Tektronix 2465B oscilloscope teardown

[sparkybg]

On page 3 BravoV posted this
 "Sad news, a failed to attempt to find the better replacement for the pesky Dallas DS1225Y BBSRAM, cause the F-RAM Ramtron FM1608-120 is not compatible. It was an impulse purchase thru a friend visiting abroad without checking the detail 1st, and also I was mis-leaded by info gathered from the Tektronix's Yahoo mailing list"
How come his chips didnt work but yours does?

The only difference between FM1608 and DS1225Y is that the FM1608 needs an impulse on CS every time the address changes in order to latch it and put the data at this address on the output. The DS1225Y does not need this - it can run with continuous CS low, and every time the address changes, the data at this address is available at the output. This is called "fast reading" or something like this. But, the 2465B processor is not using a fast reads and every time it reads or writes from/to the NVRAM it generates an impulse on CS. That's why it works.

P.S: Here is a link to the document: http://www.worldphaco.com/uploads/TEKTRONIX_2465b_OSCILLOSCOPE_CALIBRATION___REPOWERING_THE_DS1225.pdf

From my point of view, reading this document, the FM16W08 is the best possible replacement - it just needs an SOIC28-DIP28 PCB and nothing more. The STK12C68 is another one. The read/write cycles then will be equal to powering up/downs of the scope. With 1000000 store cycles and 100 years data retention, in practice this will last forever. Although it needs a PCB with some additional parts on it.

[malch]

But it didnt work in BravoV's case. ?

[sparkybg]

But it didnt work in BravoV's case. ?

I don't think Bravov ever tried it. This is stated on the datasheet as the only difference between classic SRAM ind this particular FRAM.

Anyway, let him tell us what was the case, but I can confirm it works - my scope is right next to me in this moment, and works flawlessly with FM18W08 on it.

[malch]

Thats one of the reasons I spent time putting a new battery in my Dallas. It works good by the way. It is supposed to last 26 years.

edit: maybe 130 years.

[grumpydoc]

Now, go to eBay, seek for "FM1608", and all the parts you will get will be Ramtron ones.

Then, do the same with "FM16W08", and you will again get some Ramtron devices.

Then, on the Cypress page, look for "FM1608", and you will get nothing. 16W08 is there.

So, in practice, buying FM1608 from eBay, you cannot know what exactly you are buying - a device with 10E14 or a device with 10E12 cycles. Buying 16W08 you are sure it is 10E14.

I'll agree with you as far as "from eBay, you cannot know what exactly you are buying" 

P.S: I would like to see that datasheet for 1608 with 10E14 rating. I cannot find it anywhere. Can you please give a link or upload it here?

Now I look it's a FM1608B http://www.cypress.com/?docID=48234, Cypress don't currently list an FM1608, as you say.

You are correct that the FM1608 is 10¹² and you are correct that buying the FM16W08 will be a higher endurance part, just that the "W" does not mean higher endurance per-se.

In fact the plain FM1608 is obsolete and the FM1608B is supposed to be the replacement - so that's the normal supply range but 10E¹⁴ endurance part - there are some notes about the differences here http://www.cypress.com/?docID=47983, note that the new part is not better across the board - it has higher standby current requirements and under some conditions has lower data retention.

The other reason that you are correct in practice is that I can't see anywhere with an FM1608B in stock, whereas the FM16W08 does appear to be available.

Sorry - it's splitting hairs and for that I apologize - but you know how it is  (actually that's a bit unfair as you weren't really wrong)

[sparkybg]

Sorry - it's splitting hairs and for that I apologize - but you know how it is  (actually that's a bit unfair as you weren't really wrong)

There's nothing to apologize for.

I just finished the vertical and trigger calibration. Finaly my 2445b is in one piece. I had to do a calibration generator out of old board I constructed several years ago - it had a 10 bit DAC on it, with opamps after it to give the needed 10V for calibration. I just had to add a resistor divider network. Everything is calibrated perfectly now. I will attach some pictures of the entire process in a day or two.

[sparkybg]

Another aspect of F-RAM programming - my programmer have 1608, 1808 and 18L08 (which is older low voltage part now obsolete). When I select either 1608 and 1808, the programmer refuses to program and read the first byte (at address 0000). I suspect this is because of 10ms "power to first access" time of the 18W08. But when I select 18L08 everything is OK. So, again the "W" part gives you the flexibility of trying more different devices in your programming software, if some of them don't work. In my case, only the 18L08 worked as it should. My programmer is EasyPro 90B, and the list of devices it supports is not updated so often.

[sparkybg]

OK, here we go. I've decided to go one by one, and check if everything is OK after each step.

First , the PSU.

Nothing special, really. There was clear signs of electrolyte leak from some of the caps. And well known issue with the X/Y transparent caps. I've replaced them with EPCOS units.
Bear in mind that there are 3 bipolar electrolytic capacitors on one of the PCBs. They must be replaced with the same type. The only caps I did not replace are these bi-polar ones.

I've replaced most of the caps with Panasonic FR low ESR/long life units. On the last pictures there is a small comparison between the old ones and the new ones. As you can see, the capacitance is there even on the leaked ones, but the ESR is several times (between 5 and 20) higher than on the new ones.

Also you can replace all the 180uF and 250uF with 330uf 50 or 63V. I've replaces the 180uF with 270, and 250uF with 330.

The immediate result of this was clearer traces.

[sparkybg]

Next, the A5 board.

When I bought this scope, it had some issues with the readout. Because of the leaked caps there was pretty intense corrosion on the board. Unfortunately while removing some of the caps, the pads were heavily damaged. I had to replace some of the logic chips around the upper left part of the PCB. Well, I replaced the chips and caps back then, made a good bath with isopropyl alcohol, and all was OK.

There are some missing parts on the photo, but I forgot to take a picture before I started the restoration.

[sparkybg]

Unfortunately this was not enough. Recently the front controls began to work strangely, and the horizontal deflection of the CRT was a bit lower. When I made some measurements, the DAC reference was way out of 2.5V, and I was not able to correct it with the trimmer. There was not any reaction when I rurned the trimmer. Clearly there was a problem with it, so I decided to make a full replacement of all parts around the leakage with signs of corrosion on their legs.

The width of traces on this board is unnecessarily low, and the consequence of this is easy corrosion breaks just next to the pads the trace is connected to. So pay extra attention around the pads for broken traces. This is why I had to rotate one of the resistors. I replaced the DAC with a new one from Analog Devices.

[sparkybg]

And, here it is after the restoration. The replaced logic chips on the left are clearly visible.

[sparkybg]

After I confirmed that everything works, I wanted to replace the DS1225Y NVRAM. It was a matter of time for it to fall out of specification and loose all the calibration data.
I bought a ZD-915 desoldering gun some time ago and desoldering ot the DS1225Y was pretty easy task using it. There was not any problem reading it with my EasyPro 90B programmer.

I bought a FM18W08 F-RAM and SOIC28 - DIP28 PCB. The process was easy. I leaved two terminals stated "NC" in DS1225Y's datasheet unconnected to A5 board, and connected them to the address lines next to them. This is visible on the pictures. I was able to write the data from DS1225 in the FM18W08 flawlessly, and everything worked.

[sparkybg]

Next, full recap of A1 board, and heatsink mounting on the infamous U800.

1. You have have to desolder all the connections to back BNC connectors.

[sparkybg]

2. you have to desolder all the connections to the CRT.

[sparkybg]

3. Remove the front bezel, disconnect all the connectors to the front panel, and to the front controls (focus/brightness/etc.)

[sparkybg]

4. Remove the A5 board and power supply. While doing this, pay extra attention not to touch and bend the CRT termination.

[sparkybg]

5. Disconnect the connector to the HV power supply.

[sparkybg]

6. Remove all the screws of the front BNC connectors, remove all the screws holding the A1 board to the frame. This will also detach the delay line case from the board. You will have to desolder the delay line from the board later.

[sparkybg]

7. Remove the A1 board. You will have to pull the A1 board up to disconnect the connectors to the HV board, and next pull tha A1 board to the back of the scope to take it out. You will have to remove the small connector on the back which prevents you to pull the board up. You will have to push a delay line a bit to the front while doing this - the back BNC connectors prevents you to do it directly. Pay extra attention while removing the A1 board - there are several things you can bend or break while doing this. Pay extra attention to the focus/astig/brightness front panel connector while doing this. The back of the board needs to be pulled up and then back to remove theA1 board.

8. Once the A1 board is out, take a picture of the delay line connections, then desolder them. This is a bit tricky - you will need a pretty powerful soldering iron in order to desolder the outer mountings of the delay line. I desoldered them with the ZD-215 desoldering gun on 400C.

[sparkybg]

9. The rest is more or less easy - just replace the caps. Again, with desoldering gun this is a piece of cake. I've replaced all the caps with Panasonic FR and Panasonic EB.

10. Put a good heatsink to U800 chip. I've used an FPGA heatsink for this. I bought it from Farnell.

[sparkybg]

And, here it is, up and running again, with clear and bright traces and readout. It will last some more years for sure.

[sparkybg]

11. I went through all the adjustment procedures I could, considering the equipment I have. The vertical calibration of my unit always was was out of specification, so I decided to correct this also. I had a mixed signal board I made some time ago with MCU, DAC and OPAMPs on it, able to give 10V peak to peak amplitude on it's output. I made a resistor network in order to have 10V, 5V, 2V, 1V, 500mV, 200mV, 100mV 50mV and 20mV out of it. I hooked my Agilent U1272A DMM to this in order to be able to correct the amplitude for best possible accuracy. I went through the vertical and trigger adjustment procedure, and finally my vertical cursors began to show exactly what they should show. Here's my setup.

[malch]

Finished my U800

[malch]

Well all done. Moved the cal/nocal jumper to cal and fired it up.
The 10 v was out a little [9.98 v ] and I had to fiddle with the vertical jitter pot.
Ran test all and all passed.
What is the exerciser used for ? How do you know when it is finished? How do you stop it?

[sparkybg]

Well all done. Moved the cal/nocal jumper to cal and fired it up.
The 10 v was out a little [9.98 v ] and I had to fiddle with the vertical jitter pot.
Ran test all and all passed.
What is the exerciser used for ? How do you know when it is finished? How do you stop it?

Which exerciser? Calibration constants exerciser? If yes, on every step you have an address and the contents of this address. Just make a movie of the screen while you go from the lowest to the highest address. When you reach the highest address, it will turn back to the lowest. You can exit from exerciser at any time or any step you want.