Agilent DSO6054A noise on two channels

[dhillman]

I have recently acquired a DSO6054A that was a very early build, yet outside of the first 1000 which had mainboard issues.

What I have noticed is a very small 60Hz ripple on channels 3 and 4.  It can be seen on ranges up to 10mV/div and is roughly 1mVpp in amplitude.     
The amplitude changes correctly through 10mV/div to 2mV/div.
It is possible to see with and without 50ohm termination.

I have checked all shielding around the attenuators, and nothing looks out of place.   It strikes me that the issue may be with the DC supply leading to the ch3,4 ADC (assuming the two ADCs are from isolated sources), or the ADC has issues. 

Have any of you come across this, and if so, do you have any suggestions?

Many thanks.

[tautech]

Welcome to the forum.

As most DSO's use a SMPS for the PSU IMO it is unlikely it originates from the PSU, more likely from your environment and possibly your lighting.
Experiment with a different power source and disabling various appliances that may be providing the source of EMI.

[dhillman]

The issue is presenting on only two of the four channels which isolates the issue to within the scope.
The supply voltage is clean and the environment is clear of EMI.

Internal supply rails are within the 0.1v specification and show no ripple. 

Does anyone know if the ch1,2 and ch3,4 converters are separately fed from their own regulators?

[tautech]

60 Hz ripple can only be from the equipment surroundings.
Do you have the all channels set to the same probe attenuation?

[KE5FX]

60 Hz ripple can only be from the equipment surroundings.

Yes, such as the power-line entry filter and everything else on the line side of the SMPS.

And don't forget the AC pickoff signal that's used for line triggering.   There absolutely is some 60 Hz energy floating around inside the enclosure. 

See all the fancy metalwork in the Rigol SA that Shahriar tore down?  https://youtu.be/0wVlH1lPZdE?t=146  From the point of view of a 60 Hz H-field, it might as well not even be there.  (Thanks to the lack of any actual line filtering, the same is true of any E-field RF that comes surfing in on the power cord.)

Do you have the all channels set to the same probe attenuation?

Also, any chance of posting a screenshot?

[dhillman]

Certainly a valid point on the 60Hz pick-off point.   It is on the filter board prior to the switchmode, and drives an optoisolator prior to meeting up with the main board.   I have disconnected this without change.

To the other point, I am testing without probes.  Inputs are set identically.   50ohm external attenuation is applied.

You will see on the attached image a visual of the ripple on the two channels.   Yes, this is less than most would even care about, but in my opinion, it should not be there.

Also noted on the front end some rework that happened presumably at the time of build.  Ch1,2, and 4 CMOS switches were replaced.   Ch 3 is presumably original and has a date code that matches most other chips 0443.   The replaced chips are 0414, and hand soldered well enough.

Also of note is that this scope passed Agilent calibration in 2013.

[tautech]

Does a Self Cal or Autoset change anything?

[KE5FX]

You will see on the attached image a visual of the ripple on the two channels.   Yes, this is less than most would even care about, but in my opinion, it should not be there.

Pretty similar here, at least on channel 3:



I wouldn't lose much sleep over it, as long as it's in spec.  But if you find the cause, I'd be curious to hear it!

[Carrington]

Hi dhillman.

It is clear that your oscilloscope was repaired at some point.
Have you tried to make a "USER CAL"? You need to make a cable (see attachments).

Under normal mode: See attachments, I get something pretty similar, and here the frequency is 50Hz. I think that everything is OK with your oscilloscope.
Settings: Trigger: AUTO, 2mv/DIV, 10ms/DIV, DC coupling and 50OHM on all channel.

By the way, I wonder if you've seen this topic before:
https://www.eevblog.com/forum/testgear/agilent-mso-500060007000-anyone-hacked-these-scopes/

Another thing, please, could you take better pictures of the front end (both sides)?

Thanks and best regards,
Carrington.

PS: Edited to correct multiple errors, sorry.

[mikeselectricstuff]

Have you checked for ripple out of the main PSU? - If so it could be an aging primary-side electrolytic

[dhillman]

I haven't figured out how to use quotes, so here goes the old way.

KE5FX:  I'm glad (slightly relieved) to see that this issues is not specific to my scope.   

Tautech:  Yes, the unit does self calibrate without issue. 

Carrington:   Oh yes, I have looked at the other forum post.   It seems that I have the missing link (the 500MHz unit).  Attached are some closeups.

Mike:   There is no appreciable ripple on the 15VDC power supply.   

Thank you all for your input.   If I get to the bottom of the minor ripple, I will update the post.

[Carrington]

Thank you very much, interesting pictures. 

If you haven't already assembled it, please, could you take a few more detail pictures from the other side of the input stage, i.e the other side of the board.
Could you also measure the value of some resistors (green and black), without marking ? Without desoldering, is not necesary, on the board.

They are small resistors, so you need to take a good photograph to see its value. Please, look in attachments, and you will understand what I mean.

Thanks,
Carrington.

[nctnico]

Those green components with white stripes are most likely inductors.

[Carrington]

Those green components with white stripes are most likely inductors.

Probably, however never ever I will ask him about remove any of these to measure its value, it would be like an abuse.
He has been kind enough to take photographs, and that is enough. I will not ask for more. 

[eeadata]

Certainly a valid point on the 60Hz pick-off point.   It is on the filter board prior to the switchmode, and drives an optoisolator prior to meeting up with the main board.   I have disconnected this without change.

To the other point, I am testing without probes.  Inputs are set identically.   50ohm external attenuation is applied.

You will see on the attached image a visual of the ripple on the two channels.   Yes, this is less than most would even care about, but in my opinion, it should not be there.

Also noted on the front end some rework that happened presumably at the time of build.  Ch1,2, and 4 CMOS switches were replaced.   Ch 3 is presumably original and has a date code that matches most other chips 0443.   The replaced chips are 0414, and hand soldered well enough.

Also of note is that this scope passed Agilent calibration in 2013.

Clean the flux residue ? Maybe it helps ...

[dhillman]

Clean the flux residue ? Maybe it helps ...

I cleaned the front end of excess flux residue.   The ripple is still present.  It seems obvious that somewhere along the line the ADG451 switches were considered worthy of replacement.  I may place a replacement on channel 3 to satisfy my curiosity.   

Those green components with white stripes are most likely inductors.

Indeed, they are inductors.   As Carrington suggested, however, I will not go so far as to remove them.     I will have the unit open in the coming weeks to replace the fan and a few other items, so I'll probe them with the LCR and see if they are isolated enough to give a valid result.

If you haven't already assembled it, please, could you take a few more detail pictures from the other side of the input stage, i.e the other side of the board.
Could you also measure the value of some resistors (green and black), without marking ? Without desoldering, is not necesary, on the board.

They are small resistors, so you need to take a good photograph to see its value. Please, look in attachments, and you will understand what I mean.

Absolutely!  Stay tuned.

[eeadata]

Hunt for non rectefied direct ac operational devices near channel 3-4 . and check "line triggering facility" (Maybe a wire wrapped on ac line etc. ) of scope .

[AndyC_772]

Have you confirmed that the ripple you see actually is synchronised to the mains, and not just coincidentally more-or-less the same frequency? If it's not stationary when you set the machine to ac line trigger, then it's coming from another source.

[Carrington]

Indeed, they are inductors.   As Carrington suggested, however, I will not go so far as to remove them.     I will have the unit open in the coming weeks to replace the fan and a few other items, so I'll probe them with the LCR and see if they are isolated enough to give a valid result.

Absolutely!  Stay tuned.

Perfect thanks!

Time ago I look for a quieter fan replacement with the same flow rate/pressure curve, but I could not find any.
The factory default fan is the quietest on the market.

http://www.ncix.com/detail/nmb-mat-panaflo-92mm-high-speed-b7-13925.htm

Datasheets:
http://pdf1.alldatasheet.com/datasheet-pdf/view/167145/PANASONIC/FBA09A12H.html
http://datasheet.octopart.com/FBA09A12M1A-Panasonic-datasheet-40461.pdf


Some specialized resistors are green too:

http://www.koaspeer.com/products/resistors/surface-mount-resistors/rn73/
http://www.koaspeer.com/products/resistors/surface-mount-resistors/rk73z/

[dhillman]

Some specialized resistors are green too:

 These particular green components effectively have zero resistance, so I'd assume they are inductors. 

[Carrington]

These particular green components effectively have zero resistance, so I'd assume they are inductors.

Well, it is very probable, that seems to confirm it.


ORIGINAL FAN:

[Carrington]

These particular green components effectively have zero resistance, so I'd assume they are inductors.

Confirmed, they are inductors:



http://www.leocom.kr/MCFT000018/MULTICOMP

Datasheet: http://data.leocom.kr/datasheets/316338_360095.pdf

[dhillman]

A few pictures and some measurements of the front end on a 500MHz unit.

The inductances are too low for the meter to measure.   

[Carrington]

A few pictures and some measurements of the front end on a 500MHz unit.

Awersome.
Thank you very much! 

The inductances are too low for the meter to measure.

Yes, but at 100 kHz these inductors can be measured without problems.


I also made some measurements in the front end of an MSO6104, I attached the results, below.
I used a DER EE DE-5000, calibrated according to the different sizes (603, 1206), to compensate the deviation introduced by the SMD test clips.
The parts marked in yellow can be measured on the board, without remove them. The other components must be removed from the board to get its real value.

Note 1: The first network performs filtering and attenuation to the ranges below than 500mV (IN_1, OUT_1). It's awesome, Agilent did a good job with these machines.
Note 2: According to the inductor's datasheet it's DC resistance increases proportionally to its inductance, but this is not met by the 100nH inductor, I don't know why.

   For 33nH -> DCR =  1.35 OHM.
   For 47nH -> DCR = 1.71 OHM.
   For 100nH -> DCR = 0.7 OHM. 

In a few days I will try to do the same with my MSO6034A.

Cheers!

[Carrington]

Today I repeated (several times) all the measurements, especially for the inductance with a DCR of 0.71OHM.
My LCR almost every time read around 100nH, but occasionally it read value between 10nH and 15nH, it certainly has to do with the SMD tweezers.