开云体育

HI,
yesterday I discovered this trouble on my HP8568B spectrum analyzer. When I put ON the S.A. everything works fine , the two red leds INSTR CHECK light for a while, then come back OFF. No error message on the display.
I perform the two routines RECALL REG 8 and RECALL REG 9 and everything continues to sounds good.
But, when I go to the Error Correction Routine (SHIFT+W) then I see the failure. At the end of the routine I see on the display this message "ADJUST AMPTD CAL", instead of "COOR'D", and all the parameters (RBW, REF Level and so on) don't come back to the Preset.
You can see here what happens:



If I start again the routine, at the end I get the same result.

Any idea ? Thanks.

Art IK7JWY

it looks as if the analyzer is telling you to adjust the 2nd local
oscillator amplitude down a little to bring it into a range where the
autocal routine can handle it... check the 2nd lo cal section in your
manual

jim

No, it doesn't have anything to do with the 2nd LO. There's a
gain-adjustment stage that isn't responding properly somewhere -- either in
the IF amp section, or in the RF attenuator itself.

Try putting the calibrator onscreen (CF 20 MHz, SPAN 100 kHz, REF LEVEL -10
dBm) and hitting the up and down arrows to step the RF ATTEN setting from 10
to 70 dB. The calibrator's amplitude shouldn't vary more than a dB at any
point. If it does, then that is probably what's causing the self-cal to
fail. If it remains steady at all RF attenuator settings, look for the
problem in the IF gain-step control circuits in the display section.

-- john, KE5FX

--- In hp_agilent_equipment@..., "John Miles" <jmiles@...> wrote:

Try putting the calibrator onscreen (CF 20 MHz, SPAN 100 kHz, REF LEVEL -10
dBm) and hitting the up and down arrows to step the RF ATTEN setting from 10
to 70 dB. The calibrator's amplitude shouldn't vary more than a dB at any
point. If it does, then that is probably what's causing the self-cal to
fail.

It does :-(

CF 20 MHz
SPAN 100 kHz
RBW 1 kHz
VBW 1 kHz
REF LEVEL -10dBm

ATTEN 10 dB --> Peak at -10.60 dBm
ATTEN 20 dB --> Peak at - 7.70 dBm(out screen)
ATTEN 30 dB --> Peak at - 7.00 dBm(out screen)
ATTEN 40 dB --> Peak at -10.90 dBm
ATTEN 50 dB --> Peak at -10.90 dBm
ATTEN 60 dB --> Peak at - 7.70 dBm(out screen)
ATTEN 70 dB --> Peak at - 7.70 dBm(out screen)

Can I do anything to fix the problem ?

Art IK7JWY

It does :-(

CF 20 MHz
SPAN 100 kHz
RBW 1 kHz
VBW 1 kHz
REF LEVEL -10dBm

ATTEN 10 dB --> Peak at -10.60 dBm
ATTEN 20 dB --> Peak at - 7.70 dBm(out screen)
ATTEN 30 dB --> Peak at - 7.00 dBm(out screen)
ATTEN 40 dB --> Peak at -10.90 dBm
ATTEN 50 dB --> Peak at -10.90 dBm
ATTEN 60 dB --> Peak at - 7.70 dBm(out screen)
ATTEN 70 dB --> Peak at - 7.70 dBm(out screen)

Can I do anything to fix the problem ?

Because the error is small, it really sounds to me like an IF gain problem
rather than a bad attenuator element. I would try to go through the IF
calibration procedure in the display section service manual, and look for
signs of the problem there.

For instance, assume the 8568's attenuator has 10, 20, and 30 dB sections.
(I'm not sure what they actually are, offhand.) For RF ATTEN settings that
require a 20-dB pad to be switched in, a corresponding 20-dB amplifier stage
has to be enabled to keep the reference level constant. If the gain of that
stage is a couple dB off, that would explain what you're seeing.

Ideally you could rule out the RF attenuator by observing the amplitude at
the input to the IF section with another receiver or analyzer as you go
through the 10 dB steps. (Another way to do that is to substitute the
output of a calibrated signal generator for the RF section's IF output.)

-- john, KE5FX

Did you try adjusting the front panel AMPLITUDE CAL on the RF section for -10.00 dBm on the display while viewing recall 8 ? Does it have enough adjustment range above and below - 10 dBm?

Steve




________________________________
From: John Miles <jmiles@...>
To: hp_agilent_equipment@...
Sent: Sunday, September 27, 2009 3:29:37 PM
Subject: RE: [hp_agilent_equipment] HP8568B Error Correction Routine failure

?
No, it doesn't have anything to do with the 2nd LO. There's a
gain-adjustment stage that isn't responding properly somewhere -- either in
the IF amp section, or in the RF attenuator itself.

Try putting the calibrator onscreen (CF 20 MHz, SPAN 100 kHz, REF LEVEL -10
dBm) and hitting the up and down arrows to step the RF ATTEN setting from 10
to 70 dB. The calibrator's amplitude shouldn't vary more than a dB at any
point. If it does, then that is probably what's causing the self-cal to
fail. If it remains steady at all RF attenuator settings, look for the
problem in the IF gain-step control circuits in the display section.

-- john, KE5FX

-----Original Message-----
From: hp_agilent_equipmen t@yahoogroups. com
[mailto:hp_agilent_equipmen t@yahoogroups. com]On Behalf Of Jim Popwell Jr
Sent: Sunday, September 27, 2009 12:06 PM
To: hp_agilent_equipmen t@yahoogroups. com
Subject: Re: [hp_agilent_ equipment] HP8568B Error Correction Routine
failure


it looks as if the analyzer is telling you to adjust the 2nd local
oscillator amplitude down a little to bring it into a range where the
autocal routine can handle it... check the 2nd lo cal section in your
manual

jim


On Sep 27, 2009, at 11:46 AM, ik7jwy wrote:

HI,
yesterday I discovered this trouble on my HP8568B spectrum analyzer.
When I put ON the S.A. everything works fine , the two red leds
INSTR CHECK light for a while, then come back OFF. No error message
on the display.
I perform the two routines RECALL REG 8 and RECALL REG 9 and
everything continues to sounds good.
But, when I go to the Error Correction Routine (SHIFT+W) then I see
the failure. At the end of the routine I see on the display this
message "ADJUST AMPTD CAL", instead of "COOR'D", and all the
parameters (RBW, REF Level and so on) don't come back to the Preset.
You can see here what happens:

. tv/recorded/ 2237690

If I start again the routine, at the end I get the same result.

Any idea ? Thanks.

Art IK7JWY

[Non-text portions of this message have been removed]

Looks like a bad 20 dB element in the RF step attenuator ( It has 3 sections: 10, 20 and 40 dB for a total of 70 dB).

Steve




________________________________
From: John Miles <jmiles@...>
To: hp_agilent_equipment@...
Sent: Sunday, September 27, 2009 6:45:26 PM
Subject: RE: [hp_agilent_equipment] Re: HP8568B Error Correction Routine failure

?

It does :-(

CF 20 MHz
SPAN 100 kHz
RBW 1 kHz
VBW 1 kHz
REF LEVEL -10dBm

ATTEN 10 dB --> Peak at -10.60 dBm
ATTEN 20 dB --> Peak at - 7.70 dBm(out screen)
ATTEN 30 dB --> Peak at - 7.00 dBm(out screen)
ATTEN 40 dB --> Peak at -10.90 dBm
ATTEN 50 dB --> Peak at -10.90 dBm
ATTEN 60 dB --> Peak at - 7.70 dBm(out screen)
ATTEN 70 dB --> Peak at - 7.70 dBm(out screen)

Can I do anything to fix the problem ?

Because the error is small, it really sounds to me like an IF gain problem
rather than a bad attenuator element. I would try to go through the IF
calibration procedure in the display section service manual, and look for
signs of the problem there.

For instance, assume the 8568's attenuator has 10, 20, and 30 dB sections.
(I'm not sure what they actually are, offhand.) For RF ATTEN settings that
require a 20-dB pad to be switched in, a corresponding 20-dB amplifier stage
has to be enabled to keep the reference level constant. If the gain of that
stage is a couple dB off, that would explain what you're seeing.

Ideally you could rule out the RF attenuator by observing the amplitude at
the input to the IF section with another receiver or analyzer as you go
through the 10 dB steps. (Another way to do that is to substitute the
output of a calibrated signal generator for the RF section's IF output.)

-- john, KE5FX

It could be either. There is a good way to check. Set up the analyzer like John said to view the calibrator. Then, use the "shift-q" (that's a small "q") function which will disable the IF step gain switching.

Now push the "atten" key and step the attenuation up, 10 dB at a time. (20,30, etc.) Every time you do that, the calibrator peak should drop by 10 dB. If the drop is significantly different from 10dB, the problem is with the attenuator.

Vladan

--- In hp_agilent_equipment@..., "pianovt" <pianovt@...> wrote:

It could be either. There is a good way to check. Set up the analyzer like John said to view the calibrator. Then, use the "shift-q" (that's a small "q") function which will disable the IF step gain switching.

Now push the "atten" key and step the attenuation up, 10 dB at a time. (20,30, etc.) Every time you do that, the calibrator peak should drop by 10 dB. If the drop is significantly different from 10dB, the problem is with the attenuator.

Vladan

I did it.
This is what I see disabling the IF step switching:

ATTEN 10dB --> peak at -10,7 dBm
ATTEN 20dB --> peak at - 7,7 dBm
ATTEN 30dB --> peak at - 7,7 dBm
ATTEN 40dB --> peak at -30,9 dBm
ATTEN 50dB --> peak at -40,8 dBm
ATTEN 60dB --> peak at -30,9 dBm
ATTEN 70dB --> peak at -40,8 dBm

If the problem is the attenuator, Can I do anything by myself of I have to send the S.A. to the service ?

Thanks

73 Art IK7JWY

Sounds like the driver for the 20 dB pad is stuck on, although I'm not sure
how that accounts for the -7.7 dBm business (Why isn't it -10? Attenuators
don't fail in a way that yields +2.3 dB of unexpected gain.)

Lothar B. just posted some photos of attenuator disassembly a few days ago,
but you definitely want to check the driver signals first.

-- john, KE5FX

John Miles wrote:

Sounds like the driver for the 20 dB pad is stuck on, although I'm not sure
how that accounts for the -7.7 dBm business (Why isn't it -10? Attenuators
don't fail in a way that yields +2.3 dB of unexpected gain.)

Lothar B. just posted some photos of attenuator disassembly a few days ago,
but you definitely want to check the driver signals first.

-- john, KE5FX

In one of mine, one of the attenuator section's solenoid was stuck in one direction. There is a contact for each direction of movement that allows the solenoid current to flow momentarily, and then is opened by a plastic thingy after the solenoid moves to the end of it's travel. In mine the contacts had oxidised so it allowed no current to flow and would not move. Gently cleaning it did the trick. Mechanically the solenoid assemblies are simple enough; with the cover off it is obvious what should do what. Trouble shooting it needs only a multimeter. The actuating voltages are there all the time even though only a pulse is necessary.

John is correct though, the attenuators themselves tend to go open or higher in value when cooked, fried or otherwise mistreated...

Dan

Hi Art,

Check if the results are repeatable. What happens if you exercise the attenuator? Do you get the same results every time?

As John says, it could be the driver circuit for the attenuator, but it could also be a problem in the attenuator itself. The driver circuits are on the A12 assembly.

There are two different attenuator designs. The old one is electromechanical, the new one has some built-in control logic.

Note that the attenuator needs to work in the "pad" and "through" position. For example, if the "through" position for the 10 dB section doesn't work, it may give you 7.7dB of loss due to a bad contact, but 10 dB in the "pad" position. The 20 dB section might be stuck in the "through" mode all the time. There are several possible combinations of what could be wrong. You will probably have to dig in now and check the A12 circuits.

If the problem is the attenuator, Can I do anything by myself of
I have to send the S.A. to the service ?

Attenuator repair requires some mechanical aptitude, manual dexterity, and the ability to understand how the design works. If you feel uncertain, you may be better off sending the attenuator to Agilent for repair or for an exchange unit. If the problem is on A12, you should be able to fix it.

Vladan

--- In hp_agilent_equipment@..., "John Miles" <jmiles@...> wrote:

Sounds like the driver for the 20 dB pad is stuck on, although I'm not sure
how that accounts for the -7.7 dBm business (Why isn't it -10? Attenuators
don't fail in a way that yields +2.3 dB of unexpected gain.)

Lothar B. just posted some photos of attenuator disassembly a few days ago,
but you definitely want to check the driver signals first.

-- john, KE5FX

-----Original Message-----
From: hp_agilent_equipment@...
[mailto:hp_agilent_equipment@...]On Behalf Of ik7jwy
Sent: Monday, September 28, 2009 3:19 AM
To: hp_agilent_equipment@...
Subject: [hp_agilent_equipment] Re: HP8568B Error Correction Routine
failure


--- In hp_agilent_equipment@..., "pianovt"
<pianovt@> wrote:

It could be either. There is a good way to check. Set up the

analyzer like John said to view the calibrator. Then, use the
"shift-q" (that's a small "q") function which will disable the IF
step gain switching.

Now push the "atten" key and step the attenuation up, 10 dB at

a time. (20,30, etc.) Every time you do that, the calibrator
peak should drop by 10 dB. If the drop is significantly different
from 10dB, the problem is with the attenuator.

Vladan

I did it.
This is what I see disabling the IF step switching:

ATTEN 10dB --> peak at -10,7 dBm 10
ATTEN 20dB --> peak at - 7,7 dBm 20
ATTEN 30dB --> peak at - 7,7 dBm 10+20
ATTEN 40dB --> peak at -30,9 dBm 40
ATTEN 50dB --> peak at -40,8 dBm 10+40
ATTEN 60dB --> peak at -30,9 dBm 20+40
ATTEN 70dB --> peak at -40,8 dBm 10+20+40

If the problem is the attenuator, Can I do anything by myself of
I have to send the S.A. to the service ?

Thanks

73 Art IK7JWY

________________________________
From: John Miles <jmiles@...>
To: hp_agilent_equipment@...
Sent: Monday, September 28, 2009 12:00:50 PM
Subject: RE: [hp_agilent_equipment] Re: HP8568B Error Correction Routine failure


Sounds like the driver for the 20 dB pad is stuck on, although I'm not sure
how that accounts for the -7.7 dBm business (Why isn't it -10? Attenuators
don't fail in a way that yields +2.3 dB of unexpected gain.)

They do when the element that goes to ground goes open. My 494 had a similar problem with the 20dB pad. It's pretty hard to blow the series element, but the input resistance of the element, normally close to 50 ohms, was more like 2k, so the previous stage did not see 50ohms and the total attenuation when the 20dB pad was inserted was low by several dBs.

Didier

Sounds like the driver for the 20 dB pad is stuck on, although
I'm not sure
how that accounts for the -7.7 dBm business (Why isn't it -10?
Attenuators
don't fail in a way that yields +2.3 dB of unexpected gain.)

They do when the element that goes to ground goes open. My 494
had a similar problem with the 20dB pad. It's pretty hard to blow
the series element, but the input resistance of the element,
normally close to 50 ohms, was more like 2k, so the previous
stage did not see 50ohms and the total attenuation when the 20dB
pad was inserted was low by several dBs.

Sure, but since he sees exactly the same signal level with both 40 and 40+20
switched in, I'm not inclined to think the 20 dB section is switching at
all, open ground or not. It seems that it would have to be making a
straight-through connection for those two settings to match, wouldn't it?
Be interesting to hear what turns out to be wrong with this one.

-- john, KE5FX

________________________________
From: John Miles <jmiles@...>
To: hp_agilent_equipment@...
Sent: Monday, September 28, 2009 10:43:15 PM
Subject: RE: [hp_agilent_equipment] Re: HP8568B Error Correction Routine failure

Sounds like the driver for the 20 dB pad is stuck on, although
I'm not sure
how that accounts for the -7.7 dBm business (Why isn't it -10?
Attenuators
don't fail in a way that yields +2.3 dB of unexpected gain.)

They do when the element that goes to ground goes open. My 494
had a similar problem with the 20dB pad. It's pretty hard to blow
the series element, but the input resistance of the element,
normally close to 50 ohms, was more like 2k, so the previous
stage did not see 50ohms and the total attenuation when the 20dB
pad was inserted was low by several dBs.

Sure, but since he sees exactly the same signal level with both 40 and 40+20
switched in, I'm not inclined to think the 20 dB section is switching at
all, open ground or not. It seems that it would have to be making a
straight-through connection for those two settings to match, wouldn't it?
Be interesting to hear what turns out to be wrong with this one.

-- john, KE5FX
John,
You are most likely correct (as usual...) If the signal level does not change when the 20dB pad is supposed to be switched in, it seems unlikely a bad element could be the cause. I did not follow the entire thread, I just got in the middle of it :)

Didier KO4BB

I've worked on a lot of these in the past 25 years and have found the step attenuator to be one of the weak areas, especially since it's an electro-mechanical part. The internal O-rings age and get gummy and stick. The contacts that turn off solenoid current get intermittent. And a strange thing that I have found is the U-channel that encloses the solenoid will develop a magnetic attraction mostly on the bottom end that can stick the section. An easy fix for this is to loosen the 2 hex-head screws and carefully slide the U-shaped piece out sideways. Then affix a piece of transparent tape to the inside bottom of this piece and trim the tape edges flush.? This is also a good time to check and clean the solenoid contacts (not the attenuator contacts).
Re-assemble and test.

Steve




________________________________
From: John Miles <jmiles@...>
To: hp_agilent_equipment@...
Sent: Monday, September 28, 2009 11:43:15 PM
Subject: RE: [hp_agilent_equipment] Re: HP8568B Error Correction Routine failure

?

Sounds like the driver for the 20 dB pad is stuck on, although
I'm not sure
how that accounts for the -7.7 dBm business (Why isn't it -10?
Attenuators
don't fail in a way that yields +2.3 dB of unexpected gain.)

They do when the element that goes to ground goes open. My 494
had a similar problem with the 20dB pad. It's pretty hard to blow
the series element, but the input resistance of the element,
normally close to 50 ohms, was more like 2k, so the previous
stage did not see 50ohms and the total attenuation when the 20dB
pad was inserted was low by several dBs.

Sure, but since he sees exactly the same signal level with both 40 and 40+20
switched in, I'm not inclined to think the 20 dB section is switching at
all, open ground or not. It seems that it would have to be making a
straight-through connection for those two settings to match, wouldn't it?
Be interesting to hear what turns out to be wrong with this one.

-- john, KE5FX




[Non-text portions of this message have been removed]