Tektronix 2465B oscilloscope teardown

[med6753]

I wonder why they used such a complex fan instead of just putting a muffin fan in the panel like practically every other instrument out there.

Apparently some over zealous Tek engineer thought that the 2465 required a complex variable speed fan as shown in this schematic. And apparently later on it was discovered that a standard $20 computer type fan would do the job just as effectively. 

[jkn]

One thing I am nervous about though - this removal of the fan/mandrel. Are there any other tips anyone can give apart from 'go gently and glue back up if you have to'? I've read elsewhere that you can also push a bit on the mandrel, after loosening the nut a little. I also wondered about making some sort of jug to apply pressure in the right place. Can anyone give me advice or reassurance here?

Thanks a lot from the UK

Jon N

Your concern about the fan assembly is certainly understandable. If you apply pressure incorrectly you could break the fan itself. That would be a difficult repair.

I haven't heard about pushing the assembly forward slightly and I'm not sure if it would help because the mandrel is tapered. And all those years assembled has resulted in the fan and the mandrel "cold fusing" together. That's where the deoxit or some other penetrating oil would help. A long thin pair of needle nose pliers against the rear hub and carefully pry forward seems to be the best method that I found. The point is....don't be afraid of it....just work carefully and chances are it will come apart.     

Thanks for the info.

I think the idea about pushing is to push the screw (which is part of the mandrel?) back against the case, and separate the mandrel and fan that way. I guess you would best do this by supporting the rest of the fan with a thin spacer, or otherwise you'd put strain on the motor shaft and bearings.

Your idea about the thin pliers ... so you are suggesting levering the mandrel part of the rear of the fan? This is, what 1/2" across?

I have never seen a good picture of the fan & mandrel (removed), I am not sure I have got it right in my head.

Thanks, Jon N

[med6753]

Here's the best picture I have of the fan removed. I was mistaken when I said you could pry from the back hub. You can't. The hub of the fan extends internal to the case. But the thin needle nose can get between the case and the shaft piece and carefully pull forward. What you call a mandrel and I call a collet you can barely see extending slightly out from the end of the fan. When you tighten the nut it pulls forward and compresses against the motor shaft and also tightens it against the fan itself.

   

[Satbeginner]

It is similar how a bicycle steer is fixed.

To loosen: undo the nut a few turns, before any pulling!!, slightly tap the nut inwards, this will loosen the collet and the fan from the shaft.
Now pull the fan from the shaft.


To tighten: place the fan on the shaft, carefully tighten the nut. (A bit)

[David Hess]

I wonder why they used such a complex fan instead of just putting a muffin fan in the panel like practically every other instrument out there.

Apparently some over zealous Tek engineer thought that the 2465 required a complex variable speed fan as shown in this schematic. And apparently later on it was discovered that a standard $20 computer type fan would do the job just as effectively. 

Later they did replace them with standard tubeaxial fans which were noisier and lower performance.  It is difficult to improve on a centrifugal fan and their tubeaxial version was really good also.

Tektronix used that custom brushless fan motor and circuit for a long time going back to at least the 1970s.  Based on the various fan options in the 76xx series which used a shaded pole motor, I suspect they had issues with flux leakage from the fan motor interfering with the CRT until they settled on their own proven design.

[med6753]

It is similar how a bicycle steer is fixed.

To loosen: undo the nut a few turns, before any pulling!!, slightly tap the nut inwards, this will loosen the collet and the fan from the shaft.
Now pull the fan from the shaft.


To tighten: place the fan on the shaft, carefully tighten the nut. (A bit)

That is an excellent suggestion. 

[jkn]

Here's the best picture I have of the fan removed. I was mistaken when I said you could pry from the back hub. You can't. The hub of the fan extends internal to the case. But the thin needle nose can get between the case and the shaft piece and carefully pull forward. What you call a mandrel and I call a collet you can barely see extending slightly out from the end of the fan. When you tighten the nut it pulls forward and compresses against the motor shaft and also tightens it against the fan itself.

(...Image URL elided)

Yes, I saw that nice picture, thanks. I'm happy to call that a collet ;-). IIUC that part is separate from the fan moulding, in theory at least - is that right?

I was mistaken when I said you could pry from the back hub. You can't. The hub of the fan extends internal to the case.

I'm not sure I fully understand you here...

But the thin needle nose can get between the case and the shaft piece and carefully pull forward.

OK, I think.

Also see the post from 'Satbeginner' below:

To loosen: undo the nut a few turns, before any pulling!!, slightly tap the nut inwards, this will loosen the collet and the fan from the shaft.
Now pull the fan from the shaft.

That makes sense to me (I understand the 'quill' arrangement on a bike), but this approach doesn't seem to be mentioned much and I do wonder how slight a 'slight tap' really means... Is there a need to protect the motor shaft from the knock that you are trying to give the ... collet?

Sorry if I am seeming uncharacteristically cautious, I only have one of these scopes!

Cheers, Jon N

[jkn]

Update - managed it!

Well, more or less. I packed some cardboard between the rear of the fan and the chassis, then got a small-ish hanner and tapped the screw head of the collet.

It took a bit more force than I was expecting but then it loosened and I was able to remove the fan.

The collet was still holding inside the fan assembly; I wanted to get it out and may have made a slight error there. I turned the collet and it started coming out; however now that I have it out I see that it has a break in the middle, just where the metal screw stops. I am guessing this is a common place for it to break; it is not all the way through and I am pretty confident that a bit of epoxy resin will hold it OK once I have degreased the thing.

So now I am on with the next stage. I just have to puzzle out how to remove the black plastic mains shield. That should be easy, but...

Cheers, J^n

[med6753]

Update - managed it!

Well, more or less. I packed some cardboard between the rear of the fan and the chassis, then got a small-ish hanner and tapped the screw head of the collet.

It took a bit more force than I was expecting but then it loosened and I was able to remove the fan.

The collet was still holding inside the fan assembly; I wanted to get it out and may have made a slight error there. I turned the collet and it started coming out; however now that I have it out I see that it has a break in the middle, just where the metal screw stops. I am guessing this is a common place for it to break; it is not all the way through and I am pretty confident that a bit of epoxy resin will hold it OK once I have degreased the thing.

So now I am on with the next stage. I just have to puzzle out how to remove the black plastic mains shield. That should be easy, but...

Cheers, J^n

Good deal!        And yes, that's where they all seem to break. Epoxy will fix it.

And removing the mains shield is easy once you closely examine it.

Edit...and since you are in the UK and have 240V mains you should consider changing the line filter.

[siggi]

Apparently some over zealous Tek engineer thought that the 2465 required a complex variable speed fan as shown in this schematic. And apparently later on it was discovered that a standard $20 computer type fan would do the job just as effectively. 

I replaced the "complex variable speed fan" with a "$20 computer type fan" in my 2465 because the bearings were worn and the motor squealed like <insert wounded animal here>. After suffering this abomination for a couple of months, I took the trouble to find a NOS motor to reverse the mod. I'm also fortunate enough to own a 2467, so I've had occasion to compare the two kinds of fans.

So, from experience, the 2465 is as near silent as makes no difference - as is my 485, which has the same kind of fan. The 2467, by comparison, while perhaps not as loud as a 747 at takeoff, is annoyingly loud enough that it's only turned on at need.

All this to say that I take umbrage at the "do the job just as effectively" statement, and I'd wager you haven't had the opportunity to compare the two.
I'd also guess the squirrel cage fan design goes back a ways (witness the 485, introduced in 1972), and so was inherited by the 2465 as a tested design. There was plenty of other innovation in that scope for its time, and perhaps the "$20 computer type fan" wasn't all that common or inexpensive back in 1984 when the 2465 was introduced.

[james_s]

I suspect that any sort of retrofit is likely to be sub-optimal compared to designing the scope around an off the shelf fan in the first place. The large 4" fans used in many other instruments are not terribly noisy and some of them move a good deal of air. There also exist squirrel cage type brushless self contained fans although I don't know how common they were when this scope was made. I can certainly see some advantages to the fan they used, however I would have done it differently and at least used some kind of standard off the shelf motor.

Overall it's a small nag though and forgivable next to all the amazing engineering that Tek has done over the years. They have really gone the extra mile in many cases to make their instruments as good as they can possibly be. The fact that so many 30-40+ year old Tektronix scopes are still in regular use is a testament to that.

[med6753]

Apparently some over zealous Tek engineer thought that the 2465 required a complex variable speed fan as shown in this schematic. And apparently later on it was discovered that a standard $20 computer type fan would do the job just as effectively. 

I replaced the "complex variable speed fan" with a "$20 computer type fan" in my 2465 because the bearings were worn and the motor squealed like <insert wounded animal here>. After suffering this abomination for a couple of months, I took the trouble to find a NOS motor to reverse the mod. I'm also fortunate enough to own a 2467, so I've had occasion to compare the two kinds of fans.

So, from experience, the 2465 is as near silent as makes no difference - as is my 485, which has the same kind of fan. The 2467, by comparison, while perhaps not as loud as a 747 at takeoff, is annoyingly loud enough that it's only turned on at need.

I have a 2465 DMS with the cage fan that does the same shenanigans for about 20 minutes to about a half hour after initial power on. Then it calms down but is never whisper quiet and it makes itself known but isn't annoying. I do have a conventional fan built up and ready to go in if it gets worse. But for now it's going to remain in place. The cage fan in my other 2465 is always whisper quiet and I sometimes have to check to make sure it's still running. 
   

All this to say that I take umbrage at the "do the job just as effectively" statement, and I'd wager you haven't had the opportunity to compare the two.
I'd also guess the squirrel cage fan design goes back a ways (witness the 485, introduced in 1972), and so was inherited by the 2465 as a tested design. There was plenty of other innovation in that scope for its time, and perhaps the "$20 computer type fan" wasn't all that common or inexpensive back in 1984 when the 2465 was introduced.

Your umbrage is misplaced. I'm only speaking from a historical perspective and reality. When the 2465A was introduced the cage fan was dropped in favor of a conventional design. If the Tek engineers felt it wouldn't do the job they would have continued with the cage fan.   

[David Hess]

Your umbrage is misplaced. I'm only speaking from a historical perspective and reality. When the 2465A was introduced the cage fan was dropped in favor of a conventional design. If the Tek engineers felt it wouldn't do the job they would have continued with the cage fan.

I think it more likely that the Siemens motor they were using was discontinued.

[siggi]

I have a 2465 DMS with the cage fan that does the same shenanigans for about 20 minutes to about a half hour after initial power on. Then it calms down but is never whisper quiet and it makes itself known but isn't annoying.

My 485's fan did the same thing. Before you replace the motor, try flushing it with IPA, then lubricating it with something like 5W-30 synthetic motor oil. It might be enough to silence it for a couple of more decades . This is a bit of a pain to do, as if memory serves, you need to desolder the motor from the PCB it's mounted on to gain access to the thrust bearing screw. Once that's out, you can pour IPA through the motor a couple of times to flush out the gunked up old dust-in-oil.

Your umbrage is misplaced.

Maybe umbrage is the wrong word, I just feel my 2467 is unreasonably loud in comparison to the 485 and the 2465.

[AMR Labs]

What an excellent resource this thread is on 24xx scopes. I always wanted to get a 2465B, but shunned it a bit mostly to the part of having to deal with the Dallas RAM module replacement, and reading/programming it, etc. I am very experienced in electronics, but have not done much memory chip programming, so this will be pretty new to me. Now thanks to this great post I am very close to having all the info I need to go ahead and get one of these great scopes, and a list of all the post-purchase things that need to be taken care of. I currently have a Tek 2247A 100MHz as my main scope and I am very happy with it, I did all the required/recommended internal update work when I got it back in 2012. But.. of course one cannot have too many scopes, right?  (I also have a 222 and a 2013A).

I have been reading this post now for several hours straight from page 1 and been taking tech notes, so far I'm only up to page 20, but decided to jump ahead here to post a quick question. I have my eyes on a friend's-friend 2465BCT (SN 59xxx) and last callibrated in 2015 which he apparently no longer needs and I am told it is in good shape, although I have not seen it personally yet. My question is, what are the differences of the BCT version over the plain B type? I see that both are 400MHz, but I am not sure what the advantage would be to get this BCT type. I understand the CT part means "Counter Timer" (?). Does that mean that I can take automatic frequency/Time readings on any waveform in the same way I do it with my 2247A (as well as voltage readings, etc) which I find extremely useful features and would hate not to have them available on a 2465B (or BCT variant) scope.

Could one of you very knowledgeable members please explain what are the advantages or disadvantages of the BCT over the plain B version? The scope would see mainly hobby bench usage, but I would definitively want to know what I would be getting into to ultimately help make up my purchase decision. I looked around for a manual of the 2465BCT but so far only managed to find OM/SM manuals for the 2465B and 2467B in PDF format.

Thanks to all the people that have already contributed to this post over something like 6 years now. Simply amazing!

Well, its back to page 21. 

-Alex

[med6753]

The 2465BCT has Option 09, which is the Counter/Trigger. See attached. This is from the 1986 Product Catalog and it's the 2465 Option 09 but the 2465B is essentially the same. I have this option on my 2465 DMS and it comes in quite handy.



I have a copy of the 2465B Options Operating Manual in PDF if you want it. Send me an e-mail at seanfinn999@gmail.com. I don't have a copy of the 2465B Options Service Manual but Artek Manuals does have it for sale.   

[AMR Labs]

Thanks Med, the picture is getting clearer now about the BCT variant. I grabbed the JPG and that should be enough for now. I will contact you via email later on if I need the PDF doc.

But from what I now read this option 09 just -increases- the accuracy of the frequency and time measurement of the displayed waveform by using crystal-controlled reference. So my question here is, how does the plain 2465B take time/freq measurements without option 09? What kind of reference does it use? Or is the plain B not capable of taking frequency readings and display them on screen? I'm probably wrong but cannot understand why a sophisticated instrument like a 2465B does not have an internal high precision crystal-controlled reference to begin with. I understand the frequency counter on the B version is good up to 150-200MHz, so if its not going to be very accurate it does not make much sense to display the measurement numerically on screen.

As a comparison, in my 2247A everything time/frequency measurement based is related to the 10MHz internal oscillator, with an option to connect an external high precision 10MHz (Rubidium, etc) signal to substitute the internal one. By doing just two button presses I get dead on frequency (or time) measurements shown numerically on screen with 6 decimal places from a displayed waveform even it is only maybe half a division in amplitude, and up to 100MHz. Granted the internal 10MHz is not a OCXO and takes about 30 min to fully stabilize, but will be rock solid after that.

Does the 2465B also do frequency measuring like this?

The 2465BCT has Option 09, which is the Counter/Trigger. See attached. This is from the 1986 Product Catalog and it's the 2465 Option 09 but the 2465B is essentially the same. I have this option on my 2465 DMS and it comes in quite handy.

I have a copy of the 2465B Options Operating Manual in PDF if you want it. Send me an e-mail at seanfinn999@gmail.com. I don't have a copy of the 2465B Options Service Manual but Artek Manuals does have it for sale.

[David Hess]

But from what I now read this option 09 just -increases- the accuracy of the frequency and time measurement of the displayed waveform by using crystal-controlled reference. So my question here is, how does the plain 2465B take time/freq measurements without option 09? What kind of reference does it use? Or is the plain B not capable of taking frequency readings and display them on screen? I'm probably wrong but cannot understand why a sophisticated instrument like a 2465B does not have an internal high precision crystal-controlled reference to begin with. I understand the frequency counter on the B version is good up to 150-200MHz, so if its not going to be very accurate it does not make much sense to display the measurement numerically on screen.

As a comparison, in my 2247A everything time/frequency measurement based is related to the 10MHz internal oscillator, with an option to connect an external high precision 10MHz (Rubidium, etc) signal to substitute the internal one. By doing just two button presses I get dead on frequency (or time) measurements shown numerically on screen with 6 decimal places from a displayed waveform even it is only maybe half a division in amplitude, and up to 100MHz. Granted the internal 10MHz is not a OCXO and takes about 30 min to fully stabilize, but will be rock solid after that.

Does the 2465B also do frequency measuring like this?

You are actually asking something I have wondered about but nobody has been able to answer.  I have a 2247A and more recently acquired a 2445B so I can only answer part of your question.

The 2247A has the normal A and B triggers but also a C trigger and universal timer/counter IC which is used to make voltage and timing measurements continuously in real time.  The 2445B does the same thing using its A or B trigger (so the display blanks during measurement) and some form of microprocessor control which results in equally good voltage measurements but much lower resolution timing measurements and measurements are *not* made in real time with continuous updates; it makes a single measurement and then displays it along with the sweep.

I have no idea how the timer/counter variants of the 2465B series operate other than providing higher resolution and I have never seen a video showing them in operation.

[AMR Labs]

Thanks for the input David. Lets hope one of the Tek gurus on this post can answer our question.

BTW, don't you love your 2247A? I know I do!  (Disclaimer: no intention on hijacking this thread, not that it would even be possible) ;-)

But from what I now read this option 09 just -increases- the accuracy of the frequency and time measurement of the displayed waveform by using crystal-controlled reference. So my question here is, how does the plain 2465B take time/freq measurements without option 09? What kind of reference does it use? Or is the plain B not capable of taking frequency readings and display them on screen? I'm probably wrong but cannot understand why a sophisticated instrument like a 2465B does not have an internal high precision crystal-controlled reference to begin with. I understand the frequency counter on the B version is good up to 150-200MHz, so if its not going to be very accurate it does not make much sense to display the measurement numerically on screen.

As a comparison, in my 2247A everything time/frequency measurement based is related to the 10MHz internal oscillator, with an option to connect an external high precision 10MHz (Rubidium, etc) signal to substitute the internal one. By doing just two button presses I get dead on frequency (or time) measurements shown numerically on screen with 6 decimal places from a displayed waveform even it is only maybe half a division in amplitude, and up to 100MHz. Granted the internal 10MHz is not a OCXO and takes about 30 min to fully stabilize, but will be rock solid after that.

Does the 2465B also do frequency measuring like this?

You are actually asking something I have wondered about but nobody has been able to answer.  I have a 2247A and more recently acquired a 2445B so I can only answer part of your question.

The 2247A has the normal A and B triggers but also a C trigger and universal timer/counter IC which is used to make voltage and timing measurements continuously in real time.  The 2445B does the same thing using its A or B trigger (so the display blanks during measurement) and some form of microprocessor control which results in equally good voltage measurements but much lower resolution timing measurements and measurements are *not* made in real time with continuous updates; it makes a single measurement and then displays it along with the sweep.

I have no idea how the timer/counter variants of the 2465B series operate other than providing higher resolution and I have never seen a video showing them in operation.

[med6753]

Thanks for the input David. Lets hope one of the Tek gurus on this post can answer our question.

BTW, don't you love your 2247A? I know I do!  (Disclaimer: no intention on hijacking this thread, not that it would even be possible) ;-)

But from what I now read this option 09 just -increases- the accuracy of the frequency and time measurement of the displayed waveform by using crystal-controlled reference. So my question here is, how does the plain 2465B take time/freq measurements without option 09? What kind of reference does it use? Or is the plain B not capable of taking frequency readings and display them on screen? I'm probably wrong but cannot understand why a sophisticated instrument like a 2465B does not have an internal high precision crystal-controlled reference to begin with. I understand the frequency counter on the B version is good up to 150-200MHz, so if its not going to be very accurate it does not make much sense to display the measurement numerically on screen.

As a comparison, in my 2247A everything time/frequency measurement based is related to the 10MHz internal oscillator, with an option to connect an external high precision 10MHz (Rubidium, etc) signal to substitute the internal one. By doing just two button presses I get dead on frequency (or time) measurements shown numerically on screen with 6 decimal places from a displayed waveform even it is only maybe half a division in amplitude, and up to 100MHz. Granted the internal 10MHz is not a OCXO and takes about 30 min to fully stabilize, but will be rock solid after that.

Does the 2465B also do frequency measuring like this?

You are actually asking something I have wondered about but nobody has been able to answer.  I have a 2247A and more recently acquired a 2445B so I can only answer part of your question.

The 2247A has the normal A and B triggers but also a C trigger and universal timer/counter IC which is used to make voltage and timing measurements continuously in real time.  The 2445B does the same thing using its A or B trigger (so the display blanks during measurement) and some form of microprocessor control which results in equally good voltage measurements but much lower resolution timing measurements and measurements are *not* made in real time with continuous updates; it makes a single measurement and then displays it along with the sweep.

I have no idea how the timer/counter variants of the 2465B series operate other than providing higher resolution and I have never seen a video showing them in operation.

I'm afraid I can't answer many of your questions either. I'm just a hobbyist who happen to get real lucky and stumble across 2- 2465's real cheap (1 was $0, the other $60) this past 3 years. They are way more scope than I'll ever need and I've hardly scratched the surface of their features and capabilities. My speed is more like a 465 which I currently have on my bench undergoing restoration.  But I will answer what I know for sure.

The 2465 does not have the capability to tie in an external reference like your 2247A. And I doubt the 2465A & B have that capability either.

I do use the Counter display on the 2465 DMS when measuring waveforms. It's accuracy is right on par with my 2 frequency counters. And yes, on the 2465 it's supposedly only good out to 150MHz.

The other trigger features I have not used.

Again I'll offer the 2465B Options Operation Manual if you want it. You can study the theory and circuit description to see if it meets your needs. I can also send you the vanilla 2465 Ops manual too if you want to determine what sort of reference it uses. I don't have a copy of the vanilla 2465B Ops manual but it should be similar.

 

[David Hess]

The 2465 does not have the capability to tie in an external reference like your 2247A. And I doubt the 2465A & B have that capability either.

Option 1E on an oscilloscope with timer/counter option 06 or 09 adds an external reference input which accepts any of 1.000000, 3.579545, 4.433185, 5.000000, or 10.000000 MHz as a reference.

[med6753]

The 2465 does not have the capability to tie in an external reference like your 2247A. And I doubt the 2465A & B have that capability either.

Option 1E on an oscilloscope with timer/counter option 06 or 09 adds an external reference input which accepts any of 1.000000, 3.579545, 4.433185, 5.000000, or 10.000000 MHz as a reference.

OK, that must apply to the 2465A or B. I have the complete option list for the 2465 and Option 1E is not listed. 

[David Hess]

Option 1E on an oscilloscope with timer/counter option 06 or 09 adds an external reference input which accepts any of 1.000000, 3.579545, 4.433185, 5.000000, or 10.000000 MHz as a reference.

OK, that must apply to the 2465A or B. I have the complete option list for the 2465 and Option 1E is not listed.

I do not see it either and the timer/counter option for the 2445/2465 was added in 1985 and they were only produced for three years from 1984 to 1986.  I also do not see it for the 2445A/2455A/2465A/2467 in 1987 but the external reference is listed in the 1988 catalog so it became available part way through the 2465A series.

Oh, I did resolve one thing.  I found comment that the display update rate for the timer/counter in automatic resolution mode is twice per second or every four sweeps so it does update continuously like the 2247A timer/counter does.  So that answers my question; the 2465/A/B series timer/counter updates continuously but unlike the 2247A, the other measurements do not.

[AMR Labs]

I see answers so far have been mainly focusing on the external reference input BNC option 1E. But that still leaves the real question unanswered as to why a sophisticated instrument like the 2465 (inc A&B) did not come with a high precision internal reference to begin with (not to mention a standard external input), but instead the buyer had to add options 06 and or 10 to get a higher degree of accuracy.

I wonder if it is possible Tektronix was mainly focusing on developing a higher bandwidth scope, and not so much on other (useful in my opinion) features that could have been easily added? Which features? again, take a look at the 2247A real-time on-screen measurements (frequency, time, volts, etc).

If this is the case, I for one am a bit disappointed. I only rarely push my scope bandwidth needs beyond 100MHz, but I was instead looking forward that getting a 2465B would be (or almost) like a 2247A on steroids with 400MHz bandwidth and a somewhat nicer updated front panel functionality and design. Seems now to me that this is not quite the case, so unless someone can maybe disprove this, and with all due respect to the 2465, but I think I rather stay with my 2247A for now. No dying NVRam, no Hot deflection hybrid, and virtually no unobtanium parts. Oh, and no leaking SMD caps either.

[David Hess]

I see answers so far have been focusing on the external reference input BNC option 1E. But that still leaves the real question unanswered as to why a sophisticated instrument like the 2465 (inc A&B) did not come with a high precision internal reference to begin with (not to mention a standard external input), but instead the buyer had to add options 06 and or 10 to get a higher degree of accuracy.

I wonder if it is possible Tektronix was mainly focusing on developing a higher bandwidth scope, and not so much on other (useful in my opinion) features that could have been easily added? Which features? again, take a look at the 2247A real-time on-screen measurements (frequency, time, volts, etc).

Part of the answer is that the 2247A series and the 2465 series were designed by different groups at Tektronix and intended for different markets.  The 2247A is more for the service market and the faster 2465 series was more for the design market.

The release dates are also instructive:

2235/2236   1984   *
2445/2465   1984
2245/2246   1987
2445A/2455A/2465A/2467   1987   *
2245A/2246A/2247A   1989   *
2445B/2455B/2465B/2467B   1989

The custom timer/counter IC was first used in the 2236 and then the 2445A/2455A/2465A/2467 and then even later the 2245A/2246A/2247A.  So the 2445A/2455A/2465A/2467 had it before the 2247A and the 2247A is just a 2246A with the timer/counter option.

That timer/counter is also pretty good.  It includes a 100MHz phase locked VCO with random modulation to prevent errors from synchronous sources and it is a reciprocal counter so it delivers a constant number of digits independent of input frequency.  For some reason Tektronix never used it in an independent universal timer/counter.

If this is the case, I for one am disappointed. I only rarely push my scope bandwidth needs beyond 100MHz, but I was instead looking forward that getting a 2465B would be (or almost) like a 2247A on steroids with 400MHz bandwidth and a somewhat nicer updated front panel functionality and design. Seems now to me that this is not quite the case, so unless someone can maybe disprove this, and with all due respect to the 2465, but I think I rather stay with my 2247A for now. No dying NVRam, no Hot deflection hybrid, and virtually no unobtanium parts.

The 2246 and 2247A that I have also have a brighter and sharper CRTs than my 2445B.  Maybe that is just due to hours of operation but the CRT designs are very different to support their respective bandwidths.

I do not like the dying NVRAM issue either especially given how difficult the calibration is and as you note, there was a physical design issue with the big DIP hybrid packages.