----- Original Message -----

From: DON CRAMER

To: HP Agilent Group

Sent: Monday, November 01, 2004 2:06 AM

Subject: [hp_agilent_equipment] E4411A spectrum analyzer low gain fixed

I thought I should document repair to an E4411A (alias ESA-L1500A) 1.5Ghz spectrum analyzer in the event anyone here might find the information useful.

?

The symptom was low gain (~28dB low) but with flat response and all other functions ok.? The HP service manual only goes down to the module level, which is doubly unfortunate for this instrument because the one RF assembly module contains everything from the front panel input connector all the way back to the 2nd IF input.? The manual helped confirm the fault was in the RF assembly module, but at that point they say don't service it, get a replacement instead.? That would be ok except the replacement price is $6120 exchange.? For an analyzer that probably went for $8K new, I bet they don't get too many takers.? As the SA was initially free, I didn't have too much to lose by violating the recommendation and taking a peek inside the RF assembly.

?

The clever RF assembly comprises one PCB, with castings that sandwich the board in between providing circuit isolation.? The service manual does give some information as to what's in there via a coarse block diagram and a page or two of explanation.? I was able to find a shorted transistor (Q6) in what I believe is the 2nd mixer (could be the tail end of the 1st IF).? This was located using the diode function of a DMM; I did not have to power up the assembly.? Normal was about a 0.45V diode drop from "dot" (I thought base, but turned out to be gate) to the other leads.

?

The transistor is a "pill package" 4 lead part with "211" as the sole ID marking.? Some hunches lead to the identity of the part being most likely an HP (Agilent) ATF-21186 GaAs FET, which is used in numerous locations in the front end of the RF assy.? Regarding the 0.45V drop, what I was measuring I believe is a schottky barrier gate structure.? I don't work in RF; FETs in my line are all insulated gate (enhancement mode) devices, so this has been an interesting experience.

?

The bad news was the ATF-21186 was discontinued in 2001.? Newer parts don't use the pill package.? However, I did find a small lot of '21186 parts (9805 lot code) via a helpful broker at about $14 each, with a required lot buy of 16 pieces.? My analyzer is SN prefix "US37.." so the parts are less than a year apart in manufacture date.? Upon arrival I was pleased to find the new parts looked identical to the old.? BTW, a "-STR" suffix on the ATF-21186 designates the part packaged in a strip of 10.?

?

I did not attempt to "match" devices or anything--I just replaced the bad part with one of the new ones.? Input welcome if someone knows better.

?

One tool I found helpful in putting the RF assy castings back together with the zillion M3 screws used is a good torque screwdriver to get everything down evenly.? The tightening torque is 9 in-lbs on everything (casting screws, casting to instrument screws, and instrument case screws).? As there is no give to the housings, it is crucial all screws get tightened down the same amount.? Over tightening one screw will make the adjacent one loose for instance.? It takes two of three passes around them all to get to final torque.? I used a Proto 6106 (now apparently owned by Stanley) driver which supposedly will not over torque screws even with multiple attempts to retighten them.

?

The replacement transistor seems to have solved the problem.

?

One other hint.? The one?piece?rubber membrane key pad has thin walls where the raised key caps drop down to the sheet.? They are easily torn if you do something silly like drive the key cap sideways through overzealous cleaning for instance.? I found a reasonable fix was to build a "hinge" between the torn key cap and base sheet using a flowable silcone product made for car windshield gasket sealing: Permatex Windshield and Glass Sealer.??I found a?1.5oz tube of the stuff available at the local hardware store.

?

Don C

Beaverton, OR

?

?

?

?

-----Original Message-----
From: daveolla [mailto:grobbins@...]
Sent: Saturday, October 30, 2004 5:21 PM
To: hp_agilent_equipment@...
Subject: [hp_agilent_equipment] 3406A sampling meter/peeling faces

Greetings, recieved a 3406A sampling meter free, now I need to
brushup on? what it is good for,? and what benefits from sampling as
opposed to regular RF meters.
The question is concerning those Hp meter faces from a certain era
that peel and crack at the mirror ends (glued so well to brass
mirrored back and contraction forces). There is a clear film that
flakes up and off over the lettering and scale, the flaking stops at
the letters, as if that helps hold it down. The stuff is water
soluable, so if it is wiped it smears. My 428B meter also had this
same problem except letters below also were flaking, had to photo
copy a new one, a pain since I tried dozens of copiers around town
and found most could not copy at 100%,(laid ruler along side
meter face when copied so as to compare ), also cutting out a nice
mirror slot again is shunned.
The flaking has been nicely brushed off and I would like to overspay
(masking the mirror) with something (even water will remelt the film)
fast drying to prevent trouble below, wrinkling etc. Anyone else have
ideas to save face. Iron on films are out of the question since heat
on my 428B meter face caused fine cracks everywhere like fine art
paintings.
Hopefully that is all that is wrong with this meter since the lifting
film stopped the needle from moving......... unless this occurred
after the probe was found to be bad.
Dave Robbins
And In the Halloween spirit, might I say, BOO

-----Original Message-----
From: Scott Newell [mailto:newell@...]
Sent: Saturday, October 30, 2004 8:50 AM
To: hp_agilent_equipment@...
Subject: Re: [hp_agilent_equipment] Can't ID Equipment



At 04:30 AM 10/30/2004 +0000, gerbo04 wrote:

kind? The fan is in the front which is odd. The boards are labeled as
follows:

05345-60044 88802F
05345-60013 744414
05345-60011 88802F
05345-60050 88809F
05345-60009 73803L

Can anyone identify this thing?

As JM explained, it's certainly a 5345A.

If you do decide to part it out, let me know...I'd be interested
in the A12
(HPIB) board (which you may not have--it's not on the list you posted), a
couple of EECL chips to fix my 5370, and maybe the hot rock box.

--
newell N5TNL





Yahoo! Groups Links

-----Original Message-----
From: gerbo04 [mailto:gjwinga@...]
Sent: Friday, October 29, 2004 9:30 PM
To: hp_agilent_equipment@...
Subject: [hp_agilent_equipment] Can't ID Equipment




I came across a piece of HP gear that has the entire faceplate and
front frame missing, got it cheap. Can't identify what it is. It is a
mainframe type package with an empty space on the right side for a
plug-in. Total about 16" wide 5" high and 20" deep. Inside there are
seven plug in boards (eight slots), a big crystal in an oven, and an
impressive power supply and fan. Maybe a signal generator of some
kind? The fan is in the front which is odd. The boards are labeled as
follows:

05345-60044 88802F
05345-60013 744414
05345-60011 88802F
05345-60050 88809F
05345-60009 73803L

Can anyone identify this thing?

Gerry W.







Yahoo! Groups Links

Mike:

If you continue to have trouble with those VCOs, have you considered
an update with the "VCO on a chip" approach? A while back I had a
bit of a "wrestling match" with a Systron Donner freq. synthesizer
that refused to lock reliably, and I refused to lose that match.

A company in Philadelphia called Amplifonix makes some nice VCOs(if
a bit pricey) .
Another approach would be to use one of the Analog Devices PLL chips
that has the on-board VCO, and just use its VCO "core." The ADF4360-
7 looks like it might do the trick for you (if you ignore its
programmable PLL sections, you may need to throw in an ECL flip flop
to divide its output by two). If you go the Analog Devices route
may I also recommend the KOA KQ line of surface mount inductors (I
have found that they approach my s parameter simulator's
(Ansoft) "ideal" inductor up to around 2 GHz in a number of
circuits).

?

Mike:

If you continue to have trouble with those VCOs, have you considered
an update with the "VCO on a chip" approach?? A while back I had a
bit of a "wrestling match" with a Systron Donner freq. synthesizer
that refused to lock reliably, and I refused to lose that match.

A company in Philadelphia called Amplifonix makes some nice VCOs(if
a bit pricey)
Another approach would be to use one of the Analog Devices PLL chips
that has the on-board VCO, and just use its VCO "core."? The ADF4360-
7 looks like it might do the trick for you (if you ignore its
programmable PLL sections, you may need to throw in an ECL flip flop
to divide its output by two).? If you go the Analog Devices route
may I also recommend the KOA KQ line of surface mount inductors (I
have found that they approach my s parameter simulator's
(Ansoft) "ideal" inductor up to around 2 GHz in a number of
circuits).

Last but not least, to take a look at some very creative alternative
VCO circuits and layouts, the older Wavetek sweep generators (I'm
thinking specifically of, I belive it was the 2002A) use wideband
VCOs to get the sweeps (OK, I guess what else would you use to
generate sweeps -- but anyway if you can get a look at a manual for
one of those old generators, they have some interesting ideas).

My "update" circuits end up looking crude because, so far at least,
once I get them working, I leave them as is (I prototype mostly
using the "Jim Williams" solder to the ground plane and let
everthing else hang just above it approach, or using small boards
with manufacturer's suggested layout for particular parts but then
there are always at least a couple more parts, again, just soldered
above the ground plane), but I've always thought that updating could
be a matter of making a small custom circuit board that would be
like your own "hybrid" in essence to replace the functional block
you are updating.

Best of luck,

-Keith

--- In hp_agilent_equipment@..., "factory_peasant"
wrote:
>
> the VCOs are hand built and tested on the line, so they had to be
> sent back into the factory for repair or replacement. through the
> late 90s many of the components vital to these VCOs like the glass
> pin diodes were becoming very difficult to find and were of poor
> quality. your best bet will be to try and locate another of the
VCOs
> from a different unit. they are all identical regardless of 8662A
or
> 8663A. each unit required an identical pair of the VCOs so there's
a
> few of them out there to be sure.
>
>
> --- In hp_agilent_equipment@..., "Mike Feher"

> wrote:
> > I am trying to get my 8662A to work. I bought it knowing that it
had
> > problems.? I now have most of it working but believe that one of
> the VHF
> > (320-640 MHz) VCOs is not behaving the way it should. There are
two
> of these
> > VCOs in the unit, essentially identical. They use a pair of
2N5397
> N channel
> > JFETs in an oscillator and it is varactor tuned. There is no
info
> on them at
> > all in the manuals as it is a factory repair part only. I wonder
if
> any one
> > on here has any experience with them that they could share.
Thanks -
>? Mike??
> >
> >?
> >
> >?
> >
> > Mike B. Feher, N4FS
> >
> > 89 Arnold Blvd.
> >
> > Howell, NJ, 07731
> >
> > 732-901-9193