

Ed Breya
 

Regarding converting the old 570 series CTs for "modern" display, readout, and control, there are degrees of complexity involved, ranging from fairly simple, like providing XYZ output in a unit with CRT or HV beyond hope, to very elaborate changes to full-blown, digital display and programmable functions. I'd recommend against getting too carried away with fancy modernization of these old beauties - it's doable, with a lot of effort, but not very practical. The most common, serious issue is the need to restore operation and usability to a unit with failed display. This can be done with internal modifications to provide XYZ access, to at least put the display on an external XY scope or monitor, or possibly digitize the signals and put in an LCD display, fitted to replace the CRT face inside. Providing readout of the myriad function settings adds quite a bit of complexity. Adding remote control borders on ridiculous. Again, all of this is doable, but not advisable. If you want all that, it's best to start from scratch and build or acquire appropriate SMUs and make a fully programmable system. The ATE industry figured this all out a long time ago.

I can speak readily to the 576 issues since I'm familiar with it. The biggest issue is CRT display failure, due to HV problems, or CRT wearout. As we can see from recent discussions, HV problems can be addressed with rewound transformers or even new designs. CRTs going bad have some chance of extra usable life by rejuvenation, but ultimately they will deteriorate beyond hope. There is no direct replacement for the 576 CRT except for the existing ones that may occasionally show up NOS, or salvaged from junkers. Once your CRT goes, that's it, and finding a replacement becomes ever more difficult over time.

What would be nice is the availability of a modern, LCD replacement, like those made for certain HP gear like the 8566/85660, which suffers the same sort of problem with CRT life. In this case, the XYZ info is readily accessible, and it's already digital, so interfacing is fairly straightforward. This is not the case with the 576.

I think the most essential, and simple "upgrade" for the 576 is to provide the same display info on a fitted LCD, as on the original CRT. The trick would be in digitizing and processing the XYZ info, and presenting VGA or whatever form is appropriate for the LCD. Getting the XYZ and other info from the 576 is fairly straightforward, but of course needs some internal modifications.

So, for anyone contemplating designing and selling a 576 "upgrade," the most important thing is to replicate the original XY display in real time, as would be shown on the CRT, while ideally also storing enough info to eliminate the flicker that results from multiple stepped-curve displays. Properly handling multiple curves may be the trickiest part. Other bells and whistles can be useful, but are secondary. Assuming the basic display can be made, the next most useful function I can picture, offhand, would be to replace the graticule intensity pot (no longer needed) with a button or rotary switch to activate digital storage features - imagine storing a curve of one device, compared to another that's live (or also stored), and so on.

I looked a bit at the 576 circuits, and it appears easy to get the XYZ info, especially if the CRT is eliminated. Unfortunately, many of the control settings are implemented within the X and Y deflection amplifiers, so there are no single-ended, ground-referenced tapping points available that can capture all the required info. The most straightforward access to all the original info is to take the deflection amplifier outputs, and convert them from differential to single-ended, and eliminate the common-mode signal originally set up for the CRT deflection plates. This is all easy to do, considering the low bandwidth requirement, and that the signals are quite strong to begin with. I didn't look at the Z-axis yet, but presume it's even easier to pick off. Synchronization and setting signals for digitizing, like from line and the step generator would likely be required, and should be readily accessible. I can elaborate on all this, if anyone is interested.

BTW my main 576's display recently crapped out. I haven't looked into it much yet, but it appears to be a HV failure, even though it is of the later type with the "good" transformer. Hopefully, it will be something minor.

Ed