Is there a simple way to test a mechanical attenuator, like the
HP-355D?

Ken KC2JDY


Ken
(to reply via email
remove "zz" from address)

"Ken" wrote in message
...
Is there a simple way to test a mechanical attenuator, like the
HP-355D?

Ken KC2JDY

The 355's are a piece of cake. You can do it with DC and a voltmeter.

W4ZCB

On Fri, 26 Mar 2004 01:02:41 GMT, "Harold E. Johnson"
The 355's are a piece of cake. You can do it with DC and a voltmeter.


Hmmm. I tried it with an ohmmeter and didn't get what I expected.
Resistance increased to around 100 ohms and stayed there from around
60 dB up.

So, if I put 5VDC across one center pin and ground, I should be
reading voltages to ground at the other pin that correspond to the dB
setting on the dial?

Ken KC2JDY

Ken
(to reply via email
remove "zz" from address)

Well, 5 volts is a little higher than I would apply to a precision
attenuator (E squared /R = W) = 1/2 watt, which is a little tough on that
input quarter watt resistor.

First, terminate the output in 50 ohms. Then the input will read 50 ohms at
any setting. If it doesn't, you've got problems.

Otherwise, yes, 1 volt in, and calculate the output voltage (across 50
ohms!) for each dB step.

Personally, I use a network analyzer at the top rated frequency of the
attenuator and let the cal lab keep it calibrated.

--
Crazy George
Remove N O and S P A M imbedded in return address
"Ken" wrote in message
...
On Fri, 26 Mar 2004 01:02:41 GMT, "Harold E. Johnson"
The 355's are a piece of cake. You can do it with DC and a voltmeter.


Hmmm. I tried it with an ohmmeter and didn't get what I expected.
Resistance increased to around 100 ohms and stayed there from around
60 dB up.

So, if I put 5VDC across one center pin and ground, I should be
reading voltages to ground at the other pin that correspond to the dB
setting on the dial?

Ken KC2JDY

Ken
(to reply via email
remove "zz" from address)

Hmmm. I tried it with an ohmmeter and didn't get what I expected.
Resistance increased to around 100 ohms and stayed there from around
60 dB up.

So, if I put 5VDC across one center pin and ground, I should be
reading voltages to ground at the other pin that correspond to the dB
setting on the dial?

Sure. with 1 volt in, turn to 20 dB attenuation and the voltmeter should
read 0.1 volt.
Bear in mind, the resistors in there are rated 0.5 watt, so don't go
applying 100 volts!
Most often, a damaged step attenuator will have only the first step damaged.
It sacrifices itself and protects subsequent stages.

AHA! Also, don't forget to terminate the attenuator in 50 ohms. That may
have been where you were having trouble. Essentially, the "voltmeter" needs
to present 50 ohms to the pad or things will NOT be nice.

W4ZCB

On Fri, 26 Mar 2004 01:02:41 GMT, "Harold E. Johnson"
wrote:
The 355's are a piece of cake. You can do it with DC and a voltmeter.

Thanks.

Follow-up question: Are the connectors standard chassis mounts? Or
are they some proprietary HP item? I would like to replace one, but
they are not listed in the manual as being one of the replaceable
parts.

Ken C

Ken
(to reply via email
remove "zz" from address)

Follow-up question: Are the connectors standard chassis mounts? Or
are they some proprietary HP item? I would like to replace one, but
they are not listed in the manual as being one of the replaceable
parts.

Well, they probably suspect that some nut would overheat the precision
resistor inside, or move one of the "fudge" tabs that are in there to
provide a controlled capacity that allows them to rate the things to the
frequency they do. (Some are DC to 500 MHz and some are DC to 1 GHz.) A good
attenuator IS a precision piece of test equipment and should be well cared
for.

I've never replaced one, but the thread count is the same as some bulkhead
BNC's that I have, and would expect them to be replaceable if you took care
to NOT do at least one of the above. Look at the replacement connector and
ascertain that it is the same as the one you wish to replace before taking
that giant leap. That particularly pertains to LENGTH of the threaded
portion of the connector as well as the thread count itself. There's not a
lot of spare room on the inside. Just might require some machine shop
turning of available BNC connectors.

I have a few that were from canniballized test equipment. HP 312's, 313's
and old 8640's. some of those have SMA and SMB connectors that were swedged
into the cast frame and were immoveable without drilling out and retapping
for a larger connector. You might investigate what you have from that point
of view, since if they're already swedged BNC's, you can't install type
"N"'s as a one way replacement, there's not enough room.

W4ZCB

"Ken" wrote in message
...
On Fri, 26 Mar 2004 01:02:41 GMT, "Harold E. Johnson"
The 355's are a piece of cake. You can do it with DC and a voltmeter.


Hmmm. I tried it with an ohmmeter and didn't get what I expected.
Resistance increased to around 100 ohms and stayed there from around
60 dB up.
..... Ken KC2JDY

....does .."...on up" mean 60, 70, 80, 90. I don't remember the 355. Must
be a "D" ??

HOLD ON! Did you guys miss this part ? For a resistive attenuator, With
it set between 20dB and 60dB DC resistance will read very close to 50 ohms
from either center pin to ground. Then as the attenuation is reduced to its
minimum the resistance will climb. Both ends should read the same.


Figuring quickly (no pads or ohm meter handy) I may have this wrong, but in
true Usenet fashion will try anyway... Couldn't do it, so I cheated...

Using:
http://www.temcom.com/pages/dBCalc_en.html


Pad Rin (R1 + R3 Output open ckt)
1dB 436
3 150.5
6 83.6
10 61

355C Coaxial Step Attenuator Agilent 355
DC to 1 GHz, 0 to 12 dB in 1 dB steps, BNC connection standard.


355D Coaxial Step Attenuator Agilent 355
DC to 1 GHz, 120 dB in 10 dB steps, BNC connectors.

On Fri, 26 Mar 2004 17:45:28 GMT, "Harold E. Johnson"
wrote:
I've never replaced one, but the thread count is the same as some bulkhead
BNC's that I have, and would expect them to be replaceable if you took care
to NOT do at least one of the above. Look at the replacement connector and
ascertain that it is the same as the one you wish to replace before taking
that giant leap. That particularly pertains to LENGTH of the threaded
portion of the connector as well as the thread count itself. There's not a
lot of spare room on the inside. Just might require some machine shop
turning of available BNC connectors.

It *is* a standard bulkhead/chassis mount. Same threads and depth.
It was a little too tight until I wetted the threads with a little
WD-40. Then it went in easy as could be. Hardest part was removing
what felt like 50 little base screws. This thing is built to
withstand 50 g.

BTW, do you have a spare foot or two to sell? I am sure if I order
from HP it will be 25 cents per, plus $12 s&h.

Ken KC2JDY

Ken
(to reply via email
remove "zz" from address)

BTW, do you have a spare foot or two to sell? I am sure if I order
from HP it will be 25 cents per, plus $12 s&h.

THAT bothered me for a minute, was wondering a foot of what.

No I don't, but if you order some, we could dollar average, I'd like a
dozen.

W4ZCB