Tektronix 2246 flickering screen, ch2 not working. (Ch2 fixed) New help needed!

[grusus]

Sweep speed setting?

0,5ms

Chop/Alt mode setting?

I am changing it during video. I am intentionally showing not only crt, but you can see me changing alt/chop mode. It is in mode row where you select channels - last button in row. Default is alt mode. When button is lit it is in chop mode.

Signal frequency?

approx 1khz sine

Having a "before" and "after" video for comparison, with the exact same settings, signal, etc., would be of great help. It is becoming very time consuming to have to go back through this thread and try to locate another video for reference, that has different conditions, and try to interpret new videos.

Noted. I will in future. I am still having extreme painkillers and, my thinking is very slow and I am feeling like drunk most of the time. Sorry for that.

To locate R207 and R807, measure resistance from W2302 pin 2 to U202 pin 4 - 714 Ohms is expected. With one probe on W2302 pin 2, see if it is possible to find where the resistance drops to 357 Ohms.

W2302 pin 1 to U802 pin 1 should be somewhere between 75 Ohms and 575 Ohms, depending on pot R823 setting.

So indeed R207 is where it should be, but instead of R807 i have a jumper wire with white plastic on it like those "W" elements on board. That is what misled me.
Here is vid with R207 and "R807" removed (same settings as above):
https://photos.app.goo.gl/rWxmq21DiYAjzUvY7

Again I am changing between alt/chop. There is slight difference as it should be i suppose.

[rf+tech]

Here's a reference point: I've just replicated your test setup with a Tek 2465. In Chop mode, I see the same pattern of dots and they are more noticeable at lower intensity. In Alt mode, the dots disappear.

So it is confirmed that only Alt mode should be used for the 1 kHz test, from now on.

Next test. Can the test signal frequency be adjusted by a few Hz. to slow the marching dots to a stands-still and reverse direction?

I would recommend leaving the RO signals disconnected for now. There is a secondary effect where several dots are not as bright as others. This could be an artifact of the video camera. If this is not a video artifact, then I would like to pursue this.

No need to apologize for the medication side-effects.

[grusus]

Next test. Can the test signal frequency be adjusted by a few Hz. to slow the marching dots to a stands-still and reverse direction?

Yes. I am adjusting on precision 0,1Hz scale up and down. Freeze of dots and going different directions is responding to freq adj. Behaviour of marching dots is predictable and recurrent. Alt mode.
https://photos.app.goo.gl/hsNc2yL7Jvb1wJ1K8

There is a secondary effect where several dots are not as bright as others. This could be an artifact of the video camera. If this is not a video artifact, then I would like to pursue this.

In personal it is as in video.

[rf+tech]

This screen capture below from the last video shows 8.5 repetitions/cm



This also correlates with the number of steps in waveform 9C

Since both RO vertical and RO horizontal signals are disconnected, the anomaly must be Z-axis modulation. Set the beam intensity to clearly see the dashes in the 1 kHz sine wave. With the M07192 sweep set to 0.5 ms/cm, check waveform 7B, 7C and 7D. Is there any change when the RO intensity control rotated?

In an earlier post, you mentioned DS2702 was glowing. DS2701 is connected in series with DS2702, so if one glows, so must the other.

[grusus]

I have noticed something. There is ripple on 7B,C,D that corresponds to marching dots on 2246. It goes left, stand still or right according to dots on 2246 crt.

7B: https://photos.app.goo.gl/U7mJjkaLtiS6Uc9P6
and "zoomed": https://photos.app.goo.gl/kKCkCCTgu1mhG5FZ9

with readout (I am swiching from 0,5ms to 2us): https://photos.app.goo.gl/UVcsfYmPqSfqu4ap8

7C: https://photos.app.goo.gl/73QAbGGmrCtrqmz39

with readout(I am swiching from 0,5ms to 10us): https://photos.app.goo.gl/33eHs2Gh3EGC8JpC8

7D: https://photos.app.goo.gl/hZjLkFKhXHh9Jn386

with readout (I am swiching from 0,5ms to 20us): https://photos.app.goo.gl/bA1uwonSVkPfTazi7

+/- 1khz sine input signal. I was adjusting by 0,1Hz as before. Alt mode. Only Ch1 active. R207 and R 807 disconnected. Also R250E and R251e.

In an earlier post, you mentioned DS2702 was glowing. DS2701 is connected in series with DS2702, so if one glows, so must the other.

No. It was DS2704 and only. Gets brighter when focus is turned more CW:

Old picture.
https://photos.app.goo.gl/RPQQkWBhBpVKKWdGA

Now it doesn't.

I don't know if it is normal but when i rapidly turn focus from min to max, both DS2703 and DS2704 lights up and then goes off. They don't always behave like that. Sometimes they don't lights up.

https://photos.app.goo.gl/hibvTLhka9EuZfVWA

And with my luck my signal generator broke down :-). Lm741 is temperature sensitive. When cold works normally - when getting hot -failing. Cold spray does the trick. I must order one.

[rf+tech]

Please provide a photo showing the scope probe and ground clip connection, to the 2246.

[grusus]

7B: https://photos.app.goo.gl/qKAXj1e9eYSbDp6XA

7C: https://photos.app.goo.gl/v3T7VSd181wEG9xA7

7D: https://photos.app.goo.gl/KmwUht7EAeKV4aQt5

Remember that in SM they present screenshots of traces 7B,C,D with 1ms/div. My starting point was 0,5ms/div as you said in one of previous post.

[rf+tech]

One more photo, farther away, that shows the proximity to the HV power supply.

The "ripple" may be a probing artifact where the EM field from the HV switching supply gets picked up by the probe ground lead inductance.

See this page for some examples of minimal inductance probing:

https://www.eevblog.com/forum/beginners/diy-oscilloscope-probe-ground-spring/

[grusus]

Hmmm. I will check tomorrow if position of probe affects the traces on siemens scope.

But, HV power supply is on other end and other side of scope with metal screen between boards. HV cable is also not in that area.
In area where 7 C,D point are, there is -2,7kv in range of few centimeters.( max 2 inches)

[rf+tech]

I've gone back to an earlier video and stepped frame-by-frame while observing the random vertical lines and RO wiggle. It appears each missing dot in any given RO character is the result of the beam being deflected to some point further up the display. These vertical lines appear in rows that most likely correspond to the six right-side RO menu rows.

Further, each gap in the main sweep is also due to the beam being deflected to one of these six right-side RO menu rows. Since the horizontal and vertical RO signals are disconnected, the RO now appears as a single dot in the center of the screen. All that remains are the Z-axis "punch-outs".

Working backward (Sheet 9) from the two RO position signals, DACs U2412 and U2413 are responsible for creating an H and V analog current for every RO dot position. For a given character dot to end up on the wrong line, would imply incorrect data from Character Dot Position ROM U2408. Or, the vertical DAC U2412 itself may be missing bits on the 8-bit bus. If this were the case, I would expect the effect should be stationary. What is seen appears to be that of a free-running clock that is not synchronized.

Top left corner of Sheet 9 is where the Clock Divider/Counter is located. U2409 receives an 8 MHz clock, from U2501 Measurement Processor on Sheet 8.

An end-to-end sanity test would be to trigger on this 8 MHz clock at U2409 pin 1 (waveform 9E) and probe U2412 pin 12. This pin is the LSB and should appear synchronized with the 8 MHz clock trigger signal. While you are at it, may as well confirm the other 7 data bits to U2412 are in sync. Also, check Vref+ and Vref- on U2412, pins 14 and 15 respectively. These should both be clean and steady DC. Photos or video are not necessary, unless something unexpected is observed.

[grusus]

The "ripple" may be a probing artifact where the EM field from the HV switching supply gets picked up by the probe ground lead inductance.

You're right. It was from HV. Moving probe around is makes ripple almost dissappear.


Hmm. Clock is there but shouldn't it be a square wave?
0.2 us/div
Upper trace is clock, lower pin 12 of U2412
https://photos.app.goo.gl/4QKXrbADmqK7prbR7
and first on second:
https://photos.app.goo.gl/YCW3gyDepwZfHgAh7

as for other 7 data bits something is there with a little offset to clock as on pin 12. But they are with different amplitude and traces differ from each other.

U2412 pin 14 and 15 ripple is 100mv. On dc coupling on scope shouldn't one be positive and one negative in reference to ground? Mine both are positive +4V.

I have noticed sth today:
Nothing connected to scope coupling to gnd (no matter if AC gnd DC). Alt mode. I managed to stop trace by adjusting holdoff. Look:
1ch: https://photos.app.goo.gl/mzWKbEtigmFD7GNH6
4 ch active: https://photos.app.goo.gl/UdeeRcN33amEd3bn6

Lenght of gaps varies. It is dependent to sec/div setting.
https://photos.app.goo.gl/FmLdeRGztrMmzUY88

Maybe this is normal for this scope, but I have never seen behaviour like this on any analog scope.

[rf+tech]

The clock waveform squareness is in part, a function of the M07192 step attenuator over-compensation and scope probe compensation. A 20 MHz scope would have a 17.5 ns risetime, when properly calibrated. An 8 MHz square wave half-period is 62.5 ns. 17.5 ns is 28 percent of 62.5 ns. Adding over-compensation into the picture, what is observed is to be expected.

Cross-checking the probe/scope combination with an 8 MHz square wave from a function generator (properly terminated), would be wise.

The signal is suffieent to swing above and below logic thresholds, as evidenced by U2412 pin 2. Again, the limited bandwidth of the M07192 cannot faithfully reproduce these fast and narrow pulses. The signal is in sync with the clock and that is good.

The other data lines will only be active when an RO line is enabled across the top of the display, where a higher dot deflection voltage is required.

U2412 pin 14 and 15 ripple is 100mv. On dc coupling on scope shouldn't one be positive and one negative in reference to ground? Mine both are positive +4V.

An accurate assessment of ripple requires use of a ground spring/spike, as previously mentioned. This is even more important with fast rise-time digital signals. To understand why Vref+ and Vref- are both positive, please refer to the DAC-08 datasheet:
https://www.analog.com/media/en/technical-documentation/data-sheets/DAC08.pdf
R2415 establishes a reference bias current. One end is connected to +2.5Vp. Is this also 4 Volts?

This newly discovered behavior of the hold-off control is another good clue. It is definitely not normal. I shall look into this later in the day.

[grusus]

I have checked clock on scope itself and noticed that on higher sweep retrace is visible and part of trace is missing/dimmed:

https://photos.app.goo.gl/CUBXKJhQYkeZAVUv5
https://photos.app.goo.gl/xPsAWDXgcVEqbzpF8
https://photos.app.goo.gl/GHMqWpm6cuyGWvcn8

I cannot find pot for retrace calibration in service manual. I am just thinking if those blank spots in traces and missing trace on higher sweep + visible retrace are connected, and this only one issue that causes all things?

It looks to me that something is wrog with sweep on/off timing. Sweep is shut down after retrace, not before. This will cause a missing part of trace, and visible retrace. That just makes sense to me - maybe I am wrong.

Also I have noticed that RO flicker is differrent when sweep ends and retrace starts, till speep starts again. It can be seen on video.
https://photos.app.goo.gl/16HQSgQCLsWYF7PK8


I didn't mention that before, but if holdoff is more cw - trace is getting dimmed slightly. when more CCW trace is brighter. I don't know if it mattters and if it's normal or not.

R2415 establishes a reference bias current. One end is connected to +2.5Vp. Is this also 4 Volts?

hmm. I cannot find that resistor. My board is different. In red rectangle is U2412. Pins 14 and 15 are connected directly to chip in yellow rectangle,  pins 22&23. All that i have written on picture is what i think it is.

[grusus]

Chip in yellow rectangle is U2408.

Whoa, I made terrible mistake in pin counting... 

That screenshot i posted from pin 12 was ... pin 13 it is +v and it is around 5V. Pin 3 is -V and it is -7,5V

Pin 12 with clock:
https://photos.app.goo.gl/U35fTyrV7PAsC4fA6
https://photos.app.goo.gl/rBMKxpjbaxRmpFDC7

But, I have nothing on pin 14 and 15. 0V. Only some ac ripple which corresponds to 8mhz clk.
Found R2413 and R2415. One side is 0V both, other is 2,5V for both. Both measured 1,24k Ohms. What the heck? 2,4kohm to gnd from pin 14 for U2412 and same for U2413.

[grusus]

Update on ch2 not working.
I did some digging around AT127, U122  and U220.
I have disconnected out from At127 (pin 7) and voila - i have good signal on out of AT127.
Next i have connected it back, and disconnected out of U122(pin 8 ) and gues what - good signal present there.
So, or U220 is bad, or sth that is connected directly to it. Output is dissconnected (R250E and R251E)
Any hint? It looks like sth is pulling signal down.

[grusus]

A little update on ch2.
I was wrong.
Signal was degenerated by U220. (varied of v/div position - that somehow mislead me)
Failed TLC271 op-amp. (see picture)
Now it seems all channels will be working fine. 
I swapped one from other channel - ch2 is working perfectly.
Ordered one and waiting for package to arrive.

Still flickering issue to fix.

[grusus]

Update
After swapping tlc271 all channels are working fine. So this issue can be called fixed.

[grusus]

I just found more proper service manual for my scope. There are some minor differences between scope and schematics (no RO jitter adj) - other else looks exactly the same. Processor board matches. Also there are some differences in schematic 7 compared to previous SM (z axis, ctrl mux) that matches my scope.

http://bee.mif.pg.gda.pl/ciasteczkowypotwor/Tek/2246.pdf

From now, I suggest to reffer to this new SM.

Anyone have any hints and clues where to search for fault? I went through schematics many times, tested and I think i went into some kind of loop. Need some fresh look on this problem. If someone have knowledge, and will be kind enough to look into schematics in posted SM and give me some hint i will be greatfull. I think problem might be in digital part of scope - As those "blankings and jitter looks digital to me, not analog - but maybe I am wrong.
As RF+TECH and Breakingohmslaw are not present on forum from a while (I hope they are fine) I am stuck with my thoughts.

[grusus]

After taking some break in troubleshooting this scope, I got back to it.

After analysing service manual I got to the point that sth is messed up with clock/timing.

I think this is corelated with visible retrace.

Checked clock signal around U2409, U2508, U2507 - all seems fine. I am little concern about timer reset signal (+1VDC, no ac - only 8Mhz ripple from main clock - I don't know how it should look like), and INTR signal - also only dc no ac.

Today I made small experiment and dissconnected RO BLANK (R2419). Readouts are much brighter, grid dot is fully present, readouts are less blurry and more stable.
On Video I am connecting and dissconneting R2419.

https://photos.app.goo.gl/341Lxitp8DYbE9R97

That is why I am almost certain that there is some timing/interrupt/clock issue in this scope. It blanks when it shouldn't,

Can anybody give a hint what to check next? How to address this problem?
 

[rf+tech]

Your observation of the Hold-off control affecting the “chop” in the scope trace is a very useful clue. May I suggest reading the “DISPLAY AND TRIGGER LOGIC AND PROCESSOR INTERFACE” theory of operation, with special attention to the Chop Clock and Readout Interface descriptions.

The important detail to note is how these two circuits work to mask vertical source switching transients from Chan 1-4, and various RO areas. The Hold-off Oscillator and Calibrator signals are used to derive the vertical blanking. If the blanking is out of step, these switching transients would be visible.

To address your concern about the Timer Reset and INTR signals that seem to be missing: very narrow pulses (10s of nsec) with very slow repetition rates (in the msec range) are a challenge to observe on an analog scope. This may explain your observations.

[grusus]

INTR and timer reset are present. Thanks for explanation.

After reading "DISPLAY AND TRIGGER LOGIC AND PROCESSOR INTERFACE", I decided to check again TC, LFC, OSCOUT pins of U600. TC and OSCout looks fine to me. But. LFC influences chop clock. LFC is 1,15kHz. front panel test point for 1KHz is also 1,15kHz. I am wondering if this could be the case. Manual says it should be around 1kHz - it's 15% off. My first suspect is C935. I tested 33pF in circuit. To remove this capacitor i will have to disassembly some part of scope (CRT, Pot board for intensity and focus) It is flush mounted on pcb and I cannot desolder it from component side. Hotair is also impossible - to many plastic elements in the way. If you think this potentially can be the case of problems with this scope I will do that.

[rf+tech]

1.15 kHz is more than a little suspicious. You are definitely onto something.

C935 is 22 nF, in the parts list as well as the schematic. 33pF suggests measurement error. Nearly 3 orders of magnitude lower C leads to nearly three orders of magnitude increase in calibrator frequency.

The product of 22 nF and 22 kOhms yields a period of 484 usec, low and high, for a total of 968 usec. The reciprocal of this is 1.033 kHz. Checking the later 2246 schematic, we see R935 has been increased from 22 kohms to 24 kOhms. The calculations yield a calibrator frequency of 946 Hz. Given ceramic capacitors' well-documented decrease in C with applied DC, 24 kOhms is likely to be closer to the 1 kHz target.

For the effort required to remove C935, it would be easier to solder another capacitor on top of C935. In theory, one could temporarily disable U930B by grounding pin 6, then inject a 1 kHz square wave of suitable amplitude into the CLK 1k point, to see if the chop anomaly disappears.

[grusus]

Yeah, I bet it is a measurement error. I have checked r935 - spot on 22,1k.
C935 is not ceramic cap. It is a film capacitor in cubic plastic box. I connected 22n over C935. Frequency went down to approx 560Hz but no visible change on display in alt mode, nor in chop. I was connecting, and dissconnecting it live to see display in realtime. Main problem is somewhere else I think.
I will extract this cap tomorrow from board and check capacitance. If out of specs - I will change it. Also will change R935 to 24K if i have in stock, or add 2k resistor in series.
I checked in circuit also: R934=1k, R938= 2,2k, R939=1,2k, but 4,1V is 4,09 so... I will test all calibrator circuit parts out of circuit tomorrow.

In theory, one could temporarily disable U930B by grounding pin 6, then inject a 1 kHz square wave of suitable amplitude into the CLK 1k point, to see if the chop anomaly disappears.

Also will check that, but considering above, there propably will be no change on display.

Hmm. I forgot to mention before but I was playing around with service menu. If I go to Proc board menu -> Readout menu -> Exercise RO interfece, Flicker is much more noticable. From theory of operation This forces communication between Measurement processor and RO processor. It's AD0-7, then D0-7, then R0-7. So maybe something that is connected to those lines is playing with us. Sorry I don't have enough knowledge and experience to troubleshoot this myself. Maybe I am totally wrong.

[rf+tech]

Try 4.7 nF, 5.6 nF, 6.8 nF in parallel, whatever gets closest to 1 kHz.
Also check the effect of Hold-off, with the extra capacitance.
The objective is to remove gross error and rule out the effect, before moving on.

[grusus]

I tested whole calibrator circuit elements out of circuit. C935 was 20,5n so replaced with 22n. Also added 2K resistor in series with R935(didn't have 24K), Other elements tested fine. Now I have 1.000kHz/1.001kHz Clock at calibrator circuit. Close enough I think :-)

I think now is slightly better - judge yourself.

On Video I am adjusting holdoff, changing alt/chop and in the end when readouts gets dimmer I am connecting and dissconnecting ROBLANK R2419.
Singal on input of the scope as always 1kHz sine :-)

https://photos.app.goo.gl/9trA8izceGuTGZLe8

I think main problem lies elsewhere.

It looks like something is messed with blanking (ro blanking as examle above). + I don't know if you read this from my previous post that retrace is visible, and part of trace is missing on higher sweeps - from what I understand from service manual this is also corelated with blanking out of sync.

Vid of RO exercise:
https://photos.app.goo.gl/DeeB3DFPTxZrWYL99

I was checking clocks around U2508B and U2507.
Pin 9 of U2508B is 62,5kHz. Schematic says it should be 16kHz but in theory of operation U2508B divides 500K clk by 8 = 62,5kHz. Little confused...
Others I think are spot on.