Fluke 6061a 1GHz Synthesized Signal Generator UNCAL repair - PLL unlock (SOLVED)

[SMB784]

Howdy Ladies and Gents,

I picked up a Fluke 6061a sig gen from the surplus property office yesterday for a project in my lab and it turns out it has an UNCAL error and the signal output is all over the place. The unit itself seems to be in good physical condition.



I took it home to work on it, and found that it has a few errors that can be interpreted from the self diagnostics.  The first error obtained by holding the status button with the RF output switched on indicates that the main oscillator's PLL is unlocked.


This is corroborated by switching on the FM modulation and holding the status button to reveal a different error message pertaining to the same problem:


I decided to look for some more information on this error, so I ran the self diagnostic routine and found the following errors with the cited explanations:






The descriptions for these errors are still a bit of a mystery to me. I know that error codes get added in order to determine which error codes is displayed, however the specific error code 325 is a bit of a mystery to me (it corresponded to the BBB table), it seems to suggest that the test performed at 385MHz corresponds to both a locked and unlocked state. Not sure how that is possible, I am investigating the manual now.

Tomorrow when I get my N-Type connector I will post some outputs.

EDIT: Here is a link to a copy of the service manual for this unit: https://www.google.com/url?sa=t&source=web&rct=j&url=http://exodus.poly.edu/~kurt/manuals/manuals/Other/GIG%25206060B,%25206061A%2520Instruction.pdf&ved=0ahUKEwi2z6G4gprXAhVJ6YMKHRAyBaUQFggpMAE&usg=AOvVaw1aUh7e6P2LZDMFH9JcV5tP

[wn1fju]

Can't help you specifically, but I had to fix my Fluke 6060B and the chore was made easier by the following two observations:  1) there are several
frequency bands in this unit and sometimes things will behave differently in a bad (or good) band, and 2) the main RF synthesizer is laid out just like
the block diagram, so it pretty simple to take a scope and just work your way through.  In my case (a faulty resistor in one of the 20 dB amplifiers),
I almost didn't even have to look at the schematic - just probed along until the signal "disappeared."  And, of course, always check the power supply
first!

[SMB784]

Can't help you specifically, but I had to fix my Fluke 6060B and the chore was made easier by the following two observations:  1) there are several
frequency bands in this unit and sometimes things will behave differently in a bad (or good) band, and 2) the main RF synthesizer is laid out just like
the block diagram, so it pretty simple to take a scope and just work your way through.  In my case (a faulty resistor in one of the 20 dB amplifiers),
I almost didn't even have to look at the schematic - just probed along until the signal "disappeared."  And, of course, always check the power supply
first!

Thanks for the heads up!  I'll keep that advice in mind.

After reading the manual a bit more carefully, this unit has multiple errors across the whole spectrum of operation.  The error codes indicate that there are the following problems:

1.) When frequency modulating at 280 MHz carrier frequency and 400 Hz modulation frequency at power level of -10 dBm and "normal" modulation width (not sure what normal is just yet), the system (I think that refers to the PLL) becomes unlocked when it should be locked. (AAA code of 020)

2.) After setting the frequency to 245 MHz, the PLL remains unlocked instead of locking after 120 ms of wait time. (BBB code 001)

3.) After setting the frequency to 525 MHz, the PLL remains unlocked instead of locking after 95 ms of wait time. (BBB code 004)

4.) After setting the frequency to 385 MHz, the PLL remains unlocked instead of locking after 120 ms of wait time. (BBB code 020)

5.) After setting the frequency to 385 MHz, the PLL remains unlocked instead of locking after 200 ms of wait time. (BBB code 200)

6.) After setting the frequency to 385 MHz, the PLL locks instead of unlocking with an E0NL bit equal to 1 (no idea what that means yet) after 120 ms of wait time. (BBB code 100)

7.) The output at 1024 MHz with MIDL bit equal to 1 and HA0CTH bit equal to 0 (no idea what these bits mean yet) is level when it should be unlevel.

So it seems pretty clear that there is some tom-foolery going on with the PLL or the reference oscillator, which may explain all of the above problems.  I suspect the 10 MHz main reference oscillator has a problem, I will verify that by measuring the rear panel reference oscillator output with my oscilloscope when I get home from work.

[trevwhite]

Defpom did some repair videos on the 6060B version. Might be useful?

[SMB784]

Defpom did some repair videos on the 6060B version. Might be useful?

Was definitely useful, if only for what all the test points look like, as well as some strategy for debugging!

Anyway here is the results from last night's measurements.  I think there is something going on with the oscillator, as well as the PLL.

First thing I measured was the 10 MHz oscillator output on the back panel.  Notice the distortion at the top of the peak:

There was also some jitter in the frequency, so I turned on the trace persistence to demonstrate the jitter:


The frequency appears to be mostly correct, its around 10 MHz, however I am only measuring it with the frequency counter built into my RIGOL DS1104z.  Given the strangeness in the oscillator output, I measured the frequency output at the various settings listed in the manual snippets in my first post.  My signal generator is only capable of 100MHz bandwidth, but I am not seeing correct behavior at frequencies between 10 kHz and 100 MHz, so I don't think there is much there beyond what my scope is registering at the supposedly higher frequencies that I program in.  All measurements were taken with the power level set to 0 dBm

Here is the output at the 1050 MHz setting:


525 MHz setting:


385 MHz setting:


245 MHz setting:


100 MHz setting:


50 MHz setting:


1 MHz setting:


100 kHz setting:


10 kHz setting:


All output frequencies less than 245 MHz are fixed at ~24 MHz regardless of what I set the output to.  This is also true for frequencies between 245 MHz and 385 MHz, which outputs ~33 MHz, and for frequencies between 385 and 525 MHz (~71 Mhz), for frequencies between 525 and 1024 MHz (~67 MHz), and above 1024 MHz (~47 MHz).

This seems to indicate that something is wrong in the divide by N chip or the heterodyne mixer thingy.  All of these things could of course be explained by poor control voltages.  That will be the first thing that I investigate.

[SMB784]

I cracked open the case and opened the RF shields for the top and bottom side of the boards.

Top side:


Bottom side:


There appears to be some strange corrosion or discoloration on the top left side of the top side RF shield near the transformer.  Maybe a component exploded there?  Hard to say, I  need to investigate the power board and check the output voltages anyway.

More later on after work!

[Arkytype]

Great forum--hope to learn a lot from the members.

I noticed the link to a service manual relating to this thread.

At one time, it used to be easy to drill down and find a directory listing all the manuals by manufacturer on the http://exodus.poly.edu/~kurt/manuals/ web site.

Can someone please post the correct URL for this link?

Thanx

Lee

[technogeeky]

Defpom did some repair videos on the 6060B version. Might be useful?

Was definitely useful, if only for what all the test points look like, as well as some strategy for debugging!

Anyway here is the results from last night's measurements.  I think there is something going on with the oscillator, as well as the PLL.

First thing I measured was the 10 MHz oscillator output on the back panel.  Notice the distortion at the top of the peak:

There was also some jitter in the frequency, so I turned on the trace persistence to demonstrate the jitter:


The frequency appears to be mostly correct, its around 10 MHz, however I am only measuring it with the frequency counter built into my RIGOL DS1104z.  Given the strangeness in the oscillator output, I measured the frequency output at the various settings listed in the manual snippets in my first post.  My signal generator is only capable of 100MHz bandwidth, but I am not seeing correct behavior at frequencies between 10 kHz and 100 MHz, so I don't think there is much there beyond what my scope is registering at the supposedly higher frequencies that I program in.  All measurements were taken with the power level set to 0 dBm

I wonder if that 10 MHz reference signal is too weak. I don't think it has to be exceptionally pure (I don't think the distortion at the top and bottom are that important). But given that the distortion is at the top and bottom, and given that it's only ~ 1.5Vpp in magnitude, I wonder if it's being loaded down. I have never diagnosed a PLL myself, but I get the impression from The Signal Path's videos that a lot of times it's the amplitude of a waveform that causes them fail.

Obviously it can't really be the frequency of the reference that causes a PLL to fail -- the PLL will just adjust and assume whatever frequency it sees is correct and act accordingly. I also don't think it matters that much for the signal to be spectrally pure: for starters, sometimes a square wave is preferred (even faster risetime) and even when it's a sine wave, it's probably triggering on the zero crossing anyway so distortion at the top of the sine wave probably won't cause problems. Also, it's possible that the output you are checking at the back of the unit is generated as a square wave and filtered to a sine wave for the benefit of external devices only, so it's possible that nothing else inside the unit even sees the sine wave (and hence, the small distortion at the top).

I'm not sure if your probing settings and methods for any of these measurements (I guess in all cases you are doing direct BNC connection). I would re-try at least the 10 MHz reference using a 1:10 probe (using a BNC adapter or just, you know, holding it in there) and see if the Vpp changes significantly.

Also, I would find the 10 MHz reference on the board and see what power it starts at. Lastly, I would remove/disconnect any boards that can be easily disconnected and check the same reference and the PLL's connection of that reference to see  if the amplitude goes up when loading is decreased.

[pigrew]

I had a vaguely similar issue with my 8656B. It would never lock it's PLL, and would tend to output 25MHZ different than nominal for half of each frequency band. It took a heck of a lot of debugging to fix it. It turned out that one of the BJT-based differential amplifiers in the phase comparator was weak (it still had good diodes according to the multimeter... Just low gain).

[SMB784]

Time for an update!

So I followed the suggestions of the manual and measured the voltage outputs from the power supply PCA at the various test points.



All the voltages were in spec except for the voltage at TP7, which was about 3V below where it should be.  The manual indicated this could be due to a short or another problem loading down the supply.  So I followed the instructions and disconnected the various component boards from the power supply and remeasured the voltage.  Sure enough, disconnecting either the Synthesizer PCA or the Controller PCA caused the voltage to pop back up to near where it should be.  This lead me to believe that there is a problem on one or the other of these boards, or perhaps both.

I started by investigating the synthesizer board.  I tested the outputs of the main reference oscillator at the various test points, and found that the oscillator is functioning as intended.  I am seeing a 10 MHz output that I can control by adjusting the trimpot that controlls it, and all the voltages are within spec.  From there I moved on and started testing at the next points mentioned in the manual: the main PLL.  It turns out that the first test point mentioned (TP44) had a voltage that is wildly out of spec.  It is supposed to be between 2V and 22V, and my measurements were at approximately -0.6V.  This situation is not mentioned specifically in the manual, so I tested other points in the main PLL section of the board and got other very strange results:





Given these odd results, I am beginning to suspect something is up with the loop amp (U48 on the diagram), or something related to it, as none of the downstream test points are measuring the correct values.  More investigations to follow.

[SMB784]

Alright it's been a while, so its time for an update!  Just got back from the land of Dave, and I finally got a good frequency counter and DMM for this project.

Anyway, I started probing around the various test points and found out that the output of U48 was -2.4 V.  I checked the input signal at TP37 and TP56, and they both registered the correct voltage.  Given that the input signal to the U48 Op Amp is correct but the output is wrong, this lead me to believe that something is wrong with the rail voltage of the op amp.  Sure enough the positive rail voltage (+Vvar) measured at -0.8 V.

I followed the trace on the diagram for Vvar and found its origin at TP34:



Sure enough, Vvar measured at -0.8V.  I think that there are two possible culprits that could be causing this: the +Vhet voltage value or a faulty Q15 transistor.  I checked the voltage at Vhet and it registered the correct value of 5V, so I eliminated that as a possibility.  I also checked the input voltage to the Q15 transistor and found that it was reading the correct value at 37V.  I also measured the voltage at the output of the resistive voltage divider (between R122 & R123) that serves as the input to the transistor base, and found out that it was measuring at 29.6 V, which is in line with what the output of the resistive voltage divider should be outputting (I/O ratio of 0.8 according to the diagram, and 0.8*37 = 29.6).  So given that the input to the collector and base of the transistor are correct, but the output from the emitter is wrong (negative voltage when it should be positive from anywhere between 2 and 22 V), I believe one problem that this system has is a blown Q15 transistor.  I'm not really sure what could have caused this to fail so that the voltage swings negative, but the part is cheap and easily replaced (model # 2N3904) and I will try replacing it first to see if that fixes the problem.  Will be back soon with an update, so sit tight!

[SMB784]

 Quick update: replacing the Q15 transistor did the trick! No self test errors and after a brief investigation I can set the frequency with a mostly stable output over there whole range of frequency values.

However, the UNCAL light is still flashing when the output is turned OFF, indicating that it is in an unstable state in the off position. More detailed update coming tomorrow!

[Silveruser]

I have a Motorola branded version of the Fluke 6060 purchased as gov surplus as well. I've been meaning to post. The rear 10Mhz output on mine is a semi-square wave 5vpp. I say semi because I have a lot of ringing on the negative bottom of wave might due to no termination just the scope probe stuck in the connector. Looks like you made good progress.

Mine also flashes the uncal with the output off, I sort of accepted it as normal but maybe not. Don't think I ever found a clear answer in the manual. Maybe others could comment.

If you still have it apart you might want to replace the lithium battery on the extended memory board. It is long past it's shelf life. That was the only fault I found on mine.

[SMB784]

Alright, so here is the skinny on how the fix went down.

I replaced the Q15 transistor after discovering that it was the likely source of the fault.  Here is the offending part:



After replacing it with another 2n3904 and booting the system back up, it passed the self check with flying colors.  The voltage readings at TP7 on the power supply board also read 21.7V, which is within 0.3V of the rated specification of 22V.  I checked the frequency output at the top and bottom ranges of the device with my frequency counter (top right of the following images) and everything came out within the error digit of the counter.  Spot on!

1050 MHz output @ 0 dBm:


10 kHz output @ 0 dBm:


I did some further testing of the quality of the frequency output with the oscilloscope.

Waveform trace @ 10 kHz, 0 dBm with infinite persistence:


Waveform trace @ 10 MHz, 0 dBm with infinite persistence:


Nice to see that the amplitudes are within a few mV of eachother over the range between 10 kHz and 10 MHz.

Waveform trace @ 100 MHz, 0 dBm with infinite persistence (note lower amplitude due to bandwidth limit of scope at 100 MHz, also the built in frequency counter doesn't seem to work right beyond the bandwidth limit):


Next I did some amplitude testing at 10 MHz output, and everything seemed to check out.

-50 dBm (min output detectable with scope):


-20 dBm:


13 dBm (max output):


Traces all look stable, amplitude looks right at first pass (need to double check).

Finally I tested the AM/FM modulation capabilities at 10 MHz output with the oscilloscope, and again it performed admirably.

FM @ 10 MHz, 0 dBm, 99.9 kHz modulation width, 1 kHz modulation rate:


Simultaneous AM & FM @ 10 MHz, 0 dBm, 99.9 kHz modulation width, 1 kHz modulation rate, 90% AM depth:


That concluded the tests that I have performed on this device.  I am fairly satisfied with its functionality.  I still have no idea why the UNCAL light is flashing when the output is off, however it goes away when the output is turned on.  I performed a latch test (no idea what that really does), but it reads LATchAA as an output after the test.  I'm not sure what that means, and the manual hasn't been very helpful yet, but maybe I will learn some more later.  Regardless, I am quite happy with the outcome of this repair so far.

[technogeeky]

 

How lucky that it's such a standard part, too!

[SMB784]

I know right?? So standard that I even had a few just laying around

[kc7gr-15]

The flashing 'Uncal' when the output is turned off is perfectly normal. My 6060B does the same thing, and I found it mentioned in the manual (the copy I bought from artekmanuals.com).

[SMB784]

The flashing 'Uncal' when the output is turned off is perfectly normal. My 6060B does the same thing, and I found it mentioned in the manual (the copy I bought from artekmanuals.com).

Well since you are the second person to independently confirm this behavior as normal, I will accept this as true.

In light of this fact, I declare this device fixed! Thanks so much everyone for all of your very helpful input.  Now on to calibration.

[apel]

Good morning,
I carried out the second test as per your instructions and got the result:

025 225 000 017

I repeated this test several times, always with the same result.

Paolo

[apel]

Small update:
+15v is missing on the power supply board.
I think it may be some tantalum downstream.

[SMB784]

Ok, so according to the manual, here are the things that the error code you posted correspond to:

025 refers to the following issues in the modulation circuit:
30% depth 400 Hz amplitude modulation at 10.7 dBm unleveled at 1050 MHz
30% depth 1000 Hz amplitude modulation at 10.7 dBm unleveled at 1050 MHz
400 Hz frequency modulation at -10 dBm unlocked at 280 MHz

225 refers to the following issues in the synthesizer circuit:
PLL unlocked in 245 MHz synthesizer band after 120 ms wait
PLL unlocked in 525 MHz synthesizer band after 95 ms wait
PLL unlocked in 385 MHz synthesizer band after 120 ms wait
PLL unlocked at 385 MHz synthesizer band after 200 ms wait

000 means no issues in digital control circuit

017 refers to the following issues in the output filter circuit:
Output unleveled in 350 MHz band
Output unleveled in 512 MHz band
Output unleveled in 730 MHz band
Output unleveled in 1050 MHz band

The first thing to say is that it looks like your output transistor is blown or is getting no voltage, as you're getting output unlevel on all frequency bands.  The second thing to say is your PLL is unlocked in all frequency bands, which probably has a similar cause to the lack of output, i.e. no voltage.  The fact that you arent getting a 15V signal is highly suspect.  Trace that path backwards to the power supply and see if there's an issue there.

[miken]

Don't have my notes handy but I think they did something silly like 16V tantalums on the 15V rail. I had a short in one; definitely worth replacing anyway.

[charlyd]

i have a 6062A which had a missing 15v go and look for exploded tantal in the supply line... maybe voltage injection will help ;-) to find it fast  luckily you have schematics my tantal which shorted was nearly at the other end of the line ( and unit). 


btw SMB784   good catch...i will check mine also sometime i have been probing in the area a long while ago......but now on other unit to repair.

[apel]

Good morning,
Thank you all for your suggestions.
I have isolated the power unit and given an external load to 7815 on connector J6. The circuit works perfectly.
The connector sends the various voltages to the Synthesizer PCA (figure 8-6 circuit A2A1. Two tantalums are mounted at the input.
I will check them.