Tektronix 2465B oscilloscope teardown

[Bud]

If you have a GPIB interface, you can dump the NVRAM data directly to a computer (and reload it too).

Can you give a pointer to the procedure?

[DavidAlfa]

Thanks Mark, yep, that was my idea, never think you're perfect and no errors will happen.

Bud, "GPIB" search in this thread:
 https://www.eevblog.com/forum/testgear/tektronix-2465b-oscilloscope-teardown/msg1943710/?topicseen#msg1943710

[MarkL]

Thanks, DavidAlfa, for posting the initial link.

Those are the basic commands.  It was later confirmed to work on all the models.  Plus, an optional argument was discovered to dump ALL the NVRAM at once:

  https://www.eevblog.com/forum/testgear/tektronix-2465b-oscilloscope-teardown/msg1618642/#msg1618642

Since the procedure is somewhat scattered through threads on eevblog and TekScopes, let me try to summarize the essentials here.

In short, the procedure to capture the calibration and other NV data is to send the GPIB commands:

  key 0
  earom? f

The format of the data returned is the starting location in NVRAM (which is 0 in this case), a colon, and then comma separated values from incrementing NVRAM locations.

On models later than the 2445/2465, you're going to get 255 16-bit words (510 bytes), which maps to 0x1E00 to 0x1FFD in the actual NV SRAM chip.  The words are stored in big-endian format.  The address range covers the calibration data and other opaque data currently unknown to us mere mortals.

To restore the data, you need to send it back to the scope one word at a time.  The format of the argument is the location in NVRAM, a colon, and then the value:

  key 0
  earom 0:215;
  earom 1:1908;
  earom 2:1921;
  earom 3:1921;
  ...
  earom 253:16499;
  earom 254:14;
  earom 255:0;

Sending it back en-masse (i.e., replaying the one big line that the scope sent) does not work, at least not for me.

On the 2445/2465, you are only going to get 200 16-bit words, and not 255.  Words 0-99 are from the EAROM on the main processor board, and 100-199 are from the EAROM on the buffer board.

Keep in mind that all numbers for the EAROM command are represented in DECIMAL and not hex, as it appears in the EXER 02 display.

As an example from the dump above, EAROM 1 is 1908 and is shown in EXER 02 as 0774.  This is stored in the actual SRAM as 0x1E02: 0x07 and 0x1E03: 0x74.

[Bud]

Very useful post, thanks!

Edit: worked on 2467B like a charm, got 255 words (512 bytes) 

[Mark Krass]

Hi,

Does anybody know what is the jumper J504 on A5 board 2465B & 2467B? (SN B050000 & above) How it can help in troubleshooting?  No word in forums/service manual about it.

Location 5K in DMUX & analog control schematic.

Mark

[MarkL]

...
Does anybody know what is the jumper J504 on A5 board 2465B & 2467B? (SN B050000 & above) How it can help in troubleshooting?  No word in forums/service manual about it.

I've never seen it documented.  J504 allows the -1.25V reference to be provided by an imprecise resistor divider.  The -1.25V reference is usually provided by the DAC.  The jumper allows allows easier troubleshooting of the DAC and associated circuitry by removing dependency on proper operation of the DAC.

If you search for "J504" on TekScopes on groups.io, you can read how people have employed it to help in their troubleshooting.

[Mark Krass]

I've never seen it documented.  J504 allows the -1.25V reference to be provided by an imprecise resistor divider.  The -1.25V reference is usually provided by the DAC.  The jumper allows allows easier troubleshooting of the DAC and associated circuitry by removing dependency on proper operation of the DAC.

If you search for "J504" on TekScopes on groups.io, you can read how people have employed it to help in their troubleshooting.

Thanks for pointing to TekScopes on groups.io, this is where the information about J504 truly resides. You're right, this jumper provides ~1.25V reference comes from a voltage divider and not the DAC which is used to troubleshoot DAC circuitry, like you said. So I got the answer, thank you!

PS Funny thing is that google "tektronix 2465b j504" did not come up with any links to groups.io

[SynViks]

Curious how easy it'd be to add the counter timer option to a unit that doesn't already have one? IIRC those modules are automatically recognized when present. Is it just a matter pulling the PCB and hooking it up?

[AMR Labs]

Curious how easy it'd be to add the counter timer option to a unit that doesn't already have one? IIRC those modules are automatically recognized when present. Is it just a matter pulling the PCB and hooking it up?

I was able to do just that to my 2465B thanks to a member of Tekscopes that sold me the whole kit and caboodle and had done it himself more than once on other 2465B's, so he knew very well all there was to know about this option add-on.

Basically you will need the CTT card, and several ribbon and coax cables with specific connectors on their ends, plus also the rear Ext Ref and Word Recognizer jacks with their cables. The scope power supply needs to temporarily be removed, and some jumpers need to be taken off a connector where one of the new ribbon cables will plug into.

So definitively you need to make sure that you really have everything that is required, and the correct procedure to install it. And yes, the board will automatically be recognized once it is installed. Lastly you will need an accurate and stable 1MHz source to feed into the scope so the CTT function can be calibrated as explained in the service manual and which is a very short and simple procedure to perform, and also involves moving a jumper position during cal. The seller even threw in a nice TCXO to do just that as part of the package, and even added an original CTT plastic rear panel with the two additional connector holes, so that was pretty much plug and play without any drilling required. Not that it would have been very hard to do, but its definitively nice to have the connector jacks with their original labels in place, which also makes it look like it came with the scope.

BTW the upper frequency limit of the CTT counter is supposed to be 100MHz, but mine counts very reliably to 200MHz, and only stops working at about 230MHz.

CTT is definitively a very nice option to have on a 2465B, but in my opinion for anything up to or bellow 100MHz, it still does not beat my 2247A, which besides CTT functionality also comes standard with various other very handy and useful waveform voltmeter functions and some other stuff.

[factory]

Curious how easy it'd be to add the counter timer option to a unit that doesn't already have one? IIRC those modules are automatically recognized when present. Is it just a matter pulling the PCB and hooking it up?

I was able to do just that to my 2465B thanks to a member of Tekscopes that sold me the whole kit and caboodle and had done it himself more than once on other 2465B's, so he knew very well all there was to know about this option add-on.

Basically you will need the CTT card, and several ribbon and coax cables with specific connectors on their ends, plus also the rear Ext Ref and Word Recognizer jacks with their cables. The scope power supply needs to temporarily be removed, and some jumpers need to be taken off a connector where one of the new ribbon cables will plug into.

...snip...

You also need the option motherboard & frame if not already present, I'm not sure if the boards vary depending on the model (i.e. no letter, A or B).
The control panel labelling has some differences for options too.

Certainly the GPIB board does, older one plugs into option motherboard. Newer scopes with SMT A5 board, the GPIB board goes behind the A5 board (where the readout board is in older scopes).

David

[SynViks]

From looking at the price of the parts, sourcing a 2247A or 2252 might be the more practical option vs retrofitting the CCT option to a 2400 series. 

Seems like the ideal combo is to have a 2247A for general purpose and a 2465 for high bandwidth stuff, as the former has a better feature set for general bench stuff.

[AMR Labs]

...
Seems like the ideal combo is to have a 2247A for general purpose and a 2465 for high bandwidth stuff, as the former has a better feature set for general bench stuff.

In my own experience having both those models, I would have to say: ABSOLUTELY right!

[David Hess]

The measurement capabilities, especially the universal reciprocal timer counter, of the 2247A/2252 are very good.  My only experience with the 2465 series is a 2445B without the CCT so I am not entirely sure how their measurement capabilities compare.

[SynViks]

<- based on this and other videos, manual, and various comments I've come across, seems like the 2247A/2252 is better featured than the 2465B with the DMM and CT options. 2247A the measurements have real time update and are done with the scope probes.

[Runco990]

Great video on the 2247A.  I have both the 2247A and the 2465BDV.  I have to admit, I somewhat like the 2247A a bit better as a bench scope. 
So I have a general question....  which is the last and best digital or digital storage scope Tek made in this family before the TDS and TAS scopes?
Anyone?   Was the 2440 the last?

[DC1MC]

Great video on the 2247A.  I have both the 2247A and the 2465BDV.  I have to admit, I somewhat like the 2247A a bit better as a bench scope. 
So I have a general question....  which is the last and best digital or digital storage scope Tek made in this family before the TDS and TAS scopes?
Anyone?   Was the 2440 the last?

2467BHD

[Runco990]

2467BHD

Well, yes, but that's an all analog scope.  I was referring to a digital or digital STORAGE scope.  I could use this function for very slow events, or one-shots.  I have an HP and a Rigol, but really like the classic Tektronix scopes.

[OhmMyWattage]

Hi guys I’m a noob about to attempt a repair on a 2465b 57xxx. Based on the photo and serial, can someone please save me some time and tell me what caps for A5 I need? It seems a little different from the service manual I have and some of the photos I’ve seen so far, though I haven’t finished reading this thread yet. It’s gonna take some time so I’m hoping to get some parts ordered while I’m making my way through it. Looking for quality parts of course. Also, which suppliers are ideal for these components in Australia or with international shipping?

[Mark Krass]

Hi guys,

I was searching for the spare CPU for 2465B and a following point has come up:

According to service manual schematic it has Motorola MC68A02 CPU clocked by 5.0MHz source. I looked at MC68A02 datasheet and it says that max crystal freq for MC68A02 is 6.0Mhz. So far so good.

By fact on A5 board I've got MC6802P CPU installed and its datasheet says that max input freq for MC6802P is 4.0MHz. Crystal on board is 10MHz as it should be so a CPU freq input is 5.0MHz. How it could be? Where is a flaw?

Cut from datasheet attached.

[bdunham7]

So I have a general question....  which is the last and best digital or digital storage scope Tek made in this family before the TDS and TAS scopes?
Anyone?   Was the 2440 the last?

AFAIK, the 2440 was sort of the best and last of its type, although I think the reduced-feature 2439 actually came later.

[MarkL]

...
By fact on A5 board I've got MC6802P CPU installed and its datasheet says that max input freq for MC6802P is 4.0MHz. Crystal on board is 10MHz as it should be so a CPU freq input is 5.0MHz. How it could be? Where is a flaw?

You're right about the clock speed.  Perhaps your processor is a different manufacturer?  Does it have the Motorola logo on it?

It could the simplest answer that someone who wasn't paying attention to clock speed put the wrong processor in there.

All my 2445x/2465x boards have an A or B processor on them.

[Mark Krass]

You're right about the clock speed.  Perhaps your processor is a different manufacturer?  Does it have the Motorola logo on it?

It could the simplest answer that someone who wasn't paying attention to clock speed put the wrong processor in there.

All my 2445x/2465x boards have an A or B processor on them.

Yes it does have Motorola logo, see the snapshot. I double checked a soldering and it looks like factory installation. No freaking idea how it could be So the CPU is overclocked and works fine. Mistery.

I’d love to buy mil option SC67127P and swap it but no offers on ebay unfortunately.

[BravoV]

Mine, a SC67127P. 

[AMR Labs]

By fact on A5 board I've got MC6802P CPU installed and its datasheet says that max input freq for MC6802P is 4.0MHz. Crystal on board is 10MHz as it should be so a CPU freq input is 5.0MHz. How it could be? Where is a flaw?

You will most likely get an answer to this question by posting on the IO Tekscopes or Tekscopes2 groups. Some members where Tek employees and might know about the CPU/Xtal choice.

[MarkL]

Yes it does have Motorola logo, see the snapshot. I double checked a soldering and it looks like factory installation. No freaking idea how it could be So the CPU is overclocked and works fine. Mistery.
...

Interesting.  It does look very factory.

Well, it's not going to stop working at 4.01 MHz, but it may not work properly at some of the operating extremes, such as temperature and supply voltage.

You could double check that pin 39 (EXTAL) is 5MHz, but I can't imagine it would be anything else.

Maybe it's from a mis-labeled batch and Tektronix was told by Motorola?  Or maybe Tektronix qualified them in-house themselves for 5MHz knowing the scope's operating constraints?