HP41 TULIP4041 update

[MeindertKuipers]

Many of you may have noticed that Raspberry Pi have recently introduced the RP2350 microcontroller as a next generation of the RP2040. The TULIP4041 was scheduled to use the RP2040. There may be some advantages to use the newer processor and I will first test the Development Board setup with the RP2350 to find out if switching over makes any sense.
This means that I am holding off the orders for the Development Board until I have verified operation with the RP2350. If you have ordered a Development Board you should have received this information.
If I do switch over to the new processor it will most likely delay the final module version as well.

In the meantime a presentation about TULIP4041 will be given by Sylvian at the HHC conference in September, with a real live demo. I am working on the presentation and documentation as well.

[Sylvain Cote]

The board was delivered last week and the final assembly was completed early this week.

My complete test setup is:

• host computer
  
  • Hardware
    
    • Raspberry Pi 4 Model B Rev 1.5 - 8GB RAM - Cortex-A72 with 4 cores (aka RP4B)
    • Armor Case Passive Cooling or Active Cooling
    • 128GB microSDXC SAMSUNG PRO Endurance
    • Raspberry Pi 15W USB-C Power Supply
    • Optional: mouse, keyboard and monitor. -> video: micro-HDMI-male_to_standard-HDMI-female_adapter and a standard-HDMI-male-male_cable
      
  • Software
    
    • Raspberry OS 64-bit → installed on microSD card with Raspberry Pi Imager
      OS: Debian GNU/Linux 12 (bookworm) | Kernel: Linux 6.6.31+rpt-rpi-v8 | Architecture: arm64
    • Special configuration:
      
      • default headless video set to 1280x1024
      • wired network static configuration for: IP, default gateway and DNS
      • default vnc server running
      • RealVNC client installed
        
  • Optional: Development & Debugging
    
    • Raspberry Pi Debug Probe
    • Raspberry Pi Pico development tools C/C++ SDK
      installation script
      
      Code:
      

      mkdir -p ~/temp
cd ~/temp
wget https://raw.githubusercontent.com/raspberrypi/pico-setup/master/pico_setup.sh
chmod +x pico_setup.sh
./pico_setup.sh
    • testing → follow instructions given in the Raspberry Pi Debug Probe documentation
      
      • install a Raspberry Pi Pico 1 (RP2040) on a breadboard
      • connect the Raspberry Pi Debug Probe to the Pico to the SWD connector
      • connect the Raspberry Pi Debug Probe to the Pico to the RP2040 GPIO RX/TX/GND pins
      • connect the Raspberry Pi Debug Probe to the USB port of the Raspberry Pi 4
      • compiled application files location
        
        • blink → ~/temp/pico/pico-examples/build_pico/blink
        • hello world serial → ~/temp/pico/pico-examples/build_pico/hello_world/serial
          
      • notes
        
        • OpenOCD, SWD, GPIO serial port and GDB works perfectly
        • I have tested several applications and did not find any issue
          
  • Virtual HP-IL
    
    • lifutils (utilities runs natively)
      installation steps
      
      Code:
      

      mkdir -p ~/temp
cd ~/temp
git clone https://github.com/bug400/lifutils.git
cd lifutils
mkdir rp4b
cmake -DCMAKE_BUILD_TYPE=Release -S . -B rp4b
cd rp4b
make
sudo make install
    • pyilper (application run natively, no virtual machine needed, well .... apart from cpython of course)
      installation steps
      
      Code:
      

      python3 -‑version                   // installed with the OS and I get version 3.11.2
python3 -m venv ~/py311
source ~/py311/bin/activate
python -m pip install --upgrade pip
pip install pyserial
pip install pyilper
      testing:
      
      
      1. connect PIL-Box to RP4B → serial port added: /dev/ttyUSB0
      2. start pyilper → from command line
         
         Code:
         

         source ~/py311/bin/activate
pyilper -‑clean -‑instance test
      3. configure pyilper to use /dev/ttyUSB0
      4. adds and configure virtual IL devices
      5. leave pyilper
      6. restart pyilper → from command line
         
         Code:
         

         source ~/py311/bin/activate
pyilper --instance test
      7. test virtual devices with an HP-41/71B/75
         

Tulip4041

• tomorrow
  

edit: typos & formatting

[Geoff Quickfall]

Very much looking forward to the presentation!

Cheers, you two!

Geoff

[John Ioannidis]

What's the point of telling us all this if you are not posting any code?

Also, why tell us about the hosting environment? The Pico SDK runs on pretty much everything. Is it really relevant what kind of power supply your Pi4 uses or how much storage it has?

[Sylvain Cote]

(08-25-2024 11:07 PM)John Ioannidis Wrote:  What's the point of telling us all this if you are not posting any code?

Meindert is making a new product and he is just telling the peoples that have reserved the prototype boards that it will take a little more time before he can shipped them.
Also, he never said that he will publish his source code, nor that he have to.

(08-25-2024 11:07 PM)John Ioannidis Wrote:  Also, why tell us about the hosting environment? The Pico SDK runs on pretty much everything. Is it really relevant what kind of power supply your Pi4 uses or how much storage it has?

When I will report my test results, I want Meindert to know exactly what is my testing environment.
The goal of using a Raspberry Pi 4B was to have a neutral machine that is not polluted by the software I use on my main machine and that is dedicated to this task only.
The information is there to allow Meindert or anyone else to build of copy of my testing environment if he/they choose to do so.
Finally, there were some other RP4 projects on this forum that were not able to run lifutils and/or pyilper natively, so I added the step on how to build/install them.

I was planning to report my first test results, but my wife reminded me that I had to work on the house this weekend, so the next post will come later this week.

[Geoff Quickfall]

Ha, happy wife, happy life.

Merci beaucoup! for the explanation Sylvain!

Cheers[/b]

[MeindertKuipers]

(08-25-2024 11:07 PM)John Ioannidis Wrote:  What's the point of telling us all this if you are not posting any code?

No worries, there will be source code published of the version that will go out to the users of the Development Board. I do want to make certain that it is in a useable state. And my earlier offer still stands, if you are interested in the sources I will happily share it with you.

[Etienne Victoria]

(08-24-2024 02:12 AM)Sylvain Cote Wrote:  The board was delivered last week and the final assembly was completed early this week.

My complete test setup is:
[list]
[*]host computer
[list]
[*]Hardware

Thanks Sylvain!

Your description of the test setup is very useful to me and helps greatly to anticipate the arrival of the board.

Best

Etienne

[Sylvain Cote]

Tulip4041 testing has finally started on my side ... Yeah!

Startup was a little rocky at first but now everything seems to works.

The Raspberry Pi Pico (RPP) USB virtual serial ports is seen as ttyACM instead of ttyUSB on the Raspberry PI 4B.
I am able to connect on the multiple serial terminals with minicom but pyilper only uses ttyUSB devices and ignore the ttyACM ones.
Fortunately it works if I manually change pyilper configuration file and set it to "/dev/ttyACM2".
I just posted a request to Joachim to add the ttyACM devices to his PIL-Box serial interface configuration dialog.

USB virtual serial ports roles are ...

Code:

/dev/ttyACM0 → CDC port 0: console, user interface
/dev/ttyACM1 → CDC port 1: trace listing
/dev/ttyACM2 → CDC port 2: HP-IL/ PILBox frames
/dev/ttyACM3 → CDC port 3: HP-IL scope, PILBox frame monitor
/dev/ttyACM4 → CDC port 4: printer bytes

1) with minicom, I connect to the main console (/dev/ttyACM0) and control the Tulip4041 features through its commands.

Code:

 * help
        Print list of commands
 * system
        TULIP4041 system status and control
        [no argument] shows system status
        status        shows system status
        pio           shows pio status
        cdc           shows status of CDC interfaces
        cdcident      identify all CDC interfaces
        REBOOT        restart the TULIP4041 system
        BOOTSEL       restarts the TULIP4041 into BOOTSEL mode for firmware update
        poweron       drive ISA for 20 usecs to switch HP41 on
        calcreset     drive PWO to reset HP41
 * clear
        Clears the console
 * blink
        blink [b], blink the LED b times, just for testing and fun
 * dir
        dir [subdir], shows uSD card directory
 * sdcard
        SD card functions
        [no argument] shows the SD card status and mounts the card
        status        shows the SD card status and mounts the card
        mount         mounts the SD card
        unmount       unmounts the SD card
        mounted       check if the drive is mounted
        connect       connect the SD card as USB drive to the host computer
        eject         disconnect the SD card as USB drive from the host computer
 * plug
        plug functions
        [no argument] shows the current plugged ROMs
        status        shows the current plugged ROMs
        hpil          plug the HP-IL module in Page 7
        ilprinter     plug the HP-IL Printer in Page 6 (unplug HP82143A printer)
        printer       plug the HP82143A printer in Page 6 (unplug HP-IL printer)
 * unplug
        plug functions
        [no argument] shows the current plugged ROMs
        status        shows the current plugged ROMs
        hpil          unplug the HP-IL module from Page 7, unplug the HP-IL Printer
        ilprinter     unplug the HP-IL Printer from Page 6
        printer       unplug the HP82143A printer from Page 6
 * printer
        printer functions for the HP82143
        [no argument] shows the HP82143A status
        status        shows the HP82143A status
        power         toggle power
        trace         enable trace mode
        norm          enable norm mode
        man           enable man mode
        paper         toggle Out Of Paper status
        print         push PRINT button
        adv           push ADV button
 * xmem
        Extended memory functions
        status        shows the Extended Memory status
        dump          creates a dump of Extended Memory
        pattern       programs a pattern for FRAM test
        ERASE         earse all Extended Memory
        [n]           n= 0, 1, 2 
                      plugs 0, 1 or 2 Extended Memory Modules
                      do not use in combination with physical Extended Memory modules
                      this software version does NOT have the Extended Function ROM built in
                      Extended Functions must be physically plugged or available in an HP41CX
 * tracer
        tracer functions
        [no argument] shows the tracer status
        status        shows the tracer status
        trace         toggle tracer enable/disable
        sysloop       toggle tracing of system loops (RSTKB, RST05, BLINK01 and debounce)
        sysrom        toggle system rom tracing (Page 0, 1, 2, 3, 5)
        ilrom         toggle tracing of Page 6+7
        hpil          toggle HP-IL tracing to ILSCOPE USB port
        pilbox        toggle PILBox serial tracing to ILSCOPE serial port
        ilregs        toggle tracing of HP-IL registers
        save          save tracer settings

2) with minicom, I see trace listings (/dev/ttyACM1) activated by the main console

Code:

   119940  E119  1  087  0.0000000DC4.0.00  ...  R000  C0  FI0123456789ABCD  JC 16             
   119941  E129  1  09C  0.0000000DC4.0.00  ...  R000  C0  FI0123456789ABCD  R= 5              
   119942  E12A  1  310  0.0000000DC4.0.00  ...  R000  C0  FI0123456789ABCD  LD@R C            
   119943  E12B  1  390  0.0000000DC4.0.00  ...  R000  C0  FI0123456789ABCD  LD@R E            
   119944  E12C  1  350  0.0000000CC4.0.00  ...  R000  C0  FI0123456789ABCD  LD@R D            
   119945  E12D  1  36E  0.0000000CE4.0.00  ...  R000  C0  FI0123456789ABCD  ?A#C ALL          
   119946  E12E  1  12F  0.0000000CED.0.00  ...  R000  C0  FI0123456789ABCD  JC 37             
   119947  E153  1  28C  0.0000000CED.0.00  ...  R000  C0  FI0123456789ABCD  ?FSET 7           
   119948  E154  1  1BF  0.0000000CED.0.00  ...  R000  C0  FI0123456789ABCD  JC 55             
   119949  E155  1  130  0.0000000CED.0.00  ...  R000  C0  FI0123456789ABCD  LDI               
   119950  E156  0  308  0.0000000CED.0.00  ...  R000  C0  FI0123456789ABCD  308               
   119951  E157  1  236  0.0000000CED.0.00  ...  R000  C0  FI0123456789ABCD  C=C+1 XS          
   119952  E158  1  106  0.0000000CED.3.08  ...  R000  C0  FI0123456789ABCD  A=C S&X           
   119953  E159  1  3B8  0.0000000CED.4.08  ...  R000  C0  FI0123456789ABCD  READ 14(d)        
   119954  E15A  1  3B0  0.000043C048.0.01  ...  R000  C0  FI0123456789ABCD  C=C AND A         
   119955  E15B  1  366  0.000043C048.0.01  ...  R000  C0  FI0123456789ABCD  ?A#C S&X          
   119956  E15C  1  033  0.0000000000.0.00  ...  R000  C0  FI0123456789ABCD  JNC +6            
   119957  E15D  1  3B8  0.0000000000.0.00  ...  R000  C0  FI0123456789ABCD  READ 14(d)        
   119958  E15E  1  358  0.000043C048.0.01  ...  R000  C0  FI0123456789ABCD  ST=C              
   119959  E15F  1  0B0  0.000043C048.0.01  ...  R000  C0  FI0123456789ABCD  C=N

3) with pyilper connected to /dev/ttyACM2, I see all the HP-IL commands going in and out in the pyilper scope window.

Code:

DAB 10 (010) dab f0 (0f0) DAB F0 (0F0) eto    (540) LAD 01 (421) lad 01 (421)
RFC    (500) rfc    (500) RFC    (500) rfc    (500) UNL    (43F) unl    (43f)
RFC    (500) rfc    (500) RFC    (500) rfc    (500)

4) with minicom, I see PILBox trace listings (/dev/ttyACM3) activated by the main console (same listing as the scope window but in different format)

Code:

 PILBox > 0010 DAB   hi: 20 lo: 90
  PILBox < 00F0 DAB   hi: 22 lo: F0
 < 0F0 DAB
 > 0F0 DAB
  PILBox > 00F0 DAB   hi: 22 lo: F0
  PILBox < 0540 ETO   hi: 34 lo: C0
 < 540 ETO
 > 421 LAD
  PILBox > 0421 LAD   hi: 30 lo: A1
  PILBox < 0421 LAD   hi: 30 lo: A1
 < 421 LAD
 > 500 RFC
  PILBox > 0500 RFC   hi: 34 lo: 80
  PILBox < 0500 RFC   hi: 34 lo: 80
 < 500 RFC
 > 43F UNL
  PILBox > 043F UNL   hi: 30 lo: BF
  PILBox < 043F UNL   hi: 30 lo: BF
 < 43F UNL
 > 500 RFC
  PILBox > 0500 RFC   hi: 34 lo: 80
  PILBox < 0500 RFC   hi: 34 lo: 80
 < 500 RFC

5) I have not tried /dev/ttyACM4 yet.

6) CAT 2 → listing done on an HP-41, captured, transformed and sent by Tulip4041 (Raspberry Pi Pico RP2040) to pyilper 82162A printer emulator (Raspberry Pi 4B).

Code:

-TIME  2C   
-CX TIME    
-PRINTER 2E 
-MASS ST 1H 
-CTL FNS    
-MYROM L1   
-MYROM U1   
-MYROM L2   
-MYROM U2   
MELBROM 1A  
MEMORY LOST 
UST KIDDING 
-MAXX 4C    
-ZENROM 3B  
-PROGRAMMER 
-EXT FCN 2D 
-CX EXT FCN

I have reported my first small harmless bug to Meindert this morning.

So this is it for now.

Sylvain

[MeindertKuipers]

Good work Sylvian, thanks for sharing. Just a few remarks from me:
- The trace listing shows FI as high all the time. You probably have jumper JP2 open. If you close this you can trace the FI line, and you see activity there when there is HP-IL traffic or interrupts from TIME
- In the IL Scope you see double the amount of RFC frames. These are real frames, the handling of these frames in the PILBox emulation needs to be fixed.

Next items for you to test: HP82143 printing and IR printing, Extended Memory and adding the HP-IL registers to the trace (command 'tracer ilregs'). Note that the registers are shown only when there is a change.

The ILScope trace is a different format because it is mixed with PILBox serial traffic (can be switched off), and I can never remember if the IL frames in capitals are in or out, so I used the symbols > for out and < for in.

[jsi]

(08-27-2024 07:33 PM)Sylvain Cote Wrote:  I am able to connect on the multiple serial terminals with minicom but pyilper only uses ttyUSB devices and ignore the ttyACM ones.
Fortunately it works if I manually change pyilper configuration file and set it to "/dev/ttyACM2".
I just posted a request to Joachim to add the ttyACM devices to his PIL-Box serial interface configuration dialog.

The combo box entry for the PIL-Box serial device is editable. Therefore you can enter an arbitrary device name there instead of choosing a member of the list. I will be happy to add ttyACMxx devices to the combo box list in the next pyILPER version.

Regards Joachim

[Sylvain Cote]

(08-28-2024 12:44 PM)jsi Wrote:  The combo box entry for the PIL-Box serial device is editable. Therefore you can enter an arbitrary device name there instead of choosing a member of the list.

Really!!! I was so sure that it was not that I did not even tried it, funny how mind works sometimes.

(08-28-2024 12:44 PM)jsi Wrote:  I will be happy to add ttyACMxx devices to the combo box list in the next pyILPER version.

Thanks!

Sylvain

[John Ioannidis]

(08-26-2024 07:24 AM)MeindertKuipers Wrote:  

(08-25-2024 11:07 PM)John Ioannidis Wrote:  What's the point of telling us all this if you are not posting any code?

No worries, there will be source code published of the version that will go out to the users of the Development Board. I do want to make certain that it is in a useable state. And my earlier offer still stands, if you are interested in the sources I will happily share it with you.

I read what I wrote again, and now it sounded embarrassingly rude

I apologize.

Yes, please share the code. I'm 'jayeye' on GitHub.

[MeindertKuipers]

Since the sources are in a more or less stable state (pending any feedback from Sylvain) I decided to make these public already at https://github.com/mjakuipers/TULIP-DevBoard. A binary for direct downloading in the Pico is in the build directory.

The usual disclaimer applies, use at your own risk. Also keep in mind that many parts are in development and half finished. Feedback is appreciated of course.

I still have not received my Pico 2 board, and I have some trouble getting the current sources to compile under the new SDK 2.0, and I am still using the SDK 1.51 for that.

[John Ioannidis]

(08-29-2024 12:03 PM)MeindertKuipers Wrote:  Since the sources are in a more or less stable state (pending any feedback from Sylvain) I decided to make these public already at https://github.com/mjakuipers/TULIP-DevBoard. A binary for direct downloading in the Pico is in the build directory.

The usual disclaimer applies, use at your own risk. Also keep in mind that many parts are in development and half finished. Feedback is appreciated of course.

I still have not received my Pico 2 board, and I have some trouble getting the current sources to compile under the new SDK 2.0, and I am still using the SDK 1.51 for that.

Forked! Thanks!

The Pico2 is backordered on all the US sites I tried (Adafruit, Sparkfun, Vilros), so don't feel bad

[Sylvain Cote]

Testing update #2 ...

I have tested all the functionalities in multiple scenarios and send the report to Meindert yesterday.
I assume that it will take him a couple of days to go through the report and come back with answers and/or firmware fix if necessary.

Summary ...

Working:
→ almost everything: serial interfaces, control console, tracing consoles, sdcard access, virtual HP-IL, internal ROM mapping, X-Memory, etc.
→ impressive, given that the software version is: 00.01.01 BETA !!!

Minor-Issues:
→ some inconsistencies in messages and/or behavior.

Major-Issues:
→ IR-Printing and external ROM mapping.
→ That could be me, not understanding and/or not doing the things correctly and/or functionality missing in my beta firmware.

[MeindertKuipers]

(09-01-2024 02:42 PM)Sylvain Cote Wrote:  Testing update #2 ...

Minor-Issues:
→ some inconsistencies in messages and/or behavior.

Major-Issues:
→ IR-Printing and external ROM mapping.
→ That could be me, not understanding and/or not doing the things correctly and/or functionality missing in my beta firmware.

Thanks for being the guinea pig Sylvain! Thanks to your feedback some nasty bugs were already fixed.

The inconsistencies are noted.
- IR printing issues probably due to the fact that a real HP82240 printer was used, and the current firmware only supports the simulated HP82143 printer. The byte-stream over a USB serial port or IR (IR with a receiver connected to a HP82143 simulator)
- ROM mapping is currently only possible with the ROMs embedded in the firmware for HP-IL, HP-IL Printer and the HP82143 printer ROM. Support for arbitrary ROMs to follow (pending support for specific hardware).

[MeindertKuipers]

Just to keep you all up to date, I have received the Pico2 boards with the RP2350, and after some minor changes something is happening. USB ports work, but no interaction with the calculator and also the SD card does not seem to work.
This means that there is some work to do to evaluate the RP2350.

[MeindertKuipers]

Good news, TULIP4041 seems to be working on the Pico2 board. Only a few minor changes needed. Only real issue so far is that the FI line is not driven, meaning that I cannot test HP-IL. And I have done only limited functional testing so far. Really interesting is the power consumption which is supposedly better than the RP2040, probably next week.
Also important is that I do not seem to have any issues with the GPIO errata which is currently a hot item on the Pico forums.

[MeindertKuipers]

After receiving and testing the Pico2 boards I have concluded that some more testing is needed to verify if the new RP2350 CPU has enough advantages over the existing RP2040. The initial tests did prove that the Pico2 board behaves quite nice on the Development Board. For now the development of the TULIP4041 will continue with the RP2040 and the existing Pico1 board. Worst case the users need to switch to the Pico2 board later, which is not a big investment.
Today I have ordered the PCB's for the Development Board for the existing pre-orders. When I receive these I will verify them by assembling and testing one unit, and when I am happy with the results I will announce this.

The issues with the SD card I had with the Pico2 tests are resolved, it was due to an upgrade of the FatFS library, which dropped support for SDSC cards (<= 2 GByte), and I was testing with an older 2 GByte card. Only issue with the Pico2 is with one of the state machines, and this is still to be resolved. Most important wil be testing the power management of the Pico2.