HP Logic Analyzer Inverse Assemblers

[ahakman]

A question probably mostly for gslick:

Could you compile all of the available inverse assemblers with the DLL wrapper for the 169xx series analyzers if you still have the environment setup to do so? It sounds like setting up that environment is quite the chore.

I'm working on getting my 16903A up and going, and would really appreciate the dll wrapped inverse assemblers for it.
I've repaired the 2 cards I have that are compatible with the 169xx series analyzers, and I'm just waiting on the appropriate probe cables to arrive for those cards.

I'm especially interested in an IA for 68020

[gslick]

I've been meaning to build all of the common INVASM.ZIP collection of 10391B inverse assemblers into DLL versions which can be used with the 1680/1690/16800/16900 series of Windows 2000/XP/WES7 based analyzers. I should be able to get around to doing that soon.

I also need to write up notes with all of the details on how to get set up to do that yourself. I already have a good set of notes, just need to polish them up. It's not too difficult, just a bit of a chore and time consuming, as you suggest.

[ahakman]

Yes, the instructions on how to compile them wrapped in a DLL would be handy too. I read back in the thread and you mentioned you need a specific version of VisualStudio, and that probably has to run on 32 bit WindowsXP.

I happen to have a clean virtual box image of a fresh install of 32 bit WindowsXP, and I could certainly start from there.

[deanclaxton]

With enormous thanks to gslick, I was able to set up the build environment, compile the 6502.S to 6502.R and then build the .dll and installation files for it.

I've installed it, but havent actually given it a proper test yet. I'll test it out then share the installation files for anyone else who may wish to use the 65C02 IA on a 169xx logic analyzer 

[gslick]

Yes, the instructions on how to compile them wrapped in a DLL would be handy too. I read back in the thread and you mentioned you need a specific version of VisualStudio, and that probably has to run on 32 bit WindowsXP.

I happen to have a clean virtual box image of a fresh install of 32 bit WindowsXP, and I could certainly start from there.

If you already have a Windows 2000 or Windows XP VM set up, you can download iso images of the tools you need to install from here:

Microsoft Visual Studio .NET 2003 Enterprise Architect, Disc 1
https://archive.org/details/X09-46214

Microsoft Visual Studio .NET 2003 Enterprise Architect, Disc 2
https://archive.org/details/X09-46216

Microsoft Visual Studio .NET 2003 Enterprise Architect Enterprise Developer Windows Prerequisites
https://archive.org/details/X09-46245

Logic and Protocol Analyzer Software (32-bit)
https://www.keysight.com/us/en/lib/software-detail/instrument-firmware-software/logic-and-protocol-analyzer-software-32bit-2222812.html
Previous Versions
Release Date    Version     Version Description
2007-09-21      03.67.1008   16900, 16800, and 1680/90 Series Application Software (Version 03.67.1008)
Application Software CD (.iso Image, Version 03.67)
Operating System
Windows XP SP2
Windows 2000


The logic analyzer application software after Version 03.67 dropped support for building 10391B style inverse assemblers into dlls using the Analysis Add In Wizard, maybe partially because as-is it was not compatible with versions of Visual Studio after 2003, and they might not have wanted to put the effort into updating the Analysis Add In Wizard for newer versions of Visual Studio.

Only the C++ compiler component of Visual Studio 2003 needs to be installed. All of the other components can be unselected during installation.

[ahakman]

Thank you gslick!

I've downloaded the isos - I will work on getting that stuff installed into my XP VM soon and see if i can get the 68020 IA built

[ahakman]

Ok, so in my XP VM, I installed visual studio and the 3.67.1008 version of the LA software. I also installed SetupIAAnalysisAddInWizard from the LA CD, and followed the instructions in the readme in the docs folder where that installed to and copied the 3 files into the visual studio folder, but now I'm a bit stumped... what to do from here?

I don't see the Analysis Add In Wizard showing up in the LA software anywhere? Normally I'd just keep playing with this and reading the manuals to figure it out, but I really need to go to bed!

Oh, I see, you use it through visual studio, and you load in a .R file (which I have to figure out how to make). Definitely need to look more at this tomorrow after work

Also, daaaaamn, windows XP runs incredibly fast in a VM on modern hardware! Definitely something to be said for old windows before they were full of useless cruft and spyware!

[gslick]

Ok, so in my XP VM, I installed visual studio and the 3.67.1008 version of the LA software. I also installed SetupIAAnalysisAddInWizard from the LA CD, and followed the instructions in the readme in the docs folder where that installed to and copied the 3 files into the visual studio folder, but now I'm a bit stumped... what to do from here?

I don't see the Analysis Add In Wizard showing up in the LA software anywhere? Normally I'd just keep playing with this and reading the manuals to figure it out, but I really need to go to bed!

Oh, I see, you use it through visual studio, and you load in a .R file (which I have to figure out how to make). Definitely need to look more at this tomorrow after work

Yes, after installing Visual Studio 2003, then installing version 3.67.1008 of the LA software, and then separately installing version 3.60.0002 of the Analysis AddIn IA Wizard from the 3.67.1008 CD, you need to manually do this step, as you already discovered:

C:\Program Files\Agilent Technologies\Logic Development\Analysis AddIn Wizard\docs\Readme.txt:

To complete the installation of the wizard please copy the three files listed below from:
C:\Program Files\Agilent Technologies\Logic Development\Analysis AddIn Wizard

AgilentAnalysisWizard.ico
AgilentAnalysisWizard.vsdir
AgilentAnalysisWizard.vsz

to:
C:\Program Files\Microsoft Visual Studio .NET 2003\Vc7\vcprojects\

And then after doing that, the next time you start Visual Studio and create a new project, when you select Visual C++ Projects as the Project Type, one of the Template choices should now include Agilent Analysis Tool.




For the common INVASM set of inverse assemblers, there is a set of those with the .R files extracted from the inverse assemblers. See invasm_v3.zip attached to reply #19 of this thread.

[gslick]

If you have an .S source file for a 10391B style inverse assembler and you want to assemble that into an .R file, you need to install the 10391B assembler. This tool runs in a 16-bit DOS environment. You can't run it in the command prompt windows of newer versions of Windows, which lack a 16-bit environment. You should be able to run it fine in a command prompt windows of a Windows 2000 or Windows XP VM.

You can currently download the 10391B assembler here:

https://www.keysight.com/us/en/lib/software-detail/instrument-firmware-software/10391b-inverse-assembler-development-package-version-0200-sw575.html

https://www.keysight.com/us/en/assets/9018-01037/reference-guides/9018-01037.pdf

Here are some detailed notes below on how it can be installed. In the example below the X: drive is a shared network folder mapped from the the guest VM to a directory on the host VM, where the IA_Development_Disk directory was extracted from the downloaded 10391B zip file.

The INSTALL.BAT file was originally set up to install from the A: floppy drive. Instead of running INSTALL.BAT the files can be manually copied. Note that the AIAL.txt files should not have the .txt extension. Those files must be renamed to remove the .txt extension. Also note that the AIAL file must either be copied to the default location of C:\HP64700\TABLES\AIAL, or the HPTABLES environment variable must be set to the directory containing the AIAL file. If you don't do that, you will get the error "asm: Termination, Unimplemented or invalid processor name (line     0)"

Code: [Select]

C:\>md 10391B

C:\>xcopy /s X:\IA_Development_Disk 10391B
X:\IA_Development_Disk\ASM.EXE
X:\IA_Development_Disk\IALDOWN.EXE
X:\IA_Development_Disk\INSTALL.BAT
X:\IA_Development_Disk\64700\TABLES\AIAL.txt
X:\IA_Development_Disk\EXAMPLES\68010.BAT
X:\IA_Development_Disk\EXAMPLES\68010.CMD
X:\IA_Development_Disk\EXAMPLES\8085.BAT
X:\IA_Development_Disk\EXAMPLES\8085.CMD
X:\IA_Development_Disk\EXAMPLES\I68010.S
X:\IA_Development_Disk\EXAMPLES\I8085.S
X:\IA_Development_Disk\PROGRAMS\ASM.EXE
X:\IA_Development_Disk\PROGRAMS\IALDOWN.EXE
X:\IA_Development_Disk\TABLES\AIAL.txt
13 File(s) copied


C:\10391B>dir /s /b /a-d
C:\10391B\ASM.EXE
C:\10391B\IALDOWN.EXE
C:\10391B\INSTALL.BAT
C:\10391B\64700\TABLES\AIAL.txt
C:\10391B\EXAMPLES\68010.BAT
C:\10391B\EXAMPLES\68010.CMD
C:\10391B\EXAMPLES\8085.BAT
C:\10391B\EXAMPLES\8085.CMD
C:\10391B\EXAMPLES\I68010.S
C:\10391B\EXAMPLES\I8085.S
C:\10391B\PROGRAMS\ASM.EXE
C:\10391B\PROGRAMS\IALDOWN.EXE
C:\10391B\TABLES\AIAL.txt

C:\10391B>ren C:\10391B\64700\TABLES\AIAL.txt AIAL

C:\10391B>ren C:\10391B\TABLES\AIAL.txt AIAL

C:\10391B>dir /s /b /a-d
C:\10391B\ASM.EXE
C:\10391B\IALDOWN.EXE
C:\10391B\INSTALL.BAT
C:\10391B\64700\TABLES\AIAL
C:\10391B\EXAMPLES\68010.BAT
C:\10391B\EXAMPLES\68010.CMD
C:\10391B\EXAMPLES\8085.BAT
C:\10391B\EXAMPLES\8085.CMD
C:\10391B\EXAMPLES\I68010.S
C:\10391B\EXAMPLES\I8085.S
C:\10391B\PROGRAMS\ASM.EXE
C:\10391B\PROGRAMS\IALDOWN.EXE
C:\10391B\TABLES\AIAL



C:\10391B>CD EXAMPLES

C:\10391B\EXAMPLES>..\ASM I8085.S
asm: Termination, Unimplemented or invalid processor name (line     0)

C:\10391B\EXAMPLES>CD ..

C:\10391B>MD C:\HP64700

C:\10391B>XCOPY /S 64700 C:\HP64700
64700\TABLES\AIAL
1 File(s) copied

C:\10391B\EXAMPLES>DIR /S /B /A-D C:\HP64700
C:\HP64700\TABLES\AIAL

C:\10391B>CD EXAMPLES

C:\10391B\EXAMPLES>..\ASM I8085.S

C:\10391B\EXAMPLES>RD /S /Q C:\HP64700

C:\10391B\EXAMPLES>..\ASM I8085.S
asm: Termination, Unimplemented or invalid processor name (line     0)

C:\10391B\EXAMPLES>SET HPTABLES=C:\10391B\64700\TABLES\

C:\10391B\EXAMPLES>..\ASM I8085.S

C:\10391B\EXAMPLES>..\ASM /OX I8085.S > I8085.LST


[ahakman]

Thank You! I had the 10391B inverse assembler development package downloaded, but not installed yet. I think I could've figured it out from going back and reading some posts in this thread again, but as I said, it was late (like 3AM) and I had to get up for a mandatory statistics training course at 8:30 for work this morning (which taught me no statistics I didn't already know  ).

The only other thing I noticed is that the 68020 IA has a _i at the end of it's name indicating it needs special probing hardware. I found the manual for the E2426B 68020 probe shim from HP, and it looks like there's nothing special on the board other than 90k resistors in series with each of the signal lines, and a jumper to pull the cache disable pin on the processor low to disable the cache so you get a full code trace from the bus.
https://nscainc.com/wp-content/uploads/pdf/A_E2426B.pdf

That manual also mentions that there are 2 different IAs for 68020, the "regular" and the "enhanced". It says it used one vs the other depending on the version of software on the logic analyzer itself. I'm guessing the IA that's in the package here is the "enhanced" one?? Keysight has no files available for the E2426A / B probe on their site (of course)

[ahakman]

Ok, I got it built. I'm just installing the 5.90 version of the logic analyzer software in my VM to see if I can install the addin.

You really have to follow the directions in the manual very carefully. I was trying to figure out why I kept getting a bunch of path errors during the build, and was manually adding folders to the include path in visual studio, but then it got to a header file that just didn't exist, and I was like "WTF???"

Then I read the manual closer - open VS, create the project using the wizard (being sure to visit ALL 3 pages of the wizard), close VS, delete the project file, swap in the other project file, re-open VS, THEN Build. Just a slightly kludgy workaround they had there... ahhh, windows in the 90's / early 2000's - brings back not fond memories...

[gslick]

Ok, I got it built. I'm just installing the 5.90 version of the logic analyzer software in my VM to see if I can install the addin.

You really have to follow the directions in the manual very carefully. I was trying to figure out why I kept getting a bunch of path errors during the build, and was manually adding folders to the include path in visual studio, but then it got to a header file that just didn't exist, and I was like "WTF???"

Then I read the manual closer - open VS, create the project using the wizard (being sure to visit ALL 3 pages of the wizard), close VS, delete the project file, swap in the other project file, re-open VS, THEN Build. Just a slightly kludgy workaround they had there... ahhh, windows in the 90's / early 2000's - brings back not fond memories...

Yeah, a couple of other details I didn't mention yet:

First, before you use Visual Studio to create the first IAL wizard project, you need to modify this file:

C:\Program Files\Agilent Technologies\Logic Development\Analysis AddIn Wizard\wizard-files\Templates\1033\AnalysisAddIn_vcproj.txt

There are two lines in that file:

Code: [Select]

AdditionalDependencies="agAnalysisAddIn_i.idl;agAnalysisAddIn2_i.idl;"

Change both of those lines to:

Code: [Select]

AdditionalDependencies="agAnalysisAddIn_i.idl;agAnalysisAddIn2_i.idl;agAnalysisAddIn3_i.idl;"

It appears that there may be errors building an IAL wizard project if those changes are not made before using the wizard to create a new project.

And second, each time you use Visual Studio to create a new IAL wizard project, the newly created project directory should contain a .vcproj file, and a .vcproj.replace file. As the file name suggests, first close the new project, replace the .vcproj file with the .vcproj.replace file, and then reopen the new project. If you forget to do that the project will fail to build with the original .vcproj file.


After you successfully build an IAL wizard project dll, inside the project directory the IA wizard scripts should have created a directory with a create.bat file. That's what you run to create the installation directory. It will copy the IA dll to an installation directory that you specify by defining the shell variable ANALYSIS_DIRECTORY. It will also copy to that directory install.bat and uninstall.bat files, some .reg registry addition files, and RegisterAddIn.exe and UnregisterAddIn.exe utilities used by the install.bat and uninstall.bat files.

To actually install the IA dll you would run the install.bat file from the installation directory created by running the create.bat file.

As-is, the generated .reg registry addition files only work when the logic analyzer application is installed on a 32-bit version of Windows. They can be modified to work when the logic analyzer application is installed on a 64-bit version of Windows.

[ahakman]

Yes, when I was mucking around trying to install the LA suite 5.90.whatever, I had tons of issues, and ended up really borking up my VM image - good thing it's a VM image, and it was just a copy of a good image!

I did all of what you mentioned, except adding in the agAnalysisAddIn3_i.idl; dependency to the AnalysisAddIn_vcproj.txt file. The DLL still seemed to build fine.

I also figured out all the stuff about how to make the install directory from reading the batch files themselves. I copied the directory to the 16903A on a usb stick and ran the intall.bat file, and now it actually shows up in the tools menu (and my actual logic analyzer is running version 4.00.[whatever].

Now I have 2 questions:
Is is there a quick way to import the bus / pod assignments? If I try to add the IA analyzer, I get an error message saying (shortened / paraphrased) "The analysis tool has requested signal ADDR as input, but that signal isn't defined". I could define the bus layouts to the pods manually, but isn't there a way to do that automatically from some of the files in the .zip files at the beginning of this thread? It doesn't seem to like to import either the txt version of the config, or the .p binary version of the config. It's not that big of a deal - it took 2 minutes to define the buses by hand using the .txt version of the config from the zip file

The 68020 IA has 5 pods worth of signals defined. ADDR = 2 pods, DATA = 2 PODS, STAT = 1 POD. I have 2 cards that are 4 pods each, so the first card is entirely the address and data buses (as they're 32 bits each), and then the STAT bus needs to be on the first pod of the second card. So currently I have the cards independent of eachother, but with it setup that way, I can't assign signals or buses from more than one card to the IA at the same time. Is there a way to do this? Or is the only way to link the cards together with the linking cables into one larger analyzer? I would expect that I could just clock both cards from the same clock, and use them in state mode as if they're one larger card, even if they're not linked with the joiner cables, and then assign buses from pods on both cards into the signals the IA is looking for. Is this not the case??

[MarkL]

Unless you're trying to analyze two independent domains, the cards are supposed to be linked together.  This allows multiple cards to operate as one seemless unit, and it takes care of clocking, skew, pattern matching, and triggering across all the inputs.

I think you said in a previous post you had two 16752A cards.  You would need two flat-flex jumpers to combine the two cards.  If you don't have them, you can buy them as a set 16715-60001, or individually as 16715-61603.  There's a pair on ebay right now for $10 + shipping.

[ahakman]

On the 2 cards I repaired, I would have to replace the sockets on the cards for the linking cables as well - they're pretty cruddy looking and corroded.

Anyone know what the connector is on the board side so I could order new ones?

Would you trust these to work? I don't think I would... the only one that is maybe ok is card2 connector 2. The rest look pretty awful

Oddly enough I have 2 more cards here temporarily that are also broken and corroded very badly under the runners, but the linking connectors are perfect on those. Unfortunately those cards will be going back because they were sold as "Used (aka functional)" on ebay, both fail the self tests, and the seller seems to think his broken junk is worth more than I'm willing to pay for the privilege of spending hours under the microscope fixing them.

These things really are a PAIN at every step of the way...

[MarkL]

Those are not in great shape, but I've had worse.  I wouldn't plug anything into them the way they are.  You might ruin the connector on the flat-flex cable.

I would try using a soft bristle brush saturated in IPA (like a toothbrush), and try to gently scrub the green corrosion off the fingers.

If that doesn't work, and I know some people are going to hate this suggestion, you could try a very fine bristle brass brush.  I have had to resort to this in a few cases, but in my view it was a doomed connector anyway, so what did I have to lose.  I recall one connector where it was so bad I had to scrape the corrosion off with a needle finger by finger.  I made a mess of the gold plating, but it did work afterwards.

I've found the connectors on the top side of the board are usually in much better shape.  If you're good at SMD soldering, and you have good connectors on the top, you could try transplanting a pair to the bottom on one of the boards.  You have 8 connectors to work with, and you only need four to get a pair of boards connected (two on the top, and two on the bottom).

I will look around to see if I can figure out the manufacturer, but there are no identifying logos or other tell-tale marks.  The only thing I see is "RN 004" on the flat-flex side.  I seem to recall a connector manufacturer with the initials "RN" long ago, but I can't recall who it was.  But it's not clear that's even the manufacturer.

[ahakman]

I ended up doing exactly what you suggested - I spent a couple of hours this afternoon looking at board to board connectors at Digikey and Mouser and couldn't find a match, so it was on to plan B:

I spent at least a couple of hours under the microscope scraping each individual pin in the 2 connectors on the bottom side of one card with a sewing pin, and a very fine tipped "grinding pen" (if you watch Northridge Fix on youtube, you know what I'm talking about) bit just by hand. I also finished up by running a spatula with printer paper wrapped around it as a final "polish" back and forth in the central slot of the connector. That's a trick I learned many years ago - printer paper is just abrasive enough to use for cleaning up contacts without damaging any of the metal of the connector. Also used the pin and some cotton from a cotton bud and acetone to soak up all the mess, and some compressed air to blow out what was left.

First time I ran the cards together, I had 2 failing self tests, inter-chip resource bus and one of the clocking tests both failing on the slave board. Took it apart again and did a little more scraping, polishing with the printer paper on the spatula, and cleaning again with compressed air, and on the second try, it's working!

I did also use a tiny dremel sized brass brush too - again, by hand, not in any tool that would spin it, and I tried cleaning the contacts on the bottom connectors of the other board without doing any scraping first - it definitely didn't get some of the bigger "chunky" stuff off, but it looks quite a bit better. Scraping each pin individually is definitely the way to go. If I get some more boards in the future and need to use the bottom connectors on that board, I'll have to do some more manual scrapey-scrapey under the microscope to get those contacts in good shape.

All in all, the plating doesn't look perfect anymore, but it is working.

What an absolute journey it's been to get to this!



2 Cards linked together into one "analyzer", the IA compiled and inserted, and actually seeing all of the defined buses now! And 2 sets of the cables I bought showed up today too, so I'm technically ready to run this! Now I need to get the target system ready - that's a bit of a project in and of itself. Tomorrow - I'm way too tired today.

[MarkL]

Great - glad you got everything working!  Popping the stack...

Thanks for the ref on the "grinding pen".  Looks like a useful tool.   I'll have to grab one of those.

[ahakman]

On the Grinding Pen, the one Northridge sells you can find for about half price on Ali Express (of course).
I bought one from Ebay that looked like the same thing but with a different set of bits - I don't think it's the same thing at all. The bits in the ebay one seem considerably worse than the bits in the Northridge / Ali Express one.

I'm waiting on the Ali Express one to actually arrive to be able to compare it to the ebay one.

The bits from the ebay one did come in handy for this little job though...

If you want the good one and you want it reasonably quickly, just support Northridge Fix and buy it from him!

[MarkL]

...
If you want the good one and you want it reasonably quickly, just support Northridge Fix and buy it from him!

I'm going to do that.  There were a few other things in his store of interest, not to mention a nice metal bristle brush for "Circuit boards and corrosion removal".  Just what was needed!

Also, the name of the company with the RN logo (remembering from decades ago printed on wire-wrap IC sockets) was Robinson Nugent.  They are now owned by 3M, and inherited the 0.8mm connectors which are used on the analyzer boards for the interconnects.  See Pak 8 Plug Connectors and Pak 8 Socket Connectors here:

  https://www.3m.com/3M/en_US/p/c/electronics-components/interconnect-products/pcb-connectors/boardmount/i/electronics/

Or, go directly here for the sockets that are on the board:

  https://www.3m.com/3M/en_US/p/d/b30000132/

The part number looks like it would be P08-080-SLxx-A-G, where the xx depends on packaging and vacuum pickup options (see datasheet).  Quest Components has 92 available for US$5.70/ea according to octopart.com.  I've never ordered from them, so I don't know how well they cope with small orders.

Just posting the info in case anyone is ever looking for these again.

[ahakman]

Wow, amazing, you found the right connectors!

You have to be careful though, the one that Quest has for $5.70 each are P08-080-SL??-B-G - the B model is 3mm higher!
At Quest the P08-080-SL??-A that they have in stock is $38 ea.!!!! YIKES

[MarkL]

...
You have to be careful though, the one that Quest has for $5.70 each are P08-080-SL??-B-G - the B model is 3mm higher!
At Quest the P08-080-SL??-A that they have in stock is $38 ea.!!!! YIKES

Oops - thanks for noticing that on the quest listing!

An extra 3mm might be tolerated for the connectors on the top of the board, but that's not usually the ones needed.  If replacement is the only option, it might be better to look for a junked card to salvage a pair of connectors than pay $38 ea (ouch!).  16715A cards have the least capabilities in this family, and can show up fairly cheap as parts mules for the connectors and other parts in common like the comparators.

[AndersG]

On this topic, has anyone ever seen the HP1000 inverse assembler files? They seem to have existed, but apparently never released as a product and had no product number. You just had to ask for them nicely... I have some IAs for my 1631D, but they seem to be the ones that are floating around here, ie 8080, 68000 anf HPIB.

[cmax100000]

Does anyone have a learn string output from a HP 1630a with the inverse assembler loaded?  I have been looking at the configuration output from my LA and I wondered if I loaded a configuration output from another machine would work as if the IA were loaded in through the GPIB.

Thanks in advance.
I didn't see anything on GitHub. 

[AndersG]

You can load them that way. JJust flip a byte in the file. See http://www.dalton.ax/hp1631d/

If that does not work, then I can load an IA, but which one? and dump the learn string.