Tektronix 475 scope: repair required.

[PJL]

Hi Al, just replace Q1202 with a PNP transistor, it just has to be faulty but must be in a way which is not apparant with the tests you have been performing. I would have thought any 120V PNP small signal transistor is going to work. Peak dissipation is only 200mW or so.

The last test might be to measure the voltage between Q1202 BE rather than referencing ground - just in case...remember my post about the dry joint that fixed itself whenever you applied pressure to the board.

I make stupid mistakes when frustrated so just in case you have done the same, there is no chance you are fitting it incorrectly?

[Skywave]

Q1202 has been replaced 3 times with same result.

The actual voltages on Q1202 B to E correspond exactly with the measurements recently reported - which make no sense whatsoever!
The transistors are now hard wired (using solder) on short wires - about ½ inch long direct to the tracks where the transistor sockets are. These sockets are now not in use, of course.

At one point, yes, I did stand well-back and ask myself "Am I in so deep with this that I'm missing the glaringly obvious?" - as in "too close to see it?" As you imply, this does seem to be an annoying by-product of the way our minds work when faced with a truly baffling set of conflicting conditions. And yes - I have 'been there before' - sometimes with embarrassing consequences! (Haven't we all?) So I asked myself (silly . . . or maybe not so silly) questions like "Have I got the NPNs and the PNPs mixed up?" "Have I wired the transistors in wrong - i.e. BCE wires not going to the BCE correct places?" "Have I made a circuit change somewhere recently that I have since forgotten about?" (I've had a lot of interruptions whilst trying to fix this wretched 'scope). Needless to say, I could not discover / unearth anything amiss - but that doesn't mean that there isn't still a gremlin in there! Oh, and yes, since you mention it, I also have had my share of 'flexy' faults in my time.

Anyway, for the next few days I'm fully occupied on other things, so this will have to wait for a bit - again. However, looking at the circuit "d.c. wise", i.e. its 'stripped down' form with its a.c. compensation circuits removed, I am beginning to contemplate building a replica of the cct. of the "top" driver on a bread-board - just to see what then comes to light; the cct. isn't that high on component count. Things have moved on a bit with this job - it's become a question of "Why doesn't basic electronics theory work here?" rather than "I'm trying to fix this 'scope!"

I'm sure you know what I mean - it's one of those situations where, as A Man of Science, if you don't ever get an answer - and an answer that makes sense - it'll haunt you for the rest of your life .
And this job is very nearly at that point anyway! Reminds me of an old Jewish proverb: "Time or wine will see the truth out".

Al.

[dave cox]

My suggestion:
unsolder the collector lead from the upper current source and connect 50K pot plus 5K in series to -8v line. If my calcs are correct the 3mA source should swing it collector between +7v (3mA into 5K relative to -8v) and +104v (3mA at saturation) as you twiddle the pot. Best to check my calculations though !!

dc

[Skywave]

Hi Dave,

You know it's strange how great minds think alike: whilst I was out-and-about on my travels today, I had the same idea - and eventually I get home and read the immediate above. Whilst recovering from a hectic day over a cup of coffee, I did a quick cct. analysis of that transistor stage - and yes, I calculate 3 mA too.

Now as soon as I find the time . . . .

Al.

[PJL]

Just double check the following:

1. B-E on the transistor leads is -4V
2. R1202 measures 1.8K and has 0V across it
3. R1201 measures 1K and has 0V across it

If this is the case my thinking is:
1. The transistor is faulty
2. The transistor is the wrong type
3. The transistor is fitted incorrectly (perhaps the pcb markings are wrong)
4. The transistor is not a standard PNP type
5. The circuit diagram is wrong (and there is a fault)

[Skywave]

Fixed it!

As you can well imagine, there's a tale to tell about this; I'll try to be brief, but it is quite involved (surprise, surprise! ) Also some interesting lessons have been learnt . . .

First, I picked up on the suggestion of testing Q1202 (the PNP) and its design as a constant-current source. I used an external 100v d.c. PSU and built the same cct. using new resistors and fitted each PNP transistor in turn and made measurements. All of the original and the substitute PNP's that I had been using gave the same result: a current of 3 mA into a 8k2 load resistor.
So, nothing wrong with these transistors then.

Secondly, transfer all of these components to the Tek. 'scope itself - making appropriate connections to the PSU rails. Check output current: none. Hmm. O.K.: re-trace steps: what's different?

In my external test cct. (as above), I obtained the +104v (for the base feed of Q1202) via a resistor and a 6v2 zener diode. So I looked at the voltage rails with the test cct. now fitted in the 'scope: the "+104v" rail was at +108v. Hmm. Can it be that this voltage is that critical? A study of the cct. diagrams show that the +110v rail - which is produced from the main PSU - is fed to the cathode of a 6v2 zener and thus to a 47k resistor to the main +50v rail. The +104v rail is obtained from the junction of the 6v2 zener and the 47k resistor. With Q1202 out of cct., the voltage at this point was +110v: obviously wrong! Why? Because the 47k resistor was open-circuit!

Then it got really interesting. Upon replacing the 47k, I obtained +108v.: a 2-volt drop across the zener. Huh? O.K., replace zener. No difference. Hmm. But there's not much current through that zener, is there? So, replace the 47k with 22k. Result: +104v.

So, re-build the entire stage around Q1202 & Q1204 - using the original Tek. transistors and new, replacement resistors (just to be really sure). Switch on and measure voltages: all correct! And a horizontal scan as well! Do a quick re-calibration check & tweak as required; put on soak. Two hours later: running fine; no signs of anything overheating (including the 6v2 zener and its associated 22k resistor).


Conclusions.
There have been some serious lessons learnt here & a few questions that need answering.

First, my thanks must go to PJL: his observation that Q1202 is there to provide a constant-current source had escaped me. That was the critical turning point.

Secondly, I have to say, in my humble opinion, that a cct. design that is so critically-dependant on a derived voltage rail - to wit, the "+104v" - that a small shift from this value can cause a complete failure of the horizontal scan with very confusing symptoms - does not impress me. Along with this is the method whereby the +104v is obtained. In my test rig, I simply used a 6v2 zener in series with a suitable resistor connected to 0v., using the zener to lose the required 6v from the +100v source. So, why the strange design in the Tek. 'scope?

Thirdly, the way that the cct. is drawn in the Tek. manual certainly does not help. Although the +104v and +110v rails are obtained from the main +110v rail from the PSU, the components that achieve this are not clearly situated on the schematic as part of the O/P stage around Q1202. This is especially true of the +104v 'rail' - which, in effect, is really part of the constant-current source (Q1202, etc.). It's a great pity that the schematic wasn't drawn properly to reflect that. My failure to notice that short-coming significantly contributed to the trouble that I experienced in the cct. analysis & subsequent fault location.

Now don't get me wrong: I have a good deal of respect for Tektronix 'scopes - including this one - but I do feel that my observations & comments above are appropriate.

Finally (and almost needless to say) I've made notes about this saga for the next time - yes, there's always a 'next time', isn't there?

And my thanks go out to all those here who have helped me fix this very challenging fault.

Al. / Skywave.

[Patrick Dixon]

Published schematics on things of this age were often redrawn from engineer's drawings by the technical drawing department, so you don't always get things laid out ideally and logically. I recall one TD department redrawing BBC circuits as if they were a PCB layout, with the components all spaced neatly in an n by m grid and a complete rats nest of connections between them. Needless to say, we filed them in the bin and used the back-end of an envelope originals.

Well done!

[Skywave]

I know Patrick, yes, I know!

Just a quick glance at the schematic of any modern Hi-Fi system (for example - there are many others) reveals same. I just would have expected Tektronix to do things a little more professionally. Generally, the cct. descriptions and general layout of the Service Manual is well-up to the standard you'd expect from a company like Tek. It's patently clear that their intention was to produce a Manual, written by engineers for use by engineers.

Al.

[Red to black]

Well done Al,

you got there in the end

Looking again at that circuit diagram with the "odd resistor values" (close tolerance) does suggest that this part of the circuit is critical, Hindsight is a wonderful thing

well done again have a you deserve it!

cheers,
Baz

[Skywave]

Originally Posted by Red to black

Well done Al,

You got there in the end

Baz

Just another case of the Three Ps, really: Persistence, Patience and Perseverence - and little luck, of course!
Oh, yes - and Perception. Sometimes, what you're looking at & thinking about - isn't what you think it is.

Al.

[PJL]

Great result. There is nothing more disheartening than abandoning something having spent a long time on it. Peter.

[Skywave]

Yes, Peter - I know; been there! Haven't we all?

It was not just recovering the "utility factor" that was motivating me - relevant as this is for the item in question. There was also the issue of "there is something stupid going on here!" Now I don't think I'm ever going to discover some weird electronics principle that is going to revolutionise the World of Electronics . If I had simply walked away - and I was tempted - I knew it would have remained as one of those annoying memories that would keep coming back at all the wrong times, plus, of course, the doubts it would have placed in my estimation of my own abilities. Simply put, it all got a bit personal & psychological - and, yes, there were moments when that showed!

Nevertheless, I learnt a lot fixing that - useful experience and knowledge in a general sense - and also for the Tek. 475 'scope.

Al.

[jamesinnewcastl]

Congratulations - I agree that unsolved problems are like a pea under the mattress! Equally fixing a problem entitles you to 'smug' for ages!


James

[jjl]

Originally Posted by Skywave

Fixed it!

As you can well imagine, there's a tale to tell about this; I'll try to be brief, but it is quite involved (surprise, surprise! ) Also some interesting lessons have been learnt . . .

Well done Al.

I get the impression that Tek always worked at the leading edge of analogue circuit design in order to gain the best performance from the components available at the time. Their design engineers knew this; often, a cartoon wizard or similar amusing cartoon character is drawn next to a particularly clever part of a schematic in the service manuals for their 'scopes.
As a result, some circuits are a bit unusual or require very precise adjustment of supply voltages.

Having said that, I agree that running the Zener you described with such a low current by design does seem too marginal to work reilably.

John

[Guitarist28]

Hi Al,
Well done and I am so pleased that you acheived your goal and fixed the problem.

Regards

Rob

[Skywave]

Thanks Rob.

Looking back, I must say that my initial efforts to fix this left me so "I can't make head nor tail of what the is going on here!" that I did choose to try to 'farm out' the repair; despair had set in.

But what happened? Enthusiasts here - yes, like myself, but on a 'good day' - said, in effect: "Oi! You can fix that yourself - with our help!"
And they were right, too.

For which I am very grateful; thanks to all who contributed - and encouraged me to keep going! I owe you all a return favour. Maybe, here, one day . . .

Al.

[Skywave]

Originally Posted by jamesinnewcastl

Congratulations - I agree that unsolved problems are like a pea under the mattress!
James

Thanks; I could think of a stronger similie under the circumstances!

Al.

[XTC]

Well done, and definitely one well worth remembering.

Tek 475s are very good scopes but they are getting on now. Mine's gone wrong twice in the past two years but luckily nothing this obscure.

Pete.

[Skywave]

Originally Posted by XTC

. . . but they are getting on (a bit) now. . . .
Pete.

Indeed; I know the feeling well! I wish I was as young - or moved as fast - as this 'scope!

Al.

[Skywave]

This Thread has clearly reached its end. Whether it - or any parts of it - are worth archiving I leave for others to decide - but obviously it can be closed now.

Al. / Skywave.