Tektronix 2232 scope.... in need of a doctor..

[tautech]

As you complete this repair I would have the inductance of each winding measured and note it in this thread.
As time goes by these SMPS transformers will get harder to source but if each windings info is known and can be matched with n# turns, wire gauge and core parameters it shouldn't be hard to replicate DIY.

Who knows, you might even be able to start a little sideline making these and supplying the market. 

[Vince]

As you complete this repair I would have the inductance of each winding measured and note it in this thread.
As time goes by these SMPS transformers will get harder to source but if each windings info is known and can be matched with n# turns, wire gauge and core parameters it shouldn't be hard to replicate DIY.   

My intention/thoughts exactly ! 

I have ordered two coils. Will repair my scope with one of them, so I can fix it quickly, and I will use the second coil to take inductance measurements .... later.... because I don't have an RLC meter just yet ! ... and no money to buy one, have much higher priorities on my plate currently...
Well, until I can afford a decent/good RLC meter, maybe I can get a cheap/crap one for a few tens of bucks, just to get an idea...

Who knows, you might even be able to start a little sideline making these and supplying the market. 

LOL  Well I would have to sell an awful lot of them, and for a lot of money in order to make it worth the time spent winding... plus, since all the specs would be freely available on this forum, anybody could do the same as me, and profits would plummet ! LOL

But I will try and make myself a new coil, sounds like fun, I want to do that... at least once in my life. I find it educational... and being abel to buld your own stuff to repair your old gear, rather than buying old parts that eventually will run out anyway.... I find it satisfying.

[oldway]

Might buy two of them just in case I get unlucky and one of them is "noisy", as he states.

Most of them are noisy because they are faulty. When inductance is too low (internal short circuit between turns or layers) , there are high peaks currents who are stressing the ferrite core to saturation and the noise is coming from sub harmonics of magnetostriction of the core.

...and I will use the second coil to take inductance measurements .... later.... because I don't have an RLC meter just yet ! ..

How to measure an inductance without RLC meter when you have a function generator with counter, a capacitor of known value and a working oscilloscope ?
This is very easy, I let you guess the answer......

[Vince]

Most of them are noisy because they are faulty. When inductance is too low (internal short circuit between turns or layers) , there are high peaks currents who are stressing the ferrite core to saturation and the noise is coming from sub harmonics of magnetostriction of the core.

hmmm, thanks for that     So I can just measure the resistance (now I have the data from my old coil) of the windings to see if they are shorted or not...

How to measure an inductance without RLC meter when you have a function generator with counter, a capacitor of known value and a working oscilloscope ? This is very easy, I let you guess the answer......

A resonant circuit yeah why not... I am rusty on the theory/math though, I must admit... I will dig out my old text books...   1/2*PI*sqrt(LC), something like that... shame on me. I mean I studied this stuff 20 years ago, and zero practice since then... yes of course I got rusty !    No worries, I will get my old text books and refresh my memory 

I don't have a counter BTW     My TDS 544A can measure frequency though, obviously...

I have a 1Mhz analog generator, and a 2MHz "DDS" Fluke arb gen but it's in pieces... I am in the process of replacing the original LCD, with a back lit unit, because the original LCD has no backlight and it's just a real pain to read...

Anyway, I could also just wire up on a bread board, a simple first order RL filter, look for the 45° phase shift / 1/sqrt(2) amplitude reduction (the TDS544A could help with its automatic measurements) and figure out what value L is...

Well OK, in practice it would not be as accurate as measuring the frequency of a resonant circuit, I admit...

OK OK....I will do a resonant circuit.  Google refresh my memory please !

https://www.electrical4u.com/resonance-in-series-rlc-circuit/

Yeah, phew, I was correct earlier, 20 years of inactivity have not erased everything just yet, I am not as hopeless as I thought...

[oldway]

So I can just measure the resistance (now I have the data from my old coil) of the windings to see if they are shorted or not...

A short circuit between turns does not change the coil resistance enough to be measured and detected. But the inductance is reduced and even more the damping factor of the inductor associated with a capacitor in a RLC parallel ressonante circuit.
The short circuited turns works as a short circuited secondary of a transformer.
The definitive diagnostic is made "in circuit working" by measuring the current waveform on the shunt resistor and to compare it with waveform of service manual.

[tautech]

Vince
Further study for you when you get to rolling your own tranny:
https://www.eevblog.com/forum/other-blog-specific/power-transformers-the-lost-art-of-custom-winding-bsfeechannel-5/

[Vince]

Hi there, back in the game ! :-)

Thanks to who it was that posted this link :

http://ludens.cl/Electron/trafos/trafos.html

I read his other pages on his site, related to coils.. this guy really seems to enjoy coils at large, and looks like he really knows what he is talking about !
He designed some cool little tools like this one, with which you can determine the core properties... cheap and easy to build so will definitely make one and play with it, to characterize the Tek coil, so I know for sure what core to buy in ordr to wind a new coil by myself.

http://ludens.cl/Electron/lmeter/lmeter.html


I received a bunch of parts from Farnell, so I can rebuild the regulator.. 90% of it  ! :-/
I also received the two coils I ordered from a chap on Ebay in the US, as was affordable enough, 15 bucks the coil and only as much for shipping to France.. so I bought two of them, one to fix the scope quickly, and the other to play with,  so I can determine it's exact specs and build/wire my own. Unfortunately, buying two of them also caused the death of one of them ! They must have banged against one another inside the package (package itself was fine, on the outside), so one of them lost a tooth so to speak ! LOL  A chunk of the ferrite core, broke ! :-(  Rendering it pretty much useless then.. not gonna risk putting it in the scope ! Can't use it to characterize the coil either I guess... So, I contacted the seller about it, and still can't believe it but he agreed to send me a new one and even pay for shipping ! That's one honest guy, a rare sight these days !  So I think it's worth pointing people to him, if you need a coil or something else he might have :

https://www.ebay.fr/sch/shockcore/m.html?item=282530665982&ssPageName=STRK%3AMEBIDX%3AIT&rt=nc&_trksid=p2047675.l2562


Said he would send it promptly.

In the mean time, today I finally started work on the 2232. I replaced all the components that I found faulty 3 weeks ago...

I put the TL594 on a socket because accessibility is a nightmare and getting rid of the old one was a real pain, had to break it in pieces and removed the bits one by one, not fun. So should the chip blow again, I would rather be able to remove it easily and quickly, and "gracefully" !   The thing hardly runs at high frequencies so I don't think having it raised a little from the board will upset it at all... though I will see that for myself when I probe around, come time to actually power it up !

So... I am quite close to powering it again and start checking that this regulator now works as it should. But I am not ready just yet :

- Today I removed the 51V Zener diode from the crow bar, because although it tested fine in forward bias (0.7V as regular diiode would), it didn't quite test as I expected in REVERSE : should have tested "open circuit" because obviously the multimeter only puts out 2Volts or so, hardly enough to make the Zener conduct. But somehow, it DID read something ! 1+ Volt or so ! Eh ??  At that point, the SCR was still absent from the board, so I don't see how the surrounding circuitry could have affected the test. So, not wanting to take any chances with this  safety related device, I decided to remove it so I can put it on a bread board, to better test its reverse bias spec. My lab power supply can do 2x30V, so 60V in series, enough to test that 51V Zener then. But... somehow... once I had removed the Zener from the board, and laid it on the work bench... I grabbed the bread board, then went to pick the Zener on the bench.... but nowhere to be found, I lost it !!!! For the life of me, I lost sight of it for only a few seconds, and I just can't find it any more ! 
So I am screwed, must get a new one !  Hopefully I won't have to spend yet again 30+ Euros at Farnell (minimum order to get free shipping) JUST for a bloody Zener diode ! I mean I hope local electronics shops have these in stock !.... :-(

Anyway, making progress ! 

Oh, and I also need to DIY a light bulb tester before I take the risk to power it up ! I did buy a x10 pack of the appropriate fuses, but if I can save blowing all of them, the better !  And I want to build this bulb tester anyway, will be quite handy whenever I will work on a mains powered gear.

[Vince]

Found 51V Zener diodes from a "local" shop.. though 45 miles from me ! Shipping by mail is only 5 bucks, so I went for that. Should have them this week I would think, in time for this week-end any way. In the mean time I will get round to building a DIY bulb tester...

[WastelandTek]

Good work Vince!  I'm going to comment so I see how it comes out.

Somewhere is a rift in spacetime where all the surface mount components and tiny screws I have lost are all collected, don't feel bad.

[Vince]

Back again ! 

I may have lost that bloody Zener, however I found something else that compensates.... that lonely big long screw that came out of the scope when I wiggled it, that I had no clue where it was supposed to go.... I eventually figured it out I think !  When I put the main heat-sink back in to place, I screwed the FET onto it, then looked at the pair of power transistors at the bottom of it (the two trannies from the output of the inverter, that drive the primary of the main transformer). There is a plastic holder, similar to that of te FET, that grabs both transistors at the same time. The holder is screwed onto the heatsink of course... the mystery screw is a perfect fit in there ! And it's also a pig to screw because accessibility is a pain, no room to manoeuvre the wrench, was a real pain to screw back in, a 1/16th of a turn at a time...  I am 100% sure I did not have to remove that screw when I pulled the heat sink out of the case.. because I remember vividly saying to myself, when I pulledit out the other day "hey that was much easier than I thought i would be ! "  ... no way I would thought of it in these terms, had I had to painfully unscrew that thing !!!!

Anyway, I received the replacement Zener diodes from the local shop. Local shop means I obviously could not find the exact replacement for that Zener.

OEM Zener was specified as 1N978BRL  0.4W  5%
The one I could get my hands one is a BZX55C51   
It come in a DO35 package which looks a tad smaller than the OEM Zener, but somehow the power rating is a tad higher, not lower, at 0.5W vs 0.4W.
Accuracy-wise both are 5%, hardly critical. I measured the replacement Zener before hand just in case, and although it's a few tenths of a Volt above 51V, it's still waaay within 5%  so not worried here.

And over the course of the week, I also worked on building my first light bulb tester ! Had to buy the bulb holder, quite expensive at 8 buck, and another 2 bucks for the halogen bulb 77W not 100 W. Could not find a 100W one as was suggested earlier by someone. I did find some 100W old fashioned bulb, but they were the bayonet style fitting, not the screw type, and the holder/socket was only available in the screw type.  The rest of the hardware I had in already in stock (switch & mains socket combo, power cord, scrap wood).
Tried to make something decent looking and practical, so took a bit of time (two days alone just to let the 2 coats of wax/oil dry)

So, I powered up the scope at long last.... standing back and anxiously maneuvering the switch, crossing fingers..... and.... and.......

To be continued next week !

No just kidding !     

The good news : NO catastrophic failure (like when then original failure happened a month ago), no big "BANG" (pun intended), no smoke, no nothing.  So at least that's good isn't it ?! 

The bad news : it does not work either.

The light bulb does not stay on or off either... it keeps cycling between the two states at a very stable/clock-like pace. At the same time I can hear a hissing sound, presumably from one of the transformer. Of course I made sure to check every component before putting them back onto the board, and triple-checked the orientation of diodes and transistors, but I am just a human being so I guess I could still have managed to get one of them back to front...
Obviously I checked the transformer too. Core intact of course, and all windings check OK (compared to old transformer).

Made a short clip of it so you can better understand. If you crank up the volume you can hear the transformer.

https://www.youtube.com/watch?v=9jd9idAlXDY&feature=youtu.be


So here I am, some more work to do, but getting there I guess...

I am real happy I made this light bulb tester, thanks for the advice !  First, it looks cute. Second, it indeed saved me a lot of fuses already. Third, blowing a box full of fuses would not have given me as much useful information as the light build tester did : fuse would have just blown, leaving no clue as to what happened, whereas the bulb cycling tells me that it's not a catastrophic failure, because the scope keeps trying to recover over and over again, which means that some over current protection (PWM chip of crow bar) is actually being activated.

Now I guess it could one of two things : either of the current limiting protections either react to an actual overload event, or there is no overload condition but the protection triggers anyway. Not relevant in the case of the PWM chip, as it would cut off power rather than short the SMPS. However the crow bar does short things out , so maybe the crowbar is not working as it should, and shorting the 40V rail when it should not. I replaced both the SCR and the Zener diode, so maybe I screwed things up there.

so I guess the fist thing to do is, again, to remove the 0.2 ohm shunt resistor so I can isolate the inverter from the regulator... and hook it up to my lab power supply and carefully ramp up the voltage to see how the crow bar behaves.... if it's not shorting at 42V or so, then it's good and the short comes from the regulator side.  At this point I can unhook the bench power supply, still leaving hte inverter disconnected from the regulator, and power the scope from the mains... and see what the regulator is doing...

If I said anything above that makes your skin crawl "no don't that, you will blow the thing again ! ", then time to speak up !   

[Vince]

Bummer... had to remove the heatsink again so I could get access to the shunt resistor/R907 in order to isolate the inverter... grrr..;

Inverter & scope still work like a charm !   I am relieved !   So now we know it's not the crow bar that's shorting the regulator, nor the inverter nor anything down stream. So, something is going in the regulator... I tested (again) the FET : removed it's connector.. all the pins test as they should... though I did scare myself for a moment because  the drain -source was NOT open-circuit when the internal diode was biased in reverse ! That's too bad for a brand new part... too bad too, that the other same part (I ordered two of them just in case...)...  I assumed/hoped I was just doing something wrong and chasing red herrings, but at the same time I   didn't want to risk mains voltage to destroy again the scope again !  So I thought, maybe the gate, though being a power FET therefore beefy, can get charged enough just by being left floating, to make the FET partially conduct (it measured at around 50Kohms). So I put it on a bread board, shorted the gate and Source, then measured the drain-source again.... open-circuit, relieved I was ! 

So... since the crow bar, inverter and and scope aren't shorting the SMPS, since the FET is not shorted/defective... where the hell the short could be coming ?!

Was starting to think well maybe the regulator is fine and the short is upstream, in the common mode filter daughter board or main filter... or mains rectifier or filter cap. But I had already checked these, and the common mode filter well, it somehow it got shorted, I doubt it would "heal" itself in a second, then fail again, and heal again, never mind at very regular rate that the light bulb test is indicating.

So I was starting to wonder if maybe... the scope is perfectly fine, and it's just oscillating due to the presence of the bulb.  Haven't thought of this in detail, but the thermal inertia of the bulb filament, coupled with the filter cap getting charged and discharged... maybe there is some back and forth game between that filter cap and the bulb, which would explain this oscillation, and it's perfectly stable frequency ?! 

Bulbs and electronics.. all of sudden it reminds me of an experiment we did at school, where a miniature filament bulb was used as part of the feedback loop in some op-amp circuit, for automatic gain control IIRC.

The fact that an SMPS presents a big filter cap to the mains, meaning the voltage is lagging, make it prone to oscillation, the light bulb tester thing was probably "invented" before SMPSes were the norm ? With a linear PSU, hence a big transformer, the light bulb tester sees a inductive load, not capacitive, so the current is lagging, rather than the voltage.
So I guess the light bulb tester is indeed an excellent tool when restoring older tube Tek scopes that use a big transformer.. but for more modern Tek scopes with an SMPS/ capacitive " front-end " so to speak...

Anyway, that was just a crappy guess of mine, a shot in the dark to keep me going in my trouble-shooting.

So as a result, I took the risk of powering the scope withOUT the light bulb tester, just to see what would happen... thinking what the hell, there is still the mains fuse that can blow, if need be.

So... did that and... nothing happened ! But... the fuse did NOT blow either !  So maybe my reasoning above, was not 100% BS ? Maybe only 99.9% BS ! 

I checked a few voltages : there IS voltage across the mains/rectified filter cap : 300V+ DC, as you would expect ! Bear in mind this is running 220+ V not 110V.
That voltage was perfectly stable.

So looks like things are getting better now :

- Rectified DC present
- FET OK / not dumping the mains
- inverter & scope OK
- Crow bar not shorting the regulator

So at this point I am starting to assume that maybe there is no short, and I might take the risk of soldering the shunt resistor back into place, to re-establish the link between the regulator and the inverter.

But before I take that risk, I wanted to try and probe around the regulator... I probed the power supply of the PWM chip... supposed to be 14+ Volts, determined by a Zener diode. Good news is, I do get around this... bad news is that it is very unstable and keeps jumping around, going up and down, say + and - 5V around the nominal 14V. So my guess is that it's something to do with the start-up/boot-strapping mechanism used here ? I am really out of my comfort zone here as I have no experience of these things... but I guess that since I have removed the shunt resistor, the output of the FET is not connected to the 40V rail filter cap anymore (C940), hence the output voltage has no way of being smoothed out and stabilizing, so it's a chicken and egg situation and the PWM chip has no way of producing a stable supply voltage for itself. Am I having this about right ?  So basically, I have no way of testing/probing the regulator, until I have re-attached it to the inverter.. bummer, risky. At least the above means that the PWM chip is probably oscillating and trying to produce a voltage... so that's rather good news, I say ! 

OK.... I am feeling lucky today... with every I said, it looks like the scope won't blow up if I reconnect the inverter to the  regulator, and that i might even work... or at least it will be in a state where I can actually probe/troubleshoot it properly/in detail.

Taking a deep breath..... see you in a moment for the result ! Wish me luck ! 

[Vince]

Oh ! just watched the video clip I uploaded above... and I noticed that on the foiurth and last "on" moment of the bulb, the scope was actually just starting to draw a trace on the screen !!!    nothing happened in the first 3 strokes, but I guess that was just because the CRT was not warm enough.

So, indeed looks promising !

Too bad my light bulb tester made me chase red herrings !

Would still like to get that light bulb tester working here though ! I was advised 100W... my modern halogen bulb is only 77 Watts, the more powerful I could find.
How does one calculate what power rating is required in order for the gear not to oscillate ??  Does that the fact that it's a halogen bulb changes things a little bit compared to an old fashioned bulb, and that would need to be taken into account ?

For now, I guess the answer is : if it oscillates, just put a more powerful bulb in there, until the oscillation stops ?!   

I guess I will need to make a large version of my tester, so I can fit two bulb sockets and combine various bulbs to get a combination that works for a given gear.

Any way, back to the work bench, see you in a bit ! 

[tautech]

When I use a dim bulb tester I first look at the  DUT manufacturers rated wattage consumption and select a bulb greater than what's spec'ced on the label. Rarely I've had to use a 150W bulb.
For various gear you might need to use the bulb tester on it's a good idea to have a selection from ~40W and up.

[Vince]

Thanks Tautech, will go and source a more powerful bulb... even if that means I need to order it online, since local shop won't do higher than the 77W I found.

I am back from the bench... VICTORY !!! 

The light bulb tester made me chase red herrings and worry needlessly : my repair of the regulator, somehow, despite the numerous components that I replaced... worked first time !  Thanks to all who helped me through this ! 
My first decently involved repair... sure was more involved than me replacing a tactile switch in my 10 bucks alarm clock ! 

Was my first time working on a real world SMPS, so I learned a lot along the way, thank you every one.

I am so glad my little 2232 is back to life, it's my main scope and I love it to bits.

I adjusted the trimmer in the inverter section, following the calibration procedure in the service manual. They take the -8,6V rail as their reference, so I set that spot on to the required 8,60V, and all the other rails were spot on too... except for the +/- 5V rails which were way off, but I didn't care since I assumed they would comply once I would plug the digital board back in.  Sure it helped... but then all the other rails moved in unison ! They dropped enough to be out of spec ! So, I readjusted the -8.6V again to be spot on 8,60V. and all was good again. all the rails are within spec now... though the 5V rails are on the high side :  +5,17 and -5.25V (max allowed).  The 8.6V rail is also on the verge : 8,64V for an allowed maximum of... 8,64V...

So I am half-temped to drop the -8,6Volt reference just a tad (-8,56V minimum allowed) so as to bring the +8,6 and 5V rails a little bit on the safer side of things... what do you think ??

So, the scope it working again, nice... but after a quick test, I see there as still some minor issues I would like to fix, at least try to.... if it's not too involved.

1) Probe compensation signal is inverted : shows up as 0 to minus 0.5V instead of +0.5V.  Had this problem when I tested the scope without the digital board, while powering the inverted directly from the bench power supply. Well now that the digital board and everything is put back together... the problem is still there  sadly !     I tried on both channels, and they behave the same. So I hooked the probe to the compensation signal from another scope, my old Tek 5111 (100Hz 400mV signal), and the 2232 displayed it just fine ! What a relief !  So the scope itself is fine... it's just the probe compensation signal itself that's upside down... only a minor annoyance you might say, as this doesn't keep one from compensation his probes.. but still, it's not normal so if I can fix it... why not.
If it were my 5111 it would be easy. If it were my TDS 310 or TDS544A it would be easy too... the signal is generated locally on the front panel assembly, which easily pops out to let you work on it... but the accessibility of the front panel on th e2232 looks a nightmare... will check the service manual and schematics.

2) When in digital/store mode, I witness too issues : the trace is not centered anymore : it is shifter down a little bit, by roughly a small division.  Also, when I use the vertical position knob to readjust it, the traces moves rather jerkily. See video clip...

One would say that the pot just needs cleaning, expect that this very same pot works like a charm in analog mode...

https://youtu.be/ug77tJohVSA


3)  Fan is noisy... after 30 years I guess it's to be expected eh.  Makes both a low pitched rumble as well as a high pitched noise that drive my ears a little mad to be honest.  the latter noise does not come from the SMPS... my new regulator transformer is absolutely silent, the other transformers too. Any advice on a good replacement  (brand and model ) ? The 'PC' gamer nerds, IIRC, are well versed in this domain... but IIRC desktop computers use only 80mm and 120mm fans right ? So this crowd won't be much help for the Tek, as I measured the old fan and it's a smaller, 60mm unit ! Bummer. Still, they could advise on brands, I presume.

Specs are as follows : 
Brand : " Sunon "
Model " MD 1206 PTS1  "
12V DC, 1.5 Watt 
60x60x25,4mm. 

Sticker also states "Brushless", but aren't they all ? So not a criterion...


4) I am a bit worried about the orientation of two little, identical, 4 way connectors, on the edge of the digital board (CRT side). Picture below.    I am worried because although I am 100% sure that I put them back in the same orientation they were when I took the scope apart (took plenty of pics at that stage, precisely to help me with re-assembly), I noticed that the Pin #1 on the cable, is NOT matching the Pin #1 labeled on the board, on the silk screen ! 
Looking at the board view in the service manual, these are connectors J2111 and J2112.  Looks like they are directly related to some analog circuitry shaping the two channels before they are fed to the ceramic proprietary modules, one per channel. So, these connectors are bound to be important... so I would rather like to be 100% sure of their correct orientation...
Can anyone tell me how they are oriented on their scope, for those with a 2232 or similar ?
I mean, maybe it's "normal".. maybe it's a mistake from Tek, who got the Pin #1 marking wrong ? Happens even to the best I guess...
Things is, the scope has been opened before, by some previous owner, that's for sure as I found old flux on some reworked solder joints in the SMPS section of the main board. So maybe the guy who did that, put these two connectors the wrong way around, I can't know.
In the mean time, I will keep  studying the service manual closely to try to figure it out...

[Vince]

OK, I am getting there with Problem #4 above.

These two connectors  are as could be figured just by looking at the board, the two analog channels which are undergoing some final analog processing before they are fed to the ceramic modules which are obviously the ADCs, which then go to that big ceramic/gold chip which of course is the big acquisition chip.

Anyway, PDF below. First the acquisition board, where it tells us where to find where these connectors go :  sheet #2.

So, although I still don't know how Tek mean to orientate these connectors, at the least it seems that regardless, no damage can occur, as the pin-out is kinda symmetrical  : the two outer pins are both ground, and the two inner ones are a differential pair carrying the signal from the corresponding channel. It's not referenced to chassis ground any more at this point, apparently.

I am still surprised that the most vital and fragile signals in the scope... the input signals, travel from board to board across 8 inches or so, on a simple flat cable instead of a coax ? I guess that is is carried by a differential pair, surrounded by ground on both sides as well, keeps it from picking up too much noise along the way.

[Vince]

Now having a look at problem #1 above, inverted probe compensation signal.

Found it in the schematics, see below...  it's generated locally, directly within the front panel. A simple text-book op-amp astable, at least I could figure this out     However all the stuff after it,  I can't make heads or tails of it...   The waveform they give for the output of the op-amp spans 15V (makes sense, +/- 8.6V rails, minus the limited swing of the 741output which can't fully the rail, 15V or so sounds about right)   so that means all the stuff after the op-amp is supposed to reduce this to 0/500mV.

So... that analog stuff must probably take that symmetrical +/- 7.5 Volts output from the op-amp, and maybe rectify it to keep only the positive part of it, that would solve half of the problem. Then it must somehow reduce that +7.5V square, down to 0.5V. Since it's a square wave and we are not too fussy here,  it can do it either by scaling it with a voltage divider, or simply by capping it !  Now with this in mind, let's try to make sense of the schematic...

Just had a look at the front panel... the back of it is partially accessible/exposed, one can probe the top half of it, and for once I got lucky : all the circuitry for the probe compensation signal is accessible !     Will probe that...... later, 5AM here, time to go to bed, done enough for today  I think !

[BradC]

Any advice on a good replacement  (brand and model ) ? The 'PC' gamer nerds, IIRC, are well versed in this domain...

Specs are as follows : 
Brand : " Sunon "
Model " MD 1206 PTS1  "
12V DC, 1.5 Watt 
60x60x25,4mm. 

10 seconds of google :
https://www.digikey.ca/product-detail/en/sunon-fans/KDE1206PTV1.MS.A.GN/259-1358-ND/1021207

On fans. Be very wary of anything recommended by the "'PC gamer nerds". PC's don't have to last anywhere near as long as quality test equipment, and the fans are built down to a price. Airflow ratings and static pressures are generally rated in PMPO Chinese watts. Additionally, PCs don't have irreplaceable ASICs and other expensive parts that fry when your 'buck fiddy' "ultra-mega-whisper whopper-flow hoosit waffle-cone bearing super flower blade' PC fan bearing carks it and it grinds to a silent halt.

[Vince]

OK, good point (s)... noted.

So let's take this fan replacement more seriously then... Indded Digikey has lots of fans for sale, lots to choose from... well that's if I lived in the US. Choice will be much more reduced here in Frog land, obviously.

But good news is that " Sunon " the OEM, still is in busines, so I can always contact them to firgure ou a suitable replacement, and potential improvement, over their 30 year old product. Well assuming they are nice enough to reply to me, hum hum...

At the least, looks there are 3 parameters to take take into consideration :

- Fan speed at nominal voltage, 12 DC here.

- Static pressure, no idea what that means in practice... the pressure that the fan can produce before the air flow starts dropping ? If so, then I guess the higher the better.. as long as it does not compromise other factors like noise.

- Air flow .. well that one at least sounds more self explanatory..

So, armed with the OEM part number, let's try to find what these parameters are... so I can make an informed decision on what replacement to get.

The one you pointed to has a very close part number to the original unit, so at least that's a starting point.


As for the probe compensation after-opamp shaping circuitry, I am thinking maybe

1) Not all of it is meant to shape the op-amp output, come to think of it, maybe some of it could be there to protect the op-amp/scope from people accidentally applying voltage to the probe test point on the front panel, rather than hooking their probe tip on it....   Not strictly necessary, but it being a Tek / quality stuff, maybe they wanted to make the product as robust/real world-proof as possible.

2) If they drop the voltage using a voltage divider rather than just capping, then R988 and R989 might serve this purpose: they have weird values so they must be accurate... indeed parts list lists them as 1%.  That would make sense for a voltage divider. If they were part of a protection circuit then most likely they would not need to be that accurate...
On the other hand, the 200ohm resistor at the end of the chain, near the 5V Zener, is also 1% ! Hummm...     This resistor and the Zener are physically mounted next to the probe test point, away from the op-amp/rest of the circuitry.   So I guess that means the Zener and 200ohms resistor work together, and are there for input protection...

3) Diode CR989, given how its mounted... I don't see how any positive voltage could ever come out of that probe test point !   

So.... could it be that I am losing it ? Could it be that this scope has always showed me a negative compensation signal and it's only NOW that I actually happen to pay attention to it ??     I still find it weird... all my other scope give a positive signal. 

I think I am losing it.... please call the doctors !   

[tautech]

So is TP 56 waveform correct ?
If it is there's only a few components to check but I'd suspect any of the 3 diodes, including the zener.
The 2 silicon diodes are probably just signal diodes, 1N4148 or similar and the zener, well 5.1V is pretty common so sub something in there just for a check.

[Vince]

OEM Fan specs : Thank you Tek... I didn't think I would get THIS lucky, but the part list gives all the specs about the fan ! 
Does not talk about static pressure though, but all the other info is there, even tells you  what type ofb earing is used ! 

12V, 1.5W, this we already knew, was written on the fan itself
Bearing : sleeve bearing
Air flow: 19cfm
nominal speed : 4600rpm

There I go ! 

Now the question is : can I improve on this : maybe a different type of bearing would be more silent, for example ? Also, intuitively, an increased air flow might sound like a good idea, but this static pressure thing intrigues me as I don't fully understand it and it's relationship with the all important air flow. I mean, I guess the airflow is measured "unloaded"/unrestricted, but once inside a closed cabinet, by how much will the air flow drop.
A fan with a higher "nominal" air flow is no good if that flow drops by so much that it ends up lower than maybe a fan with a lower airflow but that can actually maintain that flow better than the other fan, once in the enclosure.
I guess I need a little training on cooling fans eh ?!........ 

But until I can figure this out, at least I have the specs of the old fan as a safe starting point...

[Vince]

So is TP 56 waveform correct ?

I still have not been in the lab today... but going there in a minute  !

The 2 silicon diodes are probably just signal diodes, 1N4148 or similar

Yes, just checked them on the parts list. It's the same general purpose small signal diode that they seem to throw everywhere in the scope, same that I replaced in the regulator in the SMPS.

[Vince]

OK, so after some probing... see the screen captures below. I kept the same vertical gain on all 3 pictures to make it easier to compare them at a glance : 2V/DIV
Sorry the square waves don't look really square... the old Hameg is only 60Mhz,30pF input capacitance, it's too much for the compensation range of the 500Mhz probes I got for the TDS544A.  I do intend to buy more general purpose,  slower/ switchable x1/x10 100MHz probes, to go with my slower scopes, but there are so many items on my "to buy for the lab" list, and so little money... well I will buy some, just don't know when...

I did notice while having a quick glance on Ebay though, that there are some dirt cheap Chinese 100MHz probes, like way under 10 bucks each, I wonder what problem to expect from them, or if they are actually any good. For slow stuff I guess it does not take rocket science to make a decent product ?  At any rate they are so cheap that I don't have much to lose... if I like them, I keep them... and if I don't well I can just get something else.  I guess the problem is that I don't have a function generator that can go as high as 100+ Mhz, so I would not be able to plot their attenuation vs freq response, to see if they are decently "flat" or not...
I might post on this forum later to get opinions/reviews on this subject...
... and here I just added yet another item to my list : an RF generator.... more money again ! 
No worries... I will just have to get a broken/cheap unit, and that will make for another repair topic !



Anyway, waveform #56 is fine, matches what's in the service manual. The op-amp oscillates as expected between the +/- 8,6V rails, though as expected it can't really reach the rails. Around 7,5 positive and 6.4V negative. 2.2V away from the rail, but who cares, shouldn't make any difference here. Plus, it's still well within specs, according to the datasheet I pulled at random (Texas instrument from 1998).  It even says that the minimum supply voltage ought to be 10V, not 8,6 ! So we can't really complain really... Plus, even if it were operating in the recommended voltage range, 2.2Volts from the rails would still be plenty within specs... yes, they are very, very  wide....   
So, voltage levels don't really matter at this point... the little bugger oscillates between the rails and that's all we are asking from it...


Then I probed downstream, step by step. 

First step is test point 'A' (at the anode of CR988) : here the positive half of the signal is capped. Signal now goes from -7.5V to +1.6V

Second step, test point 'B' (at the cathode of the same diode) : now the lower half of the wave is trimmed as well. Signal now spans -1.2V to +1.0V

Third and last stop, the output signal : as already mentioned, it goes from -0.5V to 0V, correct/expected amplitude, but purely negative signal, rather than the positive one one would expect...

Both signal diodes tested fine. The Zener diode however I could not test in situ : gives 200mV both ways... because R990, which is 200ohms only, is right across the Zener. The multimeter injects a calibrated 1mA current into the circuit, so obviously I get bang on 200mV across the Zener/resistor...

But at least we know it's not a dead short...

But again I just don't get how one can possibly dream of having a positive signal out of this circuit, given which way around, CR989 is mounted...   
I am really lost.

I will try to see how difficult/risky it is to pull the front panel assembly out of the scope so I can get sufficient access to play with individual components... but I guess/fear it involves removing all the knobs from the front panel, and they are so fragile that I have already broken 4 of them... so am a bit reluctant to take any chances with the remaining ones   


Hell... I looked at the user manual.. nowhere does it say that the probe compensation signal is supposed to be positive... nor negative, not symmetrical... they never, ever reference it to ground... they just say "500mV square wave"... 
Admittedly that's indeed all we care about when compensating a probe, but still... feels strange to me to have it negative ? I am starting to feel like I am losing it... where are my pills...  could people with a 2232 or at least some 22XX scope, tell me what polarity is their signal ??   

[WastelandTek]

While I don't, in general, disagree, there ARE some high quality fans targeted to the PC enthusiast market.  I would point in particular to the offerings from Noctua (particularly their industrial series), and Nidec Servo (the OEM of the well regarded "gentle typhoon")

On fans. Be very wary of anything recommended by the "'PC gamer nerds". PC's don't have to last anywhere near as long as quality test equipment, and the fans are built down to a price. Airflow ratings and static pressures are generally rated in PMPO Chinese watts. Additionally, PCs don't have irreplaceable ASICs and other expensive parts that fry when your 'buck fiddy' "ultra-mega-whisper whopper-flow hoosit waffle-cone bearing super flower blade' PC fan bearing carks it and it grinds to a silent halt.

[David Hess]

So I was starting to wonder if maybe... the scope is perfectly fine, and it's just oscillating due to the presence of the bulb.  Haven't thought of this in detail, but the thermal inertia of the bulb filament, coupled with the filter cap getting charged and discharged... maybe there is some back and forth game between that filter cap and the bulb, which would explain this oscillation, and it's perfectly stable frequency ?! 

Bulbs and electronics.. all of sudden it reminds me of an experiment we did at school, where a miniature filament bulb was used as part of the feedback loop in some op-amp circuit, for automatic gain control IIRC.

The fact that an SMPS presents a big filter cap to the mains, meaning the voltage is lagging, make it prone to oscillation, the light bulb tester thing was probably "invented" before SMPSes were the norm ? With a linear PSU, hence a big transformer, the light bulb tester sees a inductive load, not capacitive, so the current is lagging, rather than the voltage.
So I guess the light bulb tester is indeed an excellent tool when restoring older tube Tek scopes that use a big transformer.. but for more modern Tek scopes with an SMPS/ capacitive " front-end " so to speak...

The problem with dim bulb testing a switching power supply is that the input to the switching power supply has a negative resistance input characteristic; at low voltages the switching power supply draws more current and at high voltage it draws less current.  This can result in the switching power supply latching off or continuously restarting.

Would still like to get that light bulb tester working here though ! I was advised 100W... my modern halogen bulb is only 77 Watts, the more powerful I could find.
How does one calculate what power rating is required in order for the gear not to oscillate ??  Does that the fact that it's a halogen bulb changes things a little bit compared to an old fashioned bulb, and that would need to be taken into account ?

For now, I guess the answer is : if it oscillates, just put a more powerful bulb in there, until the oscillation stops ?!   

The bulb resistance needs to be low enough that the input voltage to the switching power supply does not drop below the level where the under voltage lockout trips.  The 2232 switching power supply does not have an under voltage lockout however at low voltages, the duty cycle and current of the switching power transistor goes up and eventually either trips the current limit causing shutdown as the output voltage collapses or the switching power transistor self destructs from excessive power dissipation do to high current and high duty cycle.

I am still surprised that the most vital and fragile signals in the scope... the input signals, travel from board to board across 8 inches or so, on a simple flat cable instead of a coax ? I guess that is is carried by a differential pair, surrounded by ground on both sides as well, keeps it from picking up too much noise along the way.

Exactly, this signal is differential so twisted pair or in this case a twin lead transmission line made from some ribbon cable works fine; the grounds control the impedance and reduce external coupling.  Schematic 2 shows where these signals originate; channel switches direct the differential signal from the vertical preamplifiers to either the non-storage or storage display circuits.  Note that the signal going to the storage circuits is a current and only terminated on the receiving end which you can see on schematic 14.

If you reverse the connector, then the storage display for that channel will be inverted compared to the non-storage display and trigger of that channel.

3) Diode CR989, given how its mounted... I don't see how any positive voltage could ever come out of that probe test point !   

So.... could it be that I am losing it ? Could it be that this scope has always showed me a negative compensation signal and it's only NOW that I actually happen to pay attention to it ??     I still find it weird... all my other scope give a positive signal.

The probe compensation output has always been negative.  It shifts between ground and -0.5 volts.

When the output of the operational amplifier goes positive, CR989 disconnects it from the probe output and the output settles to ground via R990.  When the output of the operational amplifier goes negative, CR988 disconnects it from R989 and the output voltage is determined by R990, CR9889, and R989.  Current through R990, CR989, and R989 becomes -2.5 milliamps.  -2.5 milliamps across R900 produces -0.5 volts at the output.  VR9900 does not conduct much in this case and can be ignored.  VR9900 is there for some type of protection but I am not sure what.

The virtue of this design is that the output voltage only depends on the ratio between R990 and R989.  The high value of R989 makes the current through CR989 relatively constant so the forward voltage drop of CR989 has little effect.

[Vince]

Hi Dave, thanks for the comments especially the operation of the probe signal circuitry, that was interesting.

I just asked on the Tek yahoo group and someone confirmed that indeed it's negative.  So I feel relieved, I can stop chasing red herrings, and stop trying to understand how this circuit can deliver a positive signal, because it obviously can't !

So that's one less problem in the list.

I removed a few more problems too, on the digital board :

- identification of the two ribbon cables : Tek did not label them, so no easy way at reassembly time, to know which is CH1 or CH2. Easy to figure out though : just swap the cables in store mode if they don't show up where they should...

- orientation of the above cables : again just enter store mode and flip the cable(s), live, 'til you get the correct polarity on the screen.. result : yes, Tek indeed did NOT match pin #1 of the cable connector, with pin #1 of the socket on the board ! How "nice" from them.... grrr...

- trace not centered in store mode : while looking at the ribbon cables, was obvious that there was some analog circuitry with trimmer, before the signal get to the ADC modules. So some of that circuitry maybe allowed to adjust the offset, which then would be of course unique to the digital board and would explain why the trace was perfectly centered in analog mode and only off-center in store mode. Looked at the service manual, indeed this analog circuitry is there to shapoe the signals prior to sampling, the usual/expected stuff : offset and  gain, as well as  "HF peak" which I don't know anything about. But whatever, I was only interested in the offset. Checked the service manuakl again to see if the adjustment procedure might be more complex than meets the eye, but it was not.
So I just tweaked the trimmers on both channels and now I am pleased to report that store mode now shows perfectly centered traces on both channels, the trace does not "jump" when I switch back and forth between store and non-store mode, it's now seamless... lovely.

So..... I am getting there !

I think I have solved all my problems  now !    The power supply is fixed, store mode is now feeling better than ever, and the probe signal was a red herring.
I just need to replace the fan because I don't feel like opening the scope again later just for that, and IU will done !    It's great to have my little 2232 back in business, it's a great all rounder for general work I find.

For the fan, I obviously can't order from Digikey, so as always I will have to do with what Farnell has to offer. 2,000 fans vs 12,000 on Digikey... not in the same league. They don't carry the same brands though, and somehow appear to have a good selection of appropriate fans. 24 items, if I include all types of bearings.
There are two Sunon models in the lot. The "KD' series, at only 10 bucks, cheap because apparently it is being phased out. The current model "Maglev" series) is more expensive at 23 bucks, but price drops down like a rock : only 8.8 bucks/Euros for just 5 units ?!  It's almost worth buying 5 of them !

So just need to get a fan, then I can put the scope back together and move  on to the next repair.... my recently acquired TDS 544A is misbehaving...

Oh, and I am disappointed... earlier I said that the transformers where not making any hissing noise... and that the fan was. No... I stopped the fan by hand and the hissing is still there... noooo... I wonder which of the transformer is doing it. I guess it's the regulator coil, because when I bypassed it to feed the inverter with the bench power supply, it was silent.  Anyway, hopefully once the cover is put back in, I won't be able to hear it that much. Plus, I already have another coil coming in the mail, should be there soon. and of course I don't give up on my (educational..) pan of wiring up my own coil/transformer, so I would have two more to try ! Maybe one of them will be silent...