Tektronix 2465B oscilloscope teardown

[Cooler Than I Look]

I am hoping for a more general writeup outlining the equipment I will need and any pitfalls for new players, I don't have any experience working with chips like this.

Neither have I so it was with some trepidation that I took this on.   I unsoldered the Dallas chip with a S-993A vacuum pump (cheap Hakko copy) after practising on some old PC boards.  I just took it slow and released one pin at a time, letting the board cool in between.  From memory there is at least one earth pin that takes some heat to get right through and I left this to last and used my normal iron to release it.     In its place I installed a low profile socket.

I read the NVRAM in a TL866 minipro (made serveral copies!) and burnt that to a 'new' Dallas chip from Ebay dated 2009 but thats much younger than the 1992 chip it replaced!  And now that the NVRAM is socketed its an easy job to replace in the future.

Reassembled the scope, held my breath and turned it on.  Worked and has done since I did this about 18 months ago.

[WastelandTek]

Thank you Cooler, I will get one of those programmers and see what I can do.

Is there any "permanent" solution for this?  or is plugging in "less old" parts the best we can do?

[Cooler Than I Look]

Is there any "permanent" solution for this?  or is plugging in "less old" parts the best we can do?

Probably but its good enough for me.  Now I have a copy of the data I can burn a new NVRAM as needed but they seem to last for decades.  In that time will the calibration data still be valid anyway?

[David Hess]

Thank you Cooler, I will get one of those programmers and see what I can do.

Is there any "permanent" solution for this?  or is plugging in "less old" parts the best we can do?

I replaced the Dallas NVSRAMs in my 2440 with EEPROM backed up SRAMs so there are alternatives.  FRAMs have some special precharge requirements but apparently work.

[WastelandTek]

What part did you use David?

[Cooler Than I Look]

Have a look at http://worldphaco.com/  about half way down "...2465B...  ...repowering the DS1225..." (title too long to type).  Hugo talks about alternatives to the Dallas NVRAM in that pdf article.

Don't get distracted by his amazing restoration work, especially on early televisions. 

OK, do get distracted.  Well worth it.

[David Hess]

What part did you use David?

Well, let's see.  I bought a couple extras and have them right here.

The 2440 used a pair of DS1230AB-120s and I replaced them with Simtek STK16C88-WF45s which being in DIP packages were direct drop-in replacements.  Simtek was bought by Cypress in 2008 which was years before I bought the chips but they were still marked Simtek.

For those who are not aware of this technology, it uses a SRAM which is backed up by floating gate memory.  The entire array is backed up in parallel when power is removed and restored in parallel when power is applied.  Some variations require an external capacitor connected to a dedicated pin to provide backup power but many like the STK16C88-WF45 I used do not.

Last time I checked, the DIP parts were discontinued but surface mount parts were still available and could be adapted with an interposer.  Third parties may have the DIP packaged parts as NOS (new old stock) for sale.

I have also heard reports of parallel FRAM memories working as direct replacements as I mentioned in an earlier post.  This have special timing requirements for their -CS pin because they require a precharge operation before every read but these Tektronix oscilloscopes apparently meet this requirement which is not unusual.

[MarkL]

...
I have also heard reports of parallel FRAM memories working as direct replacements as I mentioned in an earlier post.  This have special timing requirements for their -CS pin because they require a precharge operation before every read but these Tektronix oscilloscopes apparently meet this requirement which is not unusual.

I would not opt for the FRAM solution.  There was a discussion from user Miti having corruption problems starting here-ish:

  https://www.eevblog.com/forum/testgear/tektronix-2465b-oscilloscope-teardown/msg2003207/?topicseen#msg2003207

The short version of the related posts is that the power-down behavior of the 24xx does not meet the FRAM (in Miti's case, FM16W08) datasheet requirements for safe power down.  It seems to work for most people, but from a design perspective it's still risking the data without additional support circuitry to safely control the chip select.

The Simtek/Cypress part, now CY14E256LA, looks like a great suggestion.

[Miti]

My scope works well with FRAM after I installed the pull-up resistor. If you have a copy of the calibration constants, I think it's worth a try. Even though, replacing DS1225 with a new DS1225 would push scope's life beyond its reasonable life expectancy. I don't expect many Tek 2445, 2465 in use in another 20 years.

[WastelandTek]

I want to express my thanks for you guys veering into this discussion for me.  I'm sure it is all in the thread back there somewhere...but 60 pages...

My 2445 still boots fine, for now, but this is clearly something I need to get up to speed on and address sooner than later.

[Hamid_tehran62]

What is it?
DMI 500-2044
It is in the tek tronix power supply
How i can check it?
Anu data sheet?

[bdunham7]

What is it?
DMI 500-2044
It is in the tek tronix power supply
How i can check it?
Anu data sheet?

Is this in a 2465B?  I think it is a common-mode RF choke, but I remember seeing one like it in my 2235, but not my 2465B.  I'll be digging around in the 2465B later on and I'll look.  You can check it for continuity--it has four pins and two isolated coils.  What problem are you trying to solve?

[AMR Labs]

...
I have also heard reports of parallel FRAM memories working as direct replacements as I mentioned in an earlier post.  This have special timing requirements for their -CS pin because they require a precharge operation before every read but these Tektronix oscilloscopes apparently meet this requirement which is not unusual.

I would not opt for the FRAM solution.  There was a discussion from user Miti having corruption problems starting here-ish:

  https://www.eevblog.com/forum/testgear/tektronix-2465b-oscilloscope-teardown/msg2003207/?topicseen#msg2003207

The short version of the related posts is that the power-down behavior of the 24xx does not meet the FRAM (in Miti's case, FM16W08) datasheet requirements for safe power down.  It seems to work for most people, but from a design perspective it's still risking the data without additional support circuitry to safely control the chip select.

The Simtek/Cypress part, now CY14E256LA, looks like a great suggestion.

I am arriving late to this discussion, but wanted to confirm that the FRAM that I put into my 2465B has been working flawlessly for months now and has so far yet to show any data corruption problems as mentioned since it replaced last year the original NVRAM. And I would think there are for sure other documented examples out there successfully using a FRAM and that had no problems either. So just wanted to constructively bring this to the discussion, and point out that maybe there is still some unidentified "X-factor" out there that makes some scopes corrupt the data. But for sure mine is booting up perfectly every time, and with the same configuration before power down which is also being saved. So just saying that maybe don't blame yet the problem squarely only on the FRAM itself being incompatible with the scope.

EDIT: ok so after re-reading the above, as stated by MarkL the X-factor might be that the FRAM requirements for safe power down are not being fully met. But if it works for the majority of the cases it sounds it must be very close. Does anybody know if there is anything that can be done (that is not too complicated) in order to come closer to those power down requirements?

So if the preceding is right, sounds like I'm just one more that got lucky? Back then besides the FM16W08 I also had programmed the fresh NVRAM with the original data, and put it aside in case it is ever needed.

[MavMitchell]

Hi all, I was able to wrangle down a clean high sn#  2465B on our favorite auction site for a reasonable price. Sold as not working(out of calibration) it had minor damage on the A5 from the pesky SMD caps.

It also had a 1990 Dallas chip.

Thanks to all those who have contributed on this forum, I have been able to do a full re-cap and copy and replace the Dallas.
I got caught by the reversed C1132/C1115 issue, which was mentioned in this forum.

Thanks to all...

 

[AMR Labs]

HiMavMitchell,

Welcome to hotel 2465, where you can check in at any time, but you can never -quite- check out.

Seems like this thread must be getting really good as a reference if you managed to successfully get through all the rebuilding process and did not need to ask a single question. By any chance got any pictures of your recap/A5 rebuild? Its always nice and interesting to see other persons work, and it further adds to the knowledge base of this long running thread.

[MavMitchell]

Hi,

Pics as requested, C2113 caused the most damage. I had to remove and replace C2520,C2222 and R2016 for the cleanup.

The track repair was no issue except for the via between C2113 and C2520(completely gone). I was able to wire through the via to the pad on the other side. Its only a pad as its connected to an inner layer. I could have gone to the connector if needed as this net connects to the first pins on J251(-5 or 15v if I remember).

De-soldering the Dallas was no problem, and was able to read it fine. I had also done the Exer 02 screen capture just in case.



 

[AMR Labs]

Good job, indeed much less damage than other cases. BTW on the same A5 board remember to check the capacitor on the exact opposite top corner, which also causes the same kind of corrosion problems.

EDIT: did you take any pictures of the power supply recap?

[MavMitchell]

Yep, C2965 was also replaced. It had minor leakage that just needed cleaning of the board and C2890.

Also found leaking caps on the A3. One had completely lost one leg.

Rifas  were ok but replaced.


Very happy with my result.

[AMR Labs]

Wow that cap missing one leg was definitively not doing any good anymore. Did you notice any sharpness improvement in the traces after the recap? Frequently that is the case. How long did it take you to get everything done, including pulling out the two power supply boards?

That reminds me I still have to do my own power supply recap and replace Rifa caps, luckily I'm on 120VAC so they have not exploded yet, but already show lots of little surface cracks. Since scope is working fine have been procrastinating, but at least already have everything here that is needed for a full PS rebuild. Soon...

[AMR Labs]

My scope works well with FRAM after I installed the pull-up resistor. If you have a copy of the calibration constants, I think it's worth a try. Even though, replacing DS1225 with a new DS1225 would push scope's life beyond its reasonable life expectancy. I don't expect many Tek 2445, 2465 in use in another 20 years.

Miti, if you don't mind, what value pull-up resistor did you use, and at what pin did you connect it to?

[MavMitchell]

Actually I didn't run the scope for long when I received it. Just turned it on to make sure it worked, then shut it down. I didn't want to risk any explosions.

Removing the PSU boards is very easy, just need to remove the screws from the mains connector and move it out of the way to get access to the Voltage Selector wires. Also de-solder the fan power wire(make sure to re-connect during install - just saying).

After separating the boards, the recap took about an hour. I removed all the caps and cleaned the board then replaced with the new using the board overlay from the manual as a guide. This is where I hadn't done enough research about the error on the overlay.

After re-assembly and replacing the PSU, I discovered it was ticking on what I assumed to be over current. While checking the unregulated rails on the A3, I discovered the overlay error which I quickly fixed. Replaced the PSU and all was well.

[AMR Labs]

Actually I didn't run the scope for long when I received it. Just turned it on to make sure it worked, then shut it down. I didn't want to risk any explosions.

That was a good call given you are on 240VAC mains, those old Rifa caps can go anytime at that voltage level. Worst part is that they smell like fish after they burst! 

Removing the PSU boards is very easy, just need to remove the screws from the mains connector and move it out of the way to get access to the Voltage Selector wires. Also de-solder the fan power wire(make sure to re-connect during install - just saying).

After separating the boards, the recap took about an hour. I removed all the caps and cleaned the board then replaced with the new using the board overlay from the manual as a guide. This is where I hadn't done enough research about the error on the overlay.

After re-assembly and replacing the PSU, I discovered it was ticking on what I assumed to be over current. While checking the unregulated rails on the A3, I discovered the overlay error which I quickly fixed. Replaced the PSU and all was well.

Best way to avoid that board overlay error is to pull one cap at a time and replace before moving on to the next one, and noting the polarity of the cap itself before pulling it off the board. That is what I intend to do.

[Miti]

Miti, if you don't mind, what value pull-up resistor did you use, and at what pin did you connect it to?

1K nWE to VCC, pins 27, 28.

[AMR Labs]

OK got it. WE to VCC.
Thanks.

My FRAM so far has never had any problems, but I guess adding this resistor won't hurt anything. Will do it once I get to the LVPS recap.

[Alfons]

Super Thread, has always really helped.

I purchased a defective Tektronix 2457B with Option 9 and 10. Error was, that the device turned on and off again and again. Even the smell showed that electrolyte must have leaked out. The device stank like a trash can.

After I opened the device, I discovered the leaked SMD electrolytic capacitors on the A5 controller board. The traces were not attacked and also neighboring components were still OK. After I had cleaned everything and used new Elkos, the device worked again. The error did not occur anymore.

Then my mishap: I soldered out the NVRAM chip to read it out. From now on I made all the mistakes you can only make.  When I put the chip in Programmer, which I only had a few hours, I just wanted to test the chip and erase all the data.  All data gone.
OK, I had the video of Exerciser 02 for safety's sake. The video quality was so bad that it's a miracle that I was still able to write all the data to the hex editor. But the device did not start now, Trigger'D flickered happily, the device complained about corrupt calibration values. Ok, I stayed calm, did not smash the device, I knew I could find the mistake.

I read everything there was to read, read the service manual, nothing. Then I came up with the idea of ​​looking at the solder joints of the NVRAM under the microscope and discovered that I had severed a trace directly at the solder pad while soldering out.  That's never happened to me. I am getting old. 

After I fixed that, everything work again. A great device, I've never seen such a good Oscillscop before.

A problem that was already there before: The display, but only the beam, flickers slightly and slow, not always and only in conjunction with my digital function generator. Not for example with my old analog pulse generator. However, the signal of the function generator does not flicker with the DSO and also not with the old Tek 475. The readout I had switched off, even with Cal 08 set the display. All voltage values ​​on the board are OK and look clean, ref values ​​of the DAC are -1.248 and 1.363.