Tektronix 2445: My first vintage oscilloscope repair attempt

[@Unfinished80]

I've dealt with this in industrial equipment. The copper tangs under the relays lose tension over time. When they are new, the tangs are bent out far enough so that they will bend a little angainst the contact pads to maintain connection. However, if the relays are left in the same position most of the time and that coupled with time, temperature cycling, and vibrations cause the tang to deform and lose tension gradually until they are only floating on the contacts. This is why the paper shim was able to restore a little connection, by pushing them slightly closer together. If you want a solid connection, you'll have to re-bend the tangs on the relays. That is much easier said than done but not impossible. Perhaps you could find a tiny shim to go behind each of the tang legs? Good luck and thanks for another great video!

[@Damaniel3]

All of the pre-1990 Tek scopes are masterpieces of chassis design and circuit layout. The early tube-based CRT scopes are especially nicely laid out.

[@knifedude97]

They really are.. my profile picture is from a tubed Tek. I still cant believe how "simple" they look when the cover is off.

[@MaDmanEXE]

What happened in 1990?

[@Damaniel3]

@@MaDmanEXE Nothing in particular - the boards were still nice, but by then it was mostly ASICs and surface mount components, so there wasn't as much as aesthetically pleasing as before (when through hole components, larger individual IC counts, and even vacuum tubes dominated). That's not to say the board quality was poor in later devices. When I was working there (2000-2020), I can attest that the hardware engineers took their work extremely seriously, and they always used high quality components throughout even if the designs were fairly pedestrian by the standards of the 1960s-1980s models.

[@Edisson.]

@@Damaniel3 Yes the build is phenomenal, I have replaced several Tektronix oscilloscopes in my life, I currently have two in my home lab. Both fully restored, two channel portable model 212 with completely redesigned source section - no need to use battery as a filter - connected as a waveform tracer display unit - year of manufacture 1972 and base model 2465 with no additional digital functions - no need to worry about a battery backup because it doesn't have

[@larrybud]

It's nice to have physical buttons for functions, but obviously the new digital scopes do so much that it is just impossible to have all the settings as physical buttons.

[@jp-ny2pd]

I worked for an electronics engineer for a while in high school. He mainly used digital scopes but he always had that one bench in the corner with all the analog scopes on it. He would throw a sheet over them to keep dust off. I asked him why he keeps them and he just said "For when things get weird" and walked away.

[@ChrisB...]

I bought a pallet of about 20 of these scopes in the early 90's for $5.00. I didn't know what to do with them so I posted a notice at my university's engineering building and ended up selling most of them for $50-$100 each. Pretty nice return on investment! I miss old-style government auctions, mostly because people didn't know about them and you could get great stuff extremely CHEAP!

[@Renville80]

One time I got a pair of Amiga 1200 tower conversions for like $20 at the local city / county surplus auction, but after realizing I couldn't do much with them (several key parts were missing), I put them up on eBay hoping to at least get my money ack, and was surprised at how much they went for - MORE than recovered my original investment.

[@paulmichaelfreedman8334]

@@Renville80 I got a Laserjet II in 1992 for $10, from a company that was replacing their printers. Too bad such bargains only come along a couple of times in a lifetime.

[@RS-ls7mm]

@@paulmichaelfreedman8334 Unless you have access to an amazing dumpster like EEVlog had.

[@jamesdye4603]

Live auctions were fun, now everything is done online. Most of the old government electronics are bought up by recyclers.

[@Andrew-rc3vh]

@@paulmichaelfreedman8334 That was long enough ago when the ink cost less than the printer.

[@krandall5285]

The way I was taught to clean those gold contacts is with a small strip of white printer paper and IPA. I have serviced quite a bit of old Tek equipment. I love Adrian's show, but I would strongly advise against scratching up the pads with a steel implement...

[@russellhltn1396]

If the gold is just a coating, it's risky to try and starch it.

[@Brian_Of_Melbourne]

I too cringed when he started 'scratching'.

[@theantipope4354]

Back in the day, I always used the rough end of an ordinary rubber eraser to clean PCB edge connectors.

[@Brian_Of_Melbourne]

@@theantipope4354 I would suspect that introducing more contaminants into an attenuator like that isn't a good thing. Just Isopropyl alcohol (AKA IPA) and a Q-Tip will be fine IMHO.

[@geertkroezinga6003]

You were taught correctly... I usually clean them with ethanol, IPA or KontaktWL. Scraping them like that will for reduce the surface where the two contacts meet and maybe even remove too much material for it to conduct at all.

[@rayburke4386]

Hi Adrian , Ray Burke here , I use to work on the Tektronix Portable scope line. Most of the gold contacts shouldn't be scraped with tweesers, Tektronix in their manuals usually suggested using Isopropal Alchol and a this peace of paper in the service manual unsure about Deoxit? The problem with the CH 1 and CH2 Attenuators may be the outer mounting ring screwed to the Attenuators, using a thin open end wrench to carifully thghten them, they might be loose from connecting and disconnecting the scope probes. This could cause the signal to distort. You may contact the Tektronix support forums or the Beaverton Museum.

[@rayburke4386]

Adrian you could use a scope on the Ch1 and Ch2 input center pin for a square wave trying the ground on the outer BNC ground. Then try on the chassis ground, also use a DMM between the BNC and chassis ground. You might be able to probe with a scope through the attenuator, after proving the input waveform isnt distorted. Not through the dividers. You could get help from Tektronix of through some local scope repair centers in Beaverton. They might help with obtaining parts, Their is a source of Tektronix parts Q source Google it.

[@rayburke4386]

Adrian Ray Again I noticed the input BNC connector if it doesn't have a BNC nut on the inside (Like standard BNCs). Then the BNC ground is through the Allen screw on the bottom of the Attenuator next to the BNC, and it is probably loose. Which is what connects the BNC to the Attenuator chassis. Try tightening the Allen screw, cheap fix. I noticed the screw when you had the Attenuator upside down. Good Luck Ray.

[@user-lp3cf5yn5b]

Mr. Ray, can you help me diagnose a 2445 I just bought on eBay? I don't have it in hand yet but I'm putting out feelers already for people who may know what's going on with the scope once I have it.

[@rayburke4386]

@@user-lp3cf5yn5b Ray here, I have knowledge of the Tektronix Portable Scope line. I know lots of things about the analog and some digital scopes. I am not an expert on the high end 2445 scopes, their are more complex than the lower end scope line. You will have to get ahold of the full service manual, or find a manual on the Tektronix WIKI or forums. Do you have any experience with these scopes, DMM's, soldering irons, other scopes for troubleshooting? If you don't, I would suggest you buy a refurbished scope that is fully calibrated. Or just get a new digital scope, Rygol or others, they also come with probes. Good luck, Later Ray

[@rayburke4386]

Hey Adrian when pushing on the CH 1 and CH 2 attenuators on the outside of the probe ground, which was intermittant. The problem as I said before was the ground connection through the outside grounding Allen screw on the bottom of the chassis is loose. Which is causing the signal to loose its ground connection to the Attenuators. When you removed the Attenuators from the scope, did you loosen the Allen screws, next to the BNC's? If you didn't touch them, that could be your intermittant ground problem. Possible using a DMM and wiggle the BNC to the chassis. Good luck from Ray.

[@keithnoneya]

It time index 3:40 you talk about the duration using the coursers to figure out the frequency. If you push both buttons, the Delta V and the Delta T, at once it will give you the frequency.

[@johnglielmi6428]

Tektronix are one of the finest made precision instruments ever made. These things were built to last.

[@andrewlittleboy8532]

The inside is a work of art, absolutely beautiful layout.

[@rimmersbryggeri]

I think bending up the relay contacts will do the trick 100% since they are probably work hardned in the position they are now.

[@jimengr]

Yep.

[@dennisbartello7682]

This. I had worked on a Yamaha DX7 using these style tangs for its keybed, and just gently bending said tangs back did the trick. If you do that with the relays, be as gentle as can be.

[@adriansdigitalbasement]

Yeah others suggested that too. I may try it out as I have nothing to lose at this point... the whole assembly is insanely fragile, that's the only issue. Bending them the right amount seems easier said than done.

[@darkstatehk]

I chuckled when I saw the title of this video. Anyone that has worked in or familiar with the calibration industry will know how insane the electronics in instruments like this were. Big kudos to Adrian on this video tackling the issue! I was a young apprentice in the early 90s and I remember looking inside Tektronix scopes back then and feeling intimidated by the precision engineering. Although some might say an "easy" fix, there was a precise and logical approach to the solution. Hope that's all it was! Recently, my brain was blown by the fact that NASA had surface mount technology in the 1970s. I think I saw that somewhere on CuriousMarc's channel.

[@tekvax01]

it was the 60s in the AGC.

[@darkstatehk]

@@tekvax01 Thank you :) 😻

[@darkstatehk]

@@tekvax01 I just went back to the Apollo Guidance Computer Part 1, and yes, 8 minutes and 50 seconds on that AGC video is exactly what blew my mind, thank you for correcting me on that. That right there was something so special. We're so lucky to be able to see that footage and again I bring it back to Adrian's video here in that attempting repairs on precision tech from the past will certainly teach you a lesson you might not have thought possible.

[@captainchaos3667]

That thing is forty years old. It's _amazing_ how well put together it is, and how good it still looks.

[@JayWye52]

it's amazing the horizontal sweep still works,the H output amp IC was a common failure,a bad batch from Maxim. NO replacement available,its a "TEK-made" IC.

[@campbellmorrison8540]

Thats a lovely scope. Everybody needs an analogue scope and a digital one, just like your multimeter :) Just a little warning, I used to make thick film circuits and if you are soldering to them you should use high silver content solder or the 60/40 will absorb the silver from the film traces

[@AaronHuslage]

You should call the Vintage Tek museum in Beaverton. They will know how to fix this. Maybe do a field trip! It’s also a fabulous place to visit for sure.

[@ChrisJackson-js8rd]

definitely do a field trip!

[@Starchface]

That's a thing? Oh happy day. Perhaps a video tour with Adrian's Analog Attic can be arranged. Marking my calendar.

[@ChrisJackson-js8rd]

@@Starchface they have a youtube channel with some video tours up if youre interested. lots of neat stuff. including some vacuum tube prototypes made in old coca cola bottles

[@tekvax01]

Be very careful when touching, cleaning, and rubbing those open hybrid circuits, you can very easily damage the printed high-precision resistive ink printing, which will completely destroy the hybrid. I use to design, construct, and troubleshoot those types of hybrid circuit boards.

[@adriansdigitalbasement]

Heh well unfortunately for me, this one self broke all on its own.

[@user-lp3cf5yn5b]

Hey @tekvax01 can you help me out with something? The opto U1040 on a 2445 low voltage power supply I'm working on is blown. I think I or someone else accidentally zapped the input side coming out of U1371c. SOC123A is a number I can not find. Old common parts lists call that Tek pn 156-0885-00 a MOC1005. Also finding a reference to a TLP631 in later 2465a and b power supplies for that IC. RPR says 156-0885-05 is the replacement for -00 suffix. TLP631. Is that correct? The closest cross I seem to find is TIL117M. Would that work correctly there?

[@danielmantione]

This is one of those moments where I observe that technology has evolved backwards. This is really an amazing piece of technology of the 80s, extremely well constructed, something you rarely see in modern electronics.

[@wesley00042]

It was also $11,000 in today's money.

[@VincentGroenewold]

@@wesley00042 Exactly, you can get quality like this for that price as well. Though I do agree that engineers back in the day could use more quality components then these days.

[@graealex]

Another one of the "everything was better in the past" folks. For what this oscilloscope used to cost, you can nowadays get a modern DSO in the GHz range, with tons of features, protocol detection and in-built calculation like FFT and function generator. A proper DSO also hasn't any drawbacks, as memory depth allows to view stuff like jitter and twitching. Not even speaking about the fact that a useful DSO can nowadays be had for less than $100.

[@danielmantione]

@@graealex Not eveything was better in the past, but in this case... Modern DSO with high frequency is still very expensive. It will have a nice LCD screen and fast CPU. Is it actually a better oscilloscope?

[@graealex]

@@danielmantione Yes, it actually is. At around $500, there isn't a single metric where the DSO isn't going to be better. In particular, you will get PoV, so that imitates how a CRT works and will show things like jitter visually. Not even talking about all the neat stuff like single-shot and time lapse mode, logic analyzer and protocol decoder, that only work with DSOs. And again, at the price point of the 2445 when it came out, you can now get a top-notch DSO with all the bells and whistles. Although even today, I would rather spent $300 on a DSO, instead of the same money on a used analog one.

[@thunderbeam9166]

I had a very similar issue pop up in my LeCroy 9310 (with the gorgeous amber CRT). Long story short, after a total of 40 hours poking around in it, I decided to dump the Roms and have a friend look through them to see if we could find what controlled CH2s attenuation. Turns out that one of the 5 EPROMs had one janky bit of code and during the boot process it would internally switch to 1 Meg and there it stayed. Working on scopes of that era can be a real trip sometimes.

[@kaitlyn__L]

I’m not sure abrasion is necessary, but I’d totally consider squishing the spring pins on the relays since pressure helps the connection. Springs do weaken over time after all. It also sometimes behaves like there’s stray capacitance, so some of the pins might be losing contact but stay close enough to form some capacitance along with the circuit board etc. Either way reinforcing the pressure would be the simplest thing to try, since you noticed it’s pretty clean inside.

[@MUSTASCH1O]

As someone who is only a novice at reading circuits I was impressed by your ability to seduce the problem by working your way through the circuit diagram. Must have taken a lot of practise to get to that point!

[@zxborg9681]

Back in the 80s I bought a cheap 1950s-era Dumont Cathode Ray Oscillograph. Good up to a megahertz or so. Having something like this Tek would have been a dream, though I eventually wound up using this exact model amng others at work for most of the 90s. I've never seen the inside of one, and all I can say is holy cow, is that thing ever packed full. Glad you were able to rescue it, very cool to watch.

[@mdzacharias]

I suspect you have one or two hogged out center BNC connections on the input jacks. Explains to me why wiggling the probe plug affects it the way it does.

[@dolphhandcreme]

37:00 get a fiberglass-pen! They are perfect for tasks like this and are really soft to the material.

[@paul_k_7351]

We had these at school. As kids we had no idea what we were doing but it brought back so many memories!

[@PsRohrbaugh]

I have an old Tektronix scope that's 1000% analog - no fancy features. You've got the best of both worlds here. If I want voltage or time, I've got to count marks on the screen!

[@_droid]

Somewhere around 400V is the typical input voltage limit for any real oscilloscope, old or new. Not in 50 Ohm mode though, this is usually limited to something like 5V. Note that these voltages are the limits with a 1x probe. I have repaired tons of these Tek scopes. The ones like this full of custom ASIC's are a pain because if those ASIC die it's mega bucks to replace them if you can find them at all. I've never seen this specific problem but I have replaced whole attenuater modules due to the ceramic PCB being cracked. I recommend taking a really close look at it, the relays, and the BNC (crusty connectors is very common even if they don't look so). One of the first things I do on these is Deoxit the entire front panel. Every knob, every switch, every connector, the power switch and other mechanical stuff back there. Most of the time this fixes all intermittent issues like this.

[@adriansdigitalbasement]

I have a couple other similar vintage scope to this one, and I noticed after opening them that they also have the same type of attenuators inside. (Or they look the same at least) The others have no issues on the front end ..... so it seems a bit funny this one has this issue. Perhaps the early build date of this unit means there are some slight issues with those attenuators that developed over time?

[@JayWye52]

@@adriansdigitalbasement Plastic outgassing contaminates the plated contacts,leaves a film that interferes. it can be hard to clean off.

[@BrokeDad1]

I still have several Tek analog scopes from the 465 to the 2465 but I'm glad I no longer repair them as they were always a PIA. I remember taking the week long repair course Tektronix came and gave us technicians.

[@per-henrikpersson1884]

As a old and 8 years retired engineer in electrical (with Two exams in enginering (the second for an master of electronics) and 1.5 years studies in astronomi) this video gav me kind of years in my yes). Techtronix is This the shit, but last time I used one was in the mid seventies!!! So this was a very modern Techtronix for me😁👍! The inside of a Techtronix is real beautyfull of enginering quality. Any old engineer will Come to years and want to hug and kiss this fantastic beautyfull ❤️❤️❤️❤️😮😮😮😁😁😁👍👍👍. And as allways your videos are fantastic. And make The viewer feel good and Happy and you feels like an old really Nice friend to us. Thanks 1000 Times. And of course my old trust Amiga 1200 with a accelerator is still working just as my sons Amiga 1200 still is doing. He is also an engineer (in comp uting and software of course). And we also has Z 81s, Spectrums and Amiga 500,Amuga 500 plus and still working Commodore and old Ohilips monitor. So every time when I feel low after starting some PCs I just have to start my Amiga and feel glad and smile again. But anyhow thanks for your fantastic videos. 😁😁😁👍👍👍

[@christopherdecorte1599]

Try gently bending the contacts on the relays that meets the pcb they act like springs but the been in that position for about 40 years so with temperature changes they no longer make positive contact that should fix the flakey behavior for another 20 years. Just bend slightly.

[@dielectriccapacitor5291]

If you ever have a question on analog test equipment, you should ask Mr. Carlson. He's my go-to radio repair guy, but he owns a plethora of analog stuff. I know you're a digital guy, but Mr. C and Shango066 are the best for this sort of stuff if you need a second opinion.

[@mp-ov9dh]

Every one of these from the vacuum tube on up are built like tanks. Still great tools and are highly serviceable.

[@TheDefpom]

I only just started watching, but make sure you replace the RIFA caps in the power supply, and the main filter caps too, when those start to get weak it puts too much ripple on the rails, which then causes the yellow axial tantalum caps to fail throughout the unit and short out. I've done some videos repairing my 2432A which is the same kind of design.

[@simonscott1121]

Man that's an amazing piece of capable analog tech. My jaw dropped at some of the features.

[@Renville80]

The difference between oscilloscope CRTs and TV / monitor CRTs is in the method of deflection. Oscilloscope CRTs use two pairs of charged plates to steer the beam instead of the usual deflection yoke. There are actually some very early TVs (late 1940s era) that used electrostatic deflection CRTs. On the digital board, the Motorola IC next to the EPROMs I'm 99% sure is a custom version of the 6800.

[@Brian_Of_Melbourne]

I'm curious, why do you think it is a custom 6800?

[@JCBudro16]

My guess would have been a 68000 given the era and how ubiquitous they are.

[@miscbits6399]

@JCBudro16 A 68000 of that era came in a VERY large package (64 pins and nearly twice as wide as standard 26-28pin Eproms/40 pin CPU packages). In 1982 the most common controllers were 6809s (built in clock), Z80 or 6502 68ks were more or less brand new at that point and only found in high end computers. One of them would really have been complete and utter overkill in a scope Scrolling through the video, at 20:57 you can see the cpu is a SC67127 - which looks to be a 6800 SOC of some description (I can't find datasheets for it but it's showing up in multiple locations as the microprocessor for a number of Tek scopes It was only a couple of years after this that Atari STs were using a Z80 as embedded keyboard controller. How things have changed

[@JCBudro16]

@@miscbits6399 Ah, I thought the scope may have been from a bit later era but that makes sense. My HP 54501A is 68K based but it's a full digitizing scope.

[@Xilebo09]

This is a truly great peace of equipment then but still now.

[@robertorodriges1441]

Good day , thank you for the video , this oscilloscope of 1983 , as I understood , and there are relays in the volt division unit in it , I went to study in the 2nd grade in those years , in 1992 I worked as a repairman and I had an analog oscilloscope with a CRT kinescope , in these sensors it is necessary to measure ESR metr all electrolytic capacitors, thanks for the video, all the best

[@obsoletebutneat]

Tektronix scopes are absolute beasts.

[@stazeII]

My grandfather worked for Tek back in those days doing technical writing on the manuals. Always love seeing these old scopes. :)

[@johncasteel1780]

TEK spent nearly as much time on the manuals. A printed manual (or reproduction thereof) for a 2445B/2465B is a marvelous thing.

[@tony359]

those digitally controlled analogue scopes were marvellous. The analogue stage was just a work of art, tolerances were just impossible at those high frequencies. I had a fully analogue Tektronics many years ago, it was "spent" but ok. One day I opened it and noticed like 1000 trimmers so I didn't even think of adjusting anything! Then I stumbled into a digitally-controller Fluke 100Mhz one. I opened it up and there was not a single trimmer to adjust, everything was done digitally! The best scope ever, but lacking digital buffer to stop the signal. On cleaning those pads, I am not sure I would use DeOxit. It might leave some protection behind which is fine for a C64 SID or VIC IC but might interfere with the tiny signals at high frequency of an oscilloscope's front end. One of the cleaners I am using is not recommended for signals below 1V for example. I would probably use isopropyl only and I'm sure Tektronics used something even better, likely in clean rooms! Without knowing much about those scopes, sounds like the relays might have develop some contact issues? I'm wondering whether a bath in isopropyl might help? As you said, you have to trust your instrument so right now that scope is probably useless. Great video as usual - what's your new camera?

[@thomasmroz]

I used this model daily back in the 90's while in the Navy. Loved them.

[@AnthonyRBlacker]

Oh baby, opening that thing up looks like opening an old Ham Radio. Super interesting and extremely complicated inside! Totally different, obviously, but just as busy in there!!

[@Errr717]

I had one of these in my cubicle when I worked on microprocessor development systems in the 80's at Tektronix. Every engineer, HW and SW had them.

[@twocvbloke]

With the relay contacts being an issue, I'd go back in and give the contacts a tickle with a fibreglass pen, a very light touch though, if that doesn't do it, maybe bend the contacts very gently outwards from the relay modules to give them a bit more springiness, given over time they may have been bent inwards meaning they haven't got enough oomph to maintain a reliable connection... :)

[@mkeyser]

Great job man!! This was a huge uphill battle. Your chances of success were slim, but you did an excellent job, and you know a lot more than you give yourself credit for. Keep at it and eventually she will be 100%. It's probably the BNC itself like others have pointed out, may be missing a tine. These are the best, and it's really wonderful when someone brings one back from the dead.

[@Edisson.]

Hi, if I still had hair on my head, I would lose it terribly 😂 I completely understand your approach, if you have no experience with such a device, but you didn't even do a basic inspection of the components (or you didn't show it in the video) and stuffing it with paper sent me to the ground . 1) the input part of the attenuator was not checked (it may be mechanically damaged, e.g. by hitting the BNC), the hybrid PCB may have been damaged 2) the attenuator and motherboard connectivity was not checked (there may be cold junctions) I almost died when I saw the gold surfaces being scratched by the metal and the pressure applied to the hybrid circuit board. Even the smallest scratch is a problem that cannot be effectively removed, as someone already wrote here - very carefully polish it with a piece of paper (you can also use a soft eraser on a pencil, but it leaves microscopic residues that need to be completely removed) impregnated with IPA and then treated with a suitable means (eg Kontakchemie Gold 2000). The fault itself tells us what is happening - decay of the rectangular signal, decrease or increase in voltage = incorrect function of the input divider of the attenuator (it is only an estimate, but it is the first thing I would check) 1) decay of the rectangular signal = problem with input compensation 2) voltage drop or rise = input divider error As I already wrote, this is just an idea without specific measurements, but another fact brings me to this idea - the anchored attenuator to the structure reacts to the movement of the BNC. I could of course be wrong and the fault could be elsewhere - without measurements any advice can only be guesswork, but I am the owner of a working 2465. Good luck finding the bug and have a nice day 🙂Tom

[@Aitch-Two-Oh]

I was screaming "dry joint" all the way from 33:10, did you not hear me? Try reflowing the PCB joints. Also try swapping the modules to see if the faults transfer over.

[@Waccoon]

Man, I am so envious. I have an ancient Hitachi scope which does its job, but it's short on features and when I opened it for a good clean it's so barren inside. This Tektronix is a total work of art. I do have a modern Rigol digital scope, but I still love the instant response and feel of analog tech. It's great how much you can torture the analog stuff and have it survive.

[@TheStevenWhiting]

Probably cause I never understood it but during my first year in college in mid 90s we had to do an electronics class. I never picked anything up, struggled with it and we had to use analogue scopes. It all just seems so old fashioned and the scope just totally lost me. Now I still no nothing about electronics but the digital scopes really make me want one and to understand it. For me, the digital scope I feel changed it all. The digital scopes, when you see people using them properly and knowing what they are doing makes it so much more interesting than back in the days of college.

[@W3KW]

Wow. Glad I discovered your channel. Excellent lesson in diagnostics. I know a ton about my 2445 and you haven’t even opened the case yet! Love it. (Mine works perfectly…but)

[@JamieStuff]

I love the old Tek scopes! I had one of those on my bench when I was in the Army back in the mid-'80s. I just looked and found that scope (refurbished and guaranteed working) online for about $200. (I'm tempted to get one for my own bench, to sit next to my 500 MHz Tek DSO I snagged "untested" for $200 on Ebay some years ago.) Speaking as an EE, the triangle(ish) waveform from the square wave signal is caused by capacitance in line. Yep; cruddy contacts can do that, and they can also cause low signal. So, I think that you are on the right track. BTW, DeoxIT has a formula specifically for gold contacts.

[@TheDefpom]

@28:00 Dont assume gold connections are fine, I have fixed test equipment multiple times just by cleaning up gold connections with deoxit and a fibreglass brush, they still get oxidisation, just not as badly as other surfaces.

[@tomhajjar8257]

Use Q-tips with polishing compound on the PCB. Clean up with alcohol. Use a skinny straight pick to "tweak" the fingers to apply more tension. Use the relay slider to allow both sides to be adjusted. Clean fingers with alcohol. Use extreme caution on fingers. Two scopes fixed.

[@herbmyers805]

I worked oscopes in USAF PME Lab had a great time doing that. Great to a repair.

[@bluescooter84]

I bought a tex dual scope from Alabama Univercity trash at 2007.. They're great gadgets.. You're dominating subject very well.. Exactly..

[@DanBowkley]

I absolutely coveted that scope when i was in high school.. it's amazing how far we've come with test equipment like this lately, i now have a vastly more capable scope for probably 10% of what that Tek originally cost. I still covet the old analog stuff though!

[@miscbits6399]

That's a blast from the distant past for me. I had one on my workbench as a RF tech. It cost at least a couple month's salary when new I'll second the comments about bending the spring contacts. Ideally you need replacement relays HOWEVER you can reset their tension if you're very careful (there are tools for doing this - essentially a piece of stainless steel with a couple of slots to bend the things) - just don't overbend the spring contacts or you'll work harden them and they lose tension entirely Judging from what I saw in the video., the relay carrier has a foam pressure strip beind the relays, That may have collapsed over the last 40 years too

[@electrifyingvids3545]

After seeing that signal at 4:57, I almost immediately knew what the problem was. I have a Tek scope as well, not the same model. I have a Tek 475 from the 70s. It's fully analog and has no ditigal components. You can add a digital DM44 as an option but mine didn't have that. Back to the point. One day I noticed my 475 made some really funky signals when a square wave was applied. The signal looked identical to yours. But this only happened when the channel was coupled to AC. This was only on channel 2 though. Channel 1 worked perfect. When channel 2 was coupled to DC, the signal looked like a square wave, and not a strange triangle wave. So of course I went trouble shooting and probing around the vertical section of the scope with another oscilloscope to see exactly where the signal goes funky. On the input BNC the signal looked fine, so that rules out a bad BNC. The output of the vertical amplifier looked funky though. So it was not the CRT or any parts after the vertical amplifier. I then probed the output of the attenuators. AHA! The signal looked funky! So I opened up the shielded can and started probing there. The 475 doesn't need to have the entire attenuator module removed as you can access all the points you need to test by just removing 6 screws and a shield. This made it so I can probe the signals in the attenuators. I probes the input to the attenuators and the signals looked funky. This limited the problem down to only the very front part of the scope. The coupling switches. So checked then and found the signal going into the capacitor for the AC coupling looked fine, but the signal coming out not. My first guess was the capacitor went bad. So I replaced it and used a modern equivalent. It worked perfect. End of the story right? *W R O N G* . The AC coupling started going funky again after a few hours, so I poped it back open. The AC input cap on the 475 uses 3 pins instead of 2. One for one plate of the capacitor, one for the other, and one for the metal can surrounding it. My modern cap didn't have the shield so I thought that might be the problem so I popped the old cap back in to take some measurements to find a suitable modern replacement, but the scope started working fine again! Turns out it was just a faulty connection from the capacitor and the attenuator board. You see, the 475 and other 400 series Tek scopes used socketed devices. Eervy single transistor is socketed, every IC, and some capacitors. The contacts of the sockets where the faulty part. Apparently, nickel plated copper and gold sockets don't work as the nickel corrodes and tarnishes. All I had to do was plate the leads with a tin layer of solder and everything worked fine. I don't recommend solder plating for your scope though. But bad contacts is what caused this triangle wavefrom for a square wave.

[@jorgelotr3752]

You have repaired things with an oscilloscope many times, but this is the first time the thing you repair is an oscilloscope.

[@TeardownOZ2CPU]

thanks a lot, fantastic video, i had one of them 2445 in my lab today, very much the same kind of relay problem, attenuator some points not doing their steps correctly, but we just worked the relays many many times it slowly got better, and now it just dont fail anymore :-)

[@jeffreyyoung4104]

I have several Tektronix scopes, and I have recapped all but one, as I have not decided what I am going to do with it. I too have a 2445 scope, and I find it to be a nice, inexpensive scope! In fact, many of the different Tek scopes have come down in price, due to the upgrading many industries have made! When I recap the scopes, I focus on the electrolytics, as they dry out, leak and change value. Afterwards, I have no fear of trusting them letting out the magic smoke, or worse! I was using an HP scope without changing the caps, and it ended up shorting out one of the power supply caps, and it let out some smoke, but it didn't hurt the PC board or other components, thankfully, so I am quite anal about changing caps now.

[@An_Onion]

Always down for another episode of Adrian's Analog Attic!

[@ekenpad8482]

Don't scape with metal, just use a fiberglass pencil, followed by De-Oxit to safely polish / clean the gold plated contacts. Sraping will lead to oxidation once the protective gold layer is completely removed.

[@ernestb.2377]

By the way, if you press both delta-V and delta-t cursor control buttons it switches to frequency display 🙂 The most instructions are written on the front panel itself.

[@ralphbenthall3573]

correct..had one of these since 1992

[@BLOCKsignallingUK]

Noting that pressing the relay black carrier down appears to improve things. Could you isolate this to pressing down on only the clear casing of one or more of the relays? I am wondering if the relays are not sitting fully down onto the pcb. At 37:21 could you put two strips of paper between the black carrier and the clear relay casing along each side so the relays have slightly increased pressure to the board. Or indeed putting four rectangles of foam strip on the relay tops, between the relay tops and the aluminium can (a bit like where you have the paper, but only over the clear relay casings). At 37:37 there looks to be a strip of material there at the moment (yellowed strip) which is intended to provide some force to hold the relays in contact with the board, but might have lost its effectiveness over time (become brittle and shurnk).

[@sragga]

used these in school when i took my two year electronics technician course back in the early 90's

[@AttemptingAstro]

I love my 2230. Yes it's only 2 channel and 100MHz, but the analog/digital combo is just so cool to me. It's also the only functional 8088 machine I have haha. Seeing this 2445 was very cool!

[@seanrosedotcom]

Enjoyed. Couldn't tell from video, but a lot of those old relays aren't sealed... and the bad connection is on the inside, not the external gold pins etc. If you don't want to brave cleaning them... you could see if you have enough working ones to restore only 1 channel. Also, have that signal generator too. It's fine for audio signals, I don't recommend for higher frequencies.

[@donmoore7785]

Excellent debugging and repair

[@ChrisJackson-js8rd]

i love scopes. old or new. low end or high end. they are all fascinating (imo)

[@KHzSineWave]

I love my Tek 2211, it's a digital/analog hybrid scope so you get the best of both worlds, though it's limited to 20MHz in digital mode, 50MHz in analog. Part of me wonders if those little spring contacts on the bottom of the relays are causing the issue, they could've lost tension over time, might be worth trying to gently bend them out a bit so they make better contact with the pads on the PCB.

[@martianrays]

Well I'm sort of glad I found this video. My 2465 does the same thing only on channel 2. When you were talking about tapping it on the bottom, I was like "Hey wiggle the input just a bit!" It's a 50 minuet video, so that probably translates to 20 hours of frustration in real time. From the looks of it it almost seems like a thin piece of plastic cut from flat spot on a square bottle or something, would put better pressure on the relays. If I'm going to be aggravated for quite some hours, I think I will look into @Unfinished80 bend the tab solution. Thanks. Subbed and liked. that's the least I can do :?)

[@Y2Kxx]

Wow, this scope is absolutely insane! I didn't realize there were analog scopes that have the features that this one has! I have a HP 54645A from around the late 90s to early 2000s, and even though that one is a digital sampling scope, it's feature set is more similar to this analog scope than I expected! Great find!

[@dbmandrake]

Definitely faulty relay contacts - I've seen the same issue on plenty of other small signal relays when they get old, and they can be intermittent each time you switch them as well as with physical force applied. A combination of oxidation on the tips of the relay contacts and/or a loss of spring tension mean the moving contacts don't make reliable contact anymore. Not much you can do about it apart from replace the relays. Probably very hard to find the relays but if you could you could fix it 100%. The poor high frequency response (triangle wave) in the earlier parts of the video is an interesting one - high bandwidth scopes like this use a dual signal path - very early in the input signal path the signal is split into low frequencies and high frequencies, (with a crossover point somewhere around 10-100Khz) amplified separately in parallel by two different amplification chains using very different design techniques optimised for their respective frequency ranges and then mixed together again later in the signal chain after a significant amount of amplification has occurred. This is done because it's nearly impossible to design a single path pre-amp circuit that can work all the way from DC to 150Mhz while maintaining extremely flat frequency response, extremely high sensitivity and low noise. So the majority of that 150Mhz is handled by a pre-amp following RF amplifier design practices which can work from a few tens of Khz up to 150Mhz, the low frequency amplification is typically handled by an op amp circuit and this circuit is also responsible for precisely controlling and regulating the DC offset when you turn the vertical axis offset control. As soon as I saw that roll off at a few tens of Khz I immediately assumed the high frequency portion of the signal chain was non-functional so what you were seeing was just the low frequency (probably opamp) amplification stage working by itself. Rather than being faulty as such it looks like it was simply disconnected due to the poor contact issues.

[@oliverer3]

According to the tektronix wiki one of these would have costed ~$10,000 in todays money, so it's safe to say they were pretty high end when new. There was also a 300Mhz version of this same scope that was even more expensive.

[@tekvax01]

tektronix scopes are a challenge to repair, I have had several over the years, and use them at work as well.

[@chuckthetekkie]

I love how the PCBs are literal "Resistor Cities"

[@force4200]

So cool i have the same ;) hope you get it working. If its the model with memory its great to change the battery inside so it dosent begin to leake. Love your channel keep up the good work :)

[@wildcat189]

That's one cool piece of hardware.

[@maxtornogood]

I remember having to use something similar to this back when I did some electronics courses, these analog oscilloscopes are built like tanks!

[@simontay4851]

Bend the relay contacts out a little bit so they press down on the PCB with more force.

[@thatampguy]

Dude love the channel. Those old scopes could take some heat through the front end!

[@ed731pdh]

My first and oldest scope here, a Telequipment S43. Fully valve powered and in a case that you could drop a building onto and it will still work. Best part, you try and overload the input and the valves just hold their noses rather than give up the ghost like more modern replacements.

[@johncasteel1780]

I love my 2445Bs (3) and my 2465B (1). As far as I can tell, one of each works completely. One '45B mostly works. The other generates a stationary spot, telling me that at least the HV PS is working. The scope has two PS modules, one low voltage and one high voltage. There is a large DIP chip, U800 on the mainboard, that will overheat if you operate the scope outside the case. And eBay has parts for the attenuators and hybrids. The attenuators are going for about $40.

[@a4000t]

Nice, still use my 2465B. old workhorse!

[@TheSimTetuChannel]

Please do a followup where you remove your paper and very slightly bend the relay tabs a few mils to allow for firmer contact. In any case, awesome teardown!

[@russtrautwein6293]

flexing your relay contacts so they make a more positive connection to the module board will correct the issue without the need for the paper.

[@williamsquires3070]

Try checking the actual coax/BNC connector. There should be a nut (like on a potentiometer shaft). Make sure it’s tight, but not stupidly tight. Also, check the wire from the center conductor to the attenuator, and resolder if needed. 😊

[@SidneyCritic]

Looks like the BNC on the can is lose/broken, ie, check wire/pin connect to PCB - seen a lot of repair vids where these broken -. I guess look for new relays, but it could be the connects on the PCB that mate to the relays pins. Maybe ultrasonic the PCBs.

[@BuddhaPhi]

Great video! Love the analysis and repair. Definitely don't see these scopes around much. A couple weeks ago a friend recently gave me a 2245A scope from the mid 80s. (It's mostly identical but maxes at 100MHz.) I was amazed how much lighter weight the 2xxx series from the 80's is compared to the 70's 7603 scope I had years prior.

[@bergatrolle]

I would swap the relays between ch1 and ch2 to see if the problem switch channel. I would also check the soldering point of the two bnc connector.

[@TzOk]

Have you tried bending a little these spring contacts on the reed relays? You may also try to put some self-adhesive soft foam gaskets (like those for windows) on top of them.

[@Brian_Of_Melbourne]

Those are NOT reed relays.

[@pu4jlv]

Hello Adrian. I have a 2430A oscilloscope which is also 150MHz and has already taken some work to repair. Unlike the 2445, it only has 2 channels, but it has 2 trigger inputs (EXT TRIG 1 and EXT TRIG 2). It would be much more useful to me if it had 4 channels. Could you tell me why someone needs 2 external trigger inputs?

[@EnvAdam]

47:52 - loved and taken care of isn't immune to wear but it is great that you got it mostly back up and usable again.

[@jorgeandrade20]

I've been binge watching all your videos, so many great repair videos! Btw, none of the Ebay links mentioned above work..

[@Corn_Gone_Missing]

thank you for everything you do!

[@d18c7db]

Could it be a problem with the BNC connectors and their solder attachment to the preamp PCB? I imagine CH1 and CH2 would get the most action and repeated connect/disconnect cycles of the BNC jacks would apply enough pressure over time to cause stress fractures in the solder joints.

[@TimDyb]

...or the actual female part of the BNC connector.

[@Vermilicious]

Maybe the leads got pulled hard when connected at some point. It does seem like a mechanical issue. Perhaps you could try swapping the relays around and see if it helps. Perhaps you can get some modes working better. Otherwise try to bend the metal bits out a little bit. Maybe there are cracks in the solder on the BNC plugs too. I'm sure you can find a way, Adrian! ;)

[@FinalBaton]

Maybe the BNC jacks of the attenuation modules, are not making good contact and are the failure point that needs repairing?

[@tomsherwood4650]

My previous 2445 like my current 2245 had no atten. problems. The Tek that gave me serious atten. headaches was a 465 and saw that alot in the 465/475 units in shops.

[@williamgreen5575]

The relay pins that exit the bottom of the attenuator modules must be contacting the main PCB in some way. They look like the pins for each coil. I never saw how they connect to the main PCB, but I would definitely check/clean those contacts.

[@danmenes3143]

12:15 My garden variety Rigol 1054Z has an input voltage rating of 300V cat3, which is lower than 800V, but not bad. Input stages actually aren't that different between a digital and analog scope. Dave at the EEVBlog has an interesting videos reverse engineering a budget Rigol input stage. One thing to remember: the first thing after the DC blocking cap and 50 ohm terminator, and their assorted relays, is a 1 Megohm resistive voltage divider, so you've dropped the voltage quite a bit before you get to anything solid state.

[@danmenes3143]

Obviously, don't try to put 300V or 400V or 800V on the input if you have 50 ohm termination engaged, unless you really like the scent of Magic Smoke.

[@joefish6091]

Only on modern analog multimeters with FET input stage., old ones were purely series resistors (with germanium diode OL protection) . 5 to 100K per volt depending on quality ie meter sensitivity. 50uA being 20K etc . The more sensitive the more delicate the meter movement. Ex Gardners Radio.

[@danmenes3143]

@@joefish6091 Scopes, not multimeters.

[@ryanbrooks1671]

I wonder if a new "relay carrier" could be made to offset the locations of the relays just a little bit on one axis. This would move the contact point to a new / fresh location on that pad. Either by making the mounting holes a bit oval, or by a 3d printed replacement?