Tektronix 2465B oscilloscope teardown

[e-pirate]

Well, probably the STK12C68 is a way to go? As mentioned in document above, you'll need a special PCB and a couple of spares to make it work, but I don't think it's a real problem. Having this chip instead of DS1225 you will have 10⁶ cycles. According to it's way of operation, I'm pretty sure it's equivalent to power on count of the scope.

[e-pirate]

Hi, folks.
During my conversation with mr. Hugo he told me that it's a good idea to have another fresh lithium battery in parallel with the old flattering one in old DS1225 desoldered from dcopes before they go to the reader. The trouble is that when the battery is almost flat, it will still hold data in SRAM and the whole thing will continue to work, but when DS1225 get into the reader, some of them MAY draw small current from its pins and that may lead to voltage drop and accidentally blank DS1225.
All information concerning paralleling the battery is in document.
As I'm currently facing the same problem, I will desolder DS1225 and will try to parallel the older battery with a fresh one before plugging it to a reader. I'm going to order the Wellon VP-390 programmer and try to read it.

[iloveelectronics]

Just to give a quick update. I reinstalled the missing caps on the A5 boards, replaced the film caps on the power supply board and finally also desoldered the DS1225 and backed up the data on it. I had to get myself a Hakko 808 to desolder the chip as my other cheaper Chinese desoldering gun wasn't working very well. Made a few burn marks on the board Turned out it was merely clogged but its tip never tins properly so it's time to replace it anyway.

A new DS1225AD-200IND+ just arrived today (much sooner than I expected) so as soon as I get a bit of time I will copy the data and insert it into the newly soldered socket on the A5.

[BravoV]

Hey Franky, please take few photos if you had a chance, also the cal data zipped and attach it here as well, might be useful someday for someone, who knows.

[iloveelectronics]

Hey Franky, please take few photos if you had a chance, also the cal data zipped and attach it here as well, might be useful someday for someone, who knows.

I will take some pictures the next time I open up the case.

I just had a go with the new DS1225 but failed I'm using the MiniPRO TL866CS programmer. I thought being a brand new chip it would be all empty with all "00" or "FF" but when I read it there was already something in it, is that normal? I tried programming it a few times with the binary file I saved from the old chip but it simply didn't work...

[BravoV]

Thats weird, my experiences with DS1225 either the old one and new were without any problem at all using the minipro I purchased from you. 

Heck, I even still use the old one, while the new DS1225 which already has the cal data copied into it is in the storage, just curious how long the old one will survive. 

Where did you buy that DS1225 ? Read few posts back, there was a similar problem with new "old" stock of DS1225 bought from a surplus shop if I'm not mistaken, and problem solved by using the new DS1225 from authorized distributor.

Fyi, mine was purchased directly from Digikey with quite fresh date code, whats yours ?

[iloveelectronics]

Thats weird, my experiences with DS1225 either the old one and new were without any problem at all using the minipro I purchased from you. 

Heck, I even still use the old one, while the new DS1225 which already has the cal data copied into it is in the storage, just curious how long the old one will survive. 

Where did you buy that DS1225 ? Read few posts back, there was a similar problem with new "old" stock of DS1225 bought from a surplus shop if I'm not mistaken, and problem solved by using the new DS1225 from authorized distributor.

Fyi, mine was purchased directly from Digikey with quite fresh date code, whats yours ?

I bought mine from Mouser. Date code is 1415 so it should be about as fresh as it gets.

I read Dr. Holden's article about replacing the Dallas chip with other alternatives and saw that he had success with your RAMTRON FM1608 with the GQ-4X programmer so I also went and got myself one of those (haven't got the FM1608 yet though). I just tried using this programmer on the new DS1225 chip but still had no luck. This programmer allows me to set writing speed and it seems to be able to go a little bit further with slower speed but still no more than 3 lines of binaries max I wonder if I got a bad chip? Or did I miss any steps?

[iloveelectronics]

I just tested it some more and it looks like I can't write a "0" to the least significant bit of the high byte in each word. I can fill the chip with 10, 11, 12...1F, 30, 31, 32...3F, etc. but can't write 2x or 4x etc. 

[e-pirate]

Mr. Hugo suggests to use FM16W08 as it a newer version of FM1608 and capable of more read-write cycles. Bat it's not a direct replacement is it 4 banks of 8K x 8.

[e-pirate]

I just tested it some more and it looks like I can't write a "0" to the least significant bit of the high byte in each word. I can fill the chip with 10, 11, 12...1F, 30, 31, 32...3F, etc. but can't write 2x or 4x etc. 

BTW, did you checked old DS1225  for a multiple read sequence? It's not mentioned in document, but mr. Hugo told me that results MAY be different with data "fading" away more and more after each read attempt.
How old is your original DS1225 is?

[iloveelectronics]

BTW, did you checked old DS1225  for a multiple read sequence? It's not mentioned in document, but mr. Hugo told me that results MAY be different with data "fading" away more and more after each read attempt.
How old is your original DS1225 is?

No, I didn't try that. Perhaps I will try that when I open the case up again. My old DS1225Y has a 1990 date code, so it's 24 years of age already...

[chksum3]

Regarding your problem writing to a newly purchased DS1225: I had the same experience with a DS1225Y-200 purchased from a local supplier (Jameco) despite its having a recent date stamp.  Per advice given here, I bought another chip from Digikey - that one could be read and written to flawlessly. 

[e-pirate]

Regarding your problem writing to a newly purchased DS1225: I had the same experience with a DS1225Y-200 purchased from a local supplier (Jameco) despite its having a recent date stamp.  Per advice given here, I bought another chip from Digikey - that one could be read and written to flawlessly.

Could you provide us a photo of this two chips sitting one next to another?

[iloveelectronics]

I just replaced my DS1225Y with a Ramtron FM1608, and it works! Well, I've only tried some basic signals but everything is looking fine, just like it was with the 24-year-old DS1225Y I wrote the data onto the FM1608 using a GQ-4X programmer. The MiniPro TL866CS doesn't seem to be compatible with this chip.

Attached is the BIN file I saved from the DS1225Y chip in case anyone finds a use for it.

[BravoV]

Sigh ... Don't feel like to buy GQ-4X programmer just to program this fram, I guess I'm stuck with my TL866 and 3 pcs of unused FM1608s. 

Btw, thanks for the cal ram image. 

[iloveelectronics]

Sigh ... Don't feel like to buy GQ-4X programmer just to program this fram, I guess I'm stuck with my TL866 and 3 pcs of unused FM1608s. 

Btw, thanks for the cal ram image. 

If you don't mind mailing the chip back and forth I can help you write your data onto it.

[Bryan]

Well I think I toasted the calibration data on my Dallas.  Not sure what went wrong, but when reading it using my TL866 I just get values of 00 from Address 00000  to 0007F0 and 00 from address 001000 to 0017F0. 

When pulling off the Dallas one of the pins bent and crossed with another, so i am wondering if this somehow reset the stored values. Waiting for a 28 pin socket from Digikey before I see what error messages come up on the scope. Perhaps nothing is wrong and that is indeed my calibration data but I doubt it.

[e-pirate]

Good news, folks!
Yesterday I succeed on reading an dying cause of it respectable age DS1225Y from my TEK 2467B. As Mr. Hugo Holden noticed in his article, that this NVRAMs may get corrupted while reading attempt in programmer because of some programmer may sunk current from ICs pins, this may finish off already discharged heroic lithium battery inside DS1225 (mine is actually produces in mid 91, so it at least 23 years old with only 10 years guarantee data storage). I decided not to relay on chance and do a surgery for my DS1225Y and shunt it's internal battery with external fresh one before putting it to programmer. I spend some unpleasant hours desoldering it from A5 board, primary because if tiny traces and sucker back-kick. After DS1225 was desoldered, I replaced it with gold plated socket with round contacts and started the main act. Using Dremel with very small grinding bits I dig into compound on the bottom side of the DS1225 to reach DS1218 pin 7 shat is connected to internal battery positive terminal. Then, I tin going to install a standard CR2032 battery socket, salvaged from an old mother board. As I'm going to replace DS1225 with Ramtron 16W08 or STK12C08 (I got both, see what will do better), I decided to temporary stick socket on the top of DS1225 with a dual side sticky tape and then solder positive terminal to pin 7 of DS1218 and negative to pin 14 of DS1225. After battery was installed and wires soldered, I put DS1225 back again to see if it still holds calibration data. My scope booted without any warnings and EXERCISE 5 showed correct power hours and cycles. So, I put DS1225 to my new Wellon VP-390 programmer and successfully read data from it. Then, I just removed the battery and put DS1225 back again. I have no idea how to safely place external battery on A5 board, but now, as I have a dump, I don't really care.

[David Hess]

You description of the failure reading the Dallas NVRAMs after desoldering them from the board match what I experienced doing the same procedure on my 2440 which had a pair of DS1230 memories.  My programmer would not work with the 20+ year old memories but did with the EEPROM based replacements that I used.

While testing to determine the problem, I believe I inadvertently overwrote the NVRAMs destroying the calibration data but the 2440 series external calibration is trivial compared to that of the 2465B series so I just did that after installing sockets and the new memories.  If I had been working on a 2465B, I would have been more careful and persevered.

[Bryan]

You description of the failure reading the Dallas NVRAMs after desoldering them from the board match what I experienced doing the same procedure on my 2440 which had a pair of DS1230 memories.  My programmer would not work with the 20+ year old memories but did with the EEPROM based replacements that I used.

While testing to determine the problem, I believe I inadvertently overwrote the NVRAMs destroying the calibration data but the 2440 series external calibration is trivial compared to that of the 2465B series so I just did that after installing sockets and the new memories.  If I had been working on a 2465B, I would have been more careful and persevered.

What is strange is where one would expect a value in the address bank there is just 00's. It's almost as if the programmer couldn't read the value and substituted 00.

 

[David Hess]

What I saw was very similar if not identical.  It was like one or more of the address lines was causing the reads to be all 00 or all FF but reads from other addresses were good.

My hypothesis based on other information is that the control IC itself relies on the backup battery voltage being above a certain level or it fails to pass the chip select signal through.  Apparently at some point, the backup battery voltage gets low enough that while the SRAM contents are not immediately lost, the device cannot be accessed under any conditions even though external power is applied.

In this particular case, only some addresses were locked out for whatever reason because the backup batter voltage was just on the threshold before chip select would be completely disabled.

If I had been working on a 2465b or similar where the calibration is tedious, I would have been more careful and certainly tried accessing the backup battery directly so the calibration data could be read out.  I wonder if raising or lowering the temperature would have been enough to access the data.

[Bryan]

If I had been working on a 2465b or similar where the calibration is tedious, I would have been more careful and certainly tried accessing the backup battery directly so the calibration data could be read out.  I wonder if raising or lowering the temperature would have been enough to access the data.

Well putting the chip back in the scope and it will not work, so I suppose what you suggest could be possible. Nothing would hurt to put it in the oven or freezer for a bit and retry. Since it's likely a loss cause i could always try and de-pot the area around the battery and add a replacement one in parallel and try and read it again. Never know, just might.

Some interesting reads on the Dallas 1225 and hacks.

http://www.google.ca/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CB4QFjAA&url=http%3A%2F%2Fwww.worldphaco.com%2Fuploads%2FTEKTRONIX_2465b_OSCILLOSCOPE_CALIBRATION___REPOWERING_THE_DS1225.pdf&ei=mHm7U92SG8G7oQT6-IHADA&usg=AFQjCNGFqXCfi2ArgRR6vCgu3btbZ-9B0Q&sig2=6BETUqeTL4Vv2wFvieI1rQ&bvm=bv.70138588,d.cGU

 

[Mr Simpleton]

So for those with a lost memory and a non-working 2465B, could the scope use "generic" cal data?
I.e. how poor would the scope behave if one used cal data from an other working 2465??

[Bryan]

Tried the Dallas hack in the link I uploaded and added a external battery to the Dallas. The old one read 2.64v, unfortunately no luck in reading it on the programer, same issue. Just returns 00 in the addresses that should contain data. Oh well....

[Bryan]

So for those with a lost memory and a non-working 2465B, could the scope use "generic" cal data?
I.e. how poor would the scope behave if one used cal data from an other working 2465??

I have a TG501 time mark generator and it is pretty much bang on with the Rigol so pretty comfortable it is within calibration. Will try it on the 2565B with a generic "cal data" and see how far out it is.