Author Topic: Surface mount? In my Tek 7000 series scope?!  (Read 7746 times)

0 Members and 1 Guest are viewing this topic.

Offline TerraHertzTopic starter

  • Super Contributor
  • ***
  • Posts: 3958
  • Country: au
  • Why shouldn't we question everything?
    • It's not really a Blog
Surface mount? In my Tek 7000 series scope?!
« on: April 13, 2016, 03:05:30 pm »
There was something I'm supposed to be doing since yesterday, daunting enough to justify serious procrastination. So instead I started fixing a Tek rackmount R7633 scope I've had shelved since 2004, when it arrived from the USA badly damaged due to ridiculously inadequate packing. Pretty annoying; it was working and in great condition when the guy put it loosely in a thin cardboard box, sprinkled in a few styro peanuts, and consigned it to the courier monkeys.

Today I stripped it down to the bare frame, and most of the mangled metalwork is now fixed. Couple of little bits still to do, that need TIG welding and machining.  I'll keep this thread updated with progress, but started it just for this:

A very unusual thing turned up when disassembling this scope. A PCB with lots of surface mount parts. In a Tek 7000 series scope. It's not some latter hack either, it's an original Tek module.

Tek Wiki http://w140.com/tekwiki/wiki/7633 says:
Quote
After the 7633 and 7934 were discontinued in 1990, the 7623B (which added the reduced-scan mode to the 7623 and therefore had essentially the same specs and features as the 7633) was the last analog storage scope model in the 7000 series, available until the entire 7000 family was discontinued in 1992.

The SMD board is the display readout card, PN 389-0559-01, marked (c)1989. It has an Intel 80C31 CPU and a 27C256-20 EPROM. Some IC date codes 8914 & 8916.

My paper 7633/R7633 manual includes the first Readout board, plus a latter one: SN B240054-up. But not this one.
The manual is "Revised FEB 1989".  Apparently just too soon to include the _third_ type of display readout system in my R7633. Which was discontinued the next year.

So my broken old R7633, sitting on a shelf for over a decade, is quite a rare beast. Hope I can get it working.
Finding a manual with this readout board will probably be difficult. The PDF manuals I find online are all the same early version:
  http://w140.com/tektronix_7633_r7633.pdf  120.6MB 
    This is an early version. It only has the early Readout board, that is "SN B240053-below" in my manual.
    Schematics have blue detailing. This is the original fairly hires scan.

  http://exodus.poly.edu/~kurt/manuals/manuals/Tektronix/TEK%207633%20%20Service.pdf   41.7MB
    The exact same scan as the w140 one above, except more compression.

  http://bama.edebris.com/download/tek/7633/tek-7633.pdf  41.7MB
    Same file as exodus.

Does anyone have a late 1989 or 1990 manual for the 7633/R7633 ?  My Tek 7633 scope is SN: H705286,  made in Holland.

Edit to add: Also I attached the EPROM image. U2105.bin (in a ZIP)
Checksum reads as 88D1, which is what the label says.

Oh, and anther mystery. See pic 20160413_3338.jpg. Someone engraved "SN: 00001" in the metalwork, inside one of the slots.  Why? Could it have been the first one with the new new readout board? But you'd think if Tek did that, they'd have done a neater job?







« Last Edit: April 13, 2016, 03:24:10 pm by TerraHertz »
Collecting old scopes, logic analyzers, and unfinished projects. http://everist.org
 
The following users thanked this post: jancumps

Offline 4thDoctorWhoFan

  • Regular Contributor
  • *
  • Posts: 189
  • Country: us
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #1 on: April 13, 2016, 03:30:54 pm »
I forget the model number, but I once had a very old HP frequency counter using nixie tubes for the display that had a little board inside with about 10 or so surface mount components.  I had this unit years ago that I got directly from a military auction.  I was surprised to find that SMT board in there.
 

Offline wraper

  • Supporter
  • ****
  • Posts: 16796
  • Country: lv
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #2 on: April 13, 2016, 03:52:23 pm »
Vacuum tube-semiconductor TVs made in USSR from 1970s till 1992  :palm: contained small hybrid modules in sip form-factor with smd parts soldered on ceramic substrate. just bizzare mix of technology.

« Last Edit: April 13, 2016, 04:00:49 pm by wraper »
 
The following users thanked this post: jancumps

Offline N2IXK

  • Frequent Contributor
  • **
  • Posts: 722
  • Country: us
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #3 on: April 13, 2016, 04:30:09 pm »
Cool! I had heard about the late EPROM readout board for the 7K series, but never saw one before.  It replaced a board that was jam-packed with Tek custom character generator and ROM chips which are pretty much unobtanium these days.
"My favorite programming language is...SOLDER!"--Robert A. Pease
 

Offline TerraHertzTopic starter

  • Super Contributor
  • ***
  • Posts: 3958
  • Country: au
  • Why shouldn't we question everything?
    • It's not really a Blog
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #4 on: April 14, 2016, 02:37:25 am »
Cool! I had heard about the late EPROM readout board for the 7K series, but never saw one before.  It replaced a board that was jam-packed with Tek custom character generator and ROM chips which are pretty much unobtanium these days.

Ah, good. So it was used in other 7K-series scopes at the end. Which means there may be a chance of finding a manual with details of this board.
If anyone has or knows of such a manual, please let me know. I'd like to buy it. To add to my 'to scan' list.

Something else about this board. See the backlit pic below. The PCB was laid out by someone with no understanding of why ground planes are needed. It's rare to see such a HF-naive PCB.

There's a story about another example I've encountered, that may be relevant.

I was working for AWA wagering as a software contractor. Another project they had running at the time was being done by an outside contractor. It was developing a 68000-based board, to be used in kiosk betting machines called 'Mystery Bet'. The board had been designed and manufactured, but the software development stage bogged down. The programmers kept encountering bugs they couldn't diagnose. I had a logic analyzer with a 68000 emulator pod (and know hardware as well as software), so offered to have a look at the system and find their bugs. Management accepted.

The board looks a lot like this Tek one, except bigger. It was laid out by someone with no conception of HF effects. Double sided, no internal ground and power planes, not even any attempt to maximize ground conduction mesh.
It didn't take long to find the 'bug'. It was actually a generic hardware level failure - any time enough bits in the address and data bus transitioned at once, the ground bounce produced logic failures at multiple points on the board. The PCB was unworkable.

I wrote up a report, showed what was happening, recommended the board be redesigned by someone who knew what they were doing. Offered to do it myself if they wanted. Well within my level of experience from previous work.

But I don't think management were able to comprehend the technical points. They tried to proceed with the existing board, and in fact did (somehow) get the code to mostly work. Sold some of the Mystery Bet machines. They were a costly commercial failure, due to incessant faults and crashes. Just what I'd said would happen. Years later I found one of the machines, stuck down in the basement of an abandoned club building. So I have one of those boards. Which is good only as a conversation piece.

Anyway... I wonder if Tek had similar functional reliability problems with this board? This is a lot smaller and simpler than that other board, with only an 80C31 and so isn't much of a candidate for ground bounce glitches. But that's a really badly designed PCB.

If it was flakey, might that have had any influence on the termination of the 7000 series?
Now I'm eager to see it running. But have other things I have to do this week and next.
Collecting old scopes, logic analyzers, and unfinished projects. http://everist.org
 
The following users thanked this post: krivx

Offline tautech

  • Super Contributor
  • ***
  • Posts: 28141
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #5 on: April 14, 2016, 10:19:00 am »


That's certainly an unusual PCB.

Subscribed
« Last Edit: April 14, 2016, 10:22:36 am by tautech »
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16272
  • Country: za
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #6 on: April 14, 2016, 07:12:14 pm »
With it being mostly a single schip MCU a possible solution would be to solder some screened cable screen ( in a heatshrink tube) to join the power and ground pins of the MCU, transparent latch and the ROM. Add another link for supply and some 100n capacitors as decoupling and then the MCU would have less ground bounce. That board must suffer horribly from noise on the digital signals, not helped by the seeming attempt tu turn the board into a near perfect RF emitter.
 

Offline TerraHertzTopic starter

  • Super Contributor
  • ***
  • Posts: 3958
  • Country: au
  • Why shouldn't we question everything?
    • It's not really a Blog
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #7 on: April 15, 2016, 12:16:27 am »
With it being mostly a single schip MCU a possible solution would be to solder some screened cable screen ( in a heatshrink tube) to join the power and ground pins of the MCU, transparent latch and the ROM. Add another link for supply and some 100n capacitors as decoupling and then the MCU would have less ground bounce. That board must suffer horribly from noise on the digital signals, not helped by the seeming attempt tu turn the board into a near perfect RF emitter.

Yes.
It will be at least a couple of weeks before that scope goes back together, due to other chores first. Once it is reassembled, fingers crossed the tube wasn't damaged by the shipping impacts. Given the nature of damage - a bit like 'progressive crumple' in cars -  I'm hopeful about that. I wasn't game to power up the scope without the disassembly damage inspection first. Oh, and it has a simple linear power supply, for which I'm very grateful.
The most fun part is going to be getting all the cables back as they were. I took LOTS of pictures, and as a last resort there's always RTFM.

Also I hope to find the schematic and operation info on that board, if it exists.
After that, even if it seems to be working OK, I want to do an analysis of the noise margins, to see if the board may have been prone to flakey operation, intermittent crashes, etc. Very interesting detail, since this was happening around the time the 'old school Tektronix' was going down the toilet. Call me paranoid, but I've seen other cases where leading technological companies seem to have been poisoned from within, by infiltrators pushing through fatally bad decisions. This board by itself isn't evidence of that, but I'm keeping the potential in mind.

Finally, the board will get an anti-noise treatment. The more I look at the layout, the more disgusted I am.

SeanB, it doesn't need to be braid. Simply adding a reasonable mesh of kynar wire substituting for the missing ground cross-linking would help. Better match to the tiny pads too. Plus it sure does need some more decoupling caps, once there's something to decouple _to_. At the extreme, an actual copper foil ground plane, insulated from the board with a plastic sheet, and linked to the board at appropriate points; plus some more supply rail meshing.
That power rail 'giant C' around the board periphery but not quite completing the loop, is amazing. Whhhhyyyyy?

I wish it was possible to learn the history of how Tektronix came to make a board like this. It was Tektronix Holland, which probably means the engineer(s?) involved were not the same as had been responsible for previous 7000 series design. One certainly hopes so anyway. And in the USA, maybe those guys were all retired (or let go?) by then? Didn't Tek have _any_ competent engineering staff left at that stage, who should have vetted this board before it went into production?
Collecting old scopes, logic analyzers, and unfinished projects. http://everist.org
 

Offline PA2HK

  • Contributor
  • Posts: 12
  • Country: nl
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #8 on: April 18, 2016, 09:31:44 pm »
Greetings,

Yes, that is the late model Tek7000 readout board. I have the same board in a late model Tek 7904A mainframe. Schematics are known and you can find them on Tekwiki w140.com at the end of this 7904A service manual: http://w140.com/Tektronix_7904A_OCRed_by_Tabalabs.pdf This board has been used in the last incarnations of the R7903, R7844, 7104 and 7904A mainframes build in and after 1990.

73,

« Last Edit: April 18, 2016, 09:35:46 pm by PA2HK »
HamRadio PA2HK // Keep Calm and Carry On.
 

Offline TerraHertzTopic starter

  • Super Contributor
  • ***
  • Posts: 3958
  • Country: au
  • Why shouldn't we question everything?
    • It's not really a Blog
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #9 on: April 21, 2016, 05:40:41 pm »
Yes, that is the late model Tek7000 readout board. I have the same board in a late model Tek 7904A mainframe. Schematics are known and you can find them on Tekwiki w140.com at the end of this 7904A service manual: http://w140.com/Tektronix_7904A_OCRed_by_Tabalabs.pdf This board has been used in the last incarnations of the R7903, R7844, 7104 and 7904A mainframes build in and after 1990.

Excellent! Thanks. Yes, that's it, complete with theory of operation.
http://w140.com/Tektronix_7904A_OCRed_by_Tabalabs.pdf
Info on the Ver 3 readout board starts at PDF page 380. Took me ages to find it, since the Ver 1 & 2 are earlier in the manual, and the Ver 3 part is a kind of addendum, which if it's mentioned in the index, I missed it.
« Last Edit: April 22, 2016, 01:56:57 am by TerraHertz »
Collecting old scopes, logic analyzers, and unfinished projects. http://everist.org
 
The following users thanked this post: chris_leyson

Offline TerraHertzTopic starter

  • Super Contributor
  • ***
  • Posts: 3958
  • Country: au
  • Why shouldn't we question everything?
    • It's not really a Blog
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #10 on: May 16, 2016, 06:02:39 am »
Updating progress: http://everist.org/NobLog/20160429_first_of_the_last.htm

It's back together, and mostly works. Next stage: diagnose the readout board, which isn't working.

Questions:
 * Does anyone have a spare one of the small Tek 7000 module alignment guides? See pic below. This scope was missing one as received, and I don't have any spare Tek 7000 frames I want to scrap.

 * The manual http://w140.com/Tektronix_7904A_OCRed_by_Tabalabs.pdf which PA2HK linked does have a section on this Readout board. But the quality is very poor. It's a horrible 'auto-OCR' mess, with lo-res images that are messed up by the OCR software. The schematic in particular is awful.
So, I'm seeking a paper manual for Tek 7R7903, R7844, 7104 or 7904A mainframes, dated 1990 or later. ie the last year or two before the series was discontinued. It should include the 11 sheets for this PCB, so I can do a nice quality scan.
Collecting old scopes, logic analyzers, and unfinished projects. http://everist.org
 
The following users thanked this post: chris_leyson

Online chris_leyson

  • Super Contributor
  • ***
  • Posts: 1541
  • Country: wales
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #11 on: May 16, 2016, 07:10:43 am »
Thanks for posting the 7904A manual, I will have a read and see if they made any significant changes to the readout. One of the mildly irritating features of the Tek 7000 readout is that the readout board still blanks the display even when the readout is switched off, I wonder if they fixed that on the later revision ?
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #12 on: May 19, 2016, 09:04:46 pm »
Besides the early analog design and the late digital design, there was a hybrid which used the control logic from the analog design but replaced the analog ROMs with digital ROMs and replaced the format, decimal point, and character position generators.

...

One of the mildly irritating features of the Tek 7000 readout is that the readout board still blanks the display even when the readout is switched off, I wonder if they fixed that on the later revision?

That is not suppose to happen.  I have several 7000 mainframes and they all stop multiplexing the display when the readout is blanked via the intensity control.
 

Offline jancumps

  • Supporter
  • ****
  • Posts: 1272
  • Country: be
  • New Low
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #13 on: May 19, 2016, 09:31:56 pm »
Cool! I had heard about the late EPROM readout board for the 7K series, but never saw one before.  It replaced a board that was jam-packed with Tek custom character generator and ROM chips which are pretty much unobtanium these days.



Something else about this board. See the backlit pic below. The PCB was laid out by someone with no understanding of why ground planes are needed. It's rare to see such a HF-naive PCB.

There's a story about another example I've encountered, that may be relevant.

I was working for AWA wagering as a software contractor. Another project they had running at the time was being done by an outside contractor. It was developing a 68000-based board, to be used in kiosk betting machines called 'Mystery Bet'. The board had been designed and manufactured, but the software development stage bogged down. The programmers kept encountering bugs they couldn't diagnose. I had a logic analyzer with a 68000 emulator pod (and know hardware as well as software), so offered to have a look at the system and find their bugs. Management accepted.

The board looks a lot like this Tek one, except bigger. It was laid out by someone with no conception of HF effects. Double sided, no internal ground...

The ground plane approach is a shortcut we all use these days, having the luxury of cheap multi layer boards.
It relieves us (below certain frequencies) to get by  without the realy though high speed practices. When we use the multilayer ground plane design we get away with a lot.

The person who designed your board obviously had to work with constraints (and without the  luxury of having a board with a ground plane). And he made a functional board that lasted for many decenia. Kudos to him.
« Last Edit: May 19, 2016, 10:05:36 pm by jancumps »
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16547
  • Country: us
  • DavidH
Re: Surface mount? In my Tek 7000 series scope?!
« Reply #14 on: May 19, 2016, 09:59:51 pm »
Tektronix had been using multilayer boards with flood fills some time before the last readout board was designed but it may have been the responsibility of their European facilities in Holland.

In any event, the relatively slow edge rates involved do not require ground plane construction if you are careful about ground returns.  I have laid out faster and more complex logic without issues but it is certainly advisable to break out your best differential probes to look for ground and supply bounce if you do this.  The big problem is if whoever is doing the layout does not read the schematic notations showing the separate ground returns for various signals.  The sensitive parts of the oscilloscope mainframe use differential signaling, common mode suppression, and local bulk decoupling so EMI is not as much of a problem as you might expect.

Tektronix also used surface mount parts here and there in earlier designs.  The HYPCON connectors used in the 7904A, 7934, 7104, and later products might be considered to be an early form of surface mount leadless chip carrier.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf