Tektronix 2465B oscilloscope teardown

[Satbeginner]

It comes out as a separate unit, see in this picture.

[MarkL]

My 2465 doesn't have dimly lit indicators either.  If you examine the schematic, the LEDs should be completely off since they are each driven by a single output of the 74F164's

Can you get the TRIG'D LED to light with normal brightness?  It shares the +5VD with the other LEDs but is driven to ground by a separate signal line.

Maybe the 74F164's are without a ground connection?  Just a guess.  It has to be something in common to all the LEDs.  james_s has an interesting suggestion to try to force all the LEDs on.

To capture the entire 48 bit sequence, you should be able to trigger on the LED CLOCK with the TDS2004B set to single.

[MarkL]

Actually, in looking more closely at your front panel photo in :

  https://www.eevblog.com/forum/testgear/tektronix-2465b-oscilloscope-teardown/msg1161319/#msg1161319

Would it be accurate to say that *some* of the LEDs are normal brightness?  For example, Ch1 and Ch2 "50ohm DC", all of the MODE LEDs, trigger READY, x10 MAG, plus some others.

Many of the LEDs share pullups to +5VD.  This is ok in normal operation since not more than one is expected to be on at a time in some of the groups.  But in this failure, if all LEDs are on via the 74F164's, some groups will appear dimmer since they're wired to a single pullup.

The LEDs I've named above have their own dedicated pullup, so if they are on they should be normal brightness.

[peter-h]

"Can you get the TRIG'D LED to light with normal brightness?  It shares the +5VD with the other LEDs but is driven to ground by a separate signal line."

Yes indeed I can; the manual says this one is driven directly.

If I bang the front panel, the brightness of the LEDs all changes. So it does look like a ground or some such has come off somewhere.

[james_s]

Download the manual for the DSO and sit down for an hour to learn to properly use it, that is time well spent and will make your life much easier. At least work out how to take single shot captures and set up the triggering. The TDS2004B is not particularly high bandwidth but it's still a very capable instrument that is well worth having and once you are proficient with it you will be able to sit down and use pretty much any other Tek DSO. Single shot captures are fantastic, you can use them to check the rise time and overshoot of power rails at turn-on, capture transient events, capture and analyze bitstreams like the LED data we are looking at here, once you learn how to do that you'll wonder how you ever got by without it.

Edit:
Just saw your latest reply. If tapping on the panel has an effect then something is definitely loose. Pop off the bezel and pop out the whole panel, usually there's a clip you push in on the side and it just pops out. Poke around with an insulated stick and see if you can find a cracked solder joint. I think you're very close to a solution here.

[peter-h]

I got the panel off



but to pull off the PCB so I can access the components on top of it I need to use an allen key on the large knobs. I have that but can't get the large (timebase) knob off even if I remove the grub screw completely.

If you look closely at that PCB you can see somebody has been there before, resoldering a row of joints along the centre of it.

[Satbeginner]

My time base knob has two Allen bolts......

[james_s]

You may not need to get the board off, look closely for cracked solder joints, start at the header where the ribbon cable plugs in and then follow the clock and data lines to the shift register. I suspect the clock is floating and picking up noise, randomly clocking at a much higher rate.

[peter-h]

I am stuck with the two smaller knobs (volts/div). They need Imperial allen keys which I don't have. So I made one by grinding down a metric one.



I should be able to plug these two PCBs into the processor board, presumably I hope they don't need additional grounding.

[peter-h]

Well, I did that



and while the pots and switches work, the LEDs are all dark.

A quick check shows ground continuity to the 74F164s but their +5V is not joined to the +5V of the HCT chips on the CPU board. There is about 200 ohms between those two +5V rails. Maybe that's intentional.

I need to be very careful now. One slip-up with the thing powered and it will likely be a write-off.

For some reason I can't find the schematic of this PCB... My scope is B057050.

[peter-h]

I think I found something:

The two +5V rails ARE supposed to be the same one



but there is about 1k between them. That is pin 20 of P652, and even across the cable itself it is broken.

[peter-h]

JOB DONE



You can see the yellow wire bringing the +5V from the CPU board, in the top right.

Now I have to put this damned thing together

THANK YOU ALL!

I must confess to reading the RTB2004 thread but the pricing is crazy. I also really love analog scopes. Mostly I have used Iwatsu which just keep on going... like this logic analyser



which was purchased new for about GBP 8k c. 1985. The two probes were stolen but I got the schematic of it and built new ones which plug into the IDC connector in the front.

[MarkL]

You can see the yellow wire bringing the +5V from the CPU board, in the top right.

Now I have to put this damned thing together

Great; glad you got it working!

But did you find the actual break or is the jumper your final solution?

If you haven't found the break, I'd really recommend that you do.  If it's the ribbon cable then replace it, or if it's a broken trace somewhere you should find out why it's broken.  Murphy says it will come back to get you at the most inopportune time.

You had mentioned previously that the AC/GND/DC/GND/DC push buttons didn't seem to work.  Are they ok now too?  Once you reach the front panel, the buttons are electrically separate from the LEDs.  The ribbon cable is one thing in common.

[peter-h]

The break was within the ribbon cable. It was at one end or the other; the IDC connection was obviously marginal (as they often are). It was not a PCB track break. I probed it on the cable itself and it was totally open. So the front panel was getting some current into its +5 rail via the other signals!

I don't have a spare 20 way cable so I just put in that +5V wire.

It is all working, including the AC/GND/DC/GND/DC push buttons. They were working all along; just very confusing without the LED indication.

One unrelated thing: is there any adjustment for the focus between the upper and lower part of the CRT? This photo doesn't show it too well but the upper stuff is less focused



I doubt there is. In the 1970s I built some oscilloscopes, using ancient ex-mil CRTs and finally with a Mullard scope tube, PDA, and solid state deflection plate drivers. I recall using the uA733 differential video chip, which still exists! More than about 20MHz was nontrivial though. But there was never any control over the focus, although in principle one could modulate the focus voltage according to the Y deflection plate drive. I think TVs did that but they used very wide deflection angles.

[james_s]

Just make a new cable, you can buy the connectors for very little and a piece of ribbon cable, I use a bench vise to crimp them, takes just a few minutes to make a cable. Glad to hear you tracked down the problem though.

Yeah DSOs are expensive, but I was under the impression you already had the TDS scope? I've always been an analog scope guy myself but since acquiring a really good older Tek DSO it has become my primary scope. Being able to capture and examine waveforms is just incredibly useful.

[MarkL]

One unrelated thing: is there any adjustment for the focus between the upper and lower part of the CRT? This photo doesn't show it too well but the upper stuff is less focused

There's some interplay between the front panel ASTIG and FOCUS controls.  I would try those first.

There's also a number of internal adjustments that are described starting on page 5-7, including an Edge Focus control which may help.

But be aware that messing with some of the CRT adjustments can alter where the beam lands in relation to the graticule.  Once you reach an optimum display there are later cal steps that recenter everything if needed.  I would carefully mark everything with a fine Sharpie before moving any adjustments in case you need to get back to the beginning.

Almost all the adjustments interact with each other.  It will drive you crazy trying to get it perfect.  I know it took me a couple of hours until it was "good enough".

[David Hess]

One unrelated thing: is there any adjustment for the focus between the upper and lower part of the CRT? This photo doesn't show it too well but the upper stuff is less focused

The 2465 has an edge focus adjustment so at the least it should be possible to even the focus out somewhat.  The focus also tracks the intensity and there is an adjustment for that but I do not think it is adjusted for beam position because that would be difficult to do at such a high bandwidth.  The theory section in the service manual should say.

Yeah DSOs are expensive, but I was under the impression you already had the TDS scope? I've always been an analog scope guy myself but since acquiring a really good older Tek DSO it has become my primary scope. Being able to capture and examine waveforms is just incredibly useful.

I have always understood the advantages of a DSO but would rather have a reliable instrument that I understand well than a list of questionable features marketing checked off that I can do without so my go-to DSO is a 2230/2232 unless I need higher bandwidth.

[peter-h]

"Just make a new cable, you can buy the connectors for very little and a piece of ribbon cable, I use a bench vise to crimp them, takes just a few minutes to make a cable. Glad to hear you tracked down the problem though."

I have put it back together now, but the problem is that IIRC that cable is soldered at the front panel end, and with a PTH board there is a very high risk of pulling off a pad on the topside, unless one can melt all 20 at the same time. I have a turret mill and could machine up a copper block with 20 holes on a 0.1" grid... Anyway next time I will know where to look. You people have been incredibly helpful. I co-run a forum too (EuroGA) and it isn't an easy job!

"Yeah DSOs are expensive, but I was under the impression you already had the TDS scope? I've always been an analog scope guy myself but since acquiring a really good older Tek DSO it has become my primary scope. Being able to capture and examine waveforms is just incredibly useful."

The TDS scope is horrible. The user interface is slow. You turn a knob and it takes close to a second before anything happens. This applies to the commonly used stuff like vertical position. Who designed this (I've been doing HW and SW since ~ 1975) should be shot. Maybe it uses a 4004 running Java? And it wasn't that cheap - about GBP 1000. It was bought for vibration measurement in a light aircraft - http://peter2000.co.uk/aviation/vibration/index.html - where I wanted a portable scope but the battery powered ones (well, the ones which wer not near-useless) cost a lot more. Also the screen resolution is rubbish so narrow pulses show up as a line of pixels hanging down from 5V, sometimes all the way down to 0V and sometimes halfway down, with the latter indicating that there is a short or bus contention etc. It's a toy... I hoped it would be generally usable but it isn't.

If I was buying a DSO it would have to be 200MHz at least and have a really high-res display. The RTB2004 looks like it would be nice but who knows whether the UI is fast.

Of course the storage aspect of a DSO is great. But then I see the way they load up the price if you want data analysis options This is off topic now - there is the long RTB2004 thread. I wonder if that one can be bought as a base model and hacked to 300MHz?

BTW does anybody want hi-res (15MB) photos of the innards of my Tek scope before I put the covers on it tonight? I can put them on a FTP site, permanently.

[peter-h]

One more Q: is the fan temperature controlled? It is very noisy. The supply voltage was measured at 9V so it is a 12V Nippon fan, but still very noisy. I routinely replace fans in PCs with lower noise versions, and the old HP/Tek gear all had very noisy fans.

The fan used is 80mm standard size but slightly unusual being 20mm thick (today most are 15 or 25mm).

[james_s]

It sounds like you have the DSO set to a low horizontal rate or a high capture rate, this will reduce the waveform rate so the waveform on the screen does not update until the next capture is ready to be displayed. There should be a small indicator on the edge that moves in real time though, then the waveform will appear in the new location on the next update. This is just how these scopes work, it's fine once you understand why it behaves the way it does. Mine is 1GHz but honestly most of the time 60MHz would be enough. It's way more than enough for diagnosing things like the LED problem you just dealt with. If you really don't like the DSO I would happily take it off your hands

Yeah I forget sometimes that not everyone has good desoldering gear. I use a Hakko 808 vacuum gun for this sort of thing and it makes it trivial to desolder stuff like that. I could pull that ribbon cable in under a minute with very little risk of damage but it's the sort of thing that is expensive if it isn't something you would use frequently.

I'm not sure about that particular scope, but none of the Tek gear I've worked with has temperature controlled fans. I suspect you'd be just fine replacing it *if* you are certain you get something that moves as much air. Those scopes are full of fancy hot-running hybrid ICs that are made of pure unobtanium.

[peter-h]

The fan in my scope is a Nidec D08A [D08A-12TH] and the data sheet is here: http://www.nidec.com/en-Global/product/fan/category/F010/G070/P2000261/

The spec is actually not bad at all, for noise and airflow. I guess it is noisy because it is mounted directly to the metal chassis without any vibration isolation.

I will happily have a play with the DSO - thanks

[james_s]

It's possible the fan is making more noise that it should be. Have a close look at the blades for built up crud, I've seen fans get so much dust and grime collected that it throws off the balance and they vibrate.

[peter-h]

Well, guess what? I found the fan is broken. One of the three plastic "legs" which hold the motor in place was cracked. Something hit it, perhaps in transit. Easy to repair with epoxy.

Would Tektronix have used a Japanese fan which is still in production today, 22 years later? I reckon somebody changed it. This scope has been repaired. Somebody has resoldered the joints to the three shaft encoders / multiposition rotary switches. I paid GBP 2k for this scope some 10-15 years ago. The cosmetic condition was (and is) as new and the internals are spotless.

[james_s]

I'm not sure what fan Tek used in the 2465B but my TDS300 and TDS400 series scopes all have a Japan Servo "DC Centaur" CUDC12D4 fan. The service manual for the 2465B probably lists the part number.

[peter-h]

With a repaired fan, and the fan retaining plate modded (by bending) so it doesn't push against the fan hard (that pushes the fan motor against the scope chassis) the fan is much quieter. The motor was rubbing against the chassis.