Tek 2445A Repair, trace drift, CH1 out of spec, long startup

[fivefish]

Hi all, I finally bought a Rigol DS1054Z, thanks to Dave's and y'all reviews and recommendations.  Like it very much, some little irritations, but for $375, can't really complain. 

Now I feel it's time to open up my Tek 2445A and fix it up.

The issues on my 2445A are:

1. I noticed it seemed to take longer and longer to "power up" and for all the lights/self-test to go through it's sequence.

2. Upon powering up, I had to let it warm up a bit before trace gets stable and not drift upward or downward.  Even then, sometimes the trace would just drift slowly up from the scope screen.  Adjusting it back to center line is futile as sometimes it will drift in the opposite direction.

3. Ch1 is out of calibration (voltage readings). Ch2 and others seem fine.

I've taken lots of pictures so I'll know how to assemble it back, posting some of them here.

I reckon I should start with the power supply section and replace all electrolytic capacitors with new ones? From the photos, it seems all the caps are still original?

Any Tek technicians here who can point me where to concentrate/look first regarding these problems? Thanks!

[fivefish]

photo 1

[fivefish]

photo 2

[fivefish]

photo 3

[tautech]

I reckon I should start with the power supply section and replace all electrolytic capacitors with new ones? From the photos, it seems all the caps are still original?

Yes PSU must be "spot on" first.
The Tek pdf in the first post in this board is a must for download and study.

Not sure about 2445, there maybe DC voltages listed in the schematics and they can be very useful at times but the scope should be setup a certain way for them to be relevant. RTFM lots. 

[MarkL]

I reckon I should start with the power supply section and replace all electrolytic capacitors with new ones? From the photos, it seems all the caps are still original?

Yes PSU must be "spot on" first.
The Tek pdf in the first post in this board is a must for download and study.

Just to be clear, that would be "yes" to start with the power supply as tautech suggests.  But NOT "yes" to start replacing capacitors just because they're old.  Let them be for now.

Check the power supply voltages with a DMM, and check that the voltages meet the ripple specs with your Rigol.

On problem #2, do both traces drift?  And if both, by the same amount?

On #3, how far is it out of calibration?  Is it off on several ranges by the same percentage?

No ideas yet on #1.

[fivefish]

Thanks tautech and MarkL for the reply.

I got hold of a service manual and performed some PSU measurements.

All the measured voltages are spot on (using Fluke 179 meter), and within tolerance per Service Manual.
My reading ----- Service Manual
+10.00V  --- 9.99 to 10.01
+87.00V   --- 85.26 to 88.74
+42.42  --- 41.55 to 43.25
+15.02V  --- 14.775 to 15.225
Digital 4.994V --- 4.85 to 5.15
Analog 4.98V --- 4.925 to 5.075
-4.992V  --- -4.965 to -5.035
-8.01V  --- -7.88 to -8.12
-14.95V --- -14.775 to -15.225


BUT measured Vpp Ripple reading for +/-15V seems marginal, and Digital/Analog +5V are on the high side (see below)
PSU ---- Vpp Service Manual ---- My reading
+15V --- 15mV  ---- ave 15.6, max 18mV
Digital +5V --- 150mV --- ave 207mV, max 220mV
Analog +5V --- 15mV ---- ave 63mV
-15V --- 10mV  --- ave 12mV
-5V --- 15mV ---- ave 14mV

The Analog +5V is really way off. 4X times as much.


Now I'm thinking replacing electrolytic caps on these sections.
For +5VDigital, I'm thinking C1110 and C1111 (both 250uf)
For +5V unreg, the C1115 (250uf)
For +15V, Im thinking C1116 (180uf)
For -15V, the C1113  (180uf)
For -5V, the C1114 (250uf)
Might as well replace C1112 (4.7u)



What do you guys think?

PS: The service manual I'm using is 2445, but my scope is 2445A. It seems the parts placement diagram look slightly different.
Does anyone have an actual 2445A service manual?

[tautech]

As MarkL says wholesale cap replacement is not always necessary. I'd be checking them all first.
In this thread: https://www.eevblog.com/forum/repair/tektronix-468-repairrefurb/ member A2AT saw first hand how increased load on a PSU may bluff you into thinking the main caps might be stuffed, when just an "exceeding design" load can send ripple levels high.

Luckily the 468 manual described how to isolate parts of the scope from the PSU and help diagnose where the excessive PSU load might be. RTFM.
Try this one:
 http://www.schematicsunlimited.com/t/tektronix/tektronix-2465a-service-manual-may-87

So check caps, sure, but it may not be the magic bullet. 

[fivefish]

What do you mean by "check caps" ? Visually or ESR meter? -- I don't have an ESR meter, might be a good excuse to get one now.

Question: Can you check ESR of caps in-circuit on the board? Assuming they're not power supply bypass caps, can I do ESR measurement with the caps still soldered in the board?

[tautech]

Check caps in several ways:
Visual: swolen, leaking
Capacitance:  failing values and IMHO value drift of ~10% downwards means replacement.
ESR:SMPS applications are more demanding on caps, usually low ESR is a prerequesite, while there is great importance on this for design and correct operation (cap life and output ripple), I generally place lesser importance on this for preliminary checks.

[MarkL]

...
PS: The service manual I'm using is 2445, but my scope is 2445A. It seems the parts placement diagram look slightly different.
Does anyone have an actual 2445A service manual?

I found this one:

  http://bee.mif.pg.gda.pl/ciasteczkowypotwor/Tek/2445a.pdf

The schematics are a little crunchy.

If you want a top quality scan and OCR'd version of the 2445A user and service set (or almost any old manual), I'd suggest these guys:

  http://artekmanuals.com

[MarkL]

I got hold of a service manual and performed some PSU measurements.

...

What do you guys think?

I agree the power supply voltages look fine, and at this stage of the game I wouldn't worry about the ripple readings.  They are not outrageously high, and it also depends on how you measured them.  It's not the power supply.

Can you describe your problem #2 and #3 in a little more detail?


Also, it seems the service manual link I just posted is missing the example waveforms.  There's also the same version here but with better download speed:

  http://w140.com/mmm/tek-2445a.pdf

But again, it's incomplete.  You should probably get the Artek version.

EDIT: I said above, "I wouldn't worry about the ripple readings." Meaning, I wouldn't worry that they are slightly out.  It's still important that you measured them.

[fivefish]

Here's channel 1 trace drifting on it's own, downards.  Nothing connected.
https://www.youtube.com/watch?v=qKYQy2SaXoc&feature=youtu.be

And here's channel 4 doing the same thing... going upwards. Nothing connected.
https://www.youtube.com/watch?v=A5xkPuXpIG4&feature=youtu.be

These videos were taken within minutes of each other.
Ch2 will do the same thing also.
I didn't get the chance to test Ch3, but I think it's doing the same thing also.

These "drifting" will stop/stabilize after several minutes of the scope running.

[fivefish]

Discovered some other things.

Ch1 - 50ohm overload protection not working,  scope shows 5v @ 1v/div but actual volts input is only 3.2 volts.

Ch2 - overload protection working ok, voltage reading is also good...5 div @ 5v input

The drift problem affects all ch 1 to 4.

[PA4TIM]

Do not buy an ESR meter but buy a decent LCR meter instead. One that can measure at least capacitance (Cs and Cp) and D at 1kHz. It will be better if it can measure at 100/120Hz for very big caps and 10kHz for small caps too. Modern ones often have a 100kHz range and have a Rac mode and that is ESR at 100kHz. If that is usefull ? It can be handy for quick checks on caps in situ. The datasheets of most caps gives ESR at 1kHz (most times in the form of D but 1/(2pifC) gives you the reactance and D times that value is the ESR) At 100kHz they give the impedance and because the reactance is near to nothing for bigger caps you are left with the ESR (but that can be a totally different value at the working frequency. ESR is not a constant and low ESR  is not even defined. http://www.pa4tim.nl/?s=ESR for more about ESR or the ESR site from EEVblog member Krypton.

Measure caps in situ only as a quick check, do not trust it 100%. (compare it to measuring resistors in situ, if a 10k measures 1M it will be bad but if it measures 1 ohm it still can be good, there is just something parallel to it )

In manuals the technical part is often divided in a few parts. most times including something like
- performance checks
- calibration
- explaining how the circuits work

The first one is the thing you have to start with. It will tell you the problems. sometimes people think there is a problem but it turns out to be a user fault. Following the manual is better because they tell you how to set up the scope and test equipment  for each test.
The calibration part often is doing the same tests but do not start adjusting things before you are sure the rest is OK.

Do not swap parts only because you assume or read somewhere they can be bad.  You can make the problems bigger or damage hard to replace parts or the pcb itself. Caps are well known for breaking but 95% of the time these are inside (cheap) consumer electronics. Most of my test equipment still has the original caps. Only Tek exception was a Tek 2710 SA and 2 caps in a 7704A but in its defense, those where the only two not original caps and from some vague brand.

Work systematic, follow the manual. Check the schematics if you are not sure what they want you to do. Read what instruments you need for each step. If you do not have them or something similar do not mess around hoping you can do it another way unless you are 100% sure that you realy know what you are doing.

Follow the steps in order because adjustments often influence other parts of the circuit. So adjusting A can influence the magnitude of B. If you first adjust B it will look like A is bad. And if you then adjust A and forget about B because you think you have done that right you can be in for a surprise because now A and B probably are way off. And some things influence a lot of circuits. For instance a reference  powerrail. Some rails are the "mother" and influence almost everything. If you adjust that just to make it perfect, you can make things overall worse.

So forget all problems until you are sure the PSU is 100% but always keep in mind a powerrail can look bad because of a fault elsewhere. That is why you first have to do the performance tests.

Repairing scopes like this is not easy. You often need lots of specialized test gear and a sound electronics knowledge.
But you can learn a lot from it.


 

[tautech]

Great post PA4TIM. 
Very good sound advice.
Yes we often assume members have a reasonable basic knowledge but that is not always the case.
Guidance is then required. 
At least fivefish has another scope to help with this repair.
Reading the many posts of repair can give one a lot of knowledge of others problems and how they have managed to overcome them. It is my favorite board. 
Who doesn't like fixing stuff.   

[PA4TIM]

Yep, I like repairing so much, I ended up doing it as a living: http://schneiderelectronicsrepair.nl/
Jim Williams had the same opinion about the educational part of repair. He used it as a training aid while he worked at MIT. He even did some sort of repair contests with students.

[siggi]

Discovered some other things.

Ch1 - 50ohm overload protection not working,  scope shows 5v @ 1v/div but actual volts input is only 3.2 volts.

Ch2 - overload protection working ok, voltage reading is also good...5 div @ 5v input

The drift problem affects all ch 1 to 4.

If your scope has the SMD version of the A5 board, there are some electrolytics on there that tend to leak. When this happens they tend to take out components related to the DAC reference, including the 10k precision, low-tempco resistor that sets the ref current and the calibration trimmer.
This will cause all kinds of trouble in the scope, as most things (like trace position) are controlled with voltages derived from the DAC. The DAC is also used to measure e.g. the 50Ohm overload thermistors, by running them through a comparator against the DAC output. So far, all the symptoms you've listed are compatible with A5 problems. If so, this is good news, as the A5 board is very repairable, and it's all obtainable components...
There are 10V, 1.36V and -1.25V test points on the A5 board that are good to check first, then it gets more involved .
My 2467 has the throughole version of the A5, and it suffered a shorted clamp zener on one of the reference voltages in the A5 board. This led to no traces, no on-screen display, and the 50Ohm overload sense always tripped.

The PSU ripple you've measured doesn't seem likely to be causing trouble, and I'd leave the PSU until you've sorted the other trouble.
Mind that when you measure the PSU ripple that the front panel is set as specified in the service manual, and that your 20MHz bandwidth limit is on. Otherwise your readings won't be comparable to the service manual specs.

Also you may want to join TekScopes on yahoo. It's a great community where some serious Tek nerds hang out.


Sent from my iPad using Tapatalk

[MarkL]

On the drift, I'm not convinced yet it's a problem on the A5 board, but it's a good place to start.

siggi has a great hint on the DAC.  I would first take a look at the DAC reference voltage on U2101 between pins 14 and 15 with a DMM.  See if the voltage drifts in relation to the traces moving around.

If that looks stable, take a look at the trace position outputs for the four channels at W511 pins 10, 9, 8 and 7 to see if they're drifting.

This is all on schematic <diamond> 2.


On the calibration problem with Ch1, I would start with the attenuator since Ch2 is working fine.

One way to tackle a bad channel is to feed the same signal into both channels simultaneously and then start comparing the waveforms on the same corresponding test point between the two channels.

A good place to start comparing would be the attenuator outputs at test point 6 (Ch1) and test point 11 (Ch2).  And you should also try different combinations of vertical gain, 50ohm termination on/off, DC/AC input coupling, and a high impedance input (like a x10 probe on each channel).  With all those combinations, you can usually get any problems with contacts or burned out sections of the attenuator to surface.

Schematic <diamond> 4.


Not forgetting about the start-up delay, but I would try to fix the above two first.  The only idea I have for the delay is to run the start-up tests manually and compare the times with someone else who has a 2445A.  Manual control of the tests is described on page 6-9.

[fivefish]

Thanks all.  This is my first attempt at scope repair, though I've bought and fixed defective amps and bench lab PSUs.

Upon powering up, I noticed both ch1 and ch4 give the same vertical offsets.  Even after I re-position them to center, turn off the scope, then back on, they always come back to these same positions. 

CH1 offset after power ON.


CH4 offset after power ON


Maybe the above 2 problems are related, maybe not. But I read the theory of operation and checked the schematics, and don't see anything common between ch1 and ch4. (The reference +10V is spot on at 10.000.)

I'll start checking test point waveforms and voltages per Service Manual.

[MarkL]

I would suggest measuring the reference voltage going into the DAC for drift; the DMM across pins 14 and 15.

If you change any settings, like the position, the scope might not store it immediately.  I know the 2465 delays a few seconds before saving any front panel changes.  Not sure if the 2445A does that or not.  Just a thought.

[siggi]

I would suggest measuring the reference voltage going into the DAC for drift; the DMM across pins 14 and 15.

The voltage across pins 14 and 15 should be as close to zero as makes no difference, and both of those pins ought to be at ground potential. The reason being that those two pins are the inv/non-inv inputs on an op-amp in negative feedback. (I also believed that resistor network to be deriving reference voltages until recently, but that's just not so).
What matters to the DAC is the reference *current* that flows through the 10K resistor from the 10V rail, plus a small amount of trim current that goes through the trimpot. The nominal DAC reference current is 1mA, hence 10V/10K.

The 1.36 and -1.25 voltages are derived thorugh the DAC, so they ought to reflect trouble in the DACs reference.

[MarkL]

Ah, you're right.  I should have looked at the AM6102 datasheet - thanks.

[fivefish]

I read the service manual and it says to enter Diagnostics, do the following:

Press and hold ?V and ?t simultaneously
Press Trigger Slope while holding down ?V and ?t
The display should now show:
DIAGNSTIC. PUSH A/B TRIG TO EXIT

On my scope, following these steps do nothing. I don't hear any relay clicking, and no change in the display.

DC calibration works though, as per Service Manual but I can't get it to go into Diagnostics mode using the above steps.

Am I missing something?

[MarkL]

That's strange.  Do the delta-V, delta-t, and SLOPE +/- switches work normally individually?

Another way to get into diag mode is to hold down any non-latching button while the scope is booting.  It will think it has a stuck switch and enter diag mode automatically.