开云体育

this is another way but i havent seen anyone using this in
centuries, the last time i saw it was on a old jerrold sweeper,
everything i had ever since was using the approach i described,
problem with the RF method is that its quite expensive to
realize, a directional coupler, sampling mixer and the oscillator
and frequency counter costs money and space and also takes away
from the RF power, another factor is that whenever you use a comb
generator you run the risk of finding your peaks "everywhere"
even in places you dont want them :)
DC or LF stuff is cheap and doesnt take up much space so i
think thats why most companies opted for this approach

J Forster <jfor@...> wrote:
lothar baier wrote:

ok there are two basic ways to create a marker in a swept

system, one is called RF blanking the other one is called
intensity modulation, ok usually you have a control voltage thats
proportional to your RF frequency, if the analyzer sweeps its
basically a ramp, each point of the ramp corresponds to a certain
frequency, you feed this signal to a comparator and compare it to
a DC voltage thats adjustable (coming from your marker pot)
anytime the ramp "hits" this threshold you get a pulse on your
comparator output, you route this pulse back into your display
and use it to crank up the intensity just a bit at this point and
there you have a marker blip on your screen :)

There is another way. You sample the RF, mix it with a (frequency
settable) oscillator, run the output through a LPF, detector, and
comparator and use this output for intensity markers.

This system has the advantage that you can count the oscillator
and have accurate digital markers, and, if you replace the
oscillator w/ a comb generator, you get nice, calibrated pips.

Since Agilent continiously ups the spare part prices and discontinues
parts it has become more and more important to go past normal repair
practices.
Step Attenuators for example can be repaired without any special
tools or cleanroom enviroment !
Lets take a look at the failure modes first, there are a few:

1.) Blown pads
Usually common with spectrum analyzers if a level past the damage
level of the attenuator was applied

2.) Cracked pads, pad has developed a hairline crack due to
mechanical stress !

3.) Improper contact, the contact in between the attenuator and
contact fingers is disturbed.

4.) Stuck solenoid
Quite common if attenuator hasnt been used in awhile

After we remove the attenuator from the instrument the first step is
to obtain the pinout and build us a little test bench consisting of a
powermeter and a signal generator, the good old 8481D or 8484A will
do nicely, if you dont have a sig gen use the 50MHz calibrator output
of the 435 or 436 or similar model, connect the cal output to the
30dB pad connect the sensor and zero and calibrate meter, the meter
will read -30dBm.
Disconnect the sensor from the attenuator and insert the Step
Attenuator, set step attenuator to 0dB, note the level on the
powermeter, now switch in the attenuator sections one by one measure
and note the level, 70dB attenuators are comprised of a 10 20 and
40dB Pad while 110dB Attenuators normally comprise of a 10 20 and two
40dB pads !
check every section individually, you should hear a click as you
enable each section, if you find a bad section note the value it
should have if all sections appear to be good repeat the test at a
higher frequency using a sig gen, if everything appears to be ok at
low frequency then most likely you have attenuator that just needs
cleaning.
Now lets take the attenuator apart, looking at the top you will see a
nut on each SMA connector, remove this nut this will allow for you to
remove the nameplate, under the plate you will see a wealth of allen
screws, remove those but note that the screws close to the SMA
connector are shorter than the other ones.
Now look at the small sideplate, not the one with the ribbon cable
but the other side - there are 4 small allen screws, remove those and
then remove the sideplate, careful not to lose the little wire mesh
here ! now slide out the u shaped cover and remove the top side.
You now have two parts, one beein the bottom side containing the
solenoids and the other one beein the top part containing the
attenuator pads.
First lets inspect the pads, you can measure them with a ohm meter,
they should measure in the vicinity of 50ohms , except of the 10dB
pad which can measure up to 70ohms or more.
Inspect the pads visually, are there any burned or cracked pads ? if
all pads look ok and check out at DC take a surgical swab and
isopropilic alcohol and carefully clean the contacts pads on the
attenuators.
Next look at the contacts, enable all solenoids and observe the
position of the contact tongues, they should be slightly above the
surface of the block, carefully clean the contact tongues with
alcohol, be careful not to bend them.
next put the solenoids back into the 0 position and look at the top
of the solenoids, you will see two little white plastic rods where
those meet the solenoid you should see little rubber rings , there
are 4 of those on each segment, usually over the years they go
briddle and then disappear, with those rings gone the attenuator wont
make contact anymore, you can buy the rings from watchmakers supply
stores as they are beein used for sprocket gaskets !
Now if your gaskets are there put the attenuator back together after
you cleaned everything and it should work now.
If you have a broken or burned pad dont stress over it the pads are
used in all HP attenuators so just find a attenuator on ebay and take
the pad you need out of it, you can also buy pads from agilent
however be prepared to pay $100-150 per
piece

Lothar,

I hate to be stubborn or dense, but where does the marker come from if the
"RF is dead on the tracking generator"?

I do have an operation and maintenance manual. I quote from the manual:

In Marker mode the scan ramp of the spectrum analyzer is stopped momentarily
at a point determined by the Model 844A Marker Position control. At the
point where the scan is stopped, a bright marker appears on the analyzer
display CRT. Simultaneously, the RF Output frequency from the tracking
generator is counted by the Model 8443A counter. If the Function switch is
set to Track Analyzer, the counter frequency indicates marker frequency,
independent of Spectrum Analyzer input signal frequency. If Function switch
is set to Restore Frequency, the counter indicates the Spectrum Analyzer
input signal frequency (as long as the marker is placed on the signal
response).

From the above statement it seems to me the generation of the marker on the

CRT display and the counting of the marker frequency are two independent
operations, i.e., you can generate a marker and still not be able to count
it.

Since I have demonstrated the counter portion of the 8443A is functional as
it will count an external signal from my signal generator and the RF output
the 8443A is functional since there is a marker on the CRT display, the
problem is in the counting of RF output frequency at the point where the
scan is stopped to generate the marker.

If I am still off base, HELP.

David

the RF output (of)
the 8443A is functional since there is a marker on the CRT display

ok there are two basic ways to create a marker in a swept system, one is called RF blanking the other one is called intensity modulation, ok usually you have a control voltage thats proportional to your RF frequency, if the analyzer sweeps its basically a ramp, each point of the ramp corresponds to a certain frequency, you feed this signal to a comparator and compare it to a DC voltage thats adjustable (coming from your marker pot) anytime the ramp "hits" this threshold you get a pulse on your comparator output, you route this pulse back into your display and use it to crank up the intensity just a bit at this point and there you have a marker blip on your screen :)

you could also use this pulse to turn the RF of your TG off for just a moment but then you would end up with a dark spot as a marker for a short time instead of beein brighter.
Your statement about you seein a bright blip on your screen as a marker along with the statement about you seen the marker even with the calibator connected to your RF section input makes me certain that your markers are intensity modulated !

ok there are two basic ways to create a marker in a swept system, one is called RF blanking the other one is called intensity modulation, ok usually you have a control voltage thats proportional to your RF frequency, if the analyzer sweeps its basically a ramp, each point of the ramp corresponds to a certain frequency, you feed this signal to a comparator and compare it to a DC voltage thats adjustable (coming from your marker pot) anytime the ramp "hits" this threshold you get a pulse on your comparator output, you route this pulse back into your display and use it to crank up the intensity just a bit at this point and there you have a marker blip on your screen :)

you could also use this pulse to turn the RF of your TG off for just a moment but then you would end up with a dark spot as a marker for a short time instead of beein brighter.
Your statement about you seein a bright blip on your screen as a marker along with the statement about you seen the marker even with the calibator connected to your RF section input makes me certain that your markers are intensity modulated !

----- Original Message -----
From: "John Day" <<mailto:johnday%40wordsnimages.com>johnday@...>
To:
<<mailto:hp_agilent_equipment%40yahoogroups.com>hp_agilent_equipment@...>
Sent: Friday, November 16, 2007 3:24 PM
Subject: Re: [hp_agilent_equipment] 3575A - useful ?

At 03:25 PM 11/15/2007, Dave Brown wrote:

Be aware there is a 3571 out there which is a later 'version' of the
3570 and looks to a first glance to be identical, but it does NOT
allow phase measurements, only relative amplitude.

The 3571A is in fact a tracking spectrum analyzer - a different
beast
altogether.

John

Hi John-
They are not really all that different. Main difference is one is dual
channel while the other is single channel.
In essence, both are dual conversion superhet receivers with
dedicated digital processing to control their measurement functions
and generate the log amplitude (dB) and (3570 only) phase angle
outputs.

<snipped>

(I had to cheat and go look at them both to read off the bandwidth
settings-no way I could remember that detail!)

Regards
DaveB, NZ

Yeah, I know. That's what I tried to remind everyone of: that you merely
asked for a set of gears to look at, then return.

No, you still misunderstand, I had no intention of returning the 8640B,
I intended to pay for it by providing gear sets, and by making the specs,
and gear sets available to anyone that wanted them.

I was willing to do the work, but wanted an 8640B /1/2/3 in return for the
effort. I don't know what you get paid, but $175 for a days work seems
like a give-away to me.