AnsweredAssumed Answered

VRf - 8153A Blues

Question asked by robert on Nov 25, 1997
Vrf

The following from Andrew is excellent.  Another technique that I frequently
use is to exercise the control panel with the Bus I/O Monitor active and then
use the captured bus traffic as a basis for designing Direct I/O.  This
technique makes use of the panel designer's experience but allows you to shed
the overhead of the panel driver itself.  You can also omit what is not
essential for your particular measurement.

Robert Reavis
Warm Springs Computer Works
Fremont, CA  94539
510-657-8431 Fax and Voice-Mail


>     VRf
>    
>     Dear Sir,
>    
>     I'm afraid that I concur with Mr. Sanders.  Do not use panel drivers. 
>     Panel drivers do a lot of this by themselves, and therefore will cause
>     the problems you have mentioned, and then some more.
>    
>     FYI, when you use a panel driver, you increase the size of your
>     executable dramatically since the whole driver, and all of it's
>     controls, are added to your program.  In addition, the panel driver's
>     are notorious for, to put it gently, improper or incomplete
>     implementation. 
>    
>     It would probably not be a good idea to use Component Drivers either. 
>     Learn about your instrument, and it will be your friend.  ( It sounds
>     like a cliche, but it will make your life mich easier.)  Control
>     everything the instrument is supposed to do with direct I/O. 
>    
>     One additional suggestion though.  As a programming tip: be complete
>     in your intialization of your equipment.  It is far too easy to send a
>     few commands to your instrument, and take adventage of it's present
>     settings to make your measurements.  As your program grows, you may in
>     fact change the instrument settings in other places in your programs. 
>     Then you will start to encounter unexplained bugs because you are not
>     configuring your instruments completely at the time when you need to
>     make a particular measurement. 
>    
>     Best wishes, and happy holiday,
>    
>     Andrew Hamilton
>     andrew_hamilton@hp.com
>
>
>______________________________ Reply Separator
_________________________________
>Subject: Re: VRf - 8153A Blues
>Author:  Non-HP-toms (toms@symca.com) at HP-PaloAlto,mimegw6
>Date:    11/26/97 9:11 AM
>
>
>    
>    
>------------------------
>  From: Paul Bryan <pbryan@abs.net>
>  Subject: VRf - 8153A Blues
>  Date: Wed, 26 Nov 1997 10:58:26 -0500
>  To: VRf Mail <hpvxd_xc@hpislsup.lvld.hp.com>
>    
>    
>>       Hello.  I'm using an 8153A Lightwave Multimeter with an 81536A Power
>> Sensor, and an external 850/1300 LED lightsource. (and HPVEE 4.01 of
>> course.) 
>>      
>> I'm trying to set up a test routine for Multimode patchcords (insertion
>> loss.)  Without using VEE, I would:
>>
>>       -set the appropriate channel (Channel B, in this instance)
>>       -connect two launch cables together between the source and the meter
>>       -set the wavelength to 850 on both, and press the Disp->Ref key
which
>> would 'zero'          the reading 
>>       -set the Units of Measure to dBm for the Absolute Reference
measurement,
>> and           dB for the test measurement
>>
>>
>> The multimeter would now be ready to perform the test, which would go as
>> follows:
>>
>>       -set the wavelength to 850nm on source and meter
>>       -connect the cable under test (via connector couplers) between the
two
>> launch                cables (this will allow me to read the sum insertion
los
>s of both
>>       connectors)
>>       -read the dB display on the multimeter and compare the measurement
to a
>> pass/fail             criteria. (ie a -0.429 dB reading with a fail
maximum as
> -0.500
>> dB)
>>       -set the wavelength to 1300nm on source and meter and repeat the two
abo
>ve
>> steps
>>               I have been told that if a cable is 'zeroed' at 850nm, a
reading
> at 1300
>>       will yield a variation in power measurement which should remain
constant
>> from cable to cable.  This    variation would be deducted from (or added
to)
>> the measurement at 1300nm.
>>       -print a label which includes test results (in dB)
>>
>>
>> Ok, that's the background for my questions.  Over the last 3 weeks I have
>> been fiddling and poking at VEE enough to learn how to control the
>> instrument to do all these things individually.  I familiar the Panel
>> Driver, and Component Driver.  I've learned how to program sliders and
>> radio buttons to control parameters such as Chanel, Wavelength, Unit of
>> Measure etc... 
>>
>> I want to have a button 'RTZ' which will take the current measurement and
>> reference it to zero (but only when prompted, and not before each cable is
>> tested.)  I programmed one using the panel driver.  It worked fine, but
>> when I added a test routine to the program (which prompted 'test/cancel',
>> read the mesurement, and output the result to a display and a data file;
>> then returned to the test/cancel prompt) it also activated the RTZ
routine.
>>  You could imagine the hassle of testing one cable, then needing to
re-zero
>> the machine, or having it happen automatically when you don't want it to.
>>
>> In speaking with someone in tech support, they told me that when a program
>> is run, all input pins on a panel driver will be activated.  Thus, even
>> though I only want to run the 'test', the Disp->Ref input is activated
>> anyway.  They told me that I should use the 'Component Driver' for the RTZ
>> routine, then my Test routine wouldn't activate it.  But I've tried using
>> individual Component Drivers, and even placing Panel Drivers into User
>> Objects to isolate them.  The trouble is, with the individual components
>> isolated, the RTZ routine doesn't have the zeroing effect on the Panel
>> driver I'm using to do the Test Routine. 
>>
>> One of the underlying problems I can't seem to escape is that by having
>> multiple Instrument Drivers, there doesn't seem to be a way to ensure that
>> each one throughout the entire program will be set to the same channel,
>> unit's of measure, wavelength, etc.
>>
>> One more question.  What would be the best way to have the Test Routine
>> prompt for test, store the reading to file, prompt for second test store
>> the reading to file THEN run a dos command line which would activate a
>> label program to pull the two previously stored records as text for the
>> label.  The only part about doing this I'm not sure of is how to let the
>> program know to gather and store two test results before activating the
>> print program.
>>
>> Hope this message is understandable, if anything isn't clear, let me know.
>>
>> Paul Bryan <pbryan@abs.net>
>> MD
>>
>>
>    
>---------------End of Original Message-----------------
>    
>Paul,
>    
>The solution is very simple, don't use panel drivers.  The panel drivers are
of
>some limited value in manually controlling
>instruments (in my opinion anyway), but are more trouble than they are worth
in
>programs.  I have found that the best
>solution is to write my own instrument driver with a set of functions using
dire
>ct I/O. 
>    
>--------------------------------------------------------
>Tom Sanders <toms@symca.com>
>Symtx, Inc.  408-343-3610  fax:408-343-3616
>Date: 11/26/97         Time: 09:11:24
>--------------------------------------------------------
>
>



Outcomes