Hi,

?

I also have a working 8133A with option 002.

No error messages.

?

For your problem #2, it's the same for me when it's been a while since I've used it.

I just tried it a few minutes ago, and the problem was present, but after a few tries on activating the output (DISABLE button), the signal appears correctly and remains stable.

I think it's probably a relay contact. When you press the DISABLE button, you hear the sound of a relay.

?

For problem #1, perhaps a bad contact...but in any case, you will have to open the instrument to check. I assume you have the service manual. (Available on Keysight.com)

?

Yves

?

?

Hello all,

I received Agilent 8133A that I really want to make use of it. It works some sort of but It has some issues. I hope someone in this forum has gone thorough it and help me to resolve these issues.

Issue #1:
The self test fails with error E008 Frequency counter error. I can see that the pulse output on channel 1 even though this error shows up. I observed that this error only shows in some frequency band. eg. 400MHz, 500 Mhz. in some frequency bands, this error does not shows up. even though it shows error, I can see correct frequency output in channel 1.

Issue #2
This unit has option 002 (channel 2). the OUTPUT seems dead, I can see some pulses, odd shaped waveform present but nowhere close to square pulse. but negative OUTPUT terminal seems work fine. I am loosing differential capability.

That's all issues I can see now. I am using sampling scope with SD-26 head to see these outputs. more issue may be discovered later.

Thanks all for reading this.

Thanks

Hi,

Here are the results of tests similar to yours: at 500, 700 or 3GHz, the signal is stable as shown in the two images "DATA out 500MHz.png" and "DATA out 3GHz.png".

PRBS mode works very well as shown in the image of a 3GHz test "PRBS DATA out.png".

My scope is 83480A option 100 with module 83483A.

My connection cables between my oscilloscope and my 8133A were long, about 4.5m (15 feet).

?

I assume your oscilloscope is a TEK CSA803C with your SD-26 module. Here are some remarks:

Check that your TRIG config is correct.

Do the TRIG on the signal and not on an internal clock.

Try inputting an external RF source to your 8133A and see if that makes a difference in your results.

?

"Regarding issue#2, I can see good waveform output in inverted side but positive side has weird waveform, but after some time, both outputs seem to be normal, maybe it is due to temperature effect?"

?

I have exactly the same problem and on the same connector (positive output), except that after a few seconds the problem disappears especially if I press the DISABLE button several times...I conclude that it It's probably a bad contact. But when the 8133A is hot, I don't see this problem anymore, it happens at power up when the 8133A is at room temperature, after the problem disappears, it doesn't come back as long as the 8133A remains powered on ...so I would say that temperature plays a role.

Every time I use my 8133A, I always tell myself that I should solve this problem...but I still haven't put the time into this task yet.

?

Can you confirm the model of your oscilloscope, I will see if I can borrow one from one of my contacts here, I will then be able to make a better comparison.

?

?

Thank you Yves,

The issue baffled me alot. There are the waveforms captured using sampling scope, the 500MHz - 700 MHz and >3.2GHz shows significant distortion as shown below:


In other frequencies, waveforms looks much better:


regarding issue#2, I can see good waveform output in inverted side but positive side has weird waveform, but after some time, both outputs seems to be normal, may be it is due to temperature effect?



Today I faced new issue, when I try to see the channel 2 output, the waveform looks normal in square mode and data (32 bit mode) but when I switch to PRBS, the waveforms is just noisy, is your unit also shows same issue?


Any debugging idea is helpful.

Thanks!

Hi,

?

I think that indeed your divider by 2 may have a problem, does the divider receive enough signal coming from the YIG over the entire frequency range?

Below 2GHz, the divider by 2 is activated, if things go badly at 500 MHz, you should normally see it at another frequency, like at 250 MHz for example, since the YIG will be at the same frequency, i.e., 4GHz.

?

From 31.3MHz (min) to 3500 MHz (max), there are therefore 34688 possible frequencies, since the dividers are configured for a 5-digit BCD division.

The dividers make jumps and a 12-bit DAC completes the adjustment at the requested frequency.

I was able to check some frequencies on the Spectrum without opening the device, by measuring on the External input connector.

?

For the TRIGGER and PRBS, I believe you need to activate the HIGH FREQUENCY ON function on your oscilloscope.

Or, activate the divider for the TRIGGER CHANNEL OUTPUT on your 8133A.

?

For the TRIGGER and PRBS, I believe you need to activate the HIGH FREQUENCY ON trigger function on your oscilloscope.

Or, activate the divider for the TRIGGER CHANNEL OUTPUT on your 8133A.

?

Your 8133A is fully functional with an external generator that's good.

But obviously it is preferable to carry out the repair, except that now since you have a solution from an external source, it becomes less urgent to embark on this repair.

?

I opened my 8133A in 2019 to change the battery only, but I should start thinking about repairing my CH2 output.

I retested my generator. I set myself to PRBS at 500 MHz. When powered on, the non-inverted CH2 out signal is defective, it seems that only fast transients succeed in passing...and nothing seems to move after 1 minute. I toggle the DISABLE button once and wait...about 10 seconds later the signal becomes normal. This problem repeats itself almost every time the generator is turned on and it was at room temperature. I haven't opened my generator yet because I'm currently working on another instrument.

?

?

?(page 52-72)

?

?

?

Yves

?

?

Thanks Yves,

Its really helpful. Today I did few tests as per your suggestion, I have found few interesting findings. By the way I am using Tektronix 11801B (same as CSA 803C, I think its just label change and some software features differences) with sampling head SD-26 and SD-24.

issue#1: jitter in 500MHz to 600MHz band.
I used normal oscilloscope (1.5Ghz) to see the waveforms on trigger OUT, and Pulse Output, I can see the output waveform is jittering when using 8133A internal clock.
?

Now when I use external clock, The jitter problem is gone, The waveform on all frequencies (I tried up to 3.6GHz!) looks good.


It tells me the either internal clock or divider is bad, what do you think? Shall I attempt to repair or using external clock is good enough? I think? 8133A regenerates pulse (<100ps rise time) based of the external clock but clock stability and jitter may depends in the external clock source. I better fix it, no?.


issue#2: Positive data output intermittent: Seems your unit and mine shares same issue. Its not a deal breaker but I will try to see if I can fix it by changing relay. I will keep you posted on my progress in this issue.

issue#3: Sampling scope not able to show data in PRBS mode. I connected normal oscilloscope Agilent 1.5Ghz Channel 1 to trigger output and channel 4 to data output of the 8133A, I set trigger of scope on channel 1. I can clearly see PRBS data and trigger waveform clean. Now I suspect there is something going with sampling oscilloscope when I try to capture PRBS data. Any thoughts?


Again, Thank you so much for your help and support.

Hi,

?

The specification for output frequency accuracy is ±0.5%.

But there will also be some modulation, for a 500 MHz test it is around 14 kHz (112 kHz for YIG at 4 GHz) and frequency measured at 499.859MHz or -0.028%.

?

I haven't read the service manual in detail, but I assume that the YIG must be checked and stabilized by some system even if it is slow, I imagine that the integrated frequency counter would help with this stabilization.

If you use an external source, the frequency counter will show you a reading corresponding to the incoming signal. But you told me about error E008 which indicates that the frequency counter has a problem...with an external source, is this the case?

If the counter is functional with an external source, then the problem would be near the YIG, or the YIG itself...assuming the frequency counter is involved in some process of stabilizing the YIG.

If you had had an E110 type error this would possibly indicate a corruption of the EPROMs and calibration constants possibly linked to the YIG, but this is not the case.

Unfortunately, it would be necessary to check the DAC output up to the YIG, and the YIG output up to the 3 RF switches S1 and S2 which serves as a bypass for the divider by 2, and S3 which switches between an internal (YIG) and external source.

I can tell you that making these measurements will not be a simple task because the PCB involved is the one completely at the bottom of the RF section.

?

------------------------

?

Disassembly of the 8133A is really very painful and requires disassembling several mechanical parts such as the side supports of the frame as well as the rear panel. But I checked what was causing the CH2 output intermittent problem. I was able to test the RF relay independently on a VNA but still under the control of the 8133A driver and the circuit preceding the RF relay which was connected to the oscilloscope.

So, I simultaneously checked the behavior of the two circuits when powering up.

And there was no doubt, it was the relay that was at fault.

in the image, we can see that the frequency response is not flat, and that all low frequencies are very attenuated.

You will either have to find a replacement part or attempt a repair (I attempted a repair, but I was unable to open it...I stopped this task for the moment, so as not to risk damage it).

?

:Yves

?

Hi Yves,
your YIG clock frequency and output frequency text file is of great help. Today I step on each output clock setting and observe the clock frequency.?

The output clock is way off (wrong frequency and sometime have jitter) when YIG frequency is in ~2000MHz to ~2585MHz AND ~3600MHz to 4000MHz.?
My oscilloscope cannot go >1Ghz so I tested that from 31Mhz to 800MHz output frequency.?

Above issue is also indicated by the startup self test failing when I set output clock to 500Mhz. If I set instrument output clock to 700Mhz and restart the system then self test passes, That means instrument is checking the clock output and comparing with what's been set during startup. I checked this in several output frequency set points.

I really want to test the output clock when >2Ghz but my scope is that that capable. So Does it suggest some issue in the YIG output? either DAC is not setting correct voltage or YIG itself is gone bad??
Any other test can be done before I tear apart the device to probe the signals??

If its YIG then its dead end to repair may be, Now I am concerned.

Thanks

Hi,

?

I found that (calibration YIG.txt) in the old Hp journal, about 8133A. It might be a good idea to check these values.

I used the command =>?????? :DIAG:CAL:YIG?

And the result was ???=>??????? +2.02000E+09,+9.31000E+02,+2.55987E+09,+1.51400E+03,+3.02676E+09,+2.01900E+03,+3.49611E+09,+2.52500E+03,+3.95988E+09,+3.02400E+03

?

?

?

When you change the frequency from 112.5 MHz to 124 MHz and the output frequency remains unchanged at 103.5 MHz...

I would say that the DAC circuitry is at fault and not the YIG, because it would have moved with this change of DC, it could have given the wrong frequency but it would have moved.

?

I would be curious to see the behavior at 225 MHz, the generator should have an output of 207 MHz, and by changing the frequency to 248 MHz the output should maintain at 207 MHz.

This situation requires the DAC to have the same values sent to the YIG, only a divisor after the YIG will change.

?

You will receive a timing board; you will be able to solve this problem.

?

---------------------------------

?

?

For my part, I will make a frequency precision curve over the entire range of the 8133A. I'm going to reopen my 8133A to try to repair the RF relay again.

?

Regards

:Yves

?

?

Thanks Yves,

I checked the operation of frequency counter on external input signal. The frequency counter indeed shows the correct input frequency feed into the external clock input.
This points that either YIG or somewhere near YIG has problem. The frequency error is not minor, for an example, if I set internal clock to112.5MHz, I measure the clock 103.5Mhz and the output frequency stays same (103.5Mhz) even I set the clock to 124Mhz, It seems that in some bands the YIG frequency is both incorrect and not responding to the requested frequency changes.

But in some band (e,g, from 80.7Mhz to 112.4 Mhz), the output frequency is both correct and stable.
I ordered a used timing board from ebay, I hope i can harvest the YIG if it is still functional. wish me a luck.

My unit does not show any other error than E008 (frequency counter error) So I strongly suspect, frequency counter is working OK but output if frequency counter is not correct, throwing error, and because input of the frequency counter is not correct.

I looked into the Channel 1 RF relay, I think we can replace with equivalent part or may be I can open it and have a look, I plan to do it when I open the unit to debug the YIG issue. I will keep you posted.

And Indeed its very difficult to open the unit to have access to the timing board, Its tons of screws :)

Thanks!

Hi,

?

For the relay problem:

- I decided to try the disassembly again and it worked correctly this time.

- step#1: remove the cover, there are two screws, one of which is hidden under the label.

- step #2: you have to dismantle the relay coil to go further and it is difficult.

- step#3: I had no choice but to cut the screw head, drill and remove the debris from the screw. The replacement screw was a 2-56.

- step #4 and #5, shows the RF base including the mechanics of the relay without the activation coil.

- step #6: dismantling the external mechanics of the relay.

- step #7: separation of the RF and mechanical part of the relay. At this point the goal was to clean the gold-plated blades and contacts and reassemble that part. Reassembly was quite difficult because it requires precision. After this step was done, I tested with the VNA and when I manually pushed on the blade above the spring that we see in steps #8 and #9, everything was going well with the VNA.

- After step #7, I redid the assembly without the cover and tested everything to realize that it was not yet working correctly.

- I ended up understanding that the problem was the distance between the blade pushing on the spring base which didn't have enough pressure. So I did the disassembly in step #6, I adjusted the thin blade attached to a metal part, so that the blade barely touches the base of the spring as shown in step #9... and this solved the problem.

?

So, here is my advice: do steps 1,2 and 3 and look at the blade shown in 8 and 9. If it does not touch the base of the spring, do step 6, adjust the blade to get a result like step 9 and reassemble everything without the cover and do a test on your workbench. The relay operates at +15V (97mA), pay attention to polarity. The test can be done with an oscilloscope, if you do not have a VNA or Spectrum. Choose a low frequency for the test first, for example 50 MHz, you will quickly see if the signal passes when the relay is activated. If unfortunately, the relay still has the same erratic behavior, then you will have to go further and do step 7...and that is not easy at all.

?

----------------------

?

I did the same test as you to compare, and it's difficult to explain this behavior.

?

--------------------

?

I also did this automated performance test, which is the error of the output frequency vs the programming of that frequency.

The acceptance limit is ±0.5%, but typically the error should be close to ±0.1%. This test checked all programmable frequencies from 31.3MHz to 3.5GHz in steps of 100kHz, therefore 34688 frequencies, so this test lasted several hours. The error result also includes that of thermal drift during the test.

?

?

-------------------------

?

Tomorrow I'm going to go see a friend to do a comparative test regarding frequency stability on another 8133A.

I will probably have the opportunity to do a test on a Tektronix 11801B with SD-26 and I will check the trigger mode with a signal from the 8133A. I will give you my comments if this test can ever be done.

?

?

?

?

Yeah,

I did ran the command and I get following, so It seems calibration is OK (it is linear)

:DIAG:CAL:YIG?
+2.02000E+09,+9.55000E+02,+2.56030E+09,+1.54500E+03,+3.02723E+09,+2.05600E+03,+3.49605E+09,+2.56700E+03,+3.95992E+09,+3.07200E+03

I re measured and plotted the set frequency and the output frequency on ch1 (using frequency counter). I see below plot. Yellow trace is the YIG frequency from 2 to 4Ghz (expected) and the blue trace is the actual frequency output, x axis is the set frequency. I measured manually from 33 Mhz to 500Mhz. Its clear that the output frequency has some issues and not linear with set frequency.


I still do not understand how 12 bit DAC translates to the YIG output, May be some bit of DAC is an issue?.

By the way, My timing board was coming from Israel, and due to what's going on over there, it is not certain if I will get it or not.?

If you end up fixing RF relay, then I will follow your pursuit and fix mine also. :)

Thanks alot.

Hi,

?

Looking at this trace on the graph, what catches my attention is the fact that the control voltage applied to the YTO makes jumps at the start and end of the YTO range, and is linear in places, but mostly the fact that part is curved.

?

Maybe the YTO driver is more complex than I imagine. Perhaps there is switching in the transconductance amp powering the YIG coils, and one or more of these circuit sections are distorting...hard to say.

?

I saw a friend of mine yesterday, and he had three 8133As available to do some quick testing. On 2 of these generators tested, there was a relay problem, one was on the CH2, normal output, and the other was on the reversed output. Their problems were exactly the same as on my 8133A which I ended up repairing. No doubt this problem is recurring on these generators.

?

I measured the FM modulation at Spectrum at 500 MHz on these two generators, and I had the same result as on mine...this would seem normal behavior.

?

I had the opportunity to use a CS803C and SD-22, and it was possible to watch the signal in PRBS mode. However, in the trigger menu it was necessary to use the EXTERNAL DIRECT trig, to switch from VARIABLE PERSISTENCE mode.

?

For the CH2 relay, if the repair fails, this part is available on eBay.

?

?

Thanks Yves,

Collecting data for frequency accuracy analysis is lengthy task. I wrote a powershell script to set frequency, read 8133A internal frequency counter reading (and the external frequency counter connected to the ch1 output.
I stepped frequency 1Mhz from 34Mhz to 3.5Ghz. The plot below shows the frequency error (set-reading from 8133A internal counter). Indeed there is some frequency issue in some bands, which does not have explanation.


--------------
Regarding ch2 relay issue, I plan to open my unit tomorrow and try to fixt it. Let's see how it goes. Thanks for your step by step instruciton.

Thanks!

Hi,

?

This generator is really complicated to repair, especially without any electrical diagram.

My friend bought some replacement relays, I will collect these faulty relays when the time comes and see if I can repair them.

?

When you receive your timing card, there may or may not be another problem, that of calibration concerning the timing of this PCB to the outputs or other cards...the signals go through the loop at the back .

Will see you when the time comes...one problem at a time!

?

L?ches pas la patate!?

?

Yves

?

?

Thanks Yves,

Today I? opened the unit to fix the relay, As you mentioned , opening relay is so tricky. I will continue to work on it in coming days.

On the another note, I received another 8133A unit (single channel, standard unit , without channel 2). All frequency bands works well in this unit so I thought I can swap the timing board and make unit with option 2 (which has output frequency issue) in working state BUT, After spending a day to open up two units (No thanks to designer who decided to make the assembly such complex! :) ) It appears that timing board for unit with option 2 is different from standard version.

There is two Coaxial coming from timing board with option 2 but there is only one coaxial coming off to the channel link (back side of the unit) on the standard unit. Here are few pictures of the units I worked today.

Hi,

?

One problem to solve, the next one will be a little more complicated, namely the timing board.

?

For my part, I also ordered an RF relay which I received last week, and it works perfectly.

I will replace it anyway even if I managed to repair the original in the 8133A.

?

?

Regards

?

Hi Yves,

Today I received the RF relay and I managed to replaced the faulty one. Indeed the RF relay was bad as you can see the transmission losses observed using VNA (my VNA can only goes up to 3Ghz). After replacing relay, the channel 2 positive output works.
RF Relay transmission loss:


New Relay Transmission loss:


Here are more pictures:



---------------
On the clock frequency issue, I do not think I will receive the timing board. it's stuck in transit. I do not have any path forward to debug or repair of the clock.?

Regards,
?