开云体育

Just correcting my misreading of the schematic for the RST7.5 input. The _input_ to U440 is 50kHz so the RST7.5 frequency should be 500Hz as measured by Tony.

Roger

The +12V and +5V values you report below seem pretty OK to me.

You didn't say anything about the -5V line ...

John,

I was initially surprised that Reset out was low but that is correct since 8080 peripheral devices have an active high reset pin!

A couple of tests for next time you are looking at the scope:
Check that RST7.5 is 50kHz not 500kHz. 500kHz would indicate that U440 is dead and would probably saturate the processor with interrupts.
Check that pin 14 of U472 is also high so that we are not looking at a failed U472
Measure S0 and S1 (pins 29 and 33), that might show that the CPU had executed a HALT instruction and had at least done something after Reset.

The GPIB interrupt stuck high does need tracking down.

Regards,

Roger

Hi group

I have 2 TEK 468 with problems in digital circuits.
First I would like to inform the group that I never worked with the 8085 microprocessor.
In the page 5.36 of TEK 468 Service Manual, to initiate Trap test we must to momentary unplug P262 and reconnect it . Microprocessor will be interrupted and the Trap Test routine will begin.

My question is, how to know if the routine was performed correctly? I have a 16 channels logic analyzer
and I can record the signature for the whole routine. However, I don't know how to check if it
was performed correctly. Could you ot somebody tell me how to check?

The same applies to the Kernel Test. How to check if the routine was fully complied with by the microprocessor?

I would like to have a step-by-step of the trap and Kernel routines to check if the routine was done correctly.
The same applies to the Kernel Test. How to check if the routine was fully complied with by the microprocessor?

Many thanks in advance

Heitor Lima ( From Rio de Janeiro - Brazil )

Thought I would pull out my 468 and see if I could find anything useful to post. Well no trace on power up :( However it passes self test. Had to replace the 1200uf cap in the power supply to bring it back to life. First repair on this scope since new. Will probably replace all the big PS filter caps in the future in this and my 465B scopes, since I bought them from Tek in the 80s and they use the same parts. Cheap insurance, maybe... The 465B has never been touched. Guess I should go fire that one up too. They have been sitting for a decade.

Self test runs with 8's and dots first then a bunch of zeros flashing around on the display, then it goes blank when the test passes.

I did a quick once around the processor to provide some general signals for a running unit.
1 5Mhz Squarewave
2 5Mhz Squarewave, half size from 5V down to ~2.5V
3 Low, short high pulse on power up
4, 5, 6 Low
7 500Hz digital signal with 2ms positive pulse
8, 9 High
10 100Hz digital signal with 10ms positive pulse
11 High, occasional negative pulse
12 to 19,k and 21 to 33 Address/data/control constant traffic
20 Gnd
34, 35, 36 High
37 132Hz squarewave
38, 39 High
40 +5V VCC

Regards
Tony

Sorry I'm too tired to reply to anything individually tonight...long day at work (things are busy when you aren't normally in the office due to covid, and then suddenly you are in the office and you have everything to do), but noted on the proposed directions to take troubleshooting...thanks!!!

I'm crack it open this weekend and take a closer look.

Sean

I would clean the capacitor and the 5M resistor with IPA
before I condemned it... but you are right about the list
of usual suspects.

-Chuck Harris

Dave Wise wrote:

About to test where Dave said but are the barely reachable points on the power supply for the digital section alright? I could snake some tools down and measure from a 5 volt and 12 volt point. 12 volt was 12.02 volts and had a small amount of high frequency noise but was mostly negligible. The 5 volt was 5.08 and had about 30ish mV of relatively high frequency ripple. Pictured here: /g/TekScopes/photo/252782/0?p=Created,,,20,2,0,0

Here are the points I mentioned:
5V: /g/TekScopes/photo/252782/1?p=Created,,,20,2,0,0
12V: /g/TekScopes/photo/252782/2?p=Created,,,20,2,0,0
Sorry for the picture quality on those I couldn't get a conventional camera to see in.

I agree with Dave. The cap, or possibly some contamination across its terminals, is the highest-probability cause of the problem.

The timing resistor has more or less a constant voltage across it, as well as a constant current through it (the circuit is the classic Miller integrator that Tek favored until the 7B92's limitations stimulated an overdue change in tradition), so it's hard to imagine what pathologies it could possess to cause the observed behavior. If it does ultimately turn out to be the resistor, I'd be very interested in some curve tracings.

Good luck with your debug!

--Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070

Hi John, there is a zip document on Bluefeather site but someone posted maybe yesterday that it was down, with all the info and bin files to programme.

I also did a compilation of all this with a conversion socket and it was on Bluefeather and maybe on Ko4bb site. I can send you both if need be. On this site in the files section there should be a 12KB file 468.zip and is just the bin files. A more complete file with pinouts and data sheets is Tek468.zip 1890KB, not sure where it is but probably Ko4bb anyway.

Below I will post some of the discussions on Tekscope that I had copied to text files and now reside in my Tek 468 folder. Losts of words of wisdom (also check our archive of previous posts) from Tekscopes member Reed Dickinson our 468 guru.

Dave

-----------------------------------------------------------------------------------------------

July 2017

468 Repair Tips - www.ko4bb.com
From the TekScopes mailing list:

I just finished a 468 that had a bad rom.

Just for the record, there is an error in the troubleshooting section relating to the testing of just the analog section of the scope. If the processor fails, you will have a hung up trace off screen on the left side of the scope.

The troubleshooting section states to disconnect the p3008 connector on the A12 trigger board, pull the p483 connector cable and jumper the pins. Then, on the A14 vertical mode switch board move the p307 connector to the p206 test mode port.

This places the scope in analog mode and displays ch1 & 2 chopped.

The only problem is that when you disconnect the p3008 connector, you kill the +5vs source that is used on A14 - p206. The +5vs comes from the power supply on the timing board and powers all the digital logic.

The work around is to use a jumper to the +5 on the interface A15 board (test point) to pin 3, 4, or 5 on the p307 cable to force it to 5 volts, then it will work. This applies +5 to the ch 1 select, ch2 select, and /xy select, and a ground to the /chop select.

I can't tell you how much time it took to find that out. Hope this helps someone in the future.

The bad rom was one of the gold top chips.

Regards, Tom
----------------------------------------------------------------------------------------------------------------------------

Subject: [TekScopes] Appraisal Tek 468
Date: Sat, 7 Apr 2007

Hi Gang:

I have found that the weakest links of a 468 are the ROM's, the 8 segment switch on the top board and the microprocessor power supplies on the rear assembly. I was able to design a circuit to read the ROM's and can now program a 27C64 with a 28 to 24 pin adapter to put these great oscilloscopes back on the air. I also wrote and programmed a ROM that reads the value of the 8 section switch and displays the value in octal on the display, most helpful for locating a problem with the data bus or a stuck level on a tri-state device. The Volume I manual has tables of check sums that will help you diagnose ills and these are indeed useful albeit a bit tiring to work with. I have also read and programmed a 27C64 for the 067-0989-XX test ROM. It uses the same 28 to 24 pin adapter as the other ROM's. If anyone needs an operational ROM or the test ROM drop me a note off line and we can discuss your problem. If you find that your 468 does not display the marching numbers on the 7 segment display during boot-up then check the 8 segment switch on the upper board. It appears to be affected by the flux used in the wave soldering operation and frequently needs replacement. I usually install a 16 pin socket in that position to accept the switch. The microprocessor power supply is a royal PITA to work on. I can offer no quick suggestions here, just use normal troubleshooting techniques. When reassembling do NOT put long screws in the two center holes of the rear bezel as they will crack the PC board in the rear and then you really will have a problem. If that happens you can send the board to me and I can repair it as I made a point-to-point wiring list. Overall the 468 is a truly magnificent instrument. It encompasses all the very best features of the 465B along with the ability to instantly change from an analog to a digital instrument. The 10 MHz digital bandpass does leave something to be desired but is satisfactory for many applications.

Reed Dickinson
reed714@...

-----------------------------------------------------------------------------------------
Read Me text from Tek468.zip

The Tektronix Scope, For The Most Part Is A Reliable Instrument. Its Main Weakness Is The Two Firmware ROMS (U565, And U575) Tek # 160-0459-01 And # 160-0759-01. These ICs , Of Course Are Custom Devices Made By Mostek In The Late 70s-Early 80s. As They Are No Longer Aavailable (It Seems From Anywhere), I Have Found The Latest Firmware Version (So I Am Told) . The Files Listed (2764.bmp, And 2764a.bmp) Are The Pinouts Of The Common 2764 EPROM From Intel, Which I Used For My Replacement ROMs. The File (MK36XXX) Is The Pinout For The Obsolete Tek Roms (U565, U575). The File (Fig. 1.bmp) Is The Hand Drawing Of The Adaptor To Convert A Programmed 2764 To The Mostek 36xxx Pinout. The File (OMK36693) Is The "Bin" Of The New "MK36693", Which Is "Burned" To A 2764. The Same For File (OMK36694). Most EPROM Programmers Should Do The Job. Good Luck On Your Textronix 468 Restoration!!!!

I agree.
And it's unlikely for a polycarbonate (C160G) or ceramic (C160M) to go leaky, but it does happen now and then. Unless R160F has become nonlinear, the caps - and the switch insulation - are pretty much the only suspects.

Dave Wise
________________________________________
From: [email protected] <[email protected]> on behalf of Albert Otten via groups.io <aodiversen@...>
Sent: Thursday, September 17, 2020 9:50 AM
To: [email protected]
Subject: Re: [TekScopes] 585A -- time base A linearity

Hmm. There's no single component that's used only on that setting. 0.5ms/div, 0.2, and 0.1 all use C160G/M,
while all 5's from 50us/div to 0.5s/div use R160F, all 2's use R160E, and all 1's use R160D.
See the schematic for "TIME-BASE A TIMING SWITCH" in the manual, available at w140.com .

That's correct Dave, but when C160G/M is leaky then this will show the most effect on non-linearity when R160F with the highest value 5M is switched in, so at 0.5 cm/div.
Albert

Status of control pins on 8085:

READY H (5V)
SID L
HOLD L
RST6.5 L
RST5.5 H (4V)
RST7.5 500kHz pulse as indicated in manual
INTR H (4V)
RESET_OUT L (1V)
RST_IN H

Hmm. There's no single component that's used only on that setting. 0.5ms/div, 0.2, and 0.1 all use C160G/M,
while all 5's from 50us/div to 0.5s/div use R160F, all 2's use R160E, and all 1's use R160D.
See the schematic for "TIME-BASE A TIMING SWITCH" in the manual, available at w140.com .

That's correct Dave, but when C160G/M is leaky then this will show the most effect on non-linearity when R160F with the highest value 5M is switched in, so at 0.5 cm/div.
Albert

Thank you. I have downloaded some information on the 8085 to review, but in the meantime you have given me some things to look at. I should perhaps have been more specific when I said signals are absent. In fact, ALE is stuck HIGH but definitely not pulsing.

I did find a marking on the circuit diagram that says "LO REMOVES INTERRUPTS" on pin 13 of U255 (74LS74). Pin 9 of same is connected to INTR pin on the 8085.Grounding pin 13 of U255 brings INTR LOW and ALE LOW, but still nothing happens. I performed a reset with pin 13 held LOW but it made no difference. I am not sure how one can tell whether data is being read, but I was intending to research that aspect.

RESET_IN is high unless I move the jumper at which point it goes low while grounded. I will check the status of READY, SID, HOLD, TRAP and the various RST pins to see whether that reveals anything. I see RST 5.5, 6.5 and 7.5 on the diagram.

I will do some further investigations and report back.

John,

I don't have any direct experience of the Tek 468 or the 8085 processor that it uses but I am fairly familiar with some of the other 8 bit micros, so this is gleaned from the 8085 manual:

If there is no ALE signal (it should pulse once for each byte to/from memory) then the processor is not stuck in a loop. There are three possibilities 1) it has executed a HALT instruction, in which case pins 29 and 33 on the processor will be at logic zero; 2) one of the interrupt sources is permanently active. The 8085 is the opposite of what I am used to, RST5.5, RST6.5, RST7.5, INTR and TRAP are all active high. If any of these inputs are stuck at logic high they will prevent the program continuing (without reading the 8085 manual more carefully I am not sure if this condition can stop the CPU trying to fetch the next instruction, in which case ALE should be switching) 3) the processor is stuck in Reset, in which case pin 36 will be stuck low, eg due to C174 being short circuit or the Run/Reset jumper being in the wrong position (there is also a jumper on TRAP that needs to be in place).

If there are any 8085 experts around then please correct any errors in the above.

Tarnishing of pins could well be the cause of errors and also if you search this site you will find references to a particular type of IC socket (I think made by TI) that is generally a poor design. Do you have any Deoxit?

Regards,

Roger

Hmm. There's no single component that's used only on that setting. 0.5ms/div, 0.2, and 0.1 all use C160G/M, while all 5's from 50us/div to 0.5s/div use R160F, all 2's use R160E, and all 1's use R160D. See the schematic for "TIME-BASE A TIMING SWITCH" in the manual, available at w140.com .

What pot are you referring to? R348 SWP CAL? Note that series resistor R349 was changed at S/N 9300.

Dave Wise
________________________________________
From: [email protected] <[email protected]> on behalf of Tom Lee via groups.io <tomlee@...>
Sent: Thursday, September 17, 2020 12:00 AM
To: [email protected]
Subject: Re: [TekScopes] 585A -- time base A linearity

That symptom could be caused by a leaky timing cap for that sweep
setting. If it starts off at about the correct speed on the left, but
slows down as it moves to the right (e.g., square waves get
progressively squished together), that would be a further indictment of
the cap.

Try subbing it with a replacement cap and see if the problem goes away.

--Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070


On 9/16/2020 23:52, [email protected] wrote:

Hi all,

Some of you may recall the 585A I got earlier this year that had a short. I was able to get it up and running. Ended up replacing a pair of totally short diodes in the power supply with new 1N4007s which cleared the "blow fuse violently when you flip it on" problem. :o) I also ended up having to replace a trimmer capacitor that was totally broken and spinning free, which was really borking the fastest time/div settings. Luckily I found a seller on ebay with OEM original replacements for it, sold in fours.

Anyway, there is one outstanding problem that is left, and that is time base A does not quite meet up to the linearity specifications in the manual, mostly at 0.5 msec/cm. One step I took which helped a lot was to replace V161 in the Miller runup circuit with a NOS 6CL6 I was able to get. This actually brought most of the time/cm settings into specification. However, the 0.5 msec/cm setting is the outlier here. The settings on either side of it are in spec, but this setting has a very apparent linearity problem; the trace is getting shorter the further right it goes. You can really see this with a square wave input. The pot that is supposed to adjust this out is all the way at the extreme already.

I figure a good first try would be to clean the time base A control, as it is rather hard to get at. Beyond that, does anyone have an ideas about why this might be the case? I realize that these are really 3% at best instruments and will not be perfect. However, this particular thing shouldn't be this bad I think.


Thanks!

Sean

BTW, I should also mention that all the pins on the two ICs that I removed were very tarnished. I tried cleaning (very carefully since they seemed rather fragile) with both IPA and Servisol switch cleaner but it made very little difference. I would have added this to the previous post but can't see a way of editing it?

Roger, that's a kind offer which I may take you up on. I will e-mail to ask particulars. However, first, here are the results of the test procedure on page 330:

All voltages are within normal parameters.

Tests 1 - 6 were successful. All timing pulses were present and PRDONE was present on U472-16 (140).

Test 7 thru 11 had no signal present, so no signal on ALE, RD or WR or wavefore on TP477 or U255-5.

I'm not sure what 12 and 13 are about and need to look them up, but U255-9 is "stuck" high.

Since the above tests were not completed satisfactorily, I didn't do on to the Trap Test Procedure or the ROM Checksum. I have yet to go through the process on P228. but I don't understand what Service/Options Switch X10000000 means? I am presuming this refers to the row of DIP switches on the top board but am I right in thinking its simply the configuration of the individual switches from, left to right? So first switch ON and all the remaining ones OFF?

I have noted your point about pin 27 on the 28C64 EEPROM.

On Wed, Sep 16, 2020 at 02:02 PM, Ray wrote:

Roger,Thanks for clarifying that.I was getting worried there for a moment

Weird. Yeah, if I go online I have to click on "view quoted text" to see what you wrote...
For some reason, your emails to the list appear as quoted text.

John,

Since you are in the UK, I can offer to program a couple of 28C64s for you, please e-mail me (rdot gdot evans at talk21 dotcom). There are ','s between my initials. If you would find having a device programmer is useful then please don't let me stop you buying one. Beware the 28C64s are EEPROMs and really need the WEn pin 27 tied to VCC to prevent inadvertant writes in circuit.

Regards,

Roger