No horizontal scan on Tek 465B

[465-owner]

Greetings All,

I recently bought a Tektronix 465B oscilloscope for eighty Euros as a source of parts for my old 465 not realising that there were such large differences between the two models.
With lots of help and encouragement from some wonderful people in this forum my original 465 is working happily again and I now ask for assistance with the newer 465B.
I don't have a manual for the scope although have downloaded three versions none of which seems to have a circuit diagram of the horizontal Miller timebase oscillator.
With the help of the pieces of manual I have I established that there is no sawtooth being generated by Q4498 and Q4598.
I can't specifically identify the collector of Q4598 but suspect it goes to a choke wound on top of a resistor which then goes to Q 4588. there is +20 volts at both ends of the choke.
The voltages on Q4498 are 0.7v, 15v and 15.7v all constant.
Other information which might be relevant:
There is a spot far to the left of the display area which can be moved nearly on screen with position control.
The beam finder will bring the spot to the screen
The vertical amplifiers are working and both make the spot in to a vertical line shown with beam finder pushed.
The X-Y mode works but display is off to the left.
The trigger LED is sometimes on and sometimes off with probe hanging on to the calibrator.

Thank you for any assistance which you can offer particularly for the Miller timebase part of the circuit diagram.

Nigel

[Radio Wrangler]

OK, it sounds like your X amplifier may be OK, and the timebase is in the reset, ready to go state.

So the timebase may be stuck, unable to ramp, or it may be waiting for a trigger which never comes.

I've got a couple of 465Bs here, but no manual as yet

David

[maninashed]

There is a free manual available for the 465B on the BAMA website.
First things first, check that you have ALL the correct power supply voltages, +110v, +55v, +15v +5v and -8v, and check the ripple with your other scope, there are quite a few logic chips in these and they don't like a lot of ripple and smoothing capacitors are a problem, especially on the +5 and -8.
The fact that you have a display of sorts means the hv circuitry is working ok. The fact that the dot is off screen to one side and you cant move it horizontally means that there is some imbalance on the X amp, possibly a power supply failure.
Check that you have a ramp voltage generated from the time-base generator, I think there is a ramp output on a BNC at the rear of the scope.
Let us know what you find.
Bill

[Alistair D]

Around the time of the 465B TEK had problems with IC sockets so it would be worth re-seating the chips on the right hand board. My own 465B suffered a no trigger problem for exactly this reason.

Al

[maninashed]

That's right Al, the early 465Bs used cheap ic sockets and they do cause problems, later versions used turned pin sockets and these were much better, similarly, the transistors should be waggled a little to remove any corrosion on the leads.
Lets see what Nigel comes back with on the psu voltages, as I suspect the -8v is down and/or ripply.
Bill

[Chris55000]

Hi!

I have these diagrams available! Try these for now - if you're still missing anything I'll ferret thro' my collection for you and post what you need!

I'll be back with fault-finding suggestions later!

Chris Williams

[Chris55000]

Hi!

Here's another page!

Use your working 465 to make sure you've got the positive rectangular "Sweep-Gate" pulse (Waveform 71) present on P4571-1 - refer to Diagram 7 for this!

In "Auto" mode there should be a recurring rectangular pulse with a amall 'pip' on the rising edge - the small 'pip' is used to initiate the start of the sweep via Q4588 and Q4598!

I'll have to come back to this with more hints later I'm afraid!

Chris Williams

[Radio Wrangler]

Before all of this, just select X/Y mode and check that the X shift knob can move the dot across the screen, and that a signal exercises the X amplifier. It can quickly eliminate a lot of circuitry from suspicion.

David

[maninashed]

If you note what Nigel says in post #1 David, he says
"The vertical amplifiers are working and both make the spot in to a vertical line shown with beam finder pushed.
The X-Y mode works but display is off to the left. "
This would suggest to me that there is an imbalance in the x amp, as if there is a supply rail missing, or one of the output pair is u/s.
He also says there is no ramp waveform. So there may well be a couple of faults in this case.
Bill

[Radio Wrangler]

Thanks, Bill, I'd missed that.

X amps are easier to sort than timebases, anyway.

David

[465-owner]

Greetings, David, Bill, Al, and Chris,

Thank you very much for your great contributions and suggestions.

I regret I haven't yet had time to do more than download the very generously donated circuit and other PDF data and have a very brief look at it.
I had previously downloaded some of the manual from the Bamba site but even the part '465bMissingPages.pdf' didn't include any circuit diagrams of the sawtooth oscillator.

On Tuesday I checked all the power supply voltages (+110v, +55v, +15v +5v and -8v) and all are within a few percent and don't have any significant ripple.
I have wriggled the chips and the transistors but none seem to be the cause of either the lack of sawtooth or extreme left horizontal shift.

I've spent a bit more time on the scope today, below are my findings.

Bill, I can move the dot with the horizontal position control but can only see it horozontaly move when pressing the Beam Finder.
There is no sawtooth out of either of the rear A or B Gate BNC sockets.

David, In X-Y mode the channel 2 position control does shift the spot with Beam finder held in but the channel 1 or Y position control does not move the spot (with Beam finder held in).
(( I checked this on my old 465 and found the channel 1 (X) position control does not move the trace horizontally either)).
In the attached photo you will see a sine wave applied to both inputs in XY mode and with Beam finder pressed to show display.

Chris, I do not have any waveforms on connector P4571. 55 volts is present on pin 2.
Pin 3 has 200mV. Pin4 has 900mV. Pin8 has 800mV. Pin 9 has -3.4v. Pin 10 has 280 mV.

Taking a fairly random location at Q4273 & Q4274 with the position control centred (3.7 v on slider) On the collector of Q4273 I have 4.31 v and on collector of Q4274 there is -268mV.
On the base of Q4285 there is 4.3 V and on emitter 122mV (collector -8v)
Q4475 has 700mV on emitter (correct) but no shape.
Q4598 21 Volts on collector & -6 V on base. (No shape).
Q4497 -284mV emitter, base -974mV. 15.6 volts on collector (pin 2 of J4487) No shapes on any leg. I note from manual I should have -1.5v on collector and -4v on base.

The troubleshooting diagram of the manual tektronix_465B_service_manual.pdf page 258) suggests "Check DC level at collector of Q4598. If 0 to 2 volts check sweep start logic. . . . If 13 Volts "Pull Q4497. Check for sweep level on front of R4475 to go to about 0 volts." I couldn't find R 4475 but checked and found no shapes on Q4498 or Q4598.
The manual troubleshooting continues: " If No Check the sweep generator circuit".
In case there is doubt about my ability to check shapes I regularly check by touching prope tip with finger to pick up mains hum.

The shapes above were looked for with calibrate signal applied to channel 1 and sweep set to 1mS / division. Trigger light ON although it usually stays ON even if probe is removed from calibrate source. At the back of the scope the channel 1 signal is taken to channel 2 on the older 465 to trigger from.

I also looked for shapes on Q4498 in single sweep mode but with the same negative result as above.

I suspect Bill, you may be right and there are two faults, one, no oscillation and the other the large horizontal position offset.

I noticed but didn't investigate P6025 on the timing board does not have anything plugged in to it! Might it be for factory calibration or testing?

Unfortunately I've spent too long on this today and will have to try catch up with other work for perhaps a day or two.

Thank you all again so much for the suggestions which I hope I have commented on sufficiently above but if I haven't please be good enough to remind me what additional information would assist diagnosis of this scope's ailments.

Best Wishes Nigel

[Radio Wrangler]

Hi Nigel,

I have a running 465B reside me.
With no signal, XY mode selected and beam finder pressed I can move the spot vertically with 'Channel 2 Position' pot. I can get 2.5 divisions up and 1.5 divisions down max deflection from the centre.

Your photo shows it limiting at about these values. Pressing beam finder reduces the bias current in the deflection amps so the amount the spot can be moved by anything is limited.
Your vertical limits match mine.

Channel 1 position does nothing, the scope uses the normal x position pot in XY mode.
Using the x position pot I can get the spot 3.3 divisions right, and 3.2 divisions left with the X position pot. A bit different to yours, but nothing massive.

Without the beamfinder pressed, still in XY mode, with no inputs I can place the spot anywhere on the screen with CH2 and X position pots. In this state the timebase is not being used at all. If you can't do this then there would appear to be a fault in the X amplifier path. Find and fix this first, because you'll need the X amp working to be able to see what the timebase is doing when you come to fix that. There is one other possibility is that a timebase fault is blanking the trace permanently, and you might be able to move the spot anywhere, though you won't see it if it is held unlit except when the beam finder overrides the blanking signal. XY mode should also override blanking as well.

David

[Leon Crampin]

Having repaired a few Tek portables, following a close scrutiny of the power supplies for voltage and ripple, I would check the balance of the X amplifier output stage.

Usually, the output devices themselves are OK as they are hermetic metal can devices, but the drivers are frequently Fairchild plastic transistors which by now tend to fail due to moisture ingress. Very often in both the X and Y stages, a driver will fail (usually open) which unbalances the amplifier and allows only limited operation from about screen centre to beyond one edge. Poor transistor socket connections are another possibility - usually brought to light when working in the area.

I use military grade BC107/BCY70 devices as appropriate - hermetic cans, gold plated leads which make good contact in the sockets and a standard TO-18 pinout. I have not measured any loss of bandwidth, even in the Y amplifier.

Leon.

[465-owner]

Greetings,

Thank you David and Leon for your comments and suggestions.
Taking the advice to first investigate the off centre trace problem I set Timebase to X-Y mode and applied calibrator signal to channel 1, input set to 200mV /division.
With the position control, set to approximately centre, 4 volts on slider I have the following readings:
Transistor, e , b , c
Q4284 ,Grnd,-0.7,-2.4
Q4285 ,-2.2,-2.8,-8
U4269B ,0.7v,30mV,4.8
U4269A ,0.7v,70mV,1.5
Q4273 ,5.4v, 4.8,-3.6
Q4274 ,1.5v, 2.2,-1v
U4269D ,-5v ,-4.3,-2.3
U4269C ,-5v ,-4.3,-0.34
U4269E ,-4.9,-4.3,-4.3
Q4169 ,-0.5,-0.5, 4.2
Q4362 ,1.2v,0.58,-6.8
Q4161 ,Grnd,0.65,4.43
Q4361 ,1.36,0.72,-2.35
Q4146 ,108v,107v,85v
Q4341 ,108v,107v,13.4
Q4150 ,4.43,4.99,85v
Q4342 ,-2.5,-1.9,13.5

You will notice the severe discrepancy bewtween the voltages on the X deflection plates (13 volts and 85 volts) however the applied calibration square wave is showing as a 5 volt p-p square wave on to both plates equally and is visable on the screen as a 1.5 division deflection both horizontally and vertically when the position control is shifted fully to the right.
You may wonder why two ends of the same print connections are different. (Q4361 collector & Q4342 emitter) I believe this is probably because the mains voltage changed a little and the 110 volt unstabilised supply also changed.
I tried the following experiments.
I unsoldered and swapped the deflection plate connections and noted the fault still existed but positioned dots to the right (instead of left).
I swapped over both sets of output transistors but the swap made no difference. (Q4146 with Q4341 & Q4150 with Q4342).

I would much appreciate any further advice that you could offer.

Best Wishes
Nigel

[Radio Wrangler]

X plates are a lot less sensitive than Y plates (nearer to screen, not needed to have as much amp bandwidth, plates further apart to pass the Y deflected beam) And the figures you quote fit with the offset on the screen, so no surprise

To handle that data dump I'll need to print the schematic and plot the values on it which makes it a job for when I have a bit of time. The answer, or at least a good clue should be in there. Norm procedure is to put shift pot central and look at DC voltages, no signal.

David

[465-owner]

Greetings,

Thank you David for your help. I couldn't expect you to take up a lot of your valuable time to print the schematic diagram and copy my measurements to it as well as helping to diagnose the problem! I should be able to print, mark up and repost without too much difficulty. The only possible problem might be posting a large enough attachment which can be printed and read. Jpgs are limited to 800 x 600 pixels. Perhaps I could save the marked up picture as a PDF, I'll try and hope to be back later today.
Again, thank you for your help.
Nigel

[Radio Wrangler]

It's printed and I'm scribbling away. I was about ready to doze off last night. Today I'd been intending welding up a shaped scoop on a long handle (to clear a friend's septic tank!) but there's a lot of wind and it'll just blow my shield gas away. That job can wait until the weather calms down.

I think you've done a couple of typos in your data. U4269A and U4269B both have bases at a few tens of mV positive, but emitters at plus 0.7v. I'd expected -0.7v on these places.


Don't be worried by the funny resistor values. The usual familiar ones fit 12 steps into a decade change 1, 1.2, 1.5. 1.8. 2.2, 2.7, 3.3, 3.9. 4.7. 5.6, 6.8. 8.2 and starts again from 10. It's an exponential series with each increment a ratio of the 12th root of 10. It's for 10% resistors and means that the tolerance band on each value just meets that of its neighbour. So any resistor they make can be sold as one value or another (cunning!) The real numbers would be irrational (apart from 1, 10, etc) but they get trimmed to two digits. It's called the E12 series.

For 5% resistors there's the E24 series and some other values get added between the known ones so you get things like 62 Ohms

For 1% it's got a lot more serious, the E96 series! 96 values in every decade, and they get trimmed to 3-digit precision. We used to routinely use 1% resistors, but kept to those values nearest to the E24 ones just to keep the varieties down. The others were still available when really needed. so that 62 ohm 5% resistor would have been 61.9 ohms in 1% where it got trimmed to 3 not 2 digits. And in E96, the increment ratio is the 96th root of 10.

Anyway, I'm back to filling in that diagram. U4269 A and B collectors look like the input stage is well swung over before we get further down the amplifier.

David

[Radio Wrangler]

Right, let's jump in at U4269C and D

Both bases are at -4.3 and both emitters are at -5v. So both NPN transistors are biassed-up and operating. Good!

U4269D has its collector at -2.3v (Tek diagram says -1.4)
U4269C has its collector at -0.34v (Tek says -1.4 again.)

Let's look at that. the average of those collector voltages is (-2.3 + (-0.34))/2 is -1.32v which is pretty close to tek's figure of -1.4v.

It looks like your amplifier is making sense at this point and is faithfully amplifying a non-central voltage from earlier.

We've just split your X amp in two and said that the latter half looks OK.

So we start at where we split it and start working backwards.

Q4273 and Q4274 now make sense The average of 2.2 and 4.8v ontheir bases is 3.5v. Tek have written 3.1v for when the amplifier is balanced. Yeah, I'll buy that.

Things get sticky at the next earlier stage. U4269A is a summer, much like an audio mixing desk where several signals get added together in variable proportions. U4269B is the straight guy for U4269A. U4269A sets a ground reference to balance U4269A

Note that although all the subsequent balanced amp stages have emitter to emitter resistance to control gain, U4269A and B do not, they get individual shunt feedback through R4371 and R4367 respectively.

Both bases should be close to ground. 30mV on U4269B is the reference. 70mV on U4269A says its base is getting a good bit more current than that of U4269B.

OK, the voltage difference looks trivial, but with no emitter to emitter resistance, this stage has a lot of voltage gain form base difference to collector difference. What this stage is is a current to voltage converter, and it's getting the sum of the currents going into U4269A base as its input. That is the sum of currents from two timebase circuits and two shift pots.

Seems to be working.

It's beginning to look like you have a timebase fault, and that fault is still passing current to the X amplifier input even when you've got XY mode selected.

I hope you've picked up on the technique of jumping straight in to the middle of an amplifier chain, looking around and seeing if things are sensible and fit what shows on the screen. It halves the amount of work needed! Cheating is an art-form.

It's a differential amplifier, all balanced to a nicety, and we can exploit this in comparing one side versus the other.

Timebases area bit harder to get your head around than straight forward amps, but they're a lot easier since the end of the tunnel diode era!

David

[maninashed]

I think David my very well be correct in his diagnosis. However the 5 transistor array U4269 is a known problem in this X amp, its a CA3046, and its cheap, I would be inclined to swap this ic. Check the resistors around U4269A and B for correct values.
Ive had problems with these scopes before, and like David, I fancy the problem lies elsewhere, and may not be as obvious as it may first seem, because of the interaction of the different circuits. Check R4570, its close to the edge of the main board, its in the B sweep start circuit, this pot can go o/c and cause this problem.
Bill

[Radio Wrangler]

The CA3046 transistor array seems to be working OK right now, so don't touch it just yet. Swapping it introduces the prospect of adding errors and problems which could spoil the fault tracing process by adding confusion. Solder splashes happen and dud semiconductors straight out the box aren't unheard of. Fakes are another risk.

The part in use has been tested in-situ and all its transistors seem to be transisting nicely. I couldn't hope for a new part to do any better. Maybe when the scope is going it might be decided to make a preemptive strike and replace a load of known weakspots before they fail.
But we get into the business of judging relative risk. Is a new one likely to last longer than a proven one? If it's infant mortality in poor parts then it may be a net increase in risk. If bad design stresses things such that only ones with higher than average breakdown voltages survive, then the proven part may be better. If it's a case of bad design or parts exhibit a progressive degradation leading to failure, then a new part ought to set the life time clock back to zero... replacing working parts is not as clear cut as replacing dead ones
The stuff I design for a living links aircraft into anti-collision systems and air traffic control. Reliability is seen to be important so repair processes need careful consideration and all failures need diagnosis and understanding, just in case they're trying to tell us something.

David