Keysight MSO7034B, Teardown and Repair [Success]

[HighVoltage]

I bought this Agilent / Keysight MSO7034B scope a while back in broken state really cheap.
Symptoms: Does not turn on at all.

So, I thought to make a teardown and repair attempt at the same time.
I never opened one of these scopes before and I have not seen a tear down before of a 7000B series scope.
To take the back cover off, remove the obvious screws, the handle and the feet.

So far I took the back cover off and found a lot of dirt on the fan and in the housing.
May be this was the killer for the PSU, although it looks pretty clean.

Here are my first pictures

Interestingly this PSU has lots of missing components.
Usually when I have repaired some Agilent gear before, the PSU was fully populated.
 

[joellinn]

Maybe someone salvaged it 
Did you try and measure voltages on the Header that goes into the scope? Are the fuses OK? What about the transformer?

[nctnico]

It seems some outputs are not used so that would explain the 'missing' components. I'd start by checking the PSU's output voltages and go from there. Don't assume something is broken; always measure to determine something needs replacing or not.

[HighVoltage]

Thanks, yes, will test the PSU next

I took the main board out first to clean it all of the dust.
Here are some pictures of the main board.

[HighVoltage]

And here are some pictures of the front end under the metal can.
Anything else I should take close up pictures of?

[SeanB]

Looks like a standard PSU board that produces multiple rails, and they only need 2 rails here ( likely 5V and 12V) so only populated those parts on the secondary side. My bets are the crinkled blue mains input cap upper left of the PSU by the mains input sent off the ceramic mains fuse as it failed short circuit. If not then suspect any high value resistors on the primary side, especially if the main transistor switch is not blown apart.

[HighVoltage]

The fuse is OK but the PSU is dead.
So, probably a resistor failure in the PSU?
I will continue the trouble shooting tomorrow...
The schematics are probably not available for the PSU ?


Just for completeness, here are some more pictures of the tear down.

[nctnico]

The fuse is OK but the PSU is dead.
So, probably a resistor failure in the PSU?

Don't guess but measure. First check all semiconductors. If those are OK apply power using a DC power supply to the low voltage part of the primary circuit (look up the datasheet for the required voltage) so you can check if the switching chip is doing something and the primary side MOSFET is switched correctly. Using a lab power supply makes this measurement safe to do. The next step is to apply power to the primary side capacitors using a different low voltage supply (30V to 40V and a low current limit setting). In this step you'll see some things happening on the secondary side as well.

Edit: this seems like some kind of self oscillating resonant converter with a current feedback from the secondary side. Applying DC at the input should give some insight in where the path ends.

[tautech]

Before you put a lot of time and effort into this is it not worth providing the voltages necessary from an external PSU to check the rest of the scope is functioning. 

[tautech]

Is that a dry joint on the inline fuse?

[CJay]

There is absolutely nothing special about that supply, it's an off the shelf Artesyn device and is replaceable for a few tens of dollars.

If you can't find the fault in a couple of hours, just buy a replacement or equivalent.

Totally unworth the effort to repair unless you are doing it for the challenge.

https://www.artesyn.com/power/assets/nfs110_ds_1213229025.pdf

15V 110W, single rail supply.

[tautech]

There is absolutely nothing special about that supply, it's an off the shelf Artesyn device and is replaceable for a few tens of dollars.

If you can't find the fault in a couple of hours, just buy a replacement or equivalent.

Totally unworth the effort to repair unless you are doing it for the challenge.

https://www.artesyn.com/power/assets/nfs110_ds_1213229025.pdf

15V 110W, single rail supply.

MTBF 125,000 hours min. 
Yeah right. 

[gamalot]

There is absolutely nothing special about that supply, it's an off the shelf Artesyn device and is replaceable for a few tens of dollars.

If you can't find the fault in a couple of hours, just buy a replacement or equivalent.

Totally unworth the effort to repair unless you are doing it for the challenge.

https://www.artesyn.com/power/assets/nfs110_ds_1213229025.pdf

15V 110W, single rail supply.

Totally agree! I‘ve bought a Meanwell 12V PSU to replace the TDK-Lambda one in my Keithley 2306.

[nctnico]

MTBF 125,000 hours min. 
Yeah right. 

Probably at 25deg Celcius; at higher temperatures the MTBF drops considerably and even then it is just a theoretical number.
It is quite interesting to find an off-the-shelve power supply in this kind of high end scope though.

[mikeselectricstuff]

By far the most common failure in SMPSs is an electrolytic cap on the startup supply - look for a low-voltage cap on the primary side and replace it, regardless of what it might measure. 
Remember the main primary cap will stay charged quite a while on a PSU in this state so take care!

[HighVoltage]

It is quite interesting to find an off-the-shelve power supply in this kind of high end scope though.

Yes, this is a big surprise to me as well.
Although I think Keysight has stopped production, it is still available for sale in Germany for Euro 17.396 + VAT = Euro 20.701 (and that is without options)
I would have expected a high end power supply.

Thanks for all your suggestions, I think I will find the problem tomorrow.

[tautech]

As Dave would say, you'll be on a winner, winner chicken dinner if it's an easy fix.
Hope it is, good luck.

[wraper]

Before you put a lot of time and effort into this is it not worth providing the voltages necessary from an external PSU to check the rest of the scope is functioning. 

Very bad advise. Never do that unless you have a very good reason and triple check that you are doing this right. Little mistake and all the expensive hardware will be dead beyond repair.

[nctnico]

Before you put a lot of time and effort into this is it not worth providing the voltages necessary from an external PSU to check the rest of the scope is functioning. 

Very bad advise. Never do that unless you have a very good reason and triple check that you are doing this right. Little mistake and all the expensive hardware will be dead beyond repair.

I have to agree. I'm also reluctant to do this because it is easy to swap + with - and there usually isn't any reverse polarity protection on the boards. Secondly there can be requirements for rise-time, power good signals, etc.

[tautech]

Before you put a lot of time and effort into this is it not worth providing the voltages necessary from an external PSU to check the rest of the scope is functioning. 

Very bad advise. Never do that unless you have a very good reason and triple check that you are doing this right. Little mistake and all the expensive hardware will be dead beyond repair.

I have to agree. I'm also reluctant to do this because it is easy to swap + with - and there usually isn't any reverse polarity protection on the boards. Secondly there can be requirements for rise-time, power good signals, etc.

Oh really guys are we all novices or something.
Oh wait isn't that what current adjustable bench PSU's are for?

[tautech]

Manuals for those following along at home.
http://www.docs-library.com/pdf/1/6/agilent-7012.html#
Hit the "Go to link" and another tab with the pdf will open and download.

PSU checks are described on P77

Basic as these PSU's are, anybody that has fixed scopes will know there'll be a Line frequency feed to the Mobo, possibly a mains gnd too and if it's a single rail supply there'll be multiple conductors to supply the current, hence the additional pins on the connector.

[CJay]

MTBF 125,000 hours min. 
Yeah right. 

Probably at 25deg Celcius; at higher temperatures the MTBF drops considerably and even then it is just a theoretical number.
It is quite interesting to find an off-the-shelve power supply in this kind of high end scope though.

~14 years MTBF continuous use, it'll be extrapolated from a series of 'torture' tests, high temperature, load etc. (Artesyn may well release data  that shows how they derate MTBF for load and temperature and that will give a cllue as to how they test) but I do have PSUs that are considerably older and have been in use for at least ten years (on equipment like gas analysers) so for a quality part like an artesyn, I'd expect it to last a good number of years

It is a bit of a surprise to find an off the shelf part but the secret sauce on a lot of these 'scopes is the acquisition so it makes sense to have good quality off the shelf bits chosen for specification.

[SeanB]

Check R1, change the power transistor with the same type, change the low value resistor in the emitter, change all the little jellybean transistors and check all resistors are not open and it will work, providing the secondary side diodes or capacitors are not shorted. Also check the crowbar thyristor CR12 is not shorted and stopping the power supply.

[nctnico]

Before you put a lot of time and effort into this is it not worth providing the voltages necessary from an external PSU to check the rest of the scope is functioning. 

Very bad advise. Never do that unless you have a very good reason and triple check that you are doing this right. Little mistake and all the expensive hardware will be dead beyond repair.

I have to agree. I'm also reluctant to do this because it is easy to swap + with - and there usually isn't any reverse polarity protection on the boards. Secondly there can be requirements for rise-time, power good signals, etc.

Oh really guys are we all novices or something.

No, wisdom gained by experience 

Oh wait isn't that what current adjustable bench PSU's are for?

At 110W you'll need a chunky bench PSU and also need to worry about power supply risetime and switch-on overshoots if the PSU is crappy enough.

BTW: it seems line triggering isn't done through the PSU but there may be a small current transformer on the mains entry board.

[HighVoltage]

I think I found the culprit

The large high voltage DC capacitor was bad.
This was a Nippon Chemi-Con 330uF 400V
But I only have a Rubicon 220uF 400V at hand.

Installed the Rubicon and the power supply works.
This power supply only has a single 15 V DC output.

I will hook up a DC load and look at the ripple on the output.

What do you all think.
Should I get a new 330uF cap or should this 220uF cap do the job?

[acbern]

Since this is probably a wide input range PSU (please check) the capacitor was selected for the lowest possible voltage (permissible input ripple voltage after rectifier). Since you are using it at 230V, it should work. I would still change it.

[HighVoltage]

Since this is probably a wide input range PSU (please check) the capacitor was selected for the lowest possible voltage (permissible input ripple voltage after rectifier). Since you are using it at 230V, it should work. I would still change it.

Hello acbern
yes, it is 100 to 240 V AC PSU
I pulled 2A out of the PSU and it seems tob e working perfectly

Installed the PSU and the scope works !
Look at this ... what a great Sunday it is !
What an easy fix that was.
Thank you all !

I will measure now the current the scope is pulling from the PSU and the ripple voltage.

[bingo600]

Great job 

Congrats with your new scope

/Bingo

[TheSteve]

Awesome! This is one of those deals and repairs we all dream of!

[HighVoltage]

Thanks,

I measured the current from the PSU to the scope.
Here are the results:
I am surprised that it does take 2.7A from the PSU.
The ripple voltage is about 10 mV AC (measured with 87V)

[nctnico]

While you still have it open: can you identify where the scope get it's signal to trigger on the AC mains (line trigger source)?

It's a great deal though! Now hack it into an MSO!

[mikeselectricstuff]

I'd definitely get the correct cap to make sure everything is working within spec. Hardly a major cost compared to the value of the scope.

[HighVoltage]

While you still have it open: can you identify where the scope get it's signal to trigger on the AC mains (line trigger source)?

Hmm, there is no other connection to the main board from AC, only the PSU +15V DC
May be the 50 /60 Hz is created on the main board, independent of the real AC ?

It's a great deal though! Now hack it into an MSO!

Yes, the hack will be next

I'd definitely get the correct cap to make sure everything is working within spec. Hardly a major cost compared to the value of the scope.

Yes, I already found an ordered the 330uF cap.
I will take some more detailed measurements, before and after changing the the cap.
Although I dont think it will make a big difference, you are correct to keep it in specs.

I just measured the AC ripple on the output of the +15V DC and it is around 2 mV AC on the 87V Fluke Multimeter

[tautech]

I measured the current from the PSU to the scope.
Here are the results:
I am surprised that it does take 2.7A from the PSU.
The ripple voltage is about 10 mV AC (measured with 87V)

Ha, who would have expected the HV DC cap as the point of failure, not common for a SMPS. 
So 2.7A (40W) is roughly what you'd expect and could have easily been supplied by a bench PSU for checks too.

Nice find HV and congrats on your cheap DSO. 

[nctnico]

To me it seems every repair technician has his (her) own list of common problems in switching power supplies but none of them apply to every SMPS out there.

[mikeselectricstuff]

This is how the MSO6000 does it - I imagine the 7000 will be similar
https://youtu.be/5lYbD9_eIko?t=366

[nctnico]

Thanks for pointing to that video! Judging from the photos it wouldn't surprise me if the power entry board in the 7000 series is the same just like the power supply.

[tautech]

This is how the MSO6000 does it - I imagine the 7000 will be similar
https://youtu.be/5lYbD9_eIko?t=366

Yep, line frequency reference from another PCB, but it's often done directly from the PSU itself however when a cheap generic SMPS is used for these Keysights you have to provide another Line source.

Does a single PCB with Line output configuration make for a more reliable setup or using 2 PCB's ?   

[HighVoltage]

Thanks Mike for this explanation.
I will have another look and make some detailed pictures, it looks it is the same setup in my 7034B

[CJay]

To me it seems every repair technician has his (her) own list of common problems in switching power supplies but none of them apply to every SMPS out there.

We do, but there's an awful lot of crossover and common faults, I think high value resistors on the primary, low value caps on the primary, leaky high value caps on the secondary and shorted secondary rectifiers would be the very first thing pretty much any tech who repairs these things will check for.

Excellent news from the OP BTW, congratulations on such an awesome score, envious doesn't really cover it.. 

[tautech]

To me it seems every repair technician has his (her) own list of common problems in switching power supplies but none of them apply to every SMPS out there.

We do, but there's an awful lot of crossover and common faults, I think high value resistors on the primary, low value caps on the primary, leaky high value caps on the secondary and shorted secondary rectifiers would be the very first thing pretty much any tech who repairs these things will check for.

And the HV DC primary side cap is low on the list ^^, right? 

[mikeselectricstuff]

To me it seems every repair technician has his (her) own list of common problems in switching power supplies but none of them apply to every SMPS out there.

We do, but there's an awful lot of crossover and common faults, I think high value resistors on the primary, low value caps on the primary, leaky high value caps on the secondary and shorted secondary rectifiers would be the very first thing pretty much any tech who repairs these things will check for.

And the HV DC primary side cap is low on the list ^^, right? 

Don't think I've ever seen it, but maybe because manufacturers realise it's a critical part and take more selecting it, as opposed to things like standby supply caps and resistors  where a jellybean part is used despite it frequently be being exposed to significant heat or voltage stress respectively.   

[CJay]

To me it seems every repair technician has his (her) own list of common problems in switching power supplies but none of them apply to every SMPS out there.

We do, but there's an awful lot of crossover and common faults, I think high value resistors on the primary, low value caps on the primary, leaky high value caps on the secondary and shorted secondary rectifiers would be the very first thing pretty much any tech who repairs these things will check for.

And the HV DC primary side cap is low on the list ^^, right? 

It really isn't a common fault in my experience, in the ten years or so I repaired PSUs as part of my job I can't remember replacing more than a handful of them that had failed in anything other than spectacular fashion

In my telly repair career, it was a different story, the multi value, single can capacitors were a bugger for failing in some sets (G11, duff rivets).

For PSU repair work I found the component tester on the Hameg 'scopes makes the 'gross' fault finding really quick (we could turn around ~40 PSUs a day per tech if we got a large batch of the same types) so the primary reservoir would be picked up really quickly if it were faulty while testing the bridge/diodes and, depending on topology, the switching transistor(s) because it would look 'wrong'.

[wraper]

Ha, who would have expected the HV DC cap as the point of failure, not common for a SMPS. 

Not something PSUs are plagued with but is not something rare either.

[HighVoltage]

For PSU repair work I found the component tester on the Hameg 'scopes makes the 'gross' fault finding really quick

This is still a feature on the small Rohde & Schwarz scopes.
I never thought about using it "in circuit", interesting idea.

[CJay]

For PSU repair work I found the component tester on the Hameg 'scopes makes the 'gross' fault finding really quick

This is still a feature on the small Rohde & Schwarz scopes.
I never thought about using it "in circuit", interesting idea.

I find it useful for things like power supplies, amplifiers, analogue power things, but I know other people aren't keen.

I guess it's just personal taste and it really depends how often you repair things like that.

It's useful for gross faults in circuit (shorted or open junctions, open circuit resistors, leaky, short or open  caps), I can 'sweep' a board in a few seconds for those with a component tester and desolder parts that look suspicious to get in deeper with a meter or component analyser and confirm or disprove suspicious traces.

They're not so great on 'subtle' things like high value resistors that have gone higher or nF caps but combined with a multimeter and some experience, they're really effective.

[HighVoltage]

While you still have it open: can you identify where the scope get it's signal to trigger on the AC mains (line trigger source)?

Ok, here are some pictures of the AC board and the connection to the mains AC trigger.
It looks like an optocoupler in between the AC and the wire to the main board.

Even this little PCB is made by Agilent and looks great and much better than the PSC PCB.

 

[HighVoltage]

Now hack it into an MSO!

And here are two pictures of the screen, before and after the hack.

Thanks to some forum members for providing the 6.16 FW version. (No longer on the Keysight website available)
The Hack is very simple, as long as the Firmware is 06.16.0001.

[HighVoltage]

Once the Service Menu is turned ON via telnet access, this new Menu is available under "License" in the "Utility" section.

[HighVoltage]

I'd definitely get the correct cap to make sure everything is working within spec. Hardly a major cost compared to the value of the scope.

OK, to conclude this thread, I just installed the new 400V 330uF capacitor and all works great.

Interestingly, the new cap is much larger

Original Nippon Chemi-Con
- 400V 330uF
- 25 mm diameter
- 40 mm tall
- 29.8 Grams

New VishayBC
- 400V 330uF
- 35 mm diameter
- 50.7 mm tall
- 64 Grams
This new cap has an ESR of 0.07 Ohm
But only measures 298 uF
May be this one is better, because it has a larger volume.
But who knows, I have never used a Vishay electrolytic capacitor before.

Thank you all for help and suggestions, it is really appreciated.

[nctnico]

Hmm, normally replacement capacitors are smaller, not bigger!

[Samogon]

This new cap has an ESR of 0.07 Ohm
But only measures 298 uF
May be this one is better, because it has a larger volume.
But who knows, I have never used a Vishay electrolytic capacitor before.

Thank you all for help and suggestions, it is really appreciated.

It is still in 20% tolerance.
Again nice deal you made with this scope. Congratulations!

[Greybeard]

Can somebody help me to identify this diode in the front end of 7034A?
The font ends of 6034A and 5034A seem to be nearly identical to 7034A.

[mikeselectricstuff]

Can somebody help me to identify this diode in the front end of 7034A?
Is the font end of 7034A identical to 6034A and 5034A?

Unlikely to be anything exotic  - just measure the one in another channel  to determine how many diodes, shottky or not, zener or not.

[Greybeard]

Unlikely to be anything exotic  - just measure the one in another channel  to determine how many diodes, shottky or not, zener or not.

I don't like the idea of unsoldering one of the working ones, that would be the last resort.

[mikeselectricstuff]

Unlikely to be anything exotic  - just measure the one in another channel  to determine how many diodes, shottky or not, zener or not.

I don't like the idea of unsoldering one of the working ones, that would be the last resort.

At least you could try measuring it in-circuit

[HighVoltage]

I still have the original high resolution pictures and cropped this area out again
But it is still not readable.

I suggest like Mike said, try to measure the part in circuit and compare it to the other channels.
 

[mikeselectricstuff]

Would also be worth looking to see if there are parts with the same marking elsewhere ( apart from other channels) - if there are this would indicate that it's probably a fairly general-purpose device rather than something exotic

[Greybeard]

HV: I tried to process your photos, but no success.

Mike: BTW, awsome homepage (tesla photos).

[Greybeard]

Found a promising photo of the 1GHz frontend in this thread:
https://www.eevblog.com/forum/repair/agilent-dso6054a-noise-on-two-channels/msg853423/#msg853423




I think it is very probable same circuitry regarding this part:



So CR202/302/402/502 should be a double PIN diode HSMP3892 (HP/Agilent/Broadcom, marking code G2) = LXP1008‐23‐2 (Microsemi drop in equivilent).

https://www.digchip.com/datasheets/download_datasheet.php?id=384631&part-number=HSMP-3892
https://www.microsemi.com/document-portal/doc_download/9457-lxp1000-sot23-series-datasheet

PS:
Confirmed at 300MHz scope: Marking G25
Confirmed at 500MHz scope: Marking G2F
5 or F may be a maker or index mark.



PPS:
Case seems to be SOT-232 (smaller than SOT-23):




So correct diode type should be HSMP389C (same G2 SMD code):