|
Hello there! Recently I bought a 8341A opt 001 and 002 synthesized sweep with a problem of 'shut down after a while' but the reality is the unit doesn't shut down at all, but leveling problems. And hard ones ...
Yesterday I was using my variac with 120 vac (the unit came from the States and I am in Spain). It worked for about four hours and don't make any noise, smell or weird thing related to the power supply, so I don't know why the seller said that.?
The +20 volt rail is spot on since the variac reaches 90 vac, and the other rails are reasonably on specs. Fan runs ok, I was thinking my 8663A are loud... this sweeper sounds like a subway train :D (maybe because of running without all panels for inspection).
My first impressions in CW mode:
- frequency chopping is pretty accurate all over the four bands, checked with my 5342A, sharing the counter's 10 MHz reference.
- leveling is an absolute disaster all over the four bands, but looking more stable requiring 0 dBm in the 2.3-7.0 GHz zone (that's band 1 according to the manual). The behavior of the entry display vs the power display is very tricky to explain :D My unit is equiped with -90 dBm attenuator at the output, but I think it works properly¡ some loop jumps randomly taking down the level a few dBm while the frequency varies. The measured level and level shown in the 'power dBm' always matches (measured with a 8481A w 436A). But when you enter a dBm value on entry displays, the actual level fly free. Sometimes I can 'catch' the real level using the knob to equal numbers of both displays and then I can take down the real level a few dBm and the again loose control.
For the moment, I cannot see the spectral purity without a broadband spectrum analyzer (only imagine the jumping signal :D) and the AM modulation because I don't have enough equipment for the moment (broadband mixer and LO)
I already tried to re-enter de calibration constants printed in an ancient papyrum placed inside the unit. I only need to modify the values of six constants:?9, 10 and 11 (YTM gain) and 50, 51 and 52 (YTM BX gain), the other constants matches with the paper. And then, after an 'Automatic Tracking Calibration', those values change again :( Saved the modifications and still UNLEVELED light on... I tried without the recommended tracking calibration and the same outcome.
Any guy or gal that have same issues or some advice on this topic?
Best regards,?
R.
|
RubenRubio -
First read through the section of the manual describing the internal diagnostics available - good idea to NOT mess with the calibration constants until you have identified a problem. You can restore the constants to their original value - see the instruction manual.
Try the following diagnostics:
PRESET[Green key], SHIFT[blue key]' M4 - General diagnostic that tries to check some DACS and other functions
Then
PRESET,SHIFT,METER[ on the lower skirt of the front panel]' rotate the RPG [knob on the front panel] and observe if the power output varies smoothly from max to min
If this works
Test the attenuator by PRESET,CW,RPG [rotate the RPG and see if the power level decreases linearly] - it there is an attenuator problem, you will find points where the power jumps up (typically around the 10 dB points - ~13, -23, etc)
Also, before doing anything check the power supply voltages and ripple
Have fun
Cheers!
bruce
Quoting RubenRubio <rubenrb2019@...>:
toggle quoted message
Show quoted text
Hello there!
Recently I bought a 8341A opt 001 and 002 synthesized sweep with a problem of 'shut down after a while' but the reality is the unit doesn't shut down at all, but leveling problems. And hard ones ...
Yesterday I was using my variac with 120 vac (the unit came from the States and I am in Spain). It worked for about four hours and don't make any noise, smell or weird thing related to the power supply, so I don't know why the seller said that.
The +20 volt rail is spot on since the variac reaches 90 vac, and the other rails are reasonably on specs. Fan runs ok, I was thinking my 8663A are loud... this sweeper sounds like a subway train :D (maybe because of running without all panels for inspection).
My first impressions in CW mode :
- *frequency* chopping is pretty accurate all over the four bands, checked with my 5342A, sharing the counter's 10 MHz reference.
- *leveling* is an absolute disaster all over the four bands, but looking more stable requiring 0 dBm in the 2.3-7.0 GHz zone (that's band 1 according to the manual). The behavior of the entry display vs the power display is very tricky to explain :D My unit is equiped with -90 dBm attenuator at the output, but I think it works properly¡ some loop jumps randomly taking down the level a few dBm while the frequency varies. The measured level and level shown in the 'power dBm' always matches (measured with a 8481A w 436A). But when you enter a dBm value on entry displays, the actual level fly free. Sometimes I can 'catch' the real level using the knob to equal numbers of both displays and then I can take down the real level a few dBm and the again loose control.
For the moment, I cannot see the spectral purity without a broadband spectrum analyzer (only imagine the jumping signal :D) and the AM modulation because I don't have enough equipment for the moment (broadband mixer and LO)
I already tried to re-enter de calibration constants printed in an ancient papyrum placed inside the unit. I only need to modify the values of six constants:?9, 10 and 11 (YTM gain) and 50, 51 and 52 (YTM BX gain), the other constants matches with the paper. And then, after an 'Automatic Tracking Calibration', those values change again :( Saved the modifications and still UNLEVELED light on... I tried without the recommended tracking calibration and the same outcome.
Any guy or gal that have same issues or some advice on this topic?
Best regards,
R.
|
Hi Bruce, thanks for your advice, I'll do that checks and take note of the results.
The paper with the constants isn't the original values? I supose it because of the date on the paper, that matches with the revision showed in the entry display at the very start moment. Anyway I knew that this first atempt was not going to be successful, it's never that easy.
I forgot to mention that during the travel from USA to Spain, the unit suffers a hit that crack the down right corner of the front frame, bend the output connector mount.?I have not disassembled the area yet, but I observed with a powerful flashlight that there was no apparent large damage. This area contains rigid coax cables and attentuator and detectors... maybe the hit takes down those guys. Later I'll disassemble the entire front panel and check this and other things, such as buttons that needs to be hitted with a hammer to pay me attention, the entry display that has all the right vertical lines in all digits off (hopefully this is a connection problem, like the digits on my 8663A).
Best regards,
R.
|
Updating: Today, at the first start, the light FAULT in red is now on ?? the entry display shows 'cal defaulted' and the max freq now showed 26.5 GHz, so my 8341A becomes a 8340A this night :D (one interesting point: during visual inspection I saw that the high band detector mounted on the output coupler, covers 2 to 26.5 gig) I think I mess up something with the cal constants yesterday ???
Preset + Shift + M4 = full diagnostic pass
Preset + Shift + Meter = power dBm display continues showing random numbers constantly changing and actual output follows that random numbers.
Because the sweeping causes my meter to show crap, I tried in CW mode (now I can reach 26.5 gig, this is amazing, but I haven't 26.5 counter or SA to check that) but the behavior is worse than yesterday.
This is getting worse and worse xD
Regards,
R.
|
Updating:
I re-enter the cal constants of the sheet inside the unit and now the FAULT indicator is off... maybe yesterday I don't save the changes or do something wrong with the supply, on today's attempt almost all of the cal constants are changed so badly even the serial number, upper limit, all the DACs, etc... good mess :D So for now I don't want to touch the cal constants :D
I'll retry the tests you said Bruce, undo the front panel and deep inspection of the smashed area, trying to straighten out the connector and so on.
|
Sorry for machine gun posts...
Updating:
The step attenuator is working, ten by ten dB, it downs the power quite exactly. Is the fine attenuation that fails, loosing relation with the actual power at output.
Best regards,?
R.
|
I would consider recovering the cal constants from EEROM - there are directions in the manual.
BEFORE doing this I would write down both the current (Working constants), the constants in the battery backed up memory,and the original constants. I would then relod the original factory constants.
Question - does the list of constants supplied with the machine match the serial number? And is the model specified as 8341 or 8340 on the constants list.
Cheers!
Bruce
Quoting RubenRubio <rubenrb2019@...>:
toggle quoted message
Show quoted text
Updating: Today, at the first start, the light FAULT in red is now on ?? the entry display shows 'cal defaulted' and the max freq now showed 26.5 GHz, so my 8341A becomes a 8340A this night :D (one interesting point: during visual inspection I saw that the high band detector mounted on the output coupler, covers 2 to 26.5 gig) I think I mess up something with the cal constants yesterday ??
Preset + Shift + M4 = full diagnostic pass
Preset + Shift + Meter = power dBm display continues showing random numbers constantly changing and actual output follows that random numbers.
Because the sweeping causes my meter to show crap, I tried in CW mode (now I can reach 26.5 gig, this is amazing, but I haven't 26.5 counter or SA to check that) but the behavior is worse than yesterday.
This is getting worse and worse xD
Regards,
R.
|
Questions:
1) Do you have a spectrum analyzer that can STORE (MAX HOLD on an 8566) 2G - 20G sweep?
2) Do you have a detector that can measure power from about 25dBm to -10 dBm (power meter or xtal detector)?
I am sorry, I forgot to tell you to use CW mode when doing the PRESET,SHIFT,METER test - my fault.
Here is a test that will try to isolate the problem to either unleveled RF output or the ALC system - it assumes you have a spectrum analyzer that can store a full band sweep.
If so, do the following:
1) Set the Spectrum Analyzer (SA) for reference level 30 dB with an attenuation of 40 dB
2) Set the 8341 for 13 GHz (or so) CW mode and press METER - This will essentially disable the ALC system
3) Confirm that the power is within the range of the SA settings (Trace is on scale)
4) Record a sweep from 2G to 20G and note the max and min power levels (Picture would be good)
Note: the best way to record this trace is to set the 8340 sweep rate to approximately 30ms and the SA sweep time to about 200 SEC - put the SA od MAX hold if your SA has that feature. Take a picture of the result.
5) Now on the 8341 do: PRESET,START=2G,STOP=20G,POWER = 5db LOWER than the lowest power recorded in step 4. Leave the SA settings as before and do a sweep from 2-20G (picture of the result would be good)
Next repeat the ALC test I described before:
1) set 8341 PRESET, CW,SHIFT,METER - Rotate the RPG CCW and watch the power output on a spectrum analyzer and watch the 8341 power meter. On the SA, the power should go smoothly from high power (about 20dBm) to low power (-10 to -15dBm), The 8341 power meter should should follow the SA once the power out gets to 5dBm (or so). Note that you will have to turn the RPG CCW and it will take several turns before it starts to take effect.
Based on what you have said so far, I suspect the ALC, but lets do these tests before trying anything else.
Cheers!
Bruce
Quoting RubenRubio <rubenrb2019@...>:
toggle quoted message
Show quoted text
Sorry for machine gun posts...
Updating:
The step attenuator is working, ten by ten dB, it downs the power quite exactly. Is the fine attenuation that fails, loosing relation with the actual power at output.
Best regards,
R.
|
Hi again
Thanks for those steps Bruce, but currently I don't have an SA, I sold my 1.8 gig last month and planned to buy an 8566B asap. For now I only have a 5342A microwave counter, power sensor 8481A along with 436A meter and 8472B crystal detector.?
The cal data paper matches series number of the machine and model too. But I think is weird because the high band detector, the coupler and the step attenuator has APC 3.5 connectors and 26.5 gig stickers :S
I want to measure the power supply ripple in all rails... I saw a table with max current for each rail, but not the maximum allowed ripple. Anyway I want to check that before do more things. This evening?I've been taking down the front panel to check the smashed area for damage and it seems, except for the output connector and frontal frame,? ok.
Tomorrow I'll check the CW behavior with METER or CRYSTAL modes, because I have a 2 gig power splitter, maybe it helps for troubleshooting. Also, I have to study block diagrams of the RF section, but for now I think that the ALC is the culprit... the two detectors are fine, because the fail happens in band 0 and the other bands too.
I'll keep you in touch, thanks for all :)
Regards,
R.
|
OK soyou can do the tests with the detector and a scope.
Drive the scope X with the SWEEP OUT from front panel and the Y from the detector.
PRESET,START=2G, STOP=20G, METER - sweep should be continuous with lots of "wiggles"
Note the lowest point on the sweep, and set the leveled output for 2 dB (or so) less than that and run the 2-20Gig sweep - should be a relatively straight line with some slope (due to cable) at the 20G end
You can also run the ALC test - PRSET,CW,SHIFT METER - use RPG to vary output which should vary smoothly from max to min (maybe -10 dBm). I'm going to guess that this test fails - remember that you need to turn RPG CCW several turns B4 it has any effect. Once it starts to have an effect, output should vary smoothly. If not, there is an ALC problem
Cheers!
Bruce
Quoting RubenRubio <rubenrb2019@...>:
toggle quoted message
Show quoted text
Hi again
Thanks for those steps Bruce, but currently I don't have an SA, I sold my 1.8 gig last month and planned to buy an 8566B asap. For now I only have a 5342A microwave counter, power sensor 8481A along with 436A meter and 8472B crystal detector.
The cal data paper matches series number of the machine and model too. But I think is weird because the high band detector, the coupler and the step attenuator has APC 3.5 connectors and 26.5 gig stickers :S
I want to measure the power supply ripple in all rails... I saw a table with max current for each rail, but not the maximum allowed ripple. Anyway I want to check that before do more things. This evening?I've been taking down the front panel to check the smashed area for damage and it seems, except for the output connector and frontal frame,? ok.
Tomorrow I'll check the CW behavior with METER or CRYSTAL modes, because I have a 2 gig power splitter, maybe it helps for troubleshooting. Also, I have to study block diagrams of the RF section, but for now I think that the ALC is the culprit... the two detectors are fine, because the fail happens in band 0 and the other bands too.
I'll keep you in touch, thanks for all :)
Regards,
R.
|
Hi Bruce!
I checked the ripple of the eight regulated rails and everyone stays between 10 to 15 mV, I think that is ok.
- On the scope-crystal setup:
My scope is digital and x-y mode is crap. So I set two multimeters and the scope in normal mode. The first test you mentioned I think is failed and the ALC I think is ok. You can judge by yourself seeing a couple of videos I recorded, one of a single 60 second sweep from 2 to 20 gig with lvl pre-METER mode set to 2 dBm (note crystal measurement doing crazy jumps from 1.x volts to almost zero), and the other video I have to turn several times the knob and then it has a smooth ride from zero to arround minus seventy milivolt.
I am a bit confused after that...?
Kind regards,
R.
|
Bonus check:
I'm trying to do the external leveling using my crystal detector for the loop and 8481A for sensing the actual power. I tried 500 MHz, 1 GHz and 2 GHz (the max of my resistive power divider for now). You can see how power display and power meter more or less agrees (with 6 dB difference because of power divider), so once again, the internal detectors of the 8341A are ok. But the UNLEVELED light still on at aparently random levels. The lower the frequency, the lower the level at which the indicator lights.
I tried to enter higher frequencies, but the power loss could be significant due to the limitation of the divider and wrong measures at this point are totally useless, maybe it's time to get 18 GHz one.
However, the band where the signal is least attenuated is still the 1 (2.3 to 7 GHz), losing control of it at about +5 dBm. As a conclusion, no matter what source you use for leveling, the behavior seems to be the same.
|
Having troubleinterpertating the results.
Lets try a single test for the modulator (main ALC power regulating component)
Picxk a frequency above 2 GHz and hook either the scope, detector or power meter (power meter prefered if it can handle 20+ dBm) - maybe use an anttnuator).
do PRESET,CW=frequency above 2 GHz, SHIFT,METER - RPG slowly CCW. Now record the POWER METER (bottom window), The attenuation setting (Top window) and the power meter (HP436) readings as a table.
This will determin in the power control components of the ALC are working.
Cheers!
Bruce
Quoting RubenRubio <rubenrb2019@...>:
toggle quoted message
Show quoted text
Hi Bruce!
I checked the ripple of the eight regulated rails and everyone stays between 10 to 15 mV, I think that is ok.
- On the scope-crystal setup:
My scope is digital and x-y mode is crap. So I set two multimeters and the scope in normal mode. The first test you mentioned I think is failed and the ALC I think is ok. You can judge by yourself seeing a couple of videos I recorded, one of a single 60 second sweep from 2 to 20 gig with lvl pre-METER mode set to 2 dBm (note crystal measurement doing crazy jumps from 1.x volts to almost zero), and the other video I have to turn several times the knob and then it has a smooth ride from zero to arround minus seventy milivolt.
Videos
( )
I am a bit confused after that...
Kind regards,
R.
|
Hi again!
This table is what I have obtained by doing the measurements as you told me, from upper limit of 9.3 dBm to bottom limit of my 8481A power sensor. I used a 3 dB attenuator along the power sensor just in case.?I am thinking that if the problem is the upper limit of power in all bands, maybe it's a power failure, Idk.?Or perhaps the full 40 dBm scale that the generator is capable of handling without the step attenuator is shifted 10 dBm down.
Regards,
R.
|
PD: forgot to mention the data on the table were obtained at 5 GHz, inside the 'best behavior band' 2.3 to 7 gig. Maybe I need to test at other frequencies, but the starting power level is lower and my sensor only reach -30 dBm.
|
OK, the results look low but reasonable. The modulator (rf output controller) seems to be working properly. The output seems a bit low - I would expect 15-20dBm (at least that is what I get on a typical 8340a/b)
Could be the unleveled power adjustments would fix that. Try the peak function at 5 GHz and see if that increases the power output.
What is the unleveled power output on the other bands? [PRESET,CW=low/midband frequency),SHIFT METER. Do this on each of the bands - highest power would be expected near the low end of the band. Doing a PEAK before the SHIFT METER might increase the power output.
It would be best to see what can be done to ge the unleveled power up before trying much else. You may have to do the SRD ias and YTF tracking adjustments to do this and that will require at least a good detector, a reasonable scope, a voltmeter capable of 5 digits, and desirably a working spectrum analyzer 8566 is a good choice, there are others.
Your power meter and sensor should be OK if they are reasonably within calibration. (Use the CAL function on the 436)
Cheers!
Bruce
Quoting RubenRubio <rubenrb2019@...>:
toggle quoted message
Show quoted text
Hi again!
This table is what I have obtained by doing the measurements as you told me, from upper limit of 9.3 dBm to bottom limit of my 8481A power sensor. I used a 3 dB attenuator along the power sensor just in case.?I am thinking that if the problem is the upper limit of power in all bands, maybe it's a power failure, Idk.?Or perhaps the full 40 dBm scale that the generator is capable of handling without the step attenuator is shifted 10 dBm down.
Regards,
R.
|
Two more tests to ty on your 8341:
1)PRESET;SHIFT;MANUAL
2)PRESET;SHIFT;EXT
MANUAL and EXT are in the trigger section of the front panel - anything that is blinking indicates a problem.
Question - have you downloaded the user and service manuals for th 8341 that match your unit - they are delightful bed time reading.
Cheers!
Bruce
Quoting RubenRubio <rubenrb2019@...>:
toggle quoted message
Show quoted text
PD: forgot to mention the data on the table were obtained at 5 GHz, inside the 'best behavior band' 2.3 to 7 gig. Maybe I need to test at other frequencies, but the starting power level is lower and my sensor only reach -30 dBm.
|
Hi Bruce
Yes, I have the 8341A operating manual and the 8340A's four manuals w/ good scans from KO4BB website. Unfortunately, I cannot find the service volumes for the 8341A itself, but the 8340A's are very similar, hope.
The PRESET;SHIFT;MANUAL and the PRESET;SHIFT;EXT checks shows no faults at all, no one is blinking. No red fault led in the ALC board is on, also the loops are locked. That's weird, maybe only a calibration is needed to fix the problem.
What you mean when you said 'do a PEAK'? SHIFT + PEAK? or PEAK only? I do the last and power don't rise up a bit.
On band 0, non readable (low) at 10 MHz, -17 dBm @ 1 GHz, -14.3 dBm @ 2.1 GHz
On band 1, + 5.4 dBm @ 2.3 GHz, +8.9 dBm @ 5 GHz, +8 dBm @ 6.9 GHz
On band 2, -11.7 dBm @ 7 GHz, -7 dBM @ 10 GHz, +4.2 dBm @ 13 GHz
On band 3,very low (no readings on both meters) at 13,5 GHz, -11 dBM @ 15 GHz, -8.5 dBm @ 18 GHz
The only low but reasonable readings are on band 1, the others are very low.
Regards,
R.
|
RubenRubio -
You are probably going to need to do an unleveled RF adjustment - but DO NOT touch any adjustments yet.
I suggest you try measuring power at the INPUT of the attenuator or the output of the SYTM - there could be a broken or damaged cable.
PRESET;PEAK runs an auto leveling program
PRESE;CW;SHIFT;METER eliminates the ALC circuitry and drives the modulator directly by the RPG - When you do this test, you get the max output unless you turn the RPG. I usually do PRESET;CW;SHIFT;METER followed by power measurements at various frequencies.
At this point, I'm going to guess there are problems (or adjustments) on the SRD bias board and also on the YIG tracking board
Cheers!
Bruce
Quoting RubenRubio <rubenrb2019@...>:
toggle quoted message
Show quoted text
Hi Bruce
Yes, I have the 8341A operating manual and the 8340A's four manuals w/ good scans from KO4BB website. Unfortunately, I cannot find the service volumes for the 8341A itself, but the 8340A's are very similar, hope.
The PRESET;SHIFT;MANUAL and the PRESET;SHIFT;EXT checks shows no faults at all, no one is blinking. No red fault led in the ALC board is on, also the loops are locked. That's weird, maybe only a calibration is needed to fix the problem.
What you mean when you said 'do a PEAK'? SHIFT + PEAK? or PEAK only? I do the last and power don't rise up a bit.
On band 0, non readable (low) at 10 MHz, -17 dBm @ 1 GHz, -14.3 dBm @ 2.1 GHz
On band 1, + 5.4 dBm @ 2.3 GHz, +8.9 dBm @ 5 GHz, +8 dBm @ 6.9 GHz
On band 2, -11.7 dBm @ 7 GHz, -7 dBM @ 10 GHz, +4.2 dBm @ 13 GHz
On band 3,very low (no readings on both meters) at 13,5 GHz, -11 dBM @ 15 GHz, -8.5 dBm @ 18 GHz
The only low but reasonable readings are on band 1, the others are very low.
Regards,
R.
|
Rbenrubio
Some other measurements that you should make before deciding on how to proceed are:
1) The RF level from the YIG at the REAR PANEL
2) The output of the Low frequency RF amplifier (10M-2.4Gig) (Top side of the 8341 - opposite side from boards)
3) The output of the HF RF amplifier (2-7Gig)
4) The output of the 3.8G heterodyne oscillator
Note that the power levels from these could be high (in the 20dBm range). The rear output and the HF amp should be measured while scanning (with the RPG) from 2.3 to 6.9GHz; the low freq amp and the 3.8 oscillator should be measured while scanning 10M - 2.4 Gig
Basically, what you are doing is confirming the drive to the SYTM and the output of the YIG.
Cheers and have fun!
Quoting RubenRubio <rubenrb2019@...>:
toggle quoted message
Show quoted text
Hi Bruce
Yes, I have the 8341A operating manual and the 8340A's four manuals w/ good scans from KO4BB website. Unfortunately, I cannot find the service volumes for the 8341A itself, but the 8340A's are very similar, hope.
The PRESET;SHIFT;MANUAL and the PRESET;SHIFT;EXT checks shows no faults at all, no one is blinking. No red fault led in the ALC board is on, also the loops are locked. That's weird, maybe only a calibration is needed to fix the problem.
What you mean when you said 'do a PEAK'? SHIFT + PEAK? or PEAK only? I do the last and power don't rise up a bit.
On band 0, non readable (low) at 10 MHz, -17 dBm @ 1 GHz, -14.3 dBm @ 2.1 GHz
On band 1, + 5.4 dBm @ 2.3 GHz, +8.9 dBm @ 5 GHz, +8 dBm @ 6.9 GHz
On band 2, -11.7 dBm @ 7 GHz, -7 dBM @ 10 GHz, +4.2 dBm @ 13 GHz
On band 3,very low (no readings on both meters) at 13,5 GHz, -11 dBM @ 15 GHz, -8.5 dBm @ 18 GHz
The only low but reasonable readings are on band 1, the others are very low.
Regards,
R.
|
RubenRubio
