Author Topic: Tektronix 2465B oscilloscope teardown  (Read 440367 times)

0 Members and 1 Guest are viewing this topic.

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1300 on: November 23, 2018, 10:40:36 am »
I have a Xeltek 5000 at work.
That big spark at power up was by design!
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1301 on: November 23, 2018, 12:02:44 pm »
I have a Xeltek 5000 at work.

Maybe its not fully compatible with the FRAM. Is the FM16W08 specifically mentioned in the list of compatible devices? From what I learned so far, there is one crucial difference between programming a D1225 and the FRAM, in that one of the control lines needs to be toggled during each individual write instead of staying in a specific state all the way throughout the whole write process of the chip. Not sure that makes sense as it is only what I remember reading, and again I am new to programmers and will use one for the first time once I have all the remaining bits here to save my 2465B cal data.

I ordered the GQ-4X4 because its the newest version hardware, and the FM16W08 is specifically listed as being supported with the current v10 software, as well as the D1225. I also went by the comments of other people that say the FRAM writing worked for them on this programmer. I will report back my own results next week.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 10963
  • Country: us
  • DavidH
Re: Tektronix 2465B oscilloscope teardown
« Reply #1302 on: November 23, 2018, 03:37:01 pm »
From what I learned so far, there is one crucial difference between programming a D1225 and the FRAM, in that one of the control lines needs to be toggled during each individual write instead of staying in a specific state all the way throughout the whole write process of the chip. Not sure that makes sense as it is only what I remember reading, and again I am new to programmers and will use one for the first time once I have all the remaining bits here to save my 2465B cal data.

FRAM uses sense amplifiers for reading which much be precharged before each read.  Precharge occurs while -CS is high so -CS must be toggled for every memory read unlike SRAM.  Usually this is not a problem because in most applications, -CS is toggled between every access anyway but a programmer might rely on holding -CS low and operating asynchronously which will not work.

Writes should not matter as either -CS or -WE may be toggled like with a standard SRAM.
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1303 on: November 23, 2018, 04:55:59 pm »
The programmer is not the problem, Ramtron FRAM is in the list of devices and it verifies after programming. Even the scope works well very first time after programming. Once I power cycle, some locations in the calibration constants area get corrupted. That doesn't happen with the DS1225. I'm surprised that it works well for some people but not for others. Well, at least for one...
That big spark at power up was by design!
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1304 on: November 24, 2018, 11:02:14 pm »
Son of a gun, I replaced U1371, now it doesn't react to the cold anymore but it stays at 14.78V instead of 15V. Before, it would drop to about 14.8 when cold but would stabilise to 14.98V after awhile. Now it is worse. The IC is from Digikey so I don't expect it to be defective. The interesting thing is that the inverting and non inverting inputs are at different levels. The positive input (pin5) is at 10.00V sharp and the negative one (pin 6) is at 9.84V. Shouldn't they be at the same level?

If your +15V PS buss is now at 14.78V it is a tiny bit out of spec, it should be between 14.775 and 15.225V with no more than 15mV ripple (check at J119 pin 6). Strange for a 317 not to output something closer to the expected 15V, with the associated circuitry unless something is still amiss. What is the voltage at the 317 reference input?

It is 1.23V below the output and it is where it should be. I don't suspect 317.

I installed a socket in place of U1371 and I tried all my collection of 324 that I bought and found in my junk box. That is one brand new TI LM324AN bought from Digikey, two National Semiconductor LM324N bought from a local store, one TI LM324N from my junk box and one TSL ??? LM324N from my junk box. Only the last one works, all the other give me between 14.6V and 14.75V at the output. I also tried a TL064 from my junk box and a new TL084 that I bought today, no problem with them. The parts are OK, I tested them all in a buffer circuit with 20V supply, 10V at the non inverting input and 10K resistor in the feedback loop.
The original IC is a NEC C324C that either has some special performances (I can't imagine what can be so special about a 324) or  there's something wrong with that part of the circuit. I know it's Tektronix and suspecting a design screwup is almost insane but....who can come up with a reasonable explanation?
Anyway, I installed the TL084 in the end and the output is a stable 14.98V.
« Last Edit: November 24, 2018, 11:10:08 pm by Miti »
That big spark at power up was by design!
 

Offline Satbeginner

  • Regular Contributor
  • *
  • Posts: 137
  • Country: es
  • Dutch, early retired, living in Spain
Re: Tektronix 2465B oscilloscope teardown
« Reply #1305 on: November 25, 2018, 12:25:52 am »
I think I had something similar, it turned out that a specific version OpAmp had a +and- 15V range power supply, and others had a +and- 18V power supply range, so only the version with the higher specs on power supply voltage would work.

It had something todo with the version letters after the 324...

I have to look into my notes for more details.
You need a scope to repair a scope, and you need many multimeters to repair another multimeter!
*Tek 2467B, Tek 2465B, Tek 2465B, Tek 485, Tek 475A,  Keithley 175A, Keithley 2000, HP 3468B, HP 3457A, HP 34401A, PM 6671, PM 5716, Fluke 45, Fluke 75, Fluke 77, Fluke 79, AFX 9660BL, KPS 605D, etc. *
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 10963
  • Country: us
  • DavidH
Re: Tektronix 2465B oscilloscope teardown
« Reply #1306 on: November 25, 2018, 01:29:35 am »
I think I had something similar, it turned out that a specific version OpAmp had a +and- 15V range power supply, and others had a +and- 18V power supply range, so only the version with the higher specs on power supply voltage would work.

It had something todo with the version letters after the 324...

I have to look into my notes for more details.

That crops up for a lot of old bipolar process operational amplifiers but you need to check the specific part numbers for each manufacturer because they are not always consistent about which ones are 36 volts and which are 40 or 44 volts.

TL series JFETs - All 36 volts
LF series JFETs - Some are 36 and some are 44 volts
101A and 201A (military and industrial) - 44 volts
301A (commercial) - 36 volts
108 and 208 (military and industrial) - 40 volts
308 (commercial) - 36 volts
741 (commercial) - 36 volts
741 (other) - 44 volts

Tektronix used to grade 36 volt 741s for operation at 44 volts.  I assume they were using plenty of them at lower voltages so did this as a cost saving measure in the few applications where a higher voltage part was needed.
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1307 on: November 25, 2018, 03:31:46 am »
But this is only supplied at 26V or less. On the other hand, only U1371B seems to be bad, or at least U1371C seems to work in all of them. It must have something to do with the way it is used in this schematic.
That big spark at power up was by design!
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1308 on: November 25, 2018, 10:06:42 pm »
It all makes sense now   8). The positive supply voltage to U1371 is 15V. Its output swing has reach the limit about 300mV too low, it simply cannot go higher on majority 324s, but TL084 can, enough to maintain regulation. One thing that made me think of that was that, using the ICs that don't regulate properly, during the boot time, the +15V drops a bit further. Probably there's an increased current during the self test and that would need higher voltage at the reference pin of LM317, while with TL084, it is rock solid 14.98V.

There are few factors contributing to this. First is the unregulated voltages being a bit lower than expected, I don't know why. If the unregulated +15V was 19.2V as shown in the schematic and not 18.4V that I measure, I wouldn't have this issue.
Second is that R1400 seems to have increased its value, I measure 108Ohm in circuit instead of 100Ohm. Being in circuit I should measure lower that nominal and not higher, that means it may be even higher than 108Ohm. That contributes with another about 270mV voltage drop. If I restore R1400 to 100Ohm I expect an increase in the supply voltage to U1371 of about 270mV.

One interesting thing is why do they expect to get 5.1V on +5VD. First I thought there's something wrong with my inverter feedback as I'm getting +5.00V on +5VD, but if you look in the schematic, there's no way to get +5.1V if that loop works well.

Can anybody measure the unregulated voltages in a working Tek 2465B or 2445B please? A marginal design in that area, thumbs down for Tektronix.

The repair adventure continues....

Edit: Actually can anyone measure all the voltages at the output of the inverter, including +5VD?
« Last Edit: November 25, 2018, 10:38:31 pm by Miti »
That big spark at power up was by design!
 

Online MarkL

  • Supporter
  • ****
  • Posts: 1632
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1309 on: November 26, 2018, 07:10:59 pm »
I'm calculating a decent operating margin for the LM324.

The LM324 is driving the LM317 through a diode drop, so the LM324 output should be around: LM317out - LM317drop - diode --> 15.0V - 1.25V - 0.7V = 13.05V.

The LM317 adjust terminal current max is 100uA, so under normal conditions that will also be the max current for the LM324 output.

The LM324 datasheet does not show a spec for V+ = 18V, so using the V+ = 30 spec, Vo to a 10k load has a drop of 30V - 27Voh = 3V, and with that there's a current of 27V/10k = 2.7mA.  That's well under the 100uA max in the Tek circuit.  According to the typical characteristics, the Voh drop is independent of V+, so the drop should not be any worse at V+ = 18V.  The typical  characteristics plot puts the drop at around 1.2V for 100uA.

But still if we use the absolute worst case, in order to drive 13.05V on the LM324 output towards the LM317, we need a V+ supply of 13.05V + 3V = 16.05V minimum.

I think you're saying you're measuring 18.4V for the unregulated supply, but what is it on the other side of R1400 (which is also the LM324 V+)?

The LM317 dropout voltage is 3V, so that is a little tight if the unregulated supply is 18.4V and the output is 15V, but is still within 400mV.  The typical dropout is 2.5V at 1.5A.

Can you measure *all* the pins on the the LM317 and LM324 and post?  Even though it works with the TL084, I still think there's something wrong.  I don't think it's a bad design and there should be enough margin.  (And the answer to a previous question is yes, Vin+ and Vin- should be equal in this circuit since LM324 section B is normally amplifying.)


Also, on the FRAM corruption, I think it's odd that it happens and then the scope continues to run.  The FRAM is also acting as the scope's operating RAM, so if there's a problem with read/write I would expect it to crash or become unstable almost immediately.

I'm wondering if there's still some kind of start-up power supply issue that the FRAM is more sensitive to that's causing corruption.  At boot, one of the first things the scope does is a write/readback check of every RAM location, *including* the cal locations.  Starting with address 0x9FFF and going to 0x8000, the scope reads the value, does a write/readback check with 0x55, then with 0xAA, and then writes back the original value.

The previous power supply issue and FRAM corruption could be completely unrelated, but I'm throwing it out there as a possibility.
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1310 on: November 27, 2018, 03:01:06 am »
Hey MarkL,

The LM324 V+ on the other side of R1400 is 15V. Not enough for safe operation.

Regarding the FRAM, if it does that...crazy thing...at boot, that explains a lot. For some reason my TL866II Plus can "almost" program this FRAM even though it is not in the device list. I select DS1225 as the device and it can program successfully 0x00 and 0xFF but when I try to program the real content of my scope's NVRAM, it fails to program few locations.....the same locations that get altered after the first power cycle.

What I mean is, if I program properly the FRAM using the Xeltek programmer (that fully supports it), the first power up is ok, it runs well or at least I didn't see any issues in the short time that I ran it. But once I power cycle, two locations in the calibration constants area, the same locations that can't be programmed with TL866II Plus, are modified. Probably once it does the 0x55 then 0xAA, it cannot put back the original content. But why does it work for some people, that beats me.
« Last Edit: November 27, 2018, 03:08:55 am by Miti »
That big spark at power up was by design!
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1311 on: November 27, 2018, 03:29:50 pm »
In preparation for the arrival of the socket and FRAM hopefully by tomorrow afternoon, the NVRAM is now off the A5 board. Not too bad, a few of the pins needed a little bit of extra care to fully free them and make them "swing", but even pin 14 fully cleared on the first try. All in all, process went rather smoothly and the chip actually fell off the board all by itself when I unsoldered the last pin.

Now on to read the chip. Will report results once that part is done as well.
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1312 on: November 27, 2018, 03:59:00 pm »
Read the NVRAM chip, apparently all went well, took 0.24 secs according to programmer and no errors. Actually just in case read it twice and saved separate BIN files. My first chip reading.

So first thing I did was to compare the values between 1E00 and 1FF0 of the file that was just read and the data I had from the Exer02 procedure. After a simple overview it is clear that the first one and a half lines have some differences (up to 1E1005 - I think that would be the right reference), but all the rest of the data seems to be equal. I've attached pictures of both the template I filled some time ago with the EXER02 data into an empty bin file template kindly provided by MarkL, and also the resulting file of the actual scope NVRAM reading just done.

Maybe I should have gone through another EXER02 routine just to double check for any changes before removing the NVRAM chip, but it just slipped my mind. In any case I will be able to verify it once the machined socket is in place. The only things I could think of that would have caused the changes are that I performed a vertical channel DC balance (by pushing both input selectors up). Other than, I don't think that the total hours and power ups are part of the data contained between 1E00 and 1FF0, so that would not had any effect. Unless of course the last resistor change of the DAC divider chain has something to do with it, and the scope just adjusted itself to the slightly higher reference voltages after I replaced the temporary 5% resistors used just to bring back the DAC to life, with the correct 10K 0.1% ones.

I do already have the new DS1225AD on hand, and wanted to write it with a copy of the template file that I filled out by hand from the Exer02 data and see what the scope says. Then I would have erased it, and put in an exact copy of the data I just read from my own NVRAM. But now in light of the differences in the data, I will wait until I can run another NVRAM data verification and compare it with my existing Exer02 data gathered some weeks ago and see if there are actual changes that match. I suspect they will.

And as it seems to be customary here on this thread, I have also attached a zipped up BIN file of the NVRAM content, along with the serial number added to the file name, to maybe provide something helpful for others to use, and at the same time have just another BU of the file as well.

EDIT: Replaced the image of the EXER02 data, as I found a mistake I made filling it in at 1F3000 which was written as FF while it had to be FD.
EDIT2: Added another picture that shows the differences in data between Exer02 and the NVRAM read.
« Last Edit: November 27, 2018, 10:28:32 pm by AMR Labs »
 

Online MarkL

  • Supporter
  • ****
  • Posts: 1632
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1313 on: November 28, 2018, 10:25:01 pm »
The LM324 V+ on the other side of R1400 is 15V. Not enough for safe operation.

Quite so.  Just out of curiosity, I measured the voltages on a 2445A on the LM324 (U1371).  The power supply design is the same as the 2465B, at least in this section.

  +15.06 --- 1  14 --- -6.30
  +15.00 --- 2  13 --- +2.651
  +15.57 --- 3  12 --- +2.501

  +16.24 --- 4  11 --- -6.95

  +10.00 --- 5  10 --- +5.00
  +10.00 --- 6   9 --- +5.01
  +13.33 --- 7   8 --- +2.916

And the LM317 (U1260):

  +13.77 --- 1
  +15.01 --- 2
  +18.73 --- 3

So, the unregulated +15 supply (+18.73) is also low on this scope.  I'm not digging into it any further since it's working, but I think I do need to modify what I said and instead say this is not such a great design.  There's too much loss through R1400 and even the typical margins become quite small when the unregulated supply is even a little low (we're only talking 0.5V here since it should be 19.2V).

My R1400 reads 106.3 ohms, so it's also a little high like yours.  But my loss through R1400 is only 2.49V, and not 3.4V like yours.  I think you have something drawing excessive current on that side of R1400, maybe from +10V REF.

Quote
Regarding the FRAM, if it does that...crazy thing...at boot, that explains a lot. For some reason my TL866II Plus can "almost" program this FRAM even though it is not in the device list. I select DS1225 as the device and it can program successfully 0x00 and 0xFF but when I try to program the real content of my scope's NVRAM, it fails to program few locations.....the same locations that get altered after the first power cycle.

What I mean is, if I program properly the FRAM using the Xeltek programmer (that fully supports it), the first power up is ok, it runs well or at least I didn't see any issues in the short time that I ran it. But once I power cycle, two locations in the calibration constants area, the same locations that can't be programmed with TL866II Plus, are modified. Probably once it does the 0x55 then 0xAA, it cannot put back the original content. But why does it work for some people, that beats me.
If it's consistently the same locations, maybe it's a bad FRAM?  Are there counterfeit FRAMs floating around?

I'm surprised it gets through the 0x55/0xAA test in the first place.  I can't explain that.
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1314 on: November 29, 2018, 12:23:11 am »
Thanks MarkL,

This is great info!
One more voltage that I was interested in was the 5VD that I don't think can measure 5.1V no matter what.
I'm not too worried about the extra ~7mA in my R1400 as long as everything else works as it should, I wonder if it isn't exactly this pushing to the maximum of the U1371B that draws it.
What I will do is to replace R1400 with a selected 100R that measures a bit below nominal and replace the diodes CR1103 - CR1106 in the preregulator with Schottky, so I gain few hundred mV there too, and continue to be disappointed in Tek's design. I think even the best can screw up sometimes, eh? I'm sure it's a known issue among the Tek gurus and probably there's a solution too but it's not worth the time looking.
The FRAM is not a knock off, it is a Cypress bought from Digikey but that doesn't mean it cannot be defective. The first one bought from Ebay gives me garbage in the readout.
However, Xeltek can program it, no problem, it's only the TL866 (which is excused as it doesn't officially support it) and the scope that messes up the content. Anyway, it passes 0x00, 0xFF, 0x55 and 0xAA in my TL866 II Plus but I can't test it "at speed".
I see that AMR Labs is waiting for his FRAM these days so I'm curious how it works out for him.

Edit: Actually I think I have the answer regarding 5VD in the level at Pin 9 U1371. Not sure what meter you used but if it has a 10M input, 5VD should be around 5.03-5.04V.
« Last Edit: November 29, 2018, 12:36:51 am by Miti »
That big spark at power up was by design!
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1315 on: November 29, 2018, 04:10:32 pm »
Parts arrived, pictures attached.

Surprisingly FRAM came in a humidity controlled anti-static bag. Anybody ordering a FRAM from Mouser (or elsewhere) got this as well? Dodn't know this part was so humidity sensitive. For the fun of documenting it took pics.

Will report back once new NVRAM is written, and scope powers up, and verify Exer02 data. Then will erase new NVRAM and write it with the Exer02 data template file.

Last will write the FRAM, and see how it goes.
 

Offline emece67

  • Regular Contributor
  • *
  • Posts: 222
  • Country: es
Re: Tektronix 2465B oscilloscope teardown
« Reply #1316 on: November 29, 2018, 07:27:40 pm »
Hi all,

Finally replaced the backup battery from my SN <499999 unit. The cal data remained intact, so the scope is now running perfectly. I decided to replace the original 750 mAh Keppler LTC-7P battery (manufactured on the 15th week of 1988, 30+ years ago) by a more easy to find and replace 2400 mAh AA cell (of the same Li-SOCl2 chemistry, obviously) from Tadiran. The cell holder was mounted on the frame, screwed to the panel behind the knobs, as seen here:



The task was more simple than expected, in fact, after the end, I wondered if the replacement for an AA cell was wise, as desoldering and soldering the LPC-7P seemed easy enough to not to worry about 10 year periodic replacements. In any case, the task was done and I decided to keep the AA cell.

During the operation I have used another AA Li-SOCl2 cell connected to wires soldered to the terminals of R2770 and CR2070 to maintain power to the RAM. Believe it or not, but during the task, with the original battery removed, I accidentally, and briefly, shortened the battery terminals  |O. A big surprise when, after powering the unit again, it worked flawlessly  :phew:. These are the new terminals:



Now thinking on the PSU recap to be completed... The next image is slightly blurred, but the real scope display is super crisp  :-+



Regards.


« Last Edit: November 29, 2018, 07:37:35 pm by emece67 »
Information must flow.
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1317 on: November 29, 2018, 08:15:35 pm »
Everything worked out well. Just to make sure the old NVRAM was still OK after all the desolder and reading procedures, inserted it into the newly installed socket and the scope booted up normally into the exact same configuration before last powering it off. I also went to Exer02 and confirmed all the changes I saw when I first read the original NVRAM data, and it matched the ones reflected on the chip data. So maybe the vertical balance adjust I did, or something else changed those first and a half lines of the data between 1E00 and 1E10, as I pointed out a few posts back.

Did one last read of the original NVRAM of the scope and programmed the new NVRAM. Scope came up like nothing happened, exactly the same configuration, as seen in one of the attached pictures. Then programmed all FFFF"s into the new NVRAM (not sure that is the correct way to erase all data, as the Erase function on the programmer itself seemed not to work and was giving me an error). Anyway, verified that the new chip was all FFs by reading it, and the proceeded to open the template BIN file with all the data from the EXER02 procedure (updated to reflect those later changes). The scope booted almost normally without any errors but as expected showed a somewhat strange configuration as seen on another picture (I also made a movie of the boot-up process, but its in 1080P so its a 70MB MOV file), and at the end of the boot process there was no trace or readout on the screen. After some fiddling I noticed the trigger was set to Single SEQ, and after changing it back to Auto a really slow and extremely bright sweep came up. Uuups quickly turned down the intensity knob. Also the trace was no where near being centered, neither horizontally or vertically. But after turning all the related knobs either way a bit things started to settle, and the readout also came back on screen. I did not do any extensive tests for accuracy, but from what I saw all seemed to be OK, and since the scope did not complain I'm guessing the cal data checksum was good. Pulled the NVRAM to read it and saved the contents to file so any changes done by the scope to the template data are recorded for future analysis if needed.

Then again wrote all FFs into the fresh 1740 datecode NVRAM and wrote the data from the last read of the original NVRAM and put it into the scope. Upon boot up all seemed fine other that one relay made a bit of a short buzzing sound instead of the usual single click, and the scope came up normally. Then pulled out the NVRAM and read the contents to file as well, again just to document and archive everything, since I noticed that quite a bit of data from 1E00 down had been changed, not sure this would be because the scope was still expecting to see the earlier Exer02 data from the last test after it settled all the front panel settings. So once again pulled out the new NVRAM and wrote FFs, then proceeded to write once again the data from the last read of the original NVRAM, and this time the scope came up normally, no buzzing of the relay (that might have been a fluke?), and it came up to the same configuration that the scope was into when the old NVRAM was still installed. But this time upon reading again the new NVRAM, there where no changes in the cal data from 1E00 and on.

Last but not least, the FRAM. Soldered it onto the converter board, read it and it was all 0000 (zeros), so proceeded to write again the same data from the last read of the original NVRAM, and pushed it into the socket. Result: scope came up normally, no errors, and in the exact same configuration as expected. Powered down the scope and on again a few times over the course of several minutes, same thing no errors, all seems OK. Waited about 15 minutes, power up and again, everything came up the same and without any errors. Now just checked again, must have been like 30 minutes since I last powered the scope, and it is still booting up perfectly.

So not sure which one of the memory modules I will leave in. Maybe leave the old original NVRAM and see how long it will take for it to fail, but then I might loose some data along the way if any operations are performed (other than a full calibration) that might alter the end result, as it happened between the period of about a month where I first took the Exer02 data down, and the last read of the NVRAM once it was desoldered. Again, I only recall doing a vertical balance by pushing up both channel 1 and 2 input selectors, and it would make sense that this change would get written into the NVRAM, although seems obviously non-critical, as I have not done any serious calibration at all.

Or I could just put away the old NVRAM, and start using the new one, and just forget about it (for a while). But the old one would then probably die silently at some point, which at this point is inevitable anyway. So I think the old NVRAM will definitively stay out.

The other option, and which is probably what I will do for now, is to leave the FRAM in order to see how it behaves over a more prolonged period of time, and put away both the old and the new DS1225's. After all the initial idea was that if the FRAM is confirmed to be stable and working fine, it would allow me to dispense with backup battery blues altogether. Maybe at some point I could swap it out with the new NVRAM if I feel like doing some more tests, and of course always keep up with any minor changes that appear in the cal data.

What do you guys think?
« Last Edit: November 29, 2018, 08:47:16 pm by AMR Labs »
 
The following users thanked this post: BravoV, MarkL

Online MarkL

  • Supporter
  • ****
  • Posts: 1632
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1318 on: November 29, 2018, 08:47:46 pm »
...
What do you guys think?
Wow - thanks for the detailed write up!  It's nice when things work as expected.

I think I would leave the FRAM in there.  You went to a lot of trouble to create it and you'll never have to worry about the NVRAM again (knock wood).

As for the changed values after the channel balance, you could repeat the balance procedure and see if those locations change again.  EXER02 location 00 (RAM: 1E00/1E01) is the checksum, so it's expected to change if any of the other constants change (EXER02 locations 0B thru E3, if I'm reading the code right).
 

Online MarkL

  • Supporter
  • ****
  • Posts: 1632
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1319 on: November 29, 2018, 11:35:15 pm »
...
Edit: Actually I think I have the answer regarding 5VD in the level at Pin 9 U1371. Not sure what meter you used but if it has a 10M input, 5VD should be around 5.03-5.04V.
It's a Fluke 79 III.  The range for measuring the 5V is 10Mohm (40V range) with 10mV resolution.  So we're talking about the least significant digit here.

I dragged the scope over to a different bench with a 3456A (6 1/2 digits).  The impedance on the 3456A on the 10V range is >10^10 ohms.  Here's a re-read:

5VD measured on J232: 5.01501V

U1371: pin 10: 5.01095V
       pin  9: 5.01193V

There's a little bit of offset, but within spec, and some of it is probably because the input impedances are not matched in this circuit.

Is your concern that the 5VD is noted as 5.1V in the schematic?
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1320 on: November 30, 2018, 12:04:38 am »
Thanks AMR Labs!

Now I'm really pissed off.... Is it possible that both my FRAMS are defective? I can understand the one from Ebay for $6 but from Digikey?
How can I test it? What can be wrong with my scope considering that it works perfect with the old and two new DS1225?

Edit: I just had a live chat with Digikey's customer service. They will send a replacement, no questions asked. Impressed!
« Last Edit: November 30, 2018, 12:33:11 am by Miti »
That big spark at power up was by design!
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1321 on: November 30, 2018, 12:36:53 am »

Is your concern that the 5VD is noted as 5.1V in the schematic?

Thanks MarkL!
I'm not concerned, I just wondered what intern was tasked with marking the voltages on that schematic and if he used a dice or a multimeter.
That big spark at power up was by design!
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1322 on: November 30, 2018, 05:09:47 am »
Thanks AMR Labs!

Now I'm really pissed off.... Is it possible that both my FRAMS are defective? I can understand the one from Ebay for $6 but from Digikey?
How can I test it? What can be wrong with my scope considering that it works perfect with the old and two new DS1225?

Edit: I just had a live chat with Digikey's customer service. They will send a replacement, no questions asked. Impressed!

Did your FRAM from Digikey arrive in a controlled humidity package like mine did? Maybe yours got a bit wet along the way!  ;)  ;D

Well, I never really liked Digikey too much, back in 2011 or 12 when I was still located in PR, they managed to piss me off when I tried to order from them for the first time a couple of common parts (a couple of 74xx and/or 40xx chips, and some other passive parts and ship them directly to PR. They where giving me some crap that they where not able to "export" these parts (PR being an associated US State!) and asking for some other BS documentation. >:( So ever since I passed on them and really get anything I need from Mouser, which I consider a high quality and very efficient supplier. At least in my experience, Digikey was back then no longer worth the hassle, so I never looked back. And not that it would matter much, but their prices are almost always a bit higher in direct comparison to Mouser.

Just in case you decide as a backup plan to order the FRAM from Mouser as well:

FM16W08-SGTR, Mouser PN 877-FM16W08-SGTR, in stock 357, $10.85
https://www.mouser.com/ProductDetail/Cypress-Semiconductor/FM16W08-SGTR?qs=sGAEpiMZZMtsPi73Z94q0OKOSVYv1Kg2z52nkh0S%252bvo%3d

28-650000-10 Aries 28P SOIC/DIP SOCKET, Mouser PN 535-28-650000-10, in stock 12, $18.55
https://www.mouser.com/ProductDetail/Aries-Electronics/28-650000-10?qs=sGAEpiMZZMs%2fSh%2fkjph1tjJZclYmfaNPlov7zqaFl0I%3d
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1323 on: November 30, 2018, 05:38:34 am »
Some time ago I built this Avalanche Picosecond Pulse Generator:

https://entangledwaves.wordpress.com/2013/07/15/avalanche-pulse-generator/

BTW the specified now obsolete 2N2369 transistor is still available at Mouser in TO-18 metal case for about $2, although a usable alternative seems to be a common 2N3904, but the former has a more desirable breakdown characteristic. I did not need to build the DC-DC converter part for the 90V avalanche bias supply, but instead used my Heathkit IG-4505 Scope Calibrator as a power source with the DC output set to the 100V position.

There is also another very well known website on the same subject:
http://www.kerrywong.com/2013/05/18/avalanche-pulse-generator-build-using-2n3904/

Wanted to share pictures of the pulse as shown on both my 2465B and the 2247A, and possibly hear opinions. Cursor time measurement are located between the 10% and 90% points of the rising edge. Thanks.
« Last Edit: November 30, 2018, 06:20:49 am by AMR Labs »
 

Online MarkL

  • Supporter
  • ****
  • Posts: 1632
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1324 on: November 30, 2018, 04:54:59 pm »
...
What I mean is, if I program properly the FRAM using the Xeltek programmer (that fully supports it), the first power up is ok, it runs well or at least I didn't see any issues in the short time that I ran it. But once I power cycle, two locations in the calibration constants area, the same locations that can't be programmed with TL866II Plus, are modified. Probably once it does the 0x55 then 0xAA, it cannot put back the original content. But why does it work for some people, that beats me.
So, you're saying the corruption shows up on the second boot.  It's possible the FRAM is getting corrupted during the scope's power-down sequence.  That would also explain why it passes the 0x55/0xAA test; the corruption has already happened.

You could use the PWR DN signal as a trigger and start looking at the behavior of the control signals into the FRAM, including the +5V power, during power down.  Maybe there's something marginal that the DS1225 tolerates.

If you have an MSO or a logic analyzer with enough inputs, you might also try setting up a trigger to capture when the corrupted address in the FRAM is accessed or written.  But keep in mind this could be an analog signal issue and a logic analyzer might not catch it.

EDIT:  One more thought before starting down this path... With good data in the FRAM, you could do one power cycle, remove the FRAM, and read it to see if corruption occurred.
« Last Edit: November 30, 2018, 04:59:20 pm by MarkL »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf