Tektronix 2430A halts

[siggi]

One of those two caps went bang on my 2430.

You're in general troubleshooting of a complex, multi-processor device. Observe what works, and what not. It seems pot scanning works, given that you have intensity control.
This means your FPP is operational (check clock and key scanning).
Your FPP is communicating with the MPU (exercise and observe FPP/MPU interrupt and registers).
You have intensity control, which means the MPU DAC works (verify by observing DAC waveforms).

Post a video where you exercise every front panel control. When you turn the vertical sensitivity & coupling controls, do relays click?

A couple of more or less random suggestions to try:
Re-seat anything in a socket, including ROMs, the waveform processor and the acquisition hybrids (careful, they're ESD sensitive)
Set the service jumper that cycles the CPU through the address space, and see whether you find trouble in address decoding.
There is a procedure to isolate and exercise the waveform processor, try that.
Change the RAM chip. Those do go flaky sometimes, and can cause this sort of trouble.



Sent from my iPad using Tapatalk

[kavea]

Well... during the lunch time i've replaced all 5 caps (picture) by Vishay and Phillips, unfortunately it changed nothing. Then i've pocked all IC's, removed and reinstalled all ROM's = nothing. And finally i've decided to do what I've proposed earlier: desolder the U850 and install a DIP socket to force /IRQ to high (picture). After turning on, the scope didn't show any text and didn't respond to any button. Then i've repeated the procedure with fixing low level at /IRQ = same result. Then i reinstalled the U850 and turned it on ... and it booted up (picture) !! Then my lunch time have finished, so i will continue after work. There is some FAIL at Self Test, but at least it boots and i can navigate through the diagnostic menu    It isn't in the troubleshooting guide, but it worked. I don't really know what was the problem, solder joins at the U850 or putting high and/or low have changed something somewhere in the circuit.

[BradC]

Awesome job! That checksum fail probably indicates it has lost its cal constants. Start with the Autocal and go from there.
They're actually not that hard to cal with the right tools, and the leveled sine generator is only really necessary if you need to calibrate the CCD's. Most of what you need can be scrounged, made or borrowed excepting the fast rise pulse generator for triggering.

I have a 2440M (I disabled the MATE board to make the GPIB do what the standard one does) and I love it.

[siggi]

Then i reinstalled the U850 and turned it on ... and it booted up (picture) !!

Awesome, congrats!
I remember a recent thread either here or in TekScopes where the problem was a bad CPU. Maybe yours is marginal in some dimension - like e.g. an output driver or some such.

[kavea]

Thank you, but there are still some problems  I can't perform the Self Cal, it halts again. Inside Extended Calibration i managed to calibrate "Triggers" , but "Attenuators" FAILs, and "Repet" halts .. so far, the better i could get are 6100 and 8700 errors (picture). Otherwise from time to time i can exit the extended calibration menu and use the scope (picture) but magnitudes are wrong and frequency counter too
 

[siggi]

Here is the thread we're to MPU was bad https://r.tapatalk.com/shareLink?url=http%3A%2F%2Fwww%2Eeevblog%2Ecom%2Fforum%2Findex%2Ephp%3Ftopic%3D46043&share_tid=46043&share_fid=25658&share_type=t.

[kavea]

Here is the thread we're to MPU was bad https://r.tapatalk.com/shareLink?url=http%3A%2F%2Fwww%2Eeevblog%2Ecom%2Fforum%2Findex%2Ephp%3Ftopic%3D46043&share_tid=46043&share_fid=25658&share_type=t.

Thank you , i have read it quickly.. and i think in my MPU is ok since everything working just fine except three points :
1/ Self calibration halts
2/ "Repet" calibration halts
3/ "Attenuator" calibration FAILs
If one of the A0-A15 lines didn't worked it would be the same disaster like in mentioned thread.
Today i've managed to get rid of NVRAM error (image), now i have only 8700 (attenuators) which should be related to 3/. I also measured attenuators TP, which are ok (image). There is also relays clicks etc. So i think the analog part is working. There is something in common between Repet and Self calibration which makes the scope halting. For the FAIL of the Attenuator calibration, i'm not pretty sure about my reference voltages which are 0.2V, 2V and 20V, if they are not precise enough this fail is normal (p. 244). The problem is that there is no tolerance to these voltages in the SM, and with my voltage supply and the DMM i have an error up to +/- 10mV, especially at 20V. But anyway, this "extended" calibration of attenuators should be done after the Self calibration. 

[siggi]

When you say "halts", what do you mean? Does the test or process fail, or is the scope becoming non-responsive?


Sent from my iPad using Tapatalk

[kavea]

When you say "halts", what do you mean? Does the test or process fail, or is the scope becoming non-responsive?

The process halts, when the cal process is launched (any cal process), there is "RUNNING" which appears and slightly blinking during the test. In the case of Self cal, there is a blinking for 1 sec, then it stop blinking and do nothing. Normally the cal time is about 10 sec (c.f. SM) , i have been waiting several times for 5 min or so and none of buttons are responsive except the display intensity knob. I know that the interruption system is ok, at each click on any button the "RUNNING" writing is blinking one time. I think the uC is waiting for something more important than front panel control. In the case of Repet cal, however , there is no any blinking, the process halts immediately. So i have to turn OFF/ON scope each time i do this.

[siggi]

The jitter correction ramps *might* be the problem. They're a time-stretching dual slope current switching machinery. On charge they measure the time from trigger to sample clock edge.
The discharge time is measured with a zero crossing comparator and a counter. If the discharge current sink or the comparator are bad, maybe the REPET calibration would get stuck.

Sent from my Nexus 5X using Tapatalk

[kavea]

The jitter correction ramps *might* be the problem. They're a time-stretching dual slope current switching machinery. On charge they measure the time from trigger to sample clock edge.
The discharge time is measured with a zero crossing comparator and a counter. If the discharge current sink or the comparator are bad, maybe the REPET calibration would get stuck.

Ok, i was measured some voltages on ramp current source (FO20, p.327) , and indeed there is a problem, SLAMP1 and /SLAMP1 signals should be at -1.6V , but i have -0.8V on one of them, so Q491 and Q492 are not driven in the same manner.

[siggi]

Ok, i was measured some voltages on ramp current source (FO20, p.327) , and indeed there is a problem, SLAMP1 and /SLAMP1 signals should be at -1.6V , but i have -0.8V on one of them, so Q491 and Q492 are not driven in the same manner.

SLAMP1 & /SLAMP1 are complementary ECL-level signals. One should be at ~-0.8V, while the other is at ~-1.6V, and they should flip while the slow ramp is draining C491. It sounds like that's what you're observing, no?
Are you seeing fast-up/slow-down ramps across C491 or at R495 as specified by oscillogram 84? I believe you have to be acquiring a triggered signal to get that same oscillogram.

If you're getting a jittery fast-up/slow-down ramp, then probably there's nothing wrong with your jitter correction. Note that this was just a guess, but as you've had unexplained lockups with this scope already, it's perhaps more likely that the original cause persists.

Did you re-seat all socketed components?

[kavea]

SLAMP1 & /SLAMP1 are complementary ECL-level signals. One should be at ~-0.8V, while the other is at ~-1.6V, and they should flip while the slow ramp is draining C491. It sounds like that's what you're observing, no?
Are you seeing fast-up/slow-down ramps across C491 or at R495 as specified by oscillogram 84? I believe you have to be acquiring a triggered signal to get that same oscillogram.

Where have you found the -0.8 and -1.6V ? Because on the schematic p327. both are -1.6V. But indeed i'm measuring -1.6 and -0.8 V. Yes, i have observed the ramp on 84, it was not triggered but i clearly see a ramp signal.

If you're getting a jittery fast-up/slow-down ramp, then probably there's nothing wrong with your jitter correction. Note that this was just a guess, but as you've had unexplained lockups with this scope already, it's perhaps more likely that the original cause persists.

On the page 84 (Ramp switching) is said that during the calibration, the JIT1 GAIN is applied to the base of Q495, but i can't see any signal, only a DC value of -5.2 V during the normal operation or Repet Cal or Self Cal. And the MUX U651 (FO12 p.311) driving this BJT through U661A doesn't provide any kind of signal on all of its outputs (except one, with -1 V DC which is strange), whereas all input signals seems to be ok.

Did you re-seat all socketed components?

Yes, twice, unfortunately nothing changes

Otherwise while the scope is stuck the output of the comparator U490 is 0 V, compared to 3.56V when it works

[siggi]

Where have you found the -0.8 and -1.6V ? Because on the schematic p327. both are -1.6V.

This is "standard" ECL and the signals clearly need to be complementary in that circuit, as this is a current switch.

On the page 84 (Ramp switching) is said that during the calibration, the JIT1 GAIN is applied to the base of Q495, but i can't see any signal, only a DC value of -5.2 V during the normal operation or Repet Cal or Self Cal. And the MUX U651 (FO12 p.311) driving this BJT through U661A doesn't provide any kind of signal on all of its outputs (except one, with -1 V DC which is strange), whereas all input signals seems to be ok.

The signal is a DC signal, though it may wobble a little during self cal.
Basically the DC voltage on the transistor decides how much current the slow "drain" current sink sinks. This in turn decides the time expansion ratio, and ISTR the service manual wants a ratio of 1250 between the fast and slow ramps.

Otherwise while the scope is stuck the output of the comparator U490 is 0 V, compared to 3.56V when it works

How are you observing this - are you using an oscilloscope or e.g. a DMM?

This sounds reasonably salient, though only if the ramp is also running, and this is a difference only in the output and not in the inputs. If the inputs to the jitter ramp are static while the scope is stuck, you need to look further "upstream".

Interesting questions to ask and answer relate to how much of the digital circuitry is operational when the scope is stuck. We want to try and isolate the problem to a block in the block diagram or to a device in a schematic.

Since you have intensity control, it seems the FPP works when the scope is stuck.
I believe the intensity signal comes through the DAC and a mux from the MCU. This implies the regular interrupt works, the DAC works and the multiplexer in question works.
Can you check those and report back?

Can you post a video where you exercise *every* front panel control while the scope is stuck?
In particular exercise the vertical channel controls and listen for relay clicks. Exercise the coupling controls and listen for same.
Try shorting the BNC input sense ring to ground, and/or plugging in a 10X probe, see whether anything changes in the readout, and so on.

[kavea]

This is "standard" ECL and the signals clearly need to be complementary in that circuit, as this is a current switch.

Ok, i didn't know

How are you observing this - are you using an oscilloscope or e.g. a DMM?

A DMM, but i will check it with a scope

Interesting questions to ask and answer relate to how much of the digital circuitry is operational when the scope is stuck. We want to try and isolate the problem to a block in the block diagram or to a device in a schematic.

Since you have intensity control, it seems the FPP works when the scope is stuck.
I believe the intensity signal comes through the DAC and a mux from the MCU. This implies the regular interrupt works, the DAC works and the multiplexer in question works.
Can you check those and report back?

Yes, it would be great to isolate the problem, i will check these.
Today i however have desoldered the MUX (U651) (which produces JIT-GAIN, GAIN-CAL and CH-BAL signals) since i had one, and without this IC the scope behave exactly in the same manner (!), no additional errors, i have signals on the screen and it halts when i'm trying to self calibrate, all this i found strange and was thinking that it's done, faulty U651... but unfortunately after replacing it by a new one,  it changed nothing. I didn't look for the moment if there is something at its output.

Can you post a video where you exercise *every* front panel control while the scope is stuck?
In particular exercise the vertical channel controls and listen for relay clicks. Exercise the coupling controls and listen for same.
Try shorting the BNC input sense ring to ground, and/or plugging in a 10X probe, see whether anything changes in the readout, and so on.

In fact when it halts nothing except intensity knob works. No clicking from attenuators or other sounds. There is just RUNNING on the screen. But i will post a video, may be tomorrow.

[offtrack2]

all it needs is a new NV ram battery. (just did this hours ago for same reason,)
the book is clear, it says and I simplify,  if CKsum fails, ignore all other errors !!!  (that is because all amp offsets are lost and all self tests will fail)
all 2430s do this and is a simple fix,
replace the battery Keeper-II (early models, like mine is)
or if no keeper bat there.  (last models are)2430a , replace the Dallas Nvrams DS-nnnnn SRAM,  they have a battery inside. I'm told some version have 2 Dallas NVram there, (look carefully for dallas)
then remove the J156 jumper
do not run with case off, or use external fan to a10 board bottom or it will BURN UP,
let the scope warm up for 10 minutes, the screen tells you that its not warm , wait for it ...  then.....
do the full diag   (chapter 5-10 to 12 covers this the book is  free from tek, need links?)
then do the calibrate.
last do the external CAL.  using only an accurate PSU, variable power-supply, and good DMM 4-1/2 digits,  (our meter today exceeds the old PG506 level gen by tekl. by 10 fold)
I used a 10k ohm and 200ohm , low PPM resister pair  (divider) to my 20vdc bench supply. (adjustable from 0 to 20v)
then I did the external cal.
levels.  do the 3 dc levels as prompted. see it pass. ch1 and ch1 connected to source
triggers, do the 3 dc levels  ext, trig 1 and both same time.
repete   (needs no inputs) its a all automatic.
then when all 3 say passed.
turn off scope see chksum pass. on next power up.
the remove J156.
done.

that halting is normal,  it goes to cold start when CKsum fails for good reasons, not to do damage. with cal tables lost.

on mine +5vd bus rail shored and I lost all cals.
ribbon cable was cut by case, pin 1, +5vd,  cure that and the blow fuse, and it cals prefectly.  32 years old running strong.
as with any electronics
we check all supplies first for spec voltages and out of spec ripple. first.
end 2340s , that all fail this way. all do and is not big deal at all.