Tektronix 454A

[Andy - G8MNM]

Hi,

Some years ago I bought the above scope in non working condition.

Overheating transistors on the vertical output board and obviously no proper display. There was evidence of some 'fiddling' as there was a disconnected wire to the CRT.
Anyway I found that the PSU output was too high (seems to be a common fault) due to transistor failure. So I fixed that and also replaced the S/C vertical drive transistors and reconnected the floating wire. It is now working on one channel only and the vertical calibration is out, and beyond setting with the front panel pre-set. Does anyone have any suggestions as to how to proceed from here? I do have the service/repair manual.

[RogerEvans]

Andy,

Step one is always to check all the low voltage DC supplies for voltage and ripple.

Do you have a sine wave or linear ramp (triangle wave) generator? If you do it makes tracking down the Y deflection error a lot easier. You need to check that the gain does not change with vertical position which is easier with the triangle or sine waves since they will appear to distort. You also need to check if the gain error is the same for all positions of the V/div switch and if the error is the same for both channels, so it may be asier to get the second channel working first.

Can you describe the problem with the second channel in more detail? If you select Alt or Chop do you see one horizontal line or two? Do they both respond to the vertical position controls? If you only see one line try pressing the Beam Finder. If you select Add does the display respond to both Ch1 and Ch2 position controls.

If you set Trigger to Ch 1 only and select X-Y on the Horiz Mode switch can you move the spot in X and Y using the Ch1 and Ch2 position controls?

I don't have a 454A but I have fixed two 454s over many years.

Regards,

Roger

[Andy - G8MNM]

Roger,

Thanks for the help. I have just got the scope on the bench.
Power rails are fine (that was one of the original faults).

-12V rail = 11.94V

12V rail = 12.00V

75V rail = 74.8V

150V rail (unregulated) = 154V

There is no appreciable ripple on any of the power rails.

On chop there are two traces only one (ch1) can be controlled and registers an input signal.

I guess it will come down to semiconductor testing on ch2 vertical deflection as being DC coupled voltage measurements won't really be of much use.

[RogerEvans]

Andy,

The situation you have is fairly straightforward to troubleshoot. Attach a probe to Ch1 with the scope set to display Ch1 only. Connect Ch2 to the calibrator output, and you can then trace the calibrator signal through the Ch2 attenuator, phase splitter, and on as far as the channel switching diodes.

Let me know how you get on.

Roger

[Leon Crampin]

Having restored a 453, (very repairable and an excellent build standard), I was able to correct a good number of faults in the X and Y preamp stages, together with some triggering faults fairly easily. By far the greatest source of trouble was caused by the small signal transistors fabricated on a white plastic base, and sealed on top by a blob of dark coloured epoxy. To save naming and shaming, see the Tek parts list for the supplier.

This design is fatally flawed with respect to thermal cycling damage, which results in the devices becoming open circuit or very noisy. To get the scope working, I used BC107 or BCY70 devices as applicable - the Tek transistor sockets being a godsend as long as the contacts are OK. Replacement transistors should have gold plated leads for this job - as mine had, being "released" from a (cough) automotive job.

To conclude, when I tested the working scope to specification, the bandwidth and triggering were all OK with my substitute devices.

Leon.

[Andy - G8MNM]

Roger,

I ended up using another scope for checking as CH1 whilst working gives an overshoot on the display (yet another fault!)
Any way another faulty transistor found 151-0271-00 which seems to be similar to a BFQ241 or MPSH69 - from what I can find n the web. I swapped the tranny with the one from CH1 and got a signal path so obviously the transistor is faulty. It is quite difficult get onto the transistor pinouts as they are in sockets. One transistor I cannot find anywhere and not all the components are numbered either!
I will attempt to get a replacement part and then continue. I think it was the faulty power supply that has blown the transistos in he first place.

Thanks.

[Craig Sawyers]

2N4261 according to the Tek common design parts book.

Spec is below. Doesn't look to be anything to special, apart from low capacitance values. No mention of fT.

http://njsemi.com/datasheets/2N4261.pdf

Craig

PS I'd bung in a generic PNP part just to get it working while trying to find something closer.

[Andy - G8MNM]

Craig,

Thanks for the info. I will sort out a replacement or as you suggest something to get it going. Sometimes I feel like giving up as at times it is so difficult finding the actual components on the pcbs. And then being able to check voltages etc. Whilst the Scope appears to have been well used in the past the CRT is bright and sharp so it would be a shame to let it go. As I said before there are some other issues like when displaying the square wave calibration signal there is an overshoot at the top of each wave which blend it into the next one. Well the journey continues...

[Andy - G8MNM]

Thanks Leon.

The struggle shall continue!

[Craig Sawyers]

Found the spec in the Tek Common Parts data.

2N4261
fT >2GHz@10V,10mA
hFE(min) 30@1V,10mA
Vceo 12V
Vcbo 15V
Vebo 4V
Icbo 100nA
Ccb 2.5pF
Vce(sat) 350mV@10mA
Ic(max)80mA

It is very difficult to find a pnp part with that sort of spec. NXP (Philips Semiconductors as was) had an excellent line of rf transistors (surface mount of course) including ones that would have fit the bill. But they obsoleted the lot in one stroke in 2019. And all their jfets too.

Bucket loads of npn RF/UHF transistors - but absolutely nothing current that comes close in pnp, with the maximum fT that I could find being 600MHz

The best you can do is find genuine 2N4216 on auction sites.

Craig

[RogerEvans]

Andy,

If you have an overshoot and a gain error on channel 1 then the most likely cuplrits are a dirty switch wafer on the attenuator or input coupling switches or a poor contact on one of the socketed transistors. You can give the transistors a wiggle while watching the display and cycling the switches a large number of times.

Roger

[Andy - G8MNM]

Craig,

Thanks for the spec and part number. I have just ordered 2 online. Will be with me hopefully middle of next month - a friend will be bring them over for me to France with some other stuff.

Roger, On further inspection I think it is rather more than that. Swapping the good transistor to channel 2 restores it to working condition but exhibits the same fault. The beam won't focus to a nice round spot instead it can only be focused to an oval bright spot with a line (which is less bright) extending from it. I have posted a couple of pics. I am wondering if there is an issue with the EHT smoothing - I had that on my Telequipment scope. In fact the smoothing capacitor had never been properly soldered from when it was new!

[RogerEvans]

Andy,

Can you post a picture of the scope displaying the calibrator square wave? It is difficult to interpret those two pictures without knowing all the relevant conditions.

Roger

[Andy - G8MNM]

Roger,

These are the best photos I can get with the camera. I hope you can see what I mean with what looks like overshoot. I did wonder if the blanking wasn't working during the flyback period and the tube wasn't being cut off. The blanking pulses are present and the tube grid bias is set right.

[loulou31]

Hi
It is very difficult to find a pnp part with that sort of spec for 2N4261

Some BFQ32 /75 76 could meet the sepecification in a different package.
Possible also to buy on Ebay 2N4261 but with a risk of fake component...

Regards
Jean-Louis

[Andy - G8MNM]

Bonjour Jean-Louis

I have bought a couple of 2N4621 on Ebay from the UK. They appear to be genuine NOS.There are plenty of Russian variants for sale at cheaper prices but not worth risking.

[PJL]

It seems to trigger OK and the 'overshoot' looks like lots of noise on the Y plates. Would be worth seeing what happens on different timebase settings? I have one of these and the spot is small/sharp and bright enough for single shot on fast timebase settings.

[Andy - G8MNM]

With CH1 connected to the 1V calibration source there is a clean square wave on the Y plates.

There is a Sawtooth with 'Rubbish'- particularly on the Left plate and to some extent on the Right plate. Where it is coming from I don't know yet? It seems to be being generated on the Horizontal Amplifier board. All the transistors check out on Fluke meter, but may have other defects that don't show up on static check of this type. Further investigation is required. Fault finding on 'semi' fancy scopes is not really my area. Fixing old Telequipment ones was a lot easier, TV's even better!

[cmjones01]

Originally Posted by Andy - G8MNM

These are the best photos I can get with the camera. I hope you can see what I mean with what looks like overshoot. I did wonder if the blanking wasn't working during the flyback period and the tube wasn't being cut off. The blanking pulses are present and the tube grid bias is set right.

That doesn't look like what I'd understand as "overshoot". Overshoot would normally be in the vertical axis, and occur after fast edges as in the attached image. It's typically seen on an oscilloscope when the probe compensation isn't adjusted correctly, or when there's a fault in the vertical amplifier.

What I see in the images you post looks more like either a blanking problem (we're seeing part of the waveform while the spot retraces to the left before starting a new sweep) or a triggering issue, where triggering is occasionally happening at the wrong point and causing a "ghost" waveform to be seen.

Thinking about is, there is another possibility: on some scopes with dual timebases, it's possible to have the main and delayed sweeps visible at the same time, and a "trace separation" control adjusts how far apart they are on the screen. The 454A has a "mixed sweep" mode which I think shows the first part of the trace at timebase A speed then the second part at timebase B speed. What's the "horiz display" switch set to? Does waggling it make a difference? Actually, does adjusting the "delay time multiplier" control affect the display?

Chris

[Andy - G8MNM]

I noticed that for a very short time after switch on what looked like flyback on the trace wasn't there. It quickly became apparent after a very short time.

I noticed there were various voltage anomalies around Q1048 which were temporarily resolved by removing Q1053. I addition having scoped the horizontal outputs to the plates there also appeared to be some sort of 'parasitic' oscillation occurring. After chasing my tail for a while and swapping over transistors from one left and right plate drives I was able narrow it down to Q1080 151-0271-00 otherwise a 2N3051. Unless I have totally confused my self with all the time I have spent trying to sort this out I do think it is the transistor that is faulty. It seems to cause the same type of oscillation in whichever of the two positions it is in. I have just ordered 2 new ones from the USA (cheapest option) and will await the arrival.

When they have arrived I will update the thread as to the state of play.

Andy