Agilent/HP 54825A - CH1 Offset problems - Help!

[oriolsancheza]

Hello guys!

I've bought recently a pair of 54824N(A?) Scopes for a decent price (About $700 for both, neither good nor bad). They are quite old and slow but for us is more than enough.

Both passed self-test without errors, and both have trigger fails in sel-cal. I guess maybe re-installing the firmware will fix that.

Well, the problem is that one of them has some offset on CH1, it's supposed to be centered on 0v but it's about 2v higher. Despite that Triggering system is working fine in auto, trigged and single.

Obviously I'd suspected some problem related with triggering system for CH1, and yes, that what it is! There is an output comparator (MC10E1652) that is receiveing? bad voltages on it gates, compared to the other channels.

The thing is I'm not sure from where these voltage come, as I'm not able to find the schematic. I just was wondering if some of you had seen this before and knows how to fix that, because the comparator looks to work perfectly. I spent two days on the scope and I don't wanna spend too much on it as other channels are working perfectly.







I also tried to swap attenuattors and Hybrid DACs with same result.

By the way, if some of you has the 4.5 firmware version, please PM me! Looks like the file is not anymore on Keysight FTP.

Thank you guys!
Oriol.

[oriolsancheza]

Nobody has repaired one of these?!

[MarkL]

Somewhere there's a DAC that generates the trigger level and offset for each channel.  It's probably one DAC for all the channels.  Since you've already tried swapping the attenuator and ADC, I would start with the DAC to make sure it's working properly.

Ch2, Ch3, Ch4 are working, so pick one as your reference.  You can then make voltage level comparisons on various pins in the analog sections between the reference channel and Ch1 as you move the traces up and down with the offset controls.

If you provided a higher resolution photo of the area above your posted photo I might be able to point out the DAC.  I found a low resolution photo somewhere else, and it might be the 24-pin SOIC near the top of the board.  If it has an HP part number on it, you can try looking it up on Keysight's "Find-A-Part" page to see if they identify it as a DAC.

There are no schematics.  Just block level diagrams in the service guide.  And the offset issue may be causing the self-cal failure.

[MarkL]

Nobody has repaired one of these?!

I haven't, but I've repaired several 16534A scope cards which use the same ADC and attenuator modules.  The architecture is almost identical.

I've seen a similar issue before with strange offsets.  In my case it was with the trigger level because one of the opamps being driven by the DAC wasn't biased properly because one of the biasing resistors went open (for no particular reason).

Comparison with a working channel (or working board, depending on the problem) is the way to go.


EDIT: Actually, now that I think about it, I had *two* boards where this happened, and it was the exact same resistor.  No physical damage.  Strange.

[MarkL]

On the version 4.5 question, there was a posting in the HP_Agilent_Equipment Yahoo group from someone who archived a copy.  The posting was May 8, "Infiniium 54846B disk has died".

I don't feel it would be proper for me to publicly re-post an article from a closed group.  Just join the group and look for it.

[oriolsancheza]

Heelo Mark!

Thank you so much for your responses! You were right!! There is a 24 soic IC with HP part number, but I'm not able to find it.

In any case I have seen there are some LT1112 over there and one of them has the CH1 trigger voltages I think that creates wrong offset values. In any case the HP IC looks like has the correct values, so maybe is the same you pointed about open resistor? I checked ones closer to this IC but looks like everything is okay!

So I attach you some pictures, may you can point me better in order to see If I have the same problem!






Thank you very much!
Oriol.

[MarkL]

The photos were very helpful.  It looks like I was wrong about the 24-pin SOIC (part #1821-1690) being the DAC.  The DAC is the AD7846.

The AD7846 is a single output DAC, but we know we need a bunch more outputs than that.

To get multiple outputs, I believe the 1821-1690 is an analog MUX with 12 or maybe 16 outputs.  On each output of the MUX you can see a 100k resistor going to a holding capacitor.  The should sequence through all the MUX outputs while setting the DAC to the value it wants for each output.  It's a multi-channel sample & hold.

Most, if not all, of the outputs go to the dual LT1112 opamps.  Looking the symmetry of the connections and components around the opamps, it appears there are 3 distinct groupings of opamp pairs.  Since it's a dual opamp, each pair of packages would be involved with all four channels for offset or trigger level (or possibly something else).  I have outlined them in the attached photo below.  All this is my first guess, subject to revision as we poke around.

I would examine the output of each opamp group (2 packages = 4 outputs) to see what function they perform and whether you can find one output in the group that's not consistent with the others.

You mention you already found one that's not consistent.  Can you be more specific?  What are the voltage levels on chips/pins?  Also, before measuring stuff, if there's a way to do it, you should set all the calibration constants to their default value.

There are 3 LT1112 that don't seem to have symmetric circuitry around them.  They may be used for one-off functions such as the calibrator output, external trigger level, etc.

Perhaps I'm not understanding the problem completely, but are you saying the trigger level is offset too much, or the trace is offset too much?  Are you able to control them?

[oriolsancheza]

The photos were very helpful.  It looks like I was wrong about the 24-pin SOIC (part #1821-1690) being the DAC.  The DAC is the AD7846.

The AD7846 is a single output DAC, but we know we need a bunch more outputs than that.

To get multiple outputs, I believe the 1821-1690 is an analog MUX with 12 or maybe 16 outputs.  On each output of the MUX you can see a 100k resistor going to a holding capacitor.  The should sequence through all the MUX outputs while setting the DAC to the value it wants for each output.  It's a multi-channel sample & hold.

You're welcome! And thanks for your quickly response!

So yes you're right again! But i've double checked the DAC Output, and as you said as it's shared, if it was broken then all channels are suposed to get wrong values, isn't it?

Most, if not all, of the outputs go to the dual LT1112 opamps.  Looking the symmetry of the connections and components around the opamps, it appears there are 3 distinct groupings of opamp pairs.  Since it's a dual opamp, each pair of packages would be involved with all four channels for offset or trigger level (or possibly something else).  I have outlined them in the attached photo below.  All this is my first guess, subject to revision as we poke around.

I would examine the output of each opamp group (2 packages = 4 outputs) to see what function they perform and whether you can find one output in the group that's not consistent with the others.

You mention you already found one that's not consistent.  Can you be more specific?  What are the voltage levels on chips/pins?  Also, before measuring stuff, if there's a way to do it, you should set all the calibration constants to their default value.

Group A.Left controls CH1 and CH2 offsets, and Group A.Right controls CH3 and CH4 offsets. I think all LT's are in a good working condition, all inputs have same values except for CH1, so giving a wrong output value.

The inputs 2 and 3 of "groupA.Left" are receiving 0.966, while all other inputs from CH2, 3 and 4 are on 1.045 aprox. So CH1 trace is not on 0v offset level as result.

I don't either think the Mux is the problem.. So I don't know at this point how to continue with that! That's weird.

There are 3 LT1112 that don't seem to have symmetric circuitry around them.  They may be used for one-off functions such as the calibrator output, external trigger level, etc.

Probably, I will check them too, but I'm sure they're not affecting to offset problem.

Perhaps I'm not understanding the problem completely, but are you saying the trigger level is offset too much, or the trace is offset too much?  Are you able to control them?

Trigger levels are correct I think, at least with all the tests I did. The problem is the trace as you said! But in any case I'm able to move it, there's no control problem, and if I put the trace on 0v offset point then I get the correct values! But obviously offset is -250mV.

I also reinstalled the firmware twice, because maybe was a software failure, but still the same! The only thing I can do maybe is to swap the board with the other working one.. to ensure that the problem is located on that board.


Thank you so much Mark!

Best,
Oriol.

[MarkL]

But i've double checked the DAC Output, and as you said as it's shared, if it was broken then all channels are suposed to get wrong values, isn't it?

Yes, not much would be working if the DAC was dead.  I was just describing how the DAC and MUX works in case it was useful to you.

Group A.Left controls CH1 and CH2 offsets, and Group A.Right controls CH3 and CH4 offsets. I think all LT's are in a good working condition, all inputs have same values except for CH1, so giving a wrong output value.

The inputs 2 and 3 of "groupA.Left" are receiving 0.966, while all other inputs from CH2, 3 and 4 are on 1.045 aprox. So CH1 trace is not on 0v offset level as result.

It would be helpful to know the outputs of the opamps too.

That's not much difference on the input, but let's chase that for a bit and see where it leads.  See below.

There are 3 LT1112 that don't seem to have symmetric circuitry around them.  They may be used for one-off functions such as the calibrator output, external trigger level, etc.

Probably, I will check them too, but I'm sure they're not affecting to offset problem.

I don't think they're involved in the problem either.  I was only pointing out that they lack duplicated circuitry, so they are probably NOT involved in functions/problems specific to only one channel.

I also reinstalled the firmware twice, because maybe was a software failure, but still the same! The only thing I can do maybe is to swap the board with the other working one.. to ensure that the problem is located on that board.

I very much doubt it's the firmware; I wouldn't waste any time on that.

I don't think swapping boards is going to help get us closer to the problem.  I think you'll find the problem follows the acquisition board.


I tried to derive a partial schematic from the photos you sent, below.  This is at least a 4-layer board, so it's impossible to tell where some connections go and I can almost guarantee I've made some mistakes.  You should use a continuity check to verify everything, including the ground connections.  (I drew it in qucs if you want to make any corrections to the source.)

I've included some suggested probe points to go with the schematic for Ch1 and Ch3.  I picked Ch1 and Ch3 since the surrounding components and probe points are in roughly the same position in relation to their respective opamps.  Again, you should verify all of this since you have the board in front of you.

[oriolsancheza]

Thank you very much for your Help Mark!! I really appreciate it so much!

So alright.. I did all the tests on test point you have marked. There is a clear problem over there with the votlages.

I decided to swap twice the LT1112 with the other ones, and same result. Also checked all resistors and capacitors, all ok. Then I decided to go further and check voltages without the opamp soldered.

And voilà! Everything looks ok! So... I think it's related with the feedback part on the opamp? But as resistors and capacitors are ok, my only guess is like some via on internal layers is burned?

Just see:





Thanks!
Oriol.

[oriolsancheza]

Alright.. problem found and fixed finally!! Was one of 20k resistors... in fact I tested them and were ok on the value!

Now.. I don't know how but I managed to distroy the trigger system for all channels.. lol

Maybe some opamp of those we said were probably for trigger part. I will order new ones and replace them I think.

Thanks!
Oriol.

[MarkL]

Alright.. problem found and fixed finally!! Was one of 20k resistors... in fact I tested them and were ok on the value!

Now.. I don't know how but I managed to distroy the trigger system for all channels.. lol

Maybe some opamp of those we said were probably for trigger part. I will order new ones and replace them I think.

Thanks!
Oriol.

You're welcome, and I'm glad you found the problem with the offset.  Does Ch1 behave as expected and are the voltages comparable to the other channels now?  Was the 20k resistor open?  Which one was it?

What is not working with the trigger?  Does it still fail self-cal?

I would advise more analysis of the problem and reverse engineering of the relevant pieces of the circuit before replacing parts.  You should be able to pinpoint the bad component or at least have a reason to strongly suspect it.  You might manage to replace the bad part by chance, but there's also a possibility of making things worse with a shotgun approach.

[oriolsancheza]

Thanks again!

So yes, now CH1 offset opamp voltages are as expected, just the same as other opamps like on CH3. Was strange because the 20k resistor it's givin 20k value itself, but looks like when the current flows troguht it, it decreases so then the 20k/20k divider is givind a different value to opamp input.

Do you remember we've been talking about what the other lt1112 do? We said there were for trigger or whatever, so you were right!

The 4 ones on the right are fro CH1 to CH4, the remaining one in the middle of those and the ofsset opamps is like the main one for reference, this one fails now. I think I did some shortcircuit while testing and messed it up.


After doing some further tests and checking with my other scope. I finally observed that there is a problem on another divider I think, the opam is in a good condition in any case.





Before messing up with that it has passed self calibration in vertical and trigger! But no it fails, in fact it doesn't work, I can't get any signal triggered. Looks like is always "off", in fact I tried to "put" it manually, just desoldering the resistor smd code 1372, I do not put the value will be easier for you to find. But in that point comparing to the good scope, there should be 0.9v, and with resistor soldered, there is 0.9v.

I will try to follow your advise tomorrow and check it better in order to see which component fails.



Thanks!
Oriol.