suddenly unable power up Tektronix 2235 (analog) oscilloscope

[mmFooD]

I have a tektronix 2235 that I have been using for a couple of months now. I forgot to turn it off for a couple of hours this morning with the probe connected to measure ripple/noise from a switch mode regulator. When finally I turned it off I didn't notice anything suspicious but after a while when I tried to start it up again it didn't work.

Normally when pressing the power button the CRT makes that same, kind of static noise that old TV's does, for a second or two. Now my scope makes this noice but repeatedly for only a tenth of a second or so and the power led is also flashing in the same manner. It goes on like this without showing anything on the display until I turn it off.

So I left it for a while before trying to open it up to see if I could find some obvious error inside but for some reason the last screw was a very small torx which I didn't have at the time so I will have to fetch one tomorrow. But I was wondering if anyone have some idea as to what could be broken? Or at least what to look for.

I'd be very thankful for any help on this issue!

[tautech]

Welcome to the forum.

Sounds like the PSU is in tick mode and some guidance can probably be found in this recent thred:
https://www.eevblog.com/forum/repair/tektronix-2232-scope-in-need-of-a-doctor/

[David Hess]

The aluminum electrolytic capacitors in the power supply have probably worn out and are in need of replacement.

[oldway]

Check all the diodes on secondary of ferrite transformer T948....probably one of them is faulty (short circuited).

[TerraHertz]

[oldway]

What do you mean with this off topic answer ?
Are you naming who answered as "bastards" ?
I did not like your answer at all....
I reported it to moderator as off topic and insulting and I will delete my present post if your post is deleted.

[EEVblog]

Are you naming who answered as "bastards" ?

No, he's not, it's an aussie joke.

[JLNY]

I was recently having a similar issue with a 2230. Checking the PSU rails revealed that the power was pulsing on and off. Re-capping the unit immediately and completely fixed the problem for me. I also recapped a Tek 2445 recently and it booted up just fine again after that. Almost all of the Tek 22XX and 24XX scopes are suffering from capacitor failure at this point, but once you fix them they work fine and are very nice units.

[oldway]

Are you naming who answered as "bastards" ?

No, he's not, it's an aussie joke.

He should take more care because it's an international forum, not everybody knows aussie joke and not everybody is fluent english speaking.

Translated in french, that's a very bad insult.... Bastard = connard....

[BradC]

Are you naming who answered as "bastards" ?

No, he's not, it's an aussie joke.

He should take more care because it's an international forum, not everybody knows aussie joke and not everybody is fluent english speaking.

Translated in french, that's a very bad insult.... Bastard = connard....

Twaddle. You should take more care to understand what is being said before taking offence. If we all spent out lives tiptoeing around global sensitivities we'd all become Californian. Thankfully we live in a world where I'm free to speak and you are free to take offence. You don't have to of course.

Interestingly "Start ya bastard" is now a Nulon product. The Aussie equivalent of the old CRC aerostart. Very high ether content (so watch for destructive knock on startup). Part number when I last checked was Nulon SYB-001.

[oldway]

Twaddle. You should take more care to understand what is being said before taking offence. If we all spent out lives tiptoeing around global sensitivities we'd all become Californian. Thankfully we live in a world where I'm free to speak and you are free to take offence. You don't have to of course.

Interestingly "Start ya bastard" is now a Nulon product. The Aussie equivalent of the old CRC aerostart. Very high ether content (so watch for destructive knock on startup). Part number when I last checked was Nulon SYB-001.

I would like to know what care I could take other than to use Google translate ?
Perhaps you want I pay a native english speaking guy to explain me why Google Translate is wrong ?
We are on a technical forum, I did not expected for an "aussie joke" because it is not the right place for jokes...We are trying to help TS to repair his Tek 2235....

When I am trying to help somebody, I don't ask for thanks, but I do not accept to take offense...

[Falkra]

Translated in french, that's a very bad insult.... Bastard = connard....

Why would you need to translate this to french ?
Anyway, the b***d would be the oscilloscope in the joke, so I don't think it hurts the oscilloscope.

[oldway]

Translated in french, that's a very bad insult.... Bastard = connard....

Why would you need to translate this to french ?
Anyway, the b***d would be the oscilloscope in the joke, so I don't think it hurts the oscilloscope.

Because I have only a very limited knowledge of English language and "bastard" is not a word that I learned on school.
It was posted just after my post, so it seems to be an answer to what I wrote.

[Falkra]

Don't worry, that was not against you, no pb.

[oldway]

Don't worry, that was not against you, no pb.

Yes, I know, Dave explained it was just a joke....Time to go back to the repair of Tek 2235....

[grumpydoc]

Translated in french, that's a very bad insult.... Bastard = connard....

What have ducks got to do with it?

[oldway]

ducks = canards, nothing to do with connards....In Belgium or in France, you name somebody "connard", you risk serious troubles 

[grumpydoc]

ducks = canards, nothing to do with connards....In Belgium or in France, you name somebody "connard", you risk serious troubles 

I know, it was an attempt at humour

[mmFooD]

Turns out that torx size isn't that common, so it will take a little longer to access the PSU board.

I found the schematic

I guess it is a good thing to replace all of the caps since it probably hasn't been done. But since I don't have an ESR-meter availible. What caps should I focus on? The 2200 uF one  (C904) I guess and then the 1000 one (C940) and the ones around the power supply.

Checking the diodes is a good tip as well!

A related question, should I use low-ESR caps or will any name brand ones due?

[David Hess]

Where did you find a schematic?  That is better than the one I have.  I have been using the 2236 schematic for the 2235.

The high ripple current capacitors are C906 and C940.  These oscilloscopes are old enough however that the other aluminum electrolytic capacitors are worn out also so they should be changed as well.  Typical low impedance capacitors intended for switching power supply applications are suitable.

The X and Y class capacitors on the input, C900, C903, C902, and C904 should also be changed for safety.

[oldway]

You are going to make a fundamental mistake  ... I have 40 years of experience in electronic project and repair and I can guarantee it for you.

Your oscilloscope has crashed suddenly and completely, it is not a "wear" of a capacitor that caused this.

If there is something wrong with a capactor, it is short circuited, not to low capacitance, nor too high ESR....

And a short circuit, you can spot easily with a multimeter.

Start by searching for the defective component.

To repair, touch the circuits as little as possible unless absolutely necessary.

By replacing a bunch of components that are not defective, you may create failures that did not exist like:
- damaging / interrupt tracks or metallizations.
- making connections and mounting errors
- cause bad contacts by stressing solders and connectors
- install capacitors that have been off for a long time and will require a very large inrush current, and this can damage the circuits
- Make polarity errors of the capacitors.

In short, you are gone to create full of unnecessary problems ...

Sometimes, you have really to recap some gears, not because capacitors are old, but because:
- they are bad quality capacitors.
- they are known as having a lot of failures (for exemple, some SMD electrolytics capacitors, some types of tantale capacitors, ...)
- they are working very hot and their livetime is reduced for this reason.

But only recap if it is absolutely necessary....Measure the capacitors you have replaced....if they are ok, it will proove you have made a great mistake. 
 

[2N3055]

Where did you find a schematic?  That is better than the one I have.  I have been using the 2236 schematic for the 2235.

There you are, from the Tek itself..

http://www.tek.com/manual/2235-service-manual

[tautech]

You are going to make a fundamental mistake  ... I have 40 years of experience in electronic project and repair and I can guarantee it for you.

Your oscilloscope has crashed suddenly and completely, it is not a "wear" of a capacitor that caused this.

If there is something wrong with a capactor, it is short circuited, not to low capacitance, nor too high ESR....

And a short circuit, you can spot easily with a multimeter.

Start by searching for the defective component.

To repair, touch the circuits as little as possible unless absolutely necessary.

By replacing a bunch of components that are not defective, you may create failures that did not exist like:
- damaging / interrupt tracks or metallizations.
- making connections and mounting errors
- cause bad contacts by stressing solders and connectors
- install capacitors that have been off for a long time and will require a very large inrush current, and this can damage the circuits
- Make polarity errors of the capacitors.

In short, you are gone to create full of unnecessary problems ...

Sometimes, you have really to recap some gears, not because capacitors are old, but because:
- they are bad quality capacitors.
- they are known as having a lot of failures (for exemple, some SMD electrolytics capacitors, some types of tantale capacitors, ...)
- they are working very hot and their livetime is reduced for this reason.

But only recap if it is absolutely necessary....Measure the capacitors you have replaced....if they are ok, it will proove you have made a great mistake. 

+1 

Some strong clues to the failure could be in the thread I linked earlier if the OP takes the time to study it.
Like this would raise a big red flag for me:

[oldway]

A rule to follow:
When you find a capacitor that has leaked, replace all other capacitors of the same brand that you find in the device even if they do not leak.

But if I can give you a good advice to prolong the life of your 2235, it is to modify it and install a fan like the 2236.

(For installation, see manual service of 2236)

Power supply without fan is working too hot

[David Hess]

You are going to make a fundamental mistake  ... I have 40 years of experience in electronic project and repair and I can guarantee it for you.

Your oscilloscope has crashed suddenly and completely, it is not a "wear" of a capacitor that caused this.

If there is something wrong with a capacitor, it is short circuited, not to low capacitance, nor too high ESR....

Without considering that the aluminum electrolytic capacitors are old enough in this model to be worn out, I disagree because:

Normally when pressing the power button the CRT makes that same, kind of static noise that old TV's does, for a second or two. Now my scope makes this noise but repeatedly for only a tenth of a second or so and the power led is also flashing in the same manner.

[David Hess]

Where did you find a schematic?  That is better than the one I have.  I have been using the 2236 schematic for the 2235.

There you are, from the Tek itself..

http://www.tek.com/manual/2235-service-manual

That is the one I have.  The schematic mmFooD posted is complete and not split across pages.

I tried merging the pages last year and they do not overlap so strips of the schematics are missing.

[JoeO]

There is an even better image here:

Click on the image (48KB) to see a larger version (566KB)

The third post tells how the person put the schematic together.

https://www.electronicspoint.com/threads/tek-2235-good-focus-but-high-intensity.276029/

[oldway]

You are going to make a fundamental mistake  ... I have 40 years of experience in electronic project and repair and I can guarantee it for you.

Your oscilloscope has crashed suddenly and completely, it is not a "wear" of a capacitor that caused this.

If there is something wrong with a capacitor, it is short circuited, not to low capacitance, nor too high ESR....

Without considering that the aluminum electrolytic capacitors are old enough in this model to be worn out, I disagree because:

Normally when pressing the power button the CRT makes that same, kind of static noise that old TV's does, for a second or two. Now my scope makes this noise but repeatedly for only a tenth of a second or so and the power led is also flashing in the same manner.

Replacing all the capacitors, systematically, without even to look for the origin of the defect, this is, as I wrote, a complete mistake.

It is very unlikely that it would be an electrolytic capacitor that make that kind of noise because once the thin Al²O3 insulation is pierced and destroyed, the space between the electrolyte and the aluminum becomes almost zero and it would be impossible that the power supply can restart to such a hig level that a led can be flashing.

But it is true that if there is a short circuited diode, it is also impossible the led beeing flashing....inverter should try to start but led would not be flashing.

On the other hand, it can be a high voltage arc ... Then it's probably visible.

But in any case, you have to do an investigation, not just replace electrolytic capacitors for nothing.

Helping repairing something is always very difficult : for exemple, what is the static noise of an old tv ? CRT colour TV ? tubes ? Transistors ? the noise of charging the 25KV's tube capacitor ?....

[grumpydoc]

David knows of what he speaks.

The PSU in the 22xx series is well known for going into "tick mode" when the electrolytics dry out and go high ESR

C925 and C943 are common culprits but it can be pretty much any of them.

[oldway]

David knows of what he speaks.

The PSU in the 22xx series is well known for going into "tick mode" when the electrolytics dry out and go high ESR

C925 and C943 are common culprits but it can be pretty much any of them.

So, it is better to recommend simply to replace these 2 capacitors, they are small capacitors of low voltage and cheap (10 and 100?F)

The principle of recap systematically audio amplifiers and receivers is used by scam artists to earn a lot of money of audiophiles.

It is better not to repeat this formula unnecessarily

My Marantz 2285B from 1978 that I use everyday has only one capacitor replaced.
And it still has the same performance as when it was new .... worn capacitors? They are 40 years old of good and loyal service and they will probably hold another 10 or 20 years .... It will always be time to think about recap when there will be repeated capacitor breakdowns.

[David Hess]

Replacing all the capacitors, systematically, without even to look for the origin of the defect, this is, as I wrote, a complete mistake.

Of course I would look for the cause of the problem including by a close detailed inspection but once I have to pull such an old piece of equipment apart to gain access, I am going to replace the aluminum electrolytic capacitors and certain other parts if they could have caused the problem.  This goes double when the aluminum electrolytic capacitors are known to be a problem 20 years later as with the 22xx series of oscilloscopes.  (1) (2) The capacitors get tested for capacitance, dissipation, and maybe leakage *after* being pulled to verify if they could have caused the problem.

Changing the X and Y class capacitors is a preventative measure.  The ones Tektronix used fail with age.  I even have some in front of me on 22xx main boards which are split open.

And if I have to order capacitors and other parts, then I would rather order them all at once than replace things individually with separate orders.

Another thing I would do it replace that connector on the TO-220 power transistor.  I use those connectors on transistors sometimes when prototyping but in a finished design, I would solder the wires directly to the transistor leads.  I thought there was a service bulletin from Tektronix about doing this but if so, I have lost track of it.

It is very unlikely that it would be an electrolytic capacitor that make that kind of noise because once the thin Al²O3 insulation is pierced and destroyed, the space between the electrolyte and the aluminum becomes almost zero and it would be impossible that the power supply can restart to such a hig level that a led can be flashing.

Low capacitance and/or high ESR from a worn out aluminum electrolytic capacitor can cause all kind of strange behavior in switching power supplies including hissing and buzzing.

My Marantz 2285B from 1978 that I use everyday has only one capacitor replaced.
And it still has the same performance as when it was new .... worn capacitors? They are 40 years old of good and loyal service and they will probably hold another 10 or 20 years .... It will always be time to think about recap when there will be repeated capacitor breakdowns.

The common case that I run across is that once one of the aluminum capacitors has failed do to less derating than the others, the others have exceeded their rated operating life by several times and are displaying low, but not zero, capacitance and high, sometimes 10 times high, dissipation.  This is not such a big deal with bulk decoupling capacitors which may only marginally affect circuit performance but input and output capacitors are more important and especially so in switching power supplies where ripple currents are higher.

(1) There is nothing wrong with the 22xx design in regard with the capacitor operating life.  That is just the way it goes with aluminum electrolytic capacitors; they have known wear out mechanisms so do not last forever and degrade even when not used.

(2) There also seems to be a failure mode where old 22xx switching power supplies self destruct and I have not ruled out bad capacitors causing this.

[oldway]

I notice that we have a completely different conception of what is repairing an electronic equipment ... Your conception is rather what is called "reconditioning" and not repairing.

I do not like preventive electronic maintenance because I believe that in electronics, everything is measurable and one can find a defect in a component by measures, even sometimes simple measurements as ripple.

In my experience, the more you want to do (and in particular replace components unnecessarily), the more you create problems.

The systematic replacement of components is often a solution for those who do not understand what they are doing (not your case, of course) or beginners .... I am totally opposed to it.

It is sometimes also a solution in the factory because the time is expensive and the repaired SMPS must pass by the same tests as a new power supplies .... one can not afford to see the power supply refused to the tests, then, one changes all the capacitors of output filters to be sure ...That's what I was doing when I was working by Cherokee Europe.   In factory, the capacitors are much cheaper than the labor.

But for a Tek 2235 oscilloscope that is old-fashioned, which is worth 100 bucks and is likely to be in the trashbin in a few years (nobody want an analog oscilloscope anymore), for me, what must be done is to look for the faulty component, replace it , and check the power supplies, ripple, correct functional operation, and make an approximate calibration and it is good for some more years of use.

And do not pull such an old piece of equipment apart If not absolutely necessary...!!!! (sometimes, no way, you have to pull it apart...See picture of repair of an Hameg HM 605 with bad solder in output horizontal amplifier).

Ten or fifteen years ago, I was repairing a lot of analog oscilloscopes of different brands, but now, there is no more analog oscilloscopes to be repair....People prefer to buy a new DSO than to spend time or money repairing old analog scopes.

I still have some success with vintage audio and I am applying my conception of repair without any problem....all my clients are very satisfied and I had no return of any problem after my repairs.

[David Hess]

I notice that we have a completely different conception of what is repairing an electronic equipment ... Your conception is rather what is called "reconditioning" and not repairing.

I would reserve reconditioning or refurbishing to cases where I do so comprehensively.  For a power supply repair, I tend to refurbish anything I have to gain access to so I only have to do so once.  Further I know old aluminum electrolytic capacitors have reliability issues and they are easy to replace.

I do not like preventive electronic maintenance because I believe that in electronics, everything is measurable and one can find a defect in a component by measures, even sometimes simple measurements as ripple.

The systematic replacement of components is often a solution for those who do not understand what they are doing (not your case, of course) or beginners .... I am totally opposed to it.

Sometimes it is less time consuming and safer to replace multiple suspected parts and if I have to pull a part to test it, I am not likely to reinstall it.  There is also a risk tradeoff to consider.

For example if I suspect a high voltage DC restorer problem, then I am likely to replace the diodes and capacitors without testing them in circuit because testing is both hazardous and very difficult.  Once I pull those parts making testing slightly easier, I am hardly going to reinstall them so they are all going to be replaced.

The 22xx switching power supplies count here because I know they have a mysterious destructive failure mode which seems to occur when the capacitors are worn out.  The only reason it may be mysterious to me is simply because I have never gotten my hands on one which has failed destructively.  All of mine work perfectly.

In my experience, the more you want to do (and in particular replace components unnecessarily), the more you create problems.

I agree and this is especially a problem when lack of experience is a factor in which case I council very limited tests and repairs.

But for a Tek 2235 oscilloscope that is old-fashioned, which is worth 100 bucks and is likely to be in the trashbin in a few years (nobody want an analog oscilloscope anymore), for me, what must be done is to look for the faulty component, replace it , and check the power supplies, ripple, correct functional operation, and make an approximate calibration and it is good for some more years of use.

I would not use a 2235 but only because I have better old oscilloscopes some of which have useful capabilities not duplicated in many modern DSOs unless you want to spend a lot of money.

I still have some success with vintage audio and I am applying my conception of repair without any problem....all my clients are very satisfied and I had no return of any problem after my repairs.

Most audio clients would not notice subtle problems.

[oldway]

This is the rate of failure of electrolytic capacitors....

If you replace a capacitor by a new one, you must be sure you are in the wear-out failure segment...And to be sure of that, you must have first several failures of the same brand of capacitor in the same equipment...

If you replace them too soon (during the random failure segments), you will have more failures than if you don't replace them.

[oldway]

And rate of failure of a new electrolytic capacitor may even be higher than of the old one. 

[mmFooD]

I found the schematic by doing a web search for "tektronix 2235 schematic" not sure what site.

So I finally removed the PSU cover and could have at least a visual inspection of all components. I can't see any signs of any leaking caps. But I can't be sure they aren't defective since I don't have the proper equipment.

I did however try and test the diodes around the T948 and it seems as if most of them are shorted. Could this be correct? If so how could that have happened? I tried two multimeters, one fluke 83 and a cheaper uni-t UT132C. They pretty much showed the same results but the uni-t one showed ".Ol" for values over 1.5 V.

	Voltage drop (mV)
#		fwd		rev
CR960		50		50
CR962		50		50
CR961		0		0
CR963		0		0
CR956		337		407
CR957		337		403
CR954		439		2527
CR955		439		2526
CR879		492		Ol
CR907		493		999
CR970		99		157
CR967		99		157

I tested even more diodes around the board(s) and found several more that were shorted. It all seems very weird to me.


I will try and read through the linked thread and see if I learn something useful.

Thanks for all the help so far.

EDIT:
If I replace the caps. What should I be looking for in terms of specs and what should I avoid. For instance should they be Japaneese named brands only? Low ESR?

[David Hess]

This is the rate of failure of electrolytic capacitors....

If you replace a capacitor by a new one, you must be sure you are in the wear-out failure segment...And to be sure of that, you must have first several failures of the same brand of capacitor in the same equipment...

If you replace them too soon (during the random failure segments), you will have more failures than if you don't replace them.

Not having the aluminum electrolytic capacitors in the wear out portion of their reliability curve is hardly going to be a problem in a piece of test equipment which at a minimum is 28 years old.  Input capacitors in linear power supplies last the longest because they end up massively derated but even they can fail at that point though corrosion of the lead connections as water or oxygen leak through the rubber seal; eventually they dry out also.

I leave 60 Hz input capacitors alone until one actually fail but the ones in my 7603 are finally going bad do to low capacitance so it is time to rebuild its power supply.  When I had one of the big input can capacitors fail in my PS503A power supplies, I bought 4 replacements and changed them all in 2 different power supplies.  I changed the other capacitors as well and they tested as worn out with a dissipation and ESR 10 times greater than their operating life specification although their capacitance was still within tolerance.  There is a Sprague 30D series aluminum electrolytic capacitor in the z-axis amplifier of the 7904 which in my case was bad *despite* testing good for capacitance and dissipation at 1000Hz; its high frequency performance had decreased.

[David Hess]

So I finally removed the PSU cover and could have at least a visual inspection of all components. I can't see any signs of any leaking caps. But I can't be sure they aren't defective since I don't have the proper equipment.

The designer properly derated them so they do not generally show external signs of leakage.

I did however try and test the diodes around the T948 and it seems as if most of them are shorted. Could this be correct? If so how could that have happened? I tried two multimeters, one fluke 83 and a cheaper uni-t UT132C. They pretty much showed the same results but the uni-t one showed ".Ol" for values over 1.5 V.

The diodes are almost certainly going to test as a short or close to it in-circuit because of the load from the rest of the circuits through the transformer windings.  I do not see any measurement in your list indicating a problem.

If I replace the caps. What should I be looking for in terms of specs and what should I avoid. For instance should they be Japaneese named brands only? Low ESR?

Look for low ESR, high frequency, and high ripple current capacitors intended for switching power supply applications.  I like the Nichicon PW series for general use but there are many options.

[oldway]

Why don't you power on your oscilloscope with the power supply cover removed (beeing carefull !!!) to listen where is the noise coming from and to look if there is no arcing ?

I use to test 2235 feeding C940  with an external power supply 42V current limited to 2A.

If the scope is functioning correctly, then you know the problem is coming from pre-regulator.

[oldway]

Not having the aluminum electrolytic capacitors in the wear out portion of their reliability curve is hardly going to be a problem in a piece of test equipment which at a minimum is 28 years old.  Input capacitors in linear power supplies last the longest because they end up massively derated but even they can fail at that point though corrosion of the lead connections as water or oxygen leak through the rubber seal; eventually they dry out also.

there are a lot of parameters influencing the lifetime of the capacitor.

Lifetime is given in hours of working...this can vary a lot....
How many years the oscilloscope old is, that does not matter.
What matter is how many hours did it worked and at what temperature...Every 10°, lifetime is reduced by 2....

As I said, my Marantz from 1978 is still working with all is electrolytics, only one has been replaced...And this receiver is 39 years old....

I have also a SANSUI AU-5500 from 1975 working very well with all his originals electrolytics...42 years old !!!

Others Sansui's AU-7700, AU-999, AU- 9500 from the same time are also still working with their originals electrolytics ....It is prooven that the fact the oscilloscope is 28 years old does not mean that the capacitors are in the wear-out  failure segment.

[oldway]

Look for low ESR, high frequency, and high ripple current capacitors intended for switching power supply applications.  I like the Nichicon PW series for general use but there are many options.

Not at all, if you replace capacitors, buy common capacitors of good brand, no need to be low ESR.
But buy those capacitors from reliable electronic parts suppliers NOT FROM EBAY...!!!!

Why it is not neccessary and even wrong to replace the capacitors by special low ESR capacitors ?

Some people suffers what I call "ESR madness".....It seems that ESR is the name of some godess and that they are repeating again and again their prays to the adoration of this godess...

1) Look at the Service manual of 2235: you have the list of replaceable electrical parts of A1 with all the specifications of the components....none specification of low ESR capacitors...!!!!

One of the rules in repairing is to respect the technical specifications of the manufacturer when you replace a component....Not better, not worse component, only the same as specified.

Could replace a component by a better one cause any harm ?

Yes, it can: I have here a Marantz Model 170dc to repair...It has beeing repair before and the "clever" guy has replaced some transistors by new ones, other types than what the manufacturer had recomended...

Those more moderne transistors has greater gain....What happen ? the power stage where he replaced the transistors is oscillating at about 2 Mhz...

2) What are the capacitors with high ripple current ? (NB: if there only low ripple current, for sure, there is no need of a low ESR capacitor)
- C906: has both ripple current of (120) 100Hz of rectified main's voltage and ripple from buck pre-regulator but ripple voltage on capacitor does not seems critical....there no specification of max ripple, nor test point on this capacitor.
- C925: has no ripple current but is very critical for leakage....should be replaced but low ESR is not needed.
- C940: Has both ripple current from the pre-regulator and from the square wave inverter T948. The square wave inverter has very low ripple current, its nearby dc current. Ripple current is limited by the inductance of T906. No high ripple current, no need to be low ESR.
- C944: very low ripple current because drive current of Q946 and Q947 is square wave. No need to be low ESR.
- C943: very low ripple current because collector current of Q946 and Q947 is square wave. No need to be low ESR.
- C942: no ripple current, no need to be low ESR.
- C968 : very low ripple current because square wave inverter and load not varying very much, not necessary to be low ESR.
- C970 : very low ripple current because square wave inverter and load not varying very much, not necessary to be low ESR.
- C960 : very low ripple current because square wave inverter and load not varying very much, not necessary to be low ESR.
- C961 : very low ripple current because square wave inverter and load not varying very much, not necessary to be low ESR.
- C956 : very low ripple current because square wave inverter and load not varying very much, not necessary to be low ESR.
- C954 : very low ripple current because square wave inverter and load not varying very much, not necessary to be low ESR.
 

[David Hess]

1) Look at the Service manual of 2235: you have the list of replaceable electrical parts of A1 with all the specifications of the components....none specification of low ESR capacitors...!!!!

Tektronix *never* mentioned ESR or ripple current in their parts lists whether it was important or not.  The only way to know is to recognize the manufacturer's part number.  They also did not distinguish between tantalum and aluminum electrolytic capacitors.

One of the rules in repairing is to respect the technical specifications of the manufacturer when you replace a component....Not better, not worse component, only the same as specified.

Unless you find the specifications based on the manufacturer's part number or maybe the entry in the Tektronix Common Parts Design Catalog which I do not have, then you will never know.  The best you can do is to reverse engineer the design and sometimes this will reveal shortcomings where better parts should have been used.

One of the rules in repairing is to respect the technical specifications of the manufacturer when you replace a component....Not better, not worse component, only the same as specified.

This gets really tricky with capacitors in old switching power supplies because modern capacitors with the same voltage rating and capacitance will often have a *worse* ripple current rating.  This is simply because they are smaller.  One way to make up for this is to voltage derate the capacitor forcing it to be physically larger.  I would actually voltage derate C906 and C940 for this reason.

I did not want to complicate matters by discussing the requirements of every capacitor in the circuit.  I used the Nichicon PW series as a low cost example which met the requirements for all of the capacitors.

The capacitors on the inverter do not require high ripple current but low ESR is a good idea for them anyway to do filtering requirements.

[mmFooD]

So there is nothing wrong with the diodes?

I am a bit curious as to what I would find if I tried to start the scope. Is there some easy way of discharging the caps after pulling the cord again? Or how long should I wait before it is safe to poke around?

Is nichicon the same as nippon chemi-con? That is the closest brand name  I Could find at my local retailer.

Skickat från min ONEPLUS A3003 via Tapatalk

[David Hess]

So there is nothing wrong with the diodes?

As far as I can tell from your measurements, there is nothing wrong with the diodes.

I am a bit curious as to what I would find if I tried to start the scope. Is there some easy way of discharging the caps after pulling the cord again? Or how long should I wait before it is safe to poke around?

In this design, the capacitors will discharge very quickly once power is removed so manually discharging them is not necessary.

Is nichicon the same as nippon chemi-con? That is the closest brand name  I Could find at my local retailer.

Nichicon and Nippon Chemi-Con are different companies.  They both make suitable capacitors.

I have been working on a spreadsheet showing recommended replacement parts for the 22xx series power supplies however I have not gotten up to the capacitors except to note which models used which values.

[mmFooD]

Turned it on and these are the buggars making the noise I think. And they are flashing.

And they are referenced as DS856, DS858, DS860 on the PCB.


The power led is flashing just the first two burst in sync with those three components but then it fades out completely.


My thinking is that it might be caused by too high ripple in the PSU and thus faulty caps in there somewhere? Am I totally confused?

[oldway]

So there is nothing wrong with the diodes?

It seems all right.

I am a bit curious as to what I would find if I tried to start the scope. Is there some easy way of discharging the caps after pulling the cord again? Or how long should I wait before it is safe to poke around?

The only dangerous capacitor is C906....It will discharge slowly, you can measure the voltage and disharge the capacitor with a 1K 5W resistor if necessary.

There may be some high voltage on the PDA connection of CRT, but no real danger of this side.

Power on this scope, and report what you see and listen.....Ok, done, we have a solution now ! 

Tektronix *never* mentioned ESR or ripple current in their parts lists whether it was important or not.  The only way to know is to recognize the manufacturer's part number.

For me, there is a good reason why Tektronix never mentioned low ESR in the part list: ranges of special low ESR electrolytic capacitors is something relatively recent :  the market demand for low ESR, high capacitance
capacitors has been driven by industry trends toward reduced circuit voltages operated at increased frequency with higher current demands than in the past....As far I remember, in 1981, I was working in a big power electronic manufacturer and whe had a problem related with ESR and there was no low ESR capacitors on the market.

It is more or less ten years later that the low ESR capacitors appeared on the market.

You know that Tektronix has always been very accurate and detailed in their technical specifications....It is not possible that they simply would have forgotten or neglected an information as important as "low ESR" if needed...

The explanation is simpler: low-esr capacitors did not exist at the time this oscilloscope was developed and manufactured.

Later, Tektronix has used OEM SMPS and did no more give detailed schematics and part list of their power supplies.

[mmFooD]

I did just that, in my post below  . Not sure what the problem is though. Do those components behavior mean anything to you?

[oldway]

I did just that, in my post below  . Not sure what the problem is though. Do those components behavior mean anything to you?

Yes, it does....they are protecting the power supply and other components if the power is not regulated anymore and voltages goes too high....

[oldway]

Have you a multimeter with min/max recording ?

[mmFooD]

Yes! Seems like the fluke does

Skickat från min ONEPLUS A3003 via Tapatalk

[oldway]

Ok, very goed, you have to check what's the max voltage of the 42V rail (on C940) and of the +8.6V  rail(on C960)

Don't let it switched on for a long time in such condition ....Try to measure these max voltages as quick as possible...report the results.

[oldway]

If I had to repair this scope, I would feed the 42V rail by an external power supply, I  would begin with 38V current limited at 2A, it is far safer to work and to protect the electronic circuits of the scope...

But I think it is not possible because you don't have such a power supply.
I have a HP 6296A  60V 3A for this use.

[mmFooD]

C960 - max 1.8 v and min 2V and C940 is showing min - 1.3V and max .Ol

Skickat från min ONEPLUS A3003 via Tapatalk

[oldway]

C960 - max 1.8 v and min 2V and C940 is showing min - 1.3V and max .Ol

Skickat från min ONEPLUS A3003 via Tapatalk

Not possible, there is something wrong with your measurements.
I was waiting for something like 42V or more, but lower than 51V because crowbar protection did not atuated.

[oldway]

Also check Q944 for short circuit collector-emitter.

[oldway]

I explain my reasoning .... ds856, ds858 and ds870 are spark gaps designed to protect the CRT and all circuits against excessive supply voltages from the inverter.

Testing with the flashing spark gaps should be avoided as it is risky for the entire oscilloscope.

But the lack of technical ressources of the TS does not leave the choice of the method of work.
In fact, it is a mistake to try to repair an oscilloscope under such conditions.

Nevertheless, I hope that the TS is aware of the risk that he takes to permanently damage the oscilloscope by its attempt to repair.

So I'm trying to confirm:
- that (for a reason to check), the inverter supplies a voltage too high to the secondary.
- that 42V rail regulation is not responsible for this problem.

[David Hess]

Tektronix *never* mentioned ESR or ripple current in their parts lists whether it was important or not.  The only way to know is to recognize the manufacturer's part number.

For me, there is a good reason why Tektronix never mentioned low ESR in the part list: ranges of special low ESR electrolytic capacitors is something relatively recent :  the market demand for low ESR, high capacitance capacitors has been driven by industry trends toward reduced circuit voltages operated at increased frequency with higher current demands than in the past....As far I remember, in 1981, I was working in a big power electronic manufacturer and when had a problem related with ESR and there was no low ESR capacitors on the market.

It is more or less ten years later that the low ESR capacitors appeared on the market.

Tektronix never listed ESR or ripple current rating even when these were specified as with the Sprague 30D series.

You know that Tektronix has always been very accurate and detailed in their technical specifications....It is not possible that they simply would have forgotten or neglected an information as important as "low ESR" if needed...

They did not mention it in the parts list.  Their Parts Catalog for capacitors (secitons 9 to 14 of catalog 5 of 6) likely listed it but I do not have it and I do not know anybody who does.

The explanation is simpler: low-esr capacitors did not exist at the time this oscilloscope was developed and manufactured.

If someone has access to the datasheets for the Sprague 17D and Nichicon ULB series, then we could know.  Often the parts existed as with the Sprague 30D series but were not labeled that way.

Some idea can be gained however.  I have some capacitors that I pulled from 22xx series power supplies here including a Nippon Chemi-con 1000uF 50V SM series used for C940.  The SM series was replaced by the SME series which has a max ESR of 0.116 ohms and a ripple current rating of 1530 milliamps.  I even have my metric calipers sitting here and can confirm that the diameter is the same and that the SME is slightly shorter which is to be expected.  The PW series capacitor that I recommended is comparable or superior in every way although I would try to find a higher voltage part which is taller with the same diameter.

When one of my 22xx power supplies needs to be rebuilt, I plan on making measurements of the ripple current for the various capacitors.

[oldway]

Tektronix never listed ESR or ripple current rating even when these were specified as with the Sprague 30D series.

Indeed, they do better than this, they gave you the waveform of the ripple current. 

On test point 39, you have the waveform of ripple current of the C941: that's a triangular waveform....

Amplitude measured on a 0.2R shunt resistor is 50 mV peak (150-100mV)

Ripple of C941 is a triangular current of 250mA peak !!!! No need of low ESR capacitor.....

[mmFooD]

Well there is definitely no  short in Q944. And as to the C940 and C960 I tried setting the fluke 83 to manual range 400V and using the min/max button. I record similar values as before but a little higher this time. About 16V on one of them but no 45V.


C960 2V max ~ 0V min
C940 16V max -2V min


The fluke specs says that it is sampling at 100ms intervals from what I can tell. I am measuring from the back of the PCB over the footprint of the capacitor (black probe to negative and red to positive). Maybe I am doing something wrong?

[mmFooD]

This is where i did the measurements

Skickat från min ONEPLUS A3003 via Tapatalk

[David Hess]

Well there is definitely no  short in Q944. And as to the C940 and C960 I tried setting the fluke 83 to manual range 400V and using the min/max button. I record similar values as before but a little higher this time. About 16V on one of them but no 45V.

C960 2V max ~ 0V min
C940 16V max -2V min

The fluke specs says that it is sampling at 100ms intervals from what I can tell. I am measuring from the back of the PCB over the footprint of the capacitor (black probe to negative and red to positive). Maybe I am doing something wrong?

Sampling for how long?  In my experience multimeter min/max modes are only useful for steady inputs.  I am sure the peak voltages are much higher than that to be illuminating the power LED.

C940 is referenced to the isolated ground on the primary (off-line) side and not chassis ground.  This point is test point TP950 in the lower left corner of the schematic.

If you want something to measure, pull C940 and check it for capacitance and leakage and if you have the test equipment, dissipation and ESR.  Messing around on the off-line side of a malfunctioning switching power supply with just a multimeter would not be my first choice.

Or if you have suitable capacitors, tack solder them in parallel with C940, C906, and maybe C925 on the back of the board. 

[oldway]

You need a power supply capable of supplying 45V 2A and another oscilloscope.

The solution is to feed the 42v rail (C940) by an external power supply set at about 38V and adjusted it to have the -8.6V at the test point ... It is then possible to check whether the potentiometer R938 acts or not

Without technical means, and in addition, at a distance, what should be very simple becomes very complicated.

[TerraHertz]

Are you naming who answered as "bastards" ?
I did not like your answer at all....

Oh, sorry for the misunderstanding. And I did not come back to this thread till now, to notice the problem.
As Dave said, it's an Aussie joke. I'll explain for all non-Oz-English speakers:

The OP used slightly unusual English in the title. A native English speaker would have said 'unable to power up Tek 2235..." The phrase "unable to start" tends to imply something like an internal combustion engine won't start.
Now, if you'd watched the video you'd see it's a parody, a fake product advertisement. It plays on the very typical Aussie sketch of some guy trying to start a reluctant lawnmower or car engine, and ending up yelling "Start, you bastard!" at it in frustration. Actually this is an abbreviation, he'd be more likely to phrase it "START, you F*CKING BASTARD!"
Anyway, the product is supposed to be able to make any engine start easily. (We all wish!) Bearing in mind that in Oz  the word 'bastard' is considered mild and commonly used. Can even be a term of friendship. Depends on the context and tone of voice.

And so, the joke - an image of OP using this product to get his scope 'started'. Funny because it's absurd.

I think if you are easily offended, better not take a holiday in Australia. If the drop bears and bunyips don't get you, a heart attack will.

Meanwhile I have a faulty Tek 2232 that I really must get around to fixing. Been too busy.

[oldway]

I would never understand that your post had something to do with the tittle of the topic without your explanation....
In Europe, I never heard about this product, we use "start pilot" to start engines by cold wheater.

I think if you are easily offended, better not take a holiday in Australia. If the drop bears and bunyips don't get you, a heart attack will.

I am only waiting to win at the lottery, but no chance this would happen, I don't bet 

Heart attack ? I allready had one, why not another one ? 

[oldway]

@mmFooD:

You wrote:

Turned it on and these are the buggars making the noise I think. And they are flashing.
And they are referenced as DS856, DS858, DS860 on the PCB.
The power led is flashing just the first two burst in sync with those three components but then it fades out completely.

1) DS860 ? I did not find it on schematics....Is it not DS870 ?

2) Are you 100% sure that all the 3 spark gaps are flashing ? Not only two of them ?

[mmFooD]

Quote from: oldway on Today at 11:44:37

@mmFooD:

You wrote:


Quote

Turned it on and these are the buggars making the noise I think. And they are flashing.
And they are referenced as DS856, DS858, DS860 on the PCB.
The power led is flashing just the first two burst in sync with those three components but then it fades out completely.

 
1) DS860 ? I did not find it on schematics....Is it not DS870 ?

2) Are you 100% sure that all the 3 spark gaps are flashing ? Not only two of them ?

You are correct. I misread the R860 and the DS870. And, of course, only two spark gaps are flashing. As shown in the pictures. DS870 is the one not flashing.

[oldway]

I am little confused...They are neon bulbs, not spark gaps....they are limiting voltage at about 60V each....

And you said they make noise ? How is it possible ? I have never heard a neon bulb making noise....Are you sure that the noise come from those neon bulbs ?

I think they are glowing because something is arcing.

[mmFooD]

Yes, even though I read and listen to english everyday without trouble it is a different matter of expressing one self. It is a great lesson in how it is not enough to just read up on a subject without practicing what you learn.

[mmFooD]

I have no idea what they are. You called them spark gaps so I figured that's what they are.


I can't say for sure that they are the ones making the noise. It is difficult to tell but it is definately coming from that area, and the noise is occuring in sync with the flashes. I'll se if I can pin point the source of the noise.

[oldway]

As I said, we should check the max output voltage of the inverter to know what is happening, we tried with the min/max function of your Fluke 83 and it did not worked.

You need another oscilloscope, using it to measure the max voltage of 8.6V (or -8.6V) supply to see if inverter outputs are going overvoltage or not.
Have you another oscilloscope ?

[mmFooD]

you are also correct in that I don't have a bench power supply . Allthough I really want to build one with a max output of min 12V (maybe it should be able to output around 50V with this experience) and 1A. More if not to expensive so if your know of any good DIY projects that I could take notes from or where I can read up on PSU theory and -configurations that would also be greatly appreciated.


With the lack of proper equipment it seems to me that replacing the caps in the PSU area might be a good thing. If only to check that one of the list of possible errors. And then do some more trial and error replacements and see if any of it helps. But at the moment buying a DSO is not within my budget. Analog osciloscopes are kind of hard to find in this part of the world, at least for around the $60 that I payed for it. So I have to make this work.

[oldway]

There is something arcing, we must know what is arcing and why....Too high voltage or isolation failure ?

We must answer this question first.

Wear-out capacitors generaly results in a too low voltage supply, not a too high voltage....so replacing the capacitors will not solve the issue.

You do not have a friend who can lend you an oscilloscope?

[mmFooD]

No, I can't think of anyone I know that would have a scope. And I read up on he fluke specs again and found that it only measures capacitance up to 5 uF so no help there either. I'll do some more voltage probing to see if I can find the 40V somewhere

[oldway]

Repairing an analog scope is a very specialized work and it need very specialized equipment....

Such a failure as your oscilloscope, I repair it in less than one hour for sure.....

But without another oscilloscope, HV probe, an isolation transformer and a power supply, how do you want to repair an analog oscilloscope with a minimum of safety?

Do not worry with pre-regulator and 42V rail anymore.....Look at the 8.6V (or -8.6V) rails.....have you an analog multimeter ?

 

[mmFooD]

I think we have some success here finally. I tried setting the range to 4000 manually on the fluke and then used the min/max mode to measure over TP900 and TP940. I recorded a maximum spike with a value of 72V. (-72V).


I then tried measuring over TP900 and one lead of W961 marked with "-8.6V"  and around the time of this measurment the flashing stopped. Suddenly I recorded -8.55V at this point and around 48V on TP940. But still nothing on the display. But then it started flashing again, i think after I turned from the side and placed it bottom down it started flashing again. And this time a trace was visible in sync with the flashes and the voltage values on TP940 and the -8.6V where sketchy again. (a couple of volts on the -8.6V point and 74.4V max spike on TP940).


I guess new is better when it comes to troubleshooting but the question remains: what is going on 

[David Hess]

I am little confused...They are neon bulbs, not spark gaps....they are limiting voltage at about 60V each....

And you said they make noise ? How is it possible ? I have never heard a neon bulb making noise....Are you sure that the noise come from those neon bulbs ?

I think they are glowing because something is arcing.

The neon bulbs usually flash on startup and shutdown and this is normal.  They are being used as low capacitance low leakage surge suppressors to prevent excessive voltages on the CRT elements.  They are inexpensive alternatives to gas discharge tubes.

[mmFooD]

Turned it on again: No flashing -53.6V on TP940 -8.5V on TP961 and it showed a trace on the display  . But it is not stable

[mmFooD]

This is at 2mV/div (1x) vertically and 10us/div horizontally

Skickat från min ONEPLUS A3003 via Tapatalk

[oldway]

Turned it on again: No flashing -53.6V on TP940 -8.5V on TP961 and it showed a trace on the display  . But it is not stable

Measure ripple on C940 with your Fluke on AC range. (directly between the + and - connections of C940)

And you can change the tittle of your topic:

suddenly unable to start Tektronix 2235 (analog) oscilloscope

[mmFooD]

on range 40 it is recording an avarage of about 1.8V and max 7.6V. Min is 0.12V

[oldway]

That's far too much, replace it.

[David Hess]

I think we have some success here finally. I tried setting the range to 4000 manually on the fluke and then used the min/max mode to measure over TP900 and TP940. I recorded a maximum spike with a value of 72V. (-72V).

Like I said, the transformer provides isolation so measurements between the primary and secondary side have little meaning.  The input voltage which matters is between TP950 and TP940 and should be +42.8 volts

I then tried measuring over TP900 and one lead of W961 marked with "-8.6V"  and around the time of this measurment the flashing stopped. Suddenly I recorded -8.55V at this point and around 48V on TP940. But still nothing on the display. But then it started flashing again, i think after I turned from the side and placed it bottom down it started flashing again. And this time a trace was visible in sync with the flashes and the voltage values on TP940 and the -8.6V where sketchy again. (a couple of volts on the -8.6V point and 74.4V max spike on TP940).

I guess new is better when it comes to troubleshooting but the question remains: what is going on 

If the +42.8 volts at TP940 is high, then SCR crowbar Q935 fires shorting out the preregulator and protecting the inverter.  This resets the preregulator which restarts after a short delay.

If C940 or maybe C906 is worn out, then the preregulator will have problems regulating the output voltage.

This is at 2mV/div (1x) vertically and 10us/div horizontally

This is switching noise from the inverter which operates at about 20 kHz.  It implies that the input voltage is just low enough to prevent regulation and the output capacitors are bad.

[oldway]

Replace first C940 and everything will be ok....

Have you some capacitor 50 or 63V 470 or 1000µF ?
You can even solder the capacitor in parallel to make a quick test

[mmFooD]

David Hess: I did a measurement between TP950 and TP940 - max 75.2V. Now that I see the TP950 on the circuit it seems obvious that that is the ref point to measure against. Thanks!


I also measured ripple over C906 - max 2.8V. Should I replace that one as well?


Off topic:
Can anyone give me a hint as to what the purpose of the regulator vs the "regular" regulator and what part linear, LDO, and switch mode regulators have to play in the matter.

I am currently trying to regulate a 12V PLC voltage to 5V to feed a MCU and some other stuff. Preferably it should be able to handle at least 500mV. First I tried a single LM7805 but it gets very hot even with the biggest heat sink I could make room for and since it will be placed in a poorly ventilated housing that is not a good solution.

So I tried the MC34063 switch mode regulator. But the calculations where made on an input voltage of 12V and I am using an arduino uno with a 12V input as PSU and it seems to have a voltage drop of 0.7V which I think messes up the operation of the MC34063. I get a output voltage of about 5.2V and the MCU and 7segment LED display is flickering and not working proberly.

Should I use a linear and switch mode regulator in pairs, with one as a preregulator for the other?

I get that this might be a little to OT. But if anyone know some good literature online or on paper on this topic I'd appreciate it a lot if you could send me a link or other info in a PM.

[mmFooD]

I have a 1000uF and several 470uF but they are a lot smaller. Only 35 V.


Another kind of OT question. Is it a good idea to have >50-63V caps as stockpile caps. The ones I have atm is 35V and 16V.

[oldway]

I also measured ripple over C906 - max 2.8V. Should I replace that one as well?

Not necessary.

I have a 1000uF and several 470uF but they are a lot smaller. Only 35 V.

For test, if you have 8 x 470µF 35V, you can make a 940µF 70V by association in serie/parallel....Don't forget 2 resistors 10K 1/2W as voltage divider.

Another kind of OT question. Is it a good idea to have >50-63V caps as stockpile caps. The ones I have atm is 35V and 16V.

You should know better than us what you need as components...

Question off topic: open a new topic.

[David Hess]

I also measured ripple over C906 - max 2.8V. Should I replace that one as well?

I would replace them all but as for the ripple across C906, I am not real clear what the designers were intending.  The ratio of capacitance between the input, C906, and output, C940, capacitors is oddly high for an offline switching power supply and later models of this power supply increases the value of C906 to 270 microfarads which seems more reasonable although they had to add a PTC for inrush current limiting as well which may explain the low value of C906 in the 2235.  C940 actually seems to have too much capacitance unless that was needed to increase the ripple current rating or lower the ESR.  (1)

The ripple across C906 should actually be higher at roughly 26 volts peak-to-peak when operating on 60Hz 120VAC power.

Off topic:
Can anyone give me a hint as to what the purpose of the regulator vs the "regular" regulator and what part linear, LDO, and switch mode regulators have to play in the matter.

I do not understand the question.

Should I use a linear and switch mode regulator in pairs, with one as a preregulator for the other?

That can be done but it is usually not required.  Linear regulators are poor at removing noise from switching regulators when better filtering is usually called for.  This is why the 2235 uses PI output filters after the transformer and no linear regulators; there are some subtle design features which produce a low noise output.

I get that this might be a little to OT. But if anyone know some good literature online or on paper on this topic I'd appreciate it a lot if you could send me a link or other info in a PM.

Start by reading through the free Linear Technology application notes which discuss switching and linear regulators.  Several discuss the trade offs between switching and linear regulators.

(1) This is why I do not care about the measured capacitance of C940 or its ripple current rating.  That capacitor is worn out because the simplest explanation is that Tektronix selected its value to meet ESR requirements and there is no way it is still within its ESR specifications after 27+ years.  The same goes for almost all of the other capacitors; they are worn out.

I have a 1000uF and several 470uF but they are a lot smaller. Only 35 V.

A pair of the 470uF capacitors in series should be good enough to verify if there is a problem.

Another kind of OT question. Is it a good idea to have >50-63V caps as stockpile caps. The ones I have atm is 35V and 16V.

If you regularly work with 50 volt power supplies, which used to be common in analog circuits, then you should have them.  Some economy in parts selection can be had by stocking higher voltage capacitors for use in lower voltage applications; in most cases you do not need capacitors with every available voltage rating if you have higher voltage ones.

[mmFooD]

I do not understand the question.

Sorry it is supposed to say "the purpose of preregulator vs 'normal' regulator". I guess the preregulator is regulating the voltage to a more suitable one for the 'real' regulator. But when are they used and what type (linear, LDO, and switch mode) are used as one or the other.

That can be done but it is usually not required.  Linear regulators are poor at removing noise from switching regulators when better filtering is usually called for.  This is why the 2235 uses PI output filters after the transformer and no linear regulators; there are some subtle design features which produce a low noise output.

Ok, I looked up PI output filter and I think that is the optional filter mentioned in the MC34063 datasheet so it would be sufficient I think.

Start by reading through the free Linear Technology application notes which discuss switching and linear regulators.  Several discuss the trade offs between switching and linear regulators.

Will do!

A pair of the 470uF capacitors in series should be good enough to verify if there is a problem.

ok, I might try that then.

[David Hess]

Sorry it is supposed to say "the purpose of preregulator vs 'normal' regulator". I guess the preregulator is regulating the voltage to a more suitable one for the 'real' regulator. But when are they used and what type (linear, LDO, and switch mode) are used as one or the other.

There are a bunch of examples in Linear Technology application note 2 including some that use SCRs for preregulation that are very similar to how the original 22xx power supply in the 2213 and 2215 worked.

Ok, I looked up PI output filter and I think that is the optional filter mentioned in the MC34063 datasheet so it would be sufficient I think.

Linear Technology application note 70 discusses noise in switching regulators.  There are all kinds of subtle issues.

[mmFooD]

Ok, so i finally replaced C940. But it didn't fix it. I did a ripple measurement on the new one and it was high as well. I didn't have the fluke with the min/max function. But I measured peaks of 12V. Did I botch the soldering?

Skickat från min ONEPLUS A3003 via Tapatalk

[David Hess]

Ok, so i finally replaced C940. But it didn't fix it. I did a ripple measurement on the new one and it was high as well. I didn't have the fluke with the min/max function. But I measured peaks of 12V. Did I botch the soldering?

So what parts have been replaced at this point?

Is the oscilloscope still blinking the power light?

[oldway]

Ok, so i finally replaced C940. But it didn't fix it. I did a ripple measurement on the new one and it was high as well. I didn't have the fluke with the min/max function. But I measured peaks of 12V. Did I botch the soldering?

Skickat från min ONEPLUS A3003 via Tapatalk

Again a T906 failure ? 

[mmFooD]

Nope, no flashing on the power led. I've only replaced C940 at this point.

Skickat från min ONEPLUS A3003 via Tapatalk

[mmFooD]

Ok, so i finally replaced C940. But it didn't fix it. I did a ripple measurement on the new one and it was high as well. I didn't have the fluke with the min/max function. But I measured peaks of 12V. Did I botch the soldering?

Skickat från min ONEPLUS A3003 via Tapatalk

Again a T906 failure ? 

You think the problem is with the transformer?

Skickat från min ONEPLUS A3003 via Tapatalk

[oldway]

Ok, so i finally replaced C940. But it didn't fix it. I did a ripple measurement on the new one and it was high as well. I didn't have the fluke with the min/max function. But I measured peaks of 12V. Did I botch the soldering?

Skickat från min ONEPLUS A3003 via Tapatalk

Again a T906 failure ? 

You think the problem is with the transformer?
Skickat från min ONEPLUS A3003 via Tapatalk

Too much ripple on a NEW C940, that's very strange....a faulty T906 could explain this.....To check it, you need an isolation transformer and another scope to check the current waveform on r907. (waveform 39)


A faulty T906 causes all kinds of phenomena:
1) Mosfet Q9070 becomes too hot
2) the operation of the inverter becomes erratic because of the ripple
3) the voltages are no longer properly stabilized
4) there is sometimes a high pitch sound that comes from T906 and which varies with the brightness setting

Another way to diagnose a faulty T906 is to feed the 42V rail by an external power supply....If all is working good with this external power supply and a new C940, you can be sure that T906 is faulty.

That's why I always use first an external power supply for testing and repair this kind of scopes

[mmFooD]

Yes, I am starting to realise that another scope and psu would be handy. I'm thinking I could buy a Rigol DS1074Z and maybe make some PSU. What do you think?

Is it better to make a big 48+ V 1-3 A PSU or a couple of 15 V 1 A PSU to put in series or parallel to get higher voltages or currents?

I no longer have access to the fluke either, so I'm thinking I need a better DMM as well to compliment my cheaper UNI-T. One that could measure at least 10mF, temperature, µA etc. The only one I've found os far with a reasonable price with good specs is the EEVblog BM235.

[David Hess]

Yes, I am starting to realise that another scope and psu would be handy. I'm thinking I could buy a Rigol DS1074Z and maybe make some PSU. What do you think?

Is it better to make a big 48+ V 1-3 A PSU or a couple of 15 V 1 A PSU to put in series or parallel to get higher voltages or currents?

I know of at least two cases now were people replaced the 22xx preregulator circuit with a 48V switching power supply brick adjusted to 42.5 volts.

I do not suggest designing and building an off-line switching power supply.

[velik_kazakov]

Maybe I wast in the row if someone is writing before for this, but have some 2236 (same power supply) repaired in last 6 months (most with problems in PS) and always remove P9070 and solder wire to Q9070. P9070 always can make some problems - I have the same trouble with one 2236 and after removing and solder the wire to Q9070 all was OK.
Tek kit for this PS components replacement is 050-2242-03. Some good information about 22xx power supply repair here. I think that have one more recommendation from Tektronix for this power supply but not sure now.
And something else - +1 install fan. It is too hot w/o fan. The board is the same as 2236 and just you need to solder the fan to the power supply. If you need, I can to make some pictures how it is installed on  2236.

[mmFooD]

Quote from: mmFooD on 2017-08-22, 09:51:46

Yes, I am starting to realise that another scope and psu would be handy. I'm thinking I could buy a Rigol DS1074Z and maybe make some PSU. What do you think?

Is it better to make a big 48+ V 1-3 A PSU or a couple of 15 V 1 A PSU to put in series or parallel to get higher voltages or currents?

I know of at least two cases now were people replaced the 22xx preregulator circuit with a 48V switching power supply brick adjusted to 42.5 volts.

I do not suggest designing and building an off-line switching power supply.

By brick, do you mean I should buy a module and then add/replace some components to get the ajd voltage?


I guess buying two such modules I could use the second one to build a  45V lab PSU unit?

[mmFooD]

Maybe I wast in the row if someone is writing before for this, but have some 2236 (same power supply) repaired in last 6 months (most with problems in PS) and always remove P9070 and solder wire to Q9070. P9070 always can make some problems - I have the same trouble with one 2236 and after removing and solder the wire to Q9070 all was OK.
Tek kit for this PS components replacement is 050-2242-03. Some good information about 22xx power supply repair here. I think that have one more recommendation from Tektronix for this power supply but not sure now.
And something else - +1 install fan. It is too hot w/o fan. The board is the same as 2236 and just you need to solder the fan to the power supply. If you need, I can to make some pictures how it is installed on  2236.

I will definateyl look into that, especially installing the fan. Where did you plug it in exactly?


I also found this in the document you linked:

If there is an audible ticking indicating that the scope is continuously starting and
shutting down there is secondary problem. The most common fault then is one or more
shorted rectifier diode(s) on the secondary windings of the transformer. Also a bad HV
Multiplier or a supply shorted in any other way would produce the power supply burst
mode.

Is that refering to CR960-3 or is it refering to CR901-4? Allthough I have already tested those and was told that the values where fine.

[velik_kazakov]

I will definateyl look into that, especially installing the fan. Where did you plug it in exactly?

Fan is connected to +5v and separate rectifier to the output of the 8.6V traf. It produce ~ -8.6v and you can install 12V fan here but need to solder 3 new components (in yellow in the schematic). In red is marked the plug in the PS board of the schematic and PCB.

[mmFooD]

Quote from: mmFooD on Today at 13:43:55

I will definateyl look into that, especially installing the fan. Where did you plug it in exactly?

Fan is connected to +5v and separate rectifier to the output of the 8.6V traf. It produce ~ -8.6v and you can install 12V fan here but need to solder 3 new components (in yellow in the schematic). In red is marked the plug in the PS board of the schematic and PCB.

Right! But isn't the schematic reversed? It seems to be connected to the +8.6 V rail. Or am I missing something?

[mmFooD]

I do not suggest designing and building an off-line switching power supply.

Out of curiousity, what is the reason for not recommending this? Is it because of EMI etc or is it just to avoid unexperienced people fingering with mains? Or is it just overly complicated / unnecessary compared to just buying a finished AC/DC converter?

[David Hess]

I do not suggest designing and building an off-line switching power supply.

Out of curiosity, what is the reason for not recommending this? Is it because of EMI etc or is it just to avoid inexperienced people fingering with mains? Or is it just overly complicated / unnecessary compared to just buying a finished AC/DC converter?

All of the above?  Plus for 10s of dollars, a suitable embedded off-line switching power supply can be purchased from Mouser or someone.

Designing and building one would be an educational experience for sure but a waste of time for most people and if they lack the proper test equipment, expensive as well.  Miniaturizing it to fit in place of the existing 2235 preregulator would add to the challenge.