HP 83592B hangs 8350B sweeper
My 8350B with 83592B plug-in turns on?with flickering displays that go blank?and hangs the 8350B with no error message,?and no response to any switches on either the plug-in or mainframe. While flickering for a few seconds the 8350B displays show -310 MHz start and 20 GHz stop.
It does the same to a second known working 8350B. All data and address lines are active when hung so I don't understand why the 8350B does not respond.
If I press CW while the displays are flickering it responds. I may have to power down and up again to then enter a CW frequency, it comes up showing CW. Once in this state I can run tests and all front panel entries work. I ran all the A3 tests which show the ROMs and address decoders are OK.
The frequency in CW is high by 13 to 47 MHz going from?50 MHz to 18 GHz (it covers 0.01 to 20 GHz).
The power is way off. The display shows -2 to +23 dBm. Is this normal? (no opt 002 attenuator). ?-2 is actually -17 dbM?and power does not increase until about +2.5 dBm. At +10 dbM?it is actually +9.5 dBm.
Tomorrow I will start a full calibration. Suggestions and tips would be appreciated. Is there a battery backed memory in the 8350B? It seems to remember the setup for quite?a while after being turned off. Is the save-n non-volatile?
Does anyone have a spare cover for the PCBs for sale? The plate that goes on top of the 83592B and has many screws. it also holds the big heat sink in place. It is obvious that someone else's fingers have been inside this plug-in.
|
In addition to what Alan writes: - have a look at the program and see if the array that V(N,2) is an element of, is declared as an integer array. If so, any fractions added to x(1,1) will be lost by equating V(N,2) to your expression. - make sure your scale factor is correctly used (you're still sending in SINT format, i.e. only a "mantissa"). Since you're at least getting the correct x(1,1) values, I guess it's ok. - I still don't understand what you're trying to achieve, especially since you're using the reciprocal of a measurement in a calculation to generate the next value, or so it seems.
Raymond
|
Re: alternatives to 54120b test set
On 10/09/2015 05:00 PM, Chris Tofu dreyfusshudson@... [hp_agilent_equipment] wrote: I think that's accurate. I've seen test sets, not saying which one
(likely the 54121a though) for around 100$. An anomaly? Perhaps, but
deals can be had. I only payed 100$ + shipping for what I got, was in
fairly good shape (though missing the top cover). The guy, seemingly a
bona fide test equipment reseller, had the voltage swtich on 220vac.
Maybe he had some kid look it over. I didn't have the heart to tell him.
Probably in his best interests though. The umbilical cable alone can run
150$. Wow, sounds like the guy had too much blood in his caffeine stream. It would be nicer to have a self contained unit like the 54110/11/12,
but the specs are so anemic by comparison. It's not the last scope I'm
going to own anyway I'm sure. Down the road I'll either splurge for the
test set and/or one of the earlier color HPs. There's always the 2215a I
payed 4$ for. Has a beam that's all I can say for now. Yeah. Just keep in mind that the 54120B is *not* a general-purpose oscilloscope. It's a pretty specialized instrument. Those NMD connectors on the test set should give that away. Hang onto that 2215A! -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
Re: alternatives to 54120b test set
?
54123a - 34 ghz
54121a - 20 ghz
54121a - 12.5 ghz <<< typo, should have been 54122a
|
Re: alternatives to 54120b test set
54123a - 34 ghz
54121a - 20 ghz
54121a - 12.5 ghz
?I think that's accurate. I've seen test sets, not saying which one (likely the 54121a though) for around 100$. An anomaly? Perhaps, but deals can be had. I only payed 100$ + shipping for what I got, was in fairly good shape (though missing the top cover). The guy, seemingly a bona fide test equipment reseller, had the voltage swtich on 220vac. Maybe he had some kid look it over. I didn't have the heart to tell him. Probably in his best interests though. The umbilical cable alone can run 150$.
?It would be nicer to have a self contained unit like the 54110/11/12, but the specs are so anemic by comparison. It's not the last scope I'm going to own anyway I'm sure. Down the road I'll either splurge for the test set and/or one of the earlier color HPs. There's always the 2215a I payed 4$ for. Has a beam that's all I can say for now.
|
Re: alternatives to 54120b test set
On Fri, Oct 9, 2015 at 2:16 PM, Dennis Czelusniak <czelusniakd@...> wrote:
Looking at Block Diagram of 4 Channel Test Set for 54120 Mainframe shows it contains Horizontal (A1) and Vertical (A2) Assembly.
Where did you see a 4 Channel test set for under $100?
That's an achievable price for a 54121A. Here's one that went for $105 + ship:
|
I will give that a try on Monday.? I might be able to tell the DVM to do some math too - but I think that slows down the output.
|
Re: alternatives to 54120b test set
Looking at Block Diagram of 4 Channel Test Set for 54120 Mainframe shows it contains Horizontal (A1) and Vertical (A2) Assembly.
Where did you see a 4 Channel test set for under $100?
Dennis
--------------------------------------------
toggle quoted message
Show quoted text
On Fri, 10/9/15, Chris Tofu dreyfusshudson@... [hp_agilent_equipment] <hp_agilent_equipment@...> wrote:
Subject: [hp_agilent_equipment] alternatives to 54120b test set
To: "Hp_agilent_equipment" <hp_agilent_equipment@...>
Date: Friday, October 9, 2015, 2:50 PM
Hello
group, just acquired a 54120b, seems to "work"
fine (was told by seller it didn't power up). Is the
designated test set the only one that will work? Are there
workarounds/juryrigs? Sounds stupid I know. My needs
aren't particularly demanding at the moment, basically
staring at square waves under 10mhz.
?I
seemed to see a message on this screen when I powered the
unit on for
the first time, something to do with calibration I think.
Does this mean
calibration info was lost (being the test set wasn't
present and
connected)? I saw something in a pdf service manual, stating
the unit
shouldn't be powered on w/o a test set connected, but up
to that point
didn't say why. It's not that I need dead specific
calibration currently, just wondering what's the scoop.
?I'm
not spending any or much more money. This thing was a
bargain, and I'll
leave it alone and use a beat up old Tek 2215a for the time
being.
Would love to start playing with it though. I realize you
can sometimes
find these test sets (54123a, but don't quote me) for
100$ on ebay. Just
that I need a car at the moment . . .
?Would love to hear of unorthodox methods of
connecting a probe, if they exist. The umbilical cord was
supplied. My guess is some kind of pre-amp is necessary, as
this unit isn't configured like the 54110 or equivalent
models. Didn't do a lot of research (or even look at the
pictures!) when choosing this item. I used a 54100 or
something many years ago in a testing lab. This unit, with
the proper test set, is so high in bandwidth, and so
inexpensive I'm hardly experiencing buyer's
remorse.
?In theory could the front end of a 54110,
54111, 54112 be retrofitted to this scope? I'm really
interested in learning everything I can about these models.
Was warned against HP scopes, as they use proprietary micros
and whatnot (mind features an extremely proprietary Motorola
68000!!!). I LOVE THE DAMEND THINGS.
?Thanks in advance for any help or info.
|
?
You are getting it! but in integer arithmetic. If
you convert to floating point at that part of the program you will will most
probably get the same result. You need to do division arithmetic before any
scaling and conversion to integers.
?
The fractional result of?1/X(1,1) is obviously
less than 1. It is also less than 1 when you use 3/X(1,1). this?can be only
scaled?before being?converted to a single precision integer. We
suspect you are probably trying to modify the wrong part of the program. But
without sight of the whole code we are working blind.
?
as a trial try...3000/X(1,1) and see what you
get.
?
Alan
G3NYK
?
----- Original Message -----
Sent: Friday, October 09, 2015 8:43
PM
Subject: [hp_agilent_equipment] Re: I
need a SINT value!
So.....?
how do I get the value of??? (1/x(1,1,))*3? that I need?
|
So.....? how do I get the value of??? (1/x(1,1,))*3? that I need?
|
To add to? Alan's answer and assuming X(1,1) is always an integer (which I'm not sure of, without knowing your program)? and V(N,2) is an integer (because it's sent as SINT):
If X(1,1)=0, 3/X(1,1) will throw a divide by zero error
Other results of 3/X(1,1) will either be 0, 1 or 3; all X(1,1) > 3 will give a result 0
So, if V(N,2) is what you send, any value of X(1,1) > 3 will send x(1,1).
Raymond
|
Re: alternatives to 54120b test set
On 10/09/2015 02:50 PM, Chris Tofu dreyfusshudson@... [hp_agilent_equipment] wrote: Hello group, just acquired a 54120b, seems to "work" fine (was told by
seller it didn't power up). Is the designated test set the only one that
will work? Are there workarounds/juryrigs? Sounds stupid I know. My
needs aren't particularly demanding at the moment, basically staring at
square waves under 10mhz. 10MHz?! The 54120B is a 20GHz sampling oscilloscope/TDR system with risetimes in the low double-digits of picoseconds. The test set is the input sampler. "Can't get there from here"...you need the test set or the oscilloscope is a brick. I'm not spending any or much more money. This thing was a bargain, and
I'll leave it alone and use a beat up old Tek 2215a for the time being.
Would love to start playing with it though. I realize you can sometimes
find these test sets (54123a, but don't quote me) for 100$ on ebay. Just
that I need a car at the moment . . . Certainly understandable prioritization, but it's probably best to put it on a shelf until you can score a test set. Would love to hear of unorthodox methods of connecting a probe, if they
exist. The umbilical cord was supplied. My guess is some kind of pre-amp
is necessary, as this unit isn't configured like the 54110 or equivalent
models. Didn't do a lot of research (or even look at the pictures!) when
choosing this item. I used a 54100 or something many years ago in a
testing lab. This unit, with the proper test set, is so high in
bandwidth, and so inexpensive I'm hardly experiencing buyer's remorse.
In theory could the front end of a 54110, 54111, 54112 be retrofitted
to this scope? Not even close. I'm really interested in learning everything I can about
these models. Was warned against HP scopes, as they use proprietary
micros and whatnot (mind features an extremely proprietary Motorola
68000!!!). I LOVE THE DAMEND THINGS. I hate to sound contrarian...but "proprietary"? They were used in hundreds of models of computers, no more proprietary than x86...but not suicide-inducing to program or try to get performance out of. ;) All this said, though, I have a 54120B, and I absolutely love it. It is a BEAST for TDR work...absolutely amazing capability and performance. With the faster test sets (which are not terribly common, unfortunately) it goes up to 34GHz. -Dave -- Dave McGuire, AK4HZ New Kensington, PA
|
alternatives to 54120b test set
Hello group, just acquired a 54120b, seems to "work" fine (was told by seller it didn't power up). Is the designated test set the only one that will work? Are there workarounds/juryrigs? Sounds stupid I know. My needs aren't particularly demanding at the moment, basically staring at square waves under 10mhz.
?I
seemed to see a message on this screen when I powered the unit on for
the first time, something to do with calibration I think. Does this mean
calibration info was lost (being the test set wasn't present and
connected)? I saw something in a pdf service manual, stating the unit
shouldn't be powered on w/o a test set connected, but up to that point
didn't say why. It's not that I need dead specific calibration currently, just wondering what's the scoop.
?I'm
not spending any or much more money. This thing was a bargain, and I'll
leave it alone and use a beat up old Tek 2215a for the time being.
Would love to start playing with it though. I realize you can sometimes
find these test sets (54123a, but don't quote me) for 100$ on ebay. Just
that I need a car at the moment . . .
?Would love to hear of unorthodox methods of connecting a probe, if they exist. The umbilical cord was supplied. My guess is some kind of pre-amp is necessary, as this unit isn't configured like the 54110 or equivalent models. Didn't do a lot of research (or even look at the pictures!) when choosing this item. I used a 54100 or something many years ago in a testing lab. This unit, with the proper test set, is so high in bandwidth, and so inexpensive I'm hardly experiencing buyer's remorse.
?In theory could the front end of a 54110, 54111, 54112 be retrofitted to this scope? I'm really interested in learning everything I can about these models. Was warned against HP scopes, as they use proprietary micros and whatnot (mind features an extremely proprietary Motorola 68000!!!). I LOVE THE DAMEND THINGS.
?Thanks in advance for any help or info.
|
?
the 3/X(1,1) "not working" is probably due to mixed
formats and 3/X(1,1) is probably correctly giving a value of zero. Dividing
intergers returns an integer and dumps any fractional residuals.
?
Alan
G3NYK?
----- Original Message -----
Sent: Friday, October 09, 2015 6:57
PM
Subject: [hp_agilent_equipment] Re: I
need a SINT value!
Thanks
to all for your help!? You pointed me in the right direction and helped
me find the many errors in my thinking.
As it turns out the values I
used in the past were not SINT at all. The 558 I added was decimal is divided
in later processing by 300 to give? the final result I needed at the time
(which added 1.86%).
So yes, I can add and multiply to the x(1,1)
number and that works!
But, the final formula of
V(N,2)=(3/X(1,1))+x(1,1)?? which seems like it should work, does not
work.? I just get the result of? the value of x(1,1).
What am
I missing now?
Thanks!
|
Thanks to all for your help!? You pointed me in the right direction and helped me find the many errors in my thinking.
As
it turns out the values I used in the past were not SINT at all. The
558 I added was decimal is divided in later processing by 300 to give?
the final result I needed at the time (which added 1.86%).
So yes, I can add and multiply to the x(1,1) number and that works!
But, the final formula of V(N,2)=(3/X(1,1))+x(1,1)?? which seems like it should work, does not work.? I just get the result of? the value of x(1,1).
What am I missing now?
Thanks!
|
We really know too little about your configuration and programs (you didn't mention your 1/X requirement any more) but here's another attempt to help, based on assumptions that me be incorrect:
The HP 3458A User Manual explains that SINT does in fact stand for Single (precision) INTeger. It uses two bytes, representing values as two's complement (-32768 to +32767). Obviously, there's a scale factor. The "ISCALE?" command is used to retrieve it.
SINT makes it possible to continuously send two-byte integers ("mantissa") and only get the "exponent" of the measurement on request. This is asking for trouble unless values stay within a known range and the HP 3458A doesn't switch ranges (automatically). Using SINT in any calculation is a bad idea.
The examples on page 100 of the HP 3458A User Manual suggest that doing calculations with the data should be easy. The SINT values can be treated like any other number. There should be no difference just because it's SINT and conversion is implied: The "MFORMAT SINT" command (page 96) tells the 3458A to output (calculated?) values as SINT.
You seem to have made changes to the existing program in the HP
3458A. A listing of the program in that area may be helpful. It may tell us how "588" represents (plus or minus?) 0.051 V.
From what I read, HPBASIC has no problem reading and correctly interpreting SINT numbers.
It may be easier for you to write a program that takes the BDAT file, does the scaling and offset calculations and writes a new BDAT file. Unfortunately, you said that you don't know anything about programming...
Raymond
|
?
Hi ok the X array is stored in SINT (I also believe
this is signed integer where numbers below 8xxx in hex are positive and above
are negative) format. BUT your offset value of ?0.051 is not an interger so
cannot be represented in SINT format.....UNLESS there is a scaling
factor?somewhere else in the program..... say this was 1000
(millivolts?)? then 51(decimal) converts to 33(hex) if the sign bit is zero
for positive.? 558(hex) converts to 1368(decimal)? So I am afraid you
need to understand a lot more about what the program is doing.
?
If you want to multiply the array value by three
the easiest way without converting to decimal is perhaps to use
X(1,1)=SHL(X(1,1)) + X(1,1)? ....3 (decimal) being 11(hex)? where SHL
is a "shift left" function equivalent to multiplying by 2. I dont know if such a
function is available, or if there are other functions for manipulating singed
integers. The reason for their use seems to be to pack the information into 2
bytes for transfer instead of 4bytes for floating point.
?
Again this assumes I understand the problem
correctly
?
Alan
G3NYK
?
----- Original Message -----
Sent: Friday, October 09, 2015 3:33
AM
Subject: Re: [hp_agilent_equipment] Re: I
need a SINT value!
Here
is the background of this golden oldie system I have to make work.:
A Viper
card (with onboard HP Basic and integrated HPIB) in a PC sends a command to
the HP DMM which causes it to respond by sending via HPIB the current voltage
in SINT format. So? X is already in SINT format. In this config, when I
just added 0.051 to X(1,1) , I did not get the expected results.
When
I? use? x(1,1)+558? I get the desired value.
SO.... I am
thinking I need x(1,1)* (the SINT value for 3)
(BTW you are right- I
don't know beans about programming- I just need to change the existing
pgm!)
The program then goes on to get another value from a different
instrument, combine to 2 values in matrix(?) format, then create a BDAT file.
?This file must be analyzed by another program for which I do not
have source code, so I am stuck with maintaining accepting SINT output from
the DMM (unless there is a way to use HPBasic to convert another format from
the DMM to SINT).
I think the original programmer wanted to maximize
the data throughput by use of SINT and BDAT.
|
Battery might be getting weak... It would have been a good idea for him to record the actual error message.
|
Re: Comment on the UGSimple USB to GPIB adapter
" The unit will drive up to just 5 instruments. I suppose it is fair value?compared with some other offerings of this function."
Reason being that each instrument sources 5V through 3k. You need to pull that below 0.8V. Say that you need .5V to be on the safe side. That means you need to sink 1.5mA per instrument on the bus.
So, how many you can handle depends on the chosen microcontroller as none of the cheap GPIB interfaces, including mine, use the dedidated SN75160/75161 chips.
|
Re: Comment on the UGSimple USB to GPIB adapter
Hi Alan, ??????????
sequence and display the responses.> there is truly fantastic freeware windows batch automation program called Autoit,for integrating disparate windows programs not designed to work together.
This would allow you to easily test sequences of scripts using the interface you describe.
Autoit was originally developed for systems admin to do automated installations, but it can do so much more now.
This allows you to? exec software and feed keystrokes into specific windows etc. It is a very full scripting language with a Basic like syntax.? It runs rings around traditional windows scripting hosts (like vbscript,jscript), for conventional scripting and then can identify windows and buttons on windows and press them etc.delay ,wait on window focus,change window focus,wait for window to popup etc etc. It even can do simple UI's.
I have used it for creating really complex scripts for integrating a number of different programs on windows, including complex remote terminal logins for networked Unix boxes and automatic import of thousands of medical records every evening.
Although writing? software like this for scripting a UI, leaves a bad taste in the mouth,it certainly is very useful for integrating different programs not designed for integration.
The price of the HP usb clones has been dropping, so now you can get the ones that look the same as real HP units for as little as $80 on ebay.? Still more? than you paid, but *if* they work with the HP drivers maybe worth the extra $.
I am now suspicious the "Agilent" unit I bought quite a while ago, may be a clone too, even though it has worked reliably with John Miles software and my own Visa software.
Tim
From: "'Alan Melia' Alan.Melia@... [hp_agilent_equipment]"
To: hp_agilent_equipment@...
Sent: Thursday, October 8, 2015 4:46 PM
Subject: [hp_agilent_equipment] Comment on the UGSimple USB to GPIB adapter
?
I have just received this unit so not a lot of time to evalute it yet.
The user manual is a bit thin but adequate in most cases. I initially
bemoaned the lack of an example but bearing on the very different commands
for different types of equipment it seems to work using the basic command
window. However it does not seem possible to stack a few commands and reads.
The problem also is the window clears the address once send command has been
pressed. If you call for a sample on the non-HP unit I was testing with, the
response data has been sent before you can reset the address and press
read-data. Maybe I am using the window incorrectly, but maybe a "send
command and read response" button would we more useful. To some extent this
is not a problem as serious use would involve writing a program to use the
API provided (for Windows) There does not seem to be any support for Linux,
which is a limitation.
It would be nice if the basic command window would read a text file command
sequence and display the responses. Above all it would make the unit easier
to check out before the effort was put into program development. It would
also make checking units with some of the "odd" early implementations
easier.
I am a novice at GPIB so perhaps I have missed something, but there is
little space in an 8 side "manual" for much information.
The unit will drive up to just 5 instruments. I suppose it is fair value
compared with some other offerings of this function.
Alan
G3NYK
|
Dave McGuire
Anders