Tek 485 repair and restoration

[c4757p]

So, I just got an old Tek 485 in need of some TLC. Hopefully with some effort this can become my new primary analog scope, a nice pair to my DS1kZ. I'll keep this thread updated as I go.

It's a pretty nice scope. Two channels at 350 MHz bandwidth, dual timebase, nice sharp CRT (seriously fuckoff high voltage going into this!), not too big, and has a nice and neat construction. Oh, and it's gorgeous, with that 70s Tek metallic look.

Here it is from the front as I got it. Some of the controls took a good whack - the vertical scale factor knobs are busted and missing the concentric variable knob, and invisibly from outside, the delay time pot is either loose or broken. Otherwise it's not awful, no connectors banged up, nothing seriously mangled.

front.jpg

It won't trigger, so no sense in taking a picture of it displaying a signal (I can barely get one to kinda-maybe stand still by fiddling with the holdoff). Here's the CRT flood-filled by a triangle wave though - nice, even brightness throughout, except for a burn-in of the center ground line - doesn't bother me.

crt-lit.jpg

And a preview of the sexy internals to come later....

b-trigger.jpg
crt-housing.jpg

I just ordered some "standard-replacement" parts from Mouser for it - with something this old, there are a few things I like to replace even if they're not yet broken: any caps working at 1kV or greater, tantalum caps, high ripple current aluminum caps, lamps, LEDs, and neon tubes. All of those age even under proper use.

More to come soon.

[tautech]

Nice, hope this goes well for you.

Sphere normally has a good selection of parts for Teks.
http://www.sphere.bc.ca/

[c4757p]

Thanks, I forgot about them. So far I have sources for everything I'm aware of needing, though we'll see how long that remains true

Quick update: yep, the pot's wrecked. The housing came off, and the moving wiper went missing in the process. Not sure how that happened, as there doesn't otherwise seem to be any internal trauma. Damn pot will cost about $30 to replace

I'll probably put a regular pot in there for testing until the next batch of replacement parts is ordered - I'm sure I have a pile of 3/4-turn 50k pots around here.

[tautech]

There's also the Tek CRO troubleshooting guide in the links sticky of this board.

Have you got an online manual link so we can all play along ? 

[c4757p]

Sure, here's my version to which I added a PDF TOC:

https://misc.c4757p.com/tek-485.pdf (31MB)

I'll probably post schematic excerpts if I dig into any interesting bits.

[BravoV]

   subbed, as I also have an offering locally here, but haven't make any decision yet.

Keep it coming Chris, thanks.

[c4757p]

Okay, first round of preemptive replacements, on the horizontal amplifier board. The following parts were replaced:

Reference	Original	New	Part
C1099,1130,1131,1139	Tantalum, 15uF 20V	Aluminum, 15uF 100V	Nichicon UPW2A150MPD
C1133,1134,1137	Tantalum, 22uF 15V	Aluminum, 22uF 50V	Nichicon UPW1H220MDD
C1611,1633,1646,1651,1652,1663,1664,1667	Ceramic, 1nF, 4kV	Ceramic, 1nF, 6.3kV	Murata DECE33J102ZC4B
C1612	Ceramic, 6.8nF, 5kV	Ceramic, 10nF, 6kV	Vishay 564R60GAS10
DS1668,1669	Neon, 60-90V	Neon, 60-90V	VCC A1A

This board has the CRT's HV section, which is built in those silver-metallized ceramic strips Tek used to use to build whole scopes with. These require soldering with silver-bearing solder to avoid stripping the metallization.

hv_before.jpg

Failures of high voltage ceramic capacitors in these circuits aren't extremely common, but they do happen, and tend to take out a lot of other things too, so I replaced all the ?1kV ceramics in this circuit. Neons most definitely age, so the two clamping neons in here were also replaced:

hv_after.jpg

Nothing in the horizontal amplifier itself needed replacement. This board also has the B trigger circuit, though, which had plenty of tants - got those:

btrigger.jpg

The 15uF caps are oversized just because I merged two BOM lines - no particular reason they're 100V-rated, other than that I needed 100V ones elsewhere.

Here's a full view of the whole board after completion:

fullview.jpg (off-forum because it's big)

[c4757p]

Not many replacements on the sweep board.

Reference	Original	New	Part
C861,902,911,962	Tantalum, 22uF 15V	Aluminum, 22uF 50V	Nichicon UPW1H220MDD

More interesting perhaps is the metallic-looking dust that was covering this board. Bit worrying... It got a good washing to clean that up.

[c4757p]

Inverter board. This is just the switching, monitoring and resonant circuit - the controller and output side are on a different board.

Reference	Original	New	Part
C1825,1827	Ceramic, 2.2nF 2kV	Ceramic, 2.2nF 250VAC class-Y2	Vishay WYO222MCMBF0KR
C1834,1841	Aluminum, 4.7uF, 35V	Aluminum, 4.7uF, 50V	Nichicon UPW1H4R7MDD
C1848	Aluminum, 2.2uF, 150V	Aluminum, 2.2uF, 350V	Nichicon UPW2V2R2MPD
DS1824	Neon, 90V, 300uA	Neon, 90V, 300uA	VCC A1B

Finished board:

inverter.jpg

And here is why I replace neons. This one's gone and metallized its whole glass envelope. It's being used in a relaxation oscillator, and the usual circuit (single R, single C, single neon) abuses the lamp somewhat with large pulses of current, making it prone to failure. It's in a noncritical circuit (it's just a safety indicator, it doesn't perform a real circuit function) so I didn't modify their circuit for longevity. If I wanted to, I could have added a series resistor with the lamp to protect it.

badneon.jpg

The gigantic screw-mount rectifier reservoir capacitors are bloody expensive to replace, so I didn't - I tested them instead. Nominally 430uF, with a test frequency of 1kHz, they read 522uF/50milliohm and 462uF/80milliohm capacitance and ESR. With a test frequency of 120Hz the resolution on ESR isn't even high enough to get a reading, they just read zero. Think I'll keep these.

That'll be enough for tonight

[BravoV]

Thanks for the update. 

At the inverter board, there are two TO-3 trannies, secured on the white block thingy that I guess as a heatsink, is that ceramic ?

I thought ceramic is bad to dissipate the heat, eg: compared to metal heatsink like aluminium ? 

[c4757p]

It is indeed ceramic, either alumina or beryllia. Its purpose is to transfer the heat to the chassis, not to dissipate it directly - and they are extremely thermally conductive. They feel very cold to the touch. (More than alumina usually does, I think they might be beryllia - don't want to scratch them )

Those transistors are running at full rectified mains voltage, so I guess they didn't want to play around with the electrical isolation there.

Here's the schematic for this board - they are Q1834 and Q1844.

[BravoV]

Ahh I see .. Beryllium Oxide ceramic, just be careful.

[c4757p]

Fuck this IC.

[SeanB]

Aside from the mangled threads and the trimmed leads, what is wrong with that hybrid.

[c4757p]

Nothing, it's just a sadistic design.

[xrunner]

Nothing, it's just a sadistic design.

LOL - yea really.

[SeanB]

Mmmm, I need to get this small circuit to work reliably, over a wide bandwidth. Discrete parts are not easy and are too big.  Think I will call the hybrid division and give them this, they already are doing 20 others for this model range.......

[c4757p]

I'm just kidding, dude

[SeanB]

I used to repair them decades ago, as they had some common failures in them in that design. Plus was the lids were soft soldered on, and most of the faults were relatively easy to change underrated transistors where you could solder in a new metal can packaged device and still have it work. Another was a DAC thick film network which would have the thin wires corrode off from ambient exposure, there it was replace the unit with a "second source" part.

Later on VW Golf/Rabbit/Bora ignition modules, which are quite famous for a diode inside failing open circuit, and where you could bodge in a short leaded part or use a MELF package diode to repair it.

[c4757p]

Well, there's why channel 2 doesn't work in 1M...

u48.jpg

Don't see it? Here, a bit closer...

lm301.jpg

This attenuator module was serviced - it has an official "Serviced at factory service center" sticker on it. Guess the service guy wasn't too careful. I wonder how that got through the tests I'd hope they would do. This chip is rather important in the 1M buffer:

1mbuffer.png

[KE5FX]

This attenuator module was serviced - it has an official "Serviced at factory service center" sticker on it. Guess the service guy wasn't too careful. I wonder how that got through the tests I'd hope they would do. This chip is rather important in the 1M buffer:

Is there a date on the attenuator service sticker?  That IC has a 1992 date code, so it would have been replaced long after the 485 itself was manufactured, and possibly after the attenuator was serviced.

The attenuators in some of the 485s had some annoying intermittent contact issues, so it's good that yours has been back to the factory for service.  Hopefully they upgraded it or otherwise brought it up to date.

Another tip: I see a lot of >100K carbon-comp resistors on that inverter board.  Could be a good idea to measure them to see if they've drifted out of tolerance.

[c4757p]

The service was also performed in 1992, perfectly matching the opamp.

I'm not yet aware of any Tek gear that doesn't have intermittent contact issues.

Yeah, I've been having a general poke around at things like that. So far everything's good.

[lowimpedance]

Just enough pressure to make contact to pass the retest.... perhaps 
With just a little bit more care with inserting the IC they may have felt something was not right. Still it only has 8 legs, its not a 40 pinner, how hard can it be !.

[c4757p]

Did some more routine replacements, power supply this time. Nothing particularly exciting:

powersupply.jpg

These are the output capacitors for the SMPS. I'm not sure exactly what type they are - they are much heavier than they look and appear to have glass seals; could be wet tantalum, or just really good aluminum. They're in excellent condition and test significantly better even than the replacements I chose for them, so I'm keeping these in:

goodcaps.jpg

Some diverse capacitors in here... I don't see the flat tantalum very often - those are 100uF.

colorfulcaps.jpg

[c4757p]

The high voltage section of the transformer board provides some CRT bias levels. As before, I changed out anything exceeding 1kV as a matter of principle. The original ceramic caps had cracks in the outer insulation that I didn't much like the look of.

hv.jpg

Line filter board. The design of this is a bit annoying, there's not much of a way to work on it away from the chassis.

linefilter.jpg

Not a lot to do on the timing switch board, except at the back where the fan motor sits:

timing.jpg

Thank the lord for modern DC fans with built-in drivers...

fandriver.jpg

[Addicted2AnalogTek]

Nice... very good info here. (and good work!) 

 I recently got a 485 and a bunch of probes for a super-steal.  There is an issue on the inverter board though.  I completely isolated it from everything else and it still blows the fuse.  I tested every transistor, diode, and cap on it and found that the 32V diac is bad and am waiting for the replacement from Greece... 

The 4.7uf 35v tantalums were bad, so I replaced them with mil-spec 4.7uf 50V M39003s.  The 2uf 150V cap on the inverter board and in the main PSU were also bad so I replaced them with 3uf 150V Sprague 30Ds.  I went around measuring the ESR of every single cap in it, and for some odd reason nearly every single 22uf 15V tantalum was bad. 

Just as a precaution - and because I like to have spare parts - I ordered replacements for every semiconductor on the inverter board.  I'd rather have them while I can get the original Tek replacement parts than be stuck hunting them down or substituting later. 

I'll definitely continue following your restoration and will start a thread for my 485 next weekend (after I get my replacement parts).

 

[c4757p]

Interesting that the DIAC failed - was the resistor in series with it okay? I can't see too many possible failure modes for that.

This thing seems to chew through tantalums - some have already been replaced. Hence why I replaced the majority with good quality aluminum instead. I left a few alone, and there were a couple that I actually replaced with new tantalums, but most became Nichicon PW series aluminum.

I wouldn't be too worried about stocking replacement parts for semiconductors on the inverter board, there's nothing too exotic. For the DIAC you probably could have just used something like this. The concern I have is the vast number of custom Tek chips in this thing. I was careful to replace anything that bridged dangerous voltages or currents to the custom SMPS controller, and with well chosen replacements, because if something in there failed short I'd be pretty screwed.

Good luck with yours

I'll have some interesting bits when I get out of the "replace every bloody cap" phase, too - I've got some plans for this thing

[Addicted2AnalogTek]

You know, I didn't bother to measure any of the resistors on the board, but now that you mention it there are several that have been replaced with metal films instead of the carbon comp. I'll go thru and check them during the week and report back any interesting findings (if any)

  I'm a stickler for using original replacement parts (or higher grade very similar ones) so I have ordered Ohmite Carbon comps to put back in - and that is also why I found the original spec diac even though it was $5.   

I've never had to test a diac though, so it's possible that I did it wrong.  When I checked it, it gave me a reading of about 12V instead of the expected 32V.  It would open at about 29V and then erratically drop down to about 12V, and measured differently when I switched the orientation.  The only other part on the inverter board that could be faulty is the 110V zener (which I do not have a PSU that can output 110V to test it), but I bought a couple just in case.

I'll be going thru this thread again and deciding if I want to replace other components in the scope before attempting to power it up again. 

[c4757p]

I'm not one for insisting on original parts, though to each his own. I figure the engineers who designed it would have had no problem with putting in modern higher-quality parts, if those existed at the time.

I've never had to test a diac though, so it's possible that I did it wrong.  When I checked it, it gave me a reading of about 12V instead of the expected 32V.  It would open at about 29V and then erratically drop down to about 12V, and measured differently when I switched the orientation.

Well, it should allow you to put up to nearly 32V (they aren't typically accurate, 29V is fine) across it, and then once you exceed the trip point, avalanche and drop down to, yeah, about 12V. That's about right. I'm not sure what you mean by 'erratically', though - if it didn't do that every time or something, it could have had a bad internal connection.

It should be symmetric, so if it measured differently when you switch the orientation, it was probably bad.

[dave_k]

Great work! Just a note about putting images into the body of your post. Use the [img] BB code instead of the [url] one

[c4757p]

Great work! Just a note about putting images into the body of your post. Use the [img] BB code instead of the [url] one

Then they take too long to load. Image-heavy threads where people have large embedded images directly are seriously annoying.

[tautech]

Great work! Just a note about putting images into the body of your post. Use the [img] BB code instead of the [url] one

Then they take too long to load. Image-heavy threads where people have large embedded images directly are seriously annoying.

It's very nice the way you're doing the images. 
Nice little preview/teaser at the bottom of page and a seriously good quality full page image with the link.
That's the way you want to present them and it's classy IMHO. 

[Addicted2AnalogTek]

I'm not one for insisting on original parts, though to each his own. I figure the engineers who designed it would have had no problem with putting in modern higher-quality parts, if those existed at the time.

Touche....   however there are many semiconductors that are specially selected from the batches to be used for specific part numbers.  That's mainly why I always try to go for original Tek numbered parts.

Well, it should allow you to put up to nearly 32V (they aren't typically accurate, 29V is fine) across it, and then once you exceed the trip point, avalanche and drop down to, yeah, about 12V. That's about right. I'm not sure what you mean by 'erratically', though - if it didn't do that every time or something, it could have had a bad internal connection.
It should be symmetric, so if it measured differently when you switch the orientation, it was probably bad.

Well now I know how a diac should operate.  It definitely was not symmetric.  I also worded my description incorrectly. It did not avalanche erratically, but the LED I had connected in the test circuit began to erratically flicker when the diac was allowed to draw enough current to push my PSU voltmeters to about 55V (out of 64V).  I had a 1K resistor in series with the LED. 

I think I will have a closer look at the line filter board and likely replace any caps on it as well as have a closer inspection of the passives on the inverter board.   The fuse blew immediately upon attempted power up, so I'll check for any shorts to ground (as I already should have....).

[SeanB]

I always check old carbon composition resistors. They invariably have drifted, mostly higher in value and if they run hot ( the shiny surface as the wax is driven out to the surface) they will be out of tolerance at a minimum or up to twice the value.

[c4757p]

Perhaps I will check them more thoroughly. They have drifted a bit as usual, but the ones I checked had not drifted enough to make me uncomfortable.

[Addicted2AnalogTek]

so I managed to figure out the problem with my 485.  The 430uf 200V 36D caps are bad.  I was measuring them at the wrong freq - their ESR at 100hz is about 5ohms.  Acc to datasheet it should be nearer to 300mOhm.      I can't seem to find any that will fit properly that are in the range of 390uf to 650uf. 

Anyone have any thoughts about this?  Maybe these will work(?):
http://www.ebay.com/itm/New-2-pcs-of-Mepco-3186-series-780uF-200V-large-can-screw-terminal-capacitors/141044068674?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D1%26asc%3D20131003132420%26meid%3D7de746185dcc4ab19e4ea15e873671a7%26pid%3D100005%26rk%3D3%26rkt%3D6%26sd%3D130973530166

I also bought a bunch of Vishay ceramic caps and new resistors for the psu type circuits.  The scope itself is mechanically and physically in near mint condition, so I'd def like to find a pair of suitable caps.  If anyone has any sources please let me know.

[tautech]

I'd forgotten about Tedds:
http://www.tedss.com/Capacitors

You might be better off to deal with them direct.

[dave_k]

It's very much worth checking the carbon comp resistors. Most will have drifted up in value due to moisture absorption over the decades.

Take care when replacing them with modern metal or carbon film types. Some Tek circuits specify only to use carbon comp resistors because they are non-inductive and/or can withstand high energy pulses. Usually this will be indicated on the schematic.

[Addicted2AnalogTek]

thanks for the heads up.    I didn't know that Tedss was more than just an ebay seller - good to have another parts supplier on the list.

I always replace carbon comp resistors with carbon comps (or ceramic comp) in the critical circuits.  I usually only go for MF if I can't get the value I need otherwise, and in that situation I almost always buy KOA Speer with the highest voltage and temp rating possible for that wattage. 

I ended up dropping about $100 on all new ceramic caps, neons, and CC resistors - I bought enough to replace every last one in the 485 if need be.  I just didn't realize that I needed to replace the caps because I was still testing them at 100Khz  ( I use a FG and scope for ESR measurements).   

Once I find suitable replacements I will drop a link to them in this thread for future references.

[tggzzz]

I've got some plans for this thing

If all else fails, a "Tektronix 485 Oscilloscope 350MHz - fully working" went for £246 +£15p&p. http://www.ebay.co.uk/itm/Tektronix-485-Oscilloscope-350MHz-fully-working-/182058849812

No, I'm not thinking of selling mine, I like it too much

[Addicted2AnalogTek]

I'll be selling a couple of fully restored and calibrated 485's in the near future.  I'm not really sure what I'm going to be asking for them.  It will likely be one of the later SN range (B198XXX) and one of the mid SN range (B142XXX)

[c4757p]

Eh - haven't run into any issues other than bloody time. It's just waiting to be finished.

[tggzzz]

I'll be selling a couple of fully restored and calibrated 485's in the near future.

Our of curiosity, what do you mean (and not mean) by "fully restored and calibrated"? Traceable to some international standard?

If you just mean "twiddled with the internal settings", then what would be the appropriate words? My interest is as someone that has done that with a 465.

[c4757p]

To be fair, the vast majority of scopes are "calibrated" simply and roughly enough that I wouldn't really worry about traceable cal, just cal by someone who knows what he's doing and can follow the instructions in the manual.

[tggzzz]

To be fair, the vast majority of scopes are "calibrated" simply and roughly enough that I wouldn't really worry about traceable cal, just cal by someone who knows what he's doing and can follow the instructions in the manual.

I agree when it comes to using an oscilloscope for my personal use. But the context is about making statements describing a machine's condition when selling it.

That's different, and is the reason I asked what was meant by "fully restored and calibrated", and asked for alternative wording.

[c4757p]

Well, if we're going to argue about the definition of "calibration" I'd rather it be done in a different thread, if anybody cares. I'll probably come back here with more updates on this particular 485...

[tggzzz]

Well, if we're going to argue about the definition of "calibration" I'd rather it be done in a different thread, if anybody cares. I'll probably come back here with more updates on this particular 485...

Who is arguing? I'm not.

I did ask for a clarification of what Addicted2AnalogTek meant, and what alternative words people might have used in a specific context in which words' meanings matter more than in this forum.

Of course, if you want an argument about what we are discussing, the next question is whether that should be a 5 minute argument or a 10 minute argument: http://www.montypython.net/scripts/argument.php

[c4757p]

I was merely suggesting that you might be kind enough as to see I've created this thread with a specific purpose in mind, and save the side conversations/arguments/whatever for another thread, preferably in General Chat where it's easily ignored.

[Addicted2AnalogTek]

I'm definitely interested in how your 485 restoration comes along using the PW's as replacements.   I've recently run into the time constraints issue myself. I have these suckers just lying here on the bench with my faulty 106, and not the slightest clue when I'll have enough time to really do anything more.

I never intended this to be a thread hijacking, but it appears it nearly turned into it..

I hope you (and I) manage to slow time enough to get these things wrapped up soon!!

just to close out the "calibration question", I intend to take one of the four scopes over to my friend in the engineering dept of the local university to verify I have calibrated it to manual specifications - if I ever have the time to even calibrate them.  I don't "twiddle" with the internal adjustments now that I've got a slight clue what I'm actually doing.  There's 'calibration' and then there's 'traceable calibration'.