Hey, wanna come over and checkout my O-scope, baby?

[TrentO]

To decouple my oscilloscope issues from the question concerning ground straps, I've decided to create yet-another topic. I hope I don't annoy Dave too much in my doing so... and perhaps this topic should belong in the beginner-wiener section, but here goes--

What we have here is a Tektronix 2445, 150Mhz analog O-scope.

Attached is a picture of the unit, with Ch.1 and Ch.2 probes attached to the calibrator. Both channels are set to 0.2V/Div and A&B Sec/Div are locked together at 0.1us. Triggering is set to Ch.2; triggering off-of Ch.1 doesn't produce anything.

The picture shows that the Ch.1 trace produces a barely noticeable square-wave pattern. Ch.2 produces one that is quite readable (today, anyway.) The scope has been warming up by a south-facing window (in the northern hemisphere) for about an hour to get Ch.1 to show anything _at all_. The scope has been on for about 10 minutes at the time the photo was taken.

I was hoping that someone can help me troubleshoot this problem, and tell me "oh yeah, it's the 7th fetzer valve" and that I can simply use "some 30-weight ball bearings and a quart of Quaker-State" to fix the problem. Or do I simply have the wrong settings applied? But seriously-- would it be plausible that I can fix this problem by simply replacing a cap, resistor, or by washing with copious amounts of WD-40?

As an FYI-- swapping the probes back-and-forth between Ch.1 and Ch.2 has no effect.

Unfortunately I don't have another O-scope at my disposal for testing, but the problem seems binary enough, not to require one.

If the attachment doesn't make it, I'll email it to anyone who wants to look at it. I haven't taken the cover screws off, yet, but I can, and try to take additional photos of things people would want to see.

Any help would be appreciated!

-Trent "deathly afraid of the B+ voltage" O.

[TrentO]

Hopefully Dave doesn't ban me as a spammer. Just let me know and I'll stop with the questions.

Perhaps Dave could do a EEVBlog episode on how to use oscilloscopes? And the differences between analog and digital scopes. And why it might be a bad idea to try and measure the trace of a bolt of lightning. But-- what if I have simply use a REALLY-BIG grounding strap in my lightning tests?

[Simon]

I think a vblog on using scopes would be good, while I'm fairly confident with them I'm sure theres always something to learn

[Andrew]

The Tek 2445 is a nice scope. There were many (doggy) ones on the market (eBay) for some time, because AFAIC it was common equipment in the US Army and they retired a lot of them at some point in time. Apparently including every broken unit they had still in some storage.

Get the service manual from http://bama.edebris.com/manuals/tek/2445/ and follow the repair procedures. Be careful, there are some deadly voltages inside. The service manual points out the dangerous areas and safety procedures, and they mean it!

IIRC there are some parts in a 2445 that like to break but are unfortunately almost impossible to get. The hybrids are, for example, Tektronix custom parts, not manufactured since ages. (Tek sold their hybrid manufacturing to Maxim, and Maxim retired them decade(s) ago). If one of these parts is broken your best bet is to buy yet another (broken) 2445 and use that for spares. However, you'll find that you could get a new DSO for a similar price.

[TrentO]

Thank you to Andrew and GeekGirl for providing links to the manuals.

After letting the scope sit there for a few days, just to make sure its not holding on to some massive charge, I took the cover off. After careful inspection, I really couldn't find any obvious issues-- I was hoping to find a charred resistor, exploded cap or something... but everything was really clean and appeared to be in order.

Based on my amateur troubleshooting, I believed the problems were limited to the analog-input sections on Ch.1 & 2; unfortunately they were contained in some pretty heavy-duty cast-aluminum heatsink / shielding 'pods' and didn't look accessible without some desoldering work. I think Andrew was speaking from experience when he said that parting-out another scope might be the only way to repair these vintage units-- peeking through a little window on these pods, I couldn't see a single discrete component inside-- it looked like a massive 'silicon chip,' with circuits that appeared to be masked on to some kind of ceramic substrate.

Poking around, I also noticed that all of the boards have "Tektronix, (c) 1980 or 1981 on them-- this sucker was designed when I was in 3rd grade-- who knew? It looked newer than that.

Andrew's link to the "Boat Anchor Manual Guide"-- really helped to set some perspective... and that perspective is-- Its not worth it.

I just ordered a Rigol 1052E from the HappySeeds dudes on eBay. And no, I didn't buy it to further troubleshoot my Tektronix.

-Trent

[Andrew]

A working 2445 is a really good analog oscilloscope, it doesn't matter that it was designed almost 30 years ago. The Tektronix guys knew what they were doing. So a broken 2445 is worth a serious repair attempt. You can only decide if a repair is possible after you found the fault.

And yes, I own a 2445 which I once had to repair (the low voltage power board blew up, lots of old dodgy caps there). My normal oscilloscope is a DSO, but I haven't retired the 2445.

You won't get such a good and complete repair instruction for a Rigol DSO compared to what you have for the 2445. Go through the step-by-step diagnostic diagrams to isolate the fault, and then measure the reference voltages (all the way observing the safety instructions in the manual) to figure out where things start to go wrong.

And don't mess with the "pods". These are the hybrids. Spare parts are almost impossible to get.  Leave them alone. All you can do by poking at them is breaking them. If your fault is not in a hybrid you would just convert a feasible repair job into an unfeasible.

[TrentO]

Andrew-- well CRAP! You're not making it easy for me to let it go.

You REALLY think this thing is worth fixing? I should be getting my Rigol in a few weeks-- cheap Chinee, for sure, but then again since the Tektronix is an OEM Sony unit, the Tektronix could be considered cheap Japanee.  

I'll look through the repair manual in detail over the next few days, and take another look at it then. Would you know, off-hand, if its possible to adjust the amplitude on Ch.1 and Ch.2? Ch.2 appears to be two orders of magnitude higher than Ch.1, with Ch.3 and Ch.4 somewhere in between.  

Did you ever notice that the first thing the manual talks about is to never service the unit alone? To always have someone else nearby that can administer first aid (I assume CPR?)

I'm not sure if my cat would know what to do... Except say--- "Uahahahah-- ALL THIS IS NOW MINE!!!


-Trent

P.S.: Would oven mitts help?

[Andrew]

Sony? So you have one with a 3xxxxxx serial number. Sony and Tektronix founded a joint venture in 1965. Oscilloscopes produced in Japan were not OEM products, but genuine Tektronix products manufactured in Japan by that joint venture.

Regarding the value of a working 2445 see eBay:  http://shop.ebay.com/i.html?_nkw=Tektronix+2445

I don't think the seller asking for US$ 649.99 will get that, but the US$300 to US$450 range is normal for a 2445 in good condition. The ones for US$200 are likely broken, or as they say in eBay-speak "AS IS", "not tested", "needs attention", "no tools to test it". Outside of eBay used equipment traders sometimes ask for US$800 for a refurbished one, but that includes a warranty by the trader and calibration.

Page 223 (09-12) and 225 (09-13b) are maybe the most interesting ones for starting to find your fault. Page 223 lists the test configuration and the voltages to measure. Page 225 is the schematic of the input stage and pre-amp, showing where to measure the voltages. The fault you describe is not likely something that is be fixable with a simple "amplitude adjustment".

And yes, there is dangerous voltage in such an oscilloscope and it can kill you. And oven mittens won't help. Only that the coroner might scratch his head what these were supposed to be good for.

[TrentO]

Andrew--

Roger on the ineffectiveness of oven-mittens. Affirmative on the 3xxxxx serial number.

You REALLY do know a lot about these scopes!

Besides the calibration stickers, everything else says Tektronix (c) 1980/1981 Made in USA. Not much of an OEM, if you ask me....

Thanks for the pointers above-- I'll keep you posted.

-Trent "the term 'clever, those Japanese' was never used in reference to me" O.

[TrentO]

Andrew--

Parsing through the "Performance Check and Functional Verification Procedure" section, its apparent that several pieces of equipment would be required to (re-) calibrate or further troubleshoot this scope-- equipment that I just don't have (except for item #18, "Digital Multimeter.") It has also become abundantly clear to me that I couldn't just 'fake-it,' and adjust some pot somewhere-- the external references, I guess, have to be exact. Until I acquire-- 1. Variable Voltage Supply, 2. Leveled Sine-Wave Generator, 3. Calibration Generator, 4. Function Generator, 5. Time-Mark Generator-- do you think it would make sense to send this unit out to have 'professionally' calibrated?  Of course there's no guarantee that the problems that I see are merely a calibration issues-- the v/div differences are so large between the channels, its hard for me to believe that.

I think I'm going to threaten it with a bat, and see what happens. Perhaps all it needs is a 'Samoan tune-up?'

-Trent

[Andrew]

I don't think a calibration will fix your problem.

[TrentO]

I think you're right about that-- I went back and documented all of my tests, and the trend is as follows--

After a 20-30 minute warm-up period, with the scope set at all recommended settings for calibration, probes attached to the Calibrator, Ch.2 is way off scale on .5v, .2v and .1v v/div ONLY. 5v, 1v, 50mv, 20mv, 10mv, 5mv and 2mv are accurate. I noticed that this seems to correspond to a series of relays 'clicks' entering/leaving that range. I would assume that these relays are contained in the attenuator hybrid units... I also assume that a resistor is being switched-in when the ch.2 div knob is rotated CW from 1v to .5v... Page 225 shows the "Auxiliary Control Register" section controlling these relays-- I'm in the process of trying to figure out what outputs are generated through IC U120 based on vDIV and/or coupling settings.

Prior to the 20-30 minute warm up period, Ch.2 will show what I can only describe as a 'sharks tooth' wave pattern in the .5v, .2v and .1v v/div range, also off-scale high. After the 20-30 minutes, it settles into a nice square-wave (but still unreasonably-high amplitude.)

Ch.1's amplitude remains low, over the complete scale-- the trigger won't pick it up on the calibrator input until 20mv v/div.

All test point voltage levels are within 1% of spec.

So still, based on my observations, I believe the attenuator for Ch.2 is bad, worse when cold; Ch.1 suffers from generalized attenuation issues.

My Rigol 1052E shipped today from China-- assuming that Customs doesn't hold it due to my H-bomb comments earlier, my first task would be to validate the test waveform exiting the attenuator hybrid.

What a DORK, huh? Buying a new, cheap-o oscilloscope to fix another pre-Madonna scope. <--- get it ?!! 

-Trent

P.S.: Andrew, I owe you a beer (or two.)

[TrentO]

Thank you to all that responded, especially Andrew!

I got my Rigol 1052E today, and after taking it apart (and putting it back together again) I used it to once again troubleshoot my aged Tektronix 2445. As suspected, the pre-amp unit on Ch.1 is bad, and the analog-switching unit on Ch.2 is bad. Either way, I did confirm that both Ch.1 and 2 have problems. Although this Tektronix was essentially given to me, I feel compelled to give it back to the Italian-guy source. Otherwise, I would have put a bow on it and shipped it off to Andrew for spare parts!

What I learned from all this-- 1. They don't make 'em like they used to. 2. There's a reason why they don't make 'em like they used to....  <--- that's actually a joke. My pre-Madonna Tektronix has craftsmanship beyond belief-- as finely engineered and tuned as an Omega watch; and most of it was probably built by hand! My new Rigol is much more advanced, but assembled worse than my Black-and-Decker toaster. Analogous to that cheap $5 no-name digital watch that keeps better time than any mechanical Omega, the Tektronix 2445 does have it's place, but the Rigol is simply better suited to my needs. But then again, if I set my watch down for more than 30 hours, it simply stops. The $5 Timex "will take a lickin' and keep on tickin" for years and years.

I did learn a lot doing this and feel a whole lot wiser for going through the process, inherent risk of electrocution aside. Thanks, everybody!

-Trent

[TrentO]

Not that anyone gives a rat's-ass, but I fixed my Tektronix 2445 today!!!

Contrary to Andrew's and the Tektronix Service Manual's recommendations, I disassembled the "Hybrid" assemblies and cleaned them-- Ch.1 actually had a moth cocoon-case inside, limiting the travel of one of the relays. Ch.1 now works fine!

Ch.2's Hybird was fine, but I noticed someone along the way dropped a solder ball onto the mainboard (next to it)-- it looked like someone replaced the Ch.2 hybrid before, but screwed up in the process-- it was the only crappy re-soldering job on the entire unit. So 1/2-hour with solder-wick and Ch.2 now works fine!

Ah yes!!! The sweetness of having fixed something yourself--

Sorry-- my camera focused in on the probe wires and not on the screen-- but for the first time in who-knows-how-long, the traces from all four channels are in full sync!  As Dave would say-- "Whoo-hoo!"

I originally didn't plan on taking it this far, but I had nothing to lose-- it was either trying, or taking to the range to shoot-up.

Thanks for everyone's help.

-Trent

P.S.: Notice my shiney-new Rigol in the background-- what I used to fix my Tektronix!

[EEVblog]

Sweet!, nice work!

Dave.