Threaded Mode | Linear Mode

rocket.scientist · 10-21-2024, 01:43 AM

I recently finished the build of my TULIP4041, and am going through the confidence tests. All has gone splendidly with the exception of the HP-IL steps. Here is the CLI output:

TULIP> plug hpil
plugged HP-IL module
CDC Port 4 [IL Scope] connected
CDC Port 3 [HPIL] connected
TULIP> plug status
HP-IL Printer Page -
HP82143 Printer Page -
HP-IL module Page 7

The question arises on the HP41C. When I key in the DIR command I see TRANSMIT ERR instead of the message NO DRIVE as advised by "TULIP4041 doocumentation.pdf". I have not created any LIF images.

I am not sure if I have an issue or not. Thoughts?

MeindertKuipers · 10-21-2024, 08:44 AM

Good to hear that things seem to be working.

If you see the TRANSMIT ERR please check with the tracer if there is any activity on FI (if you have the extra buffer installed). For tracing FI also JP2 must be closed! I have seen this sometimes after installing new firmware. Best is to disconnect from the HP41 and USB and restart everything. I will try to reproduce later today since I am building a new TULIP myself.
As an extra check (but unlikely) you could verify the physical connections of the FI signal all the way from the HP41 connector to the input buffer.

MeindertKuipers · 10-21-2024, 02:34 PM

I have messed up something. I had this issue in an earlier firmware version and thought it was fixed. It might be a corner case due to timing of the FI output. I had some changes in my VS Code and GitHub configuration and things went wrong there.
A new firmware file is now on my GitHub. Please observe the following:
- follow the instructions in the documentation to go to BOOTSEL mode and upgrade firmware
- firmware now runs at 125 MHz (was 150 MHZ)
- FI output is fixed, and HP-IL now works
- FI input is broken, so you will see all flags active in the tracer, which is not true of course, to be fixed
- I have removed the sources as this is for a non-working configuration
- make certain that JP2 is closed, this enables driving FI

rocket.scientist · 10-21-2024, 04:10 PM

I am afraid my precious HP-41CX is having a problem, and I am just now discovering it.

Looking between at J5->FI and J5-GND, I see about +5v when the 41CX is off, including when the TULIP is unpowered. When I turn the 41CX on this jumps to about +6v and has a fraction of a volt noise on it for a second or two. The voltage slowly drops to zero when the TULIP is unpowered and I pull the 41CX battery. Note U5 is installed and JP2 is in. There are no other modules inserted.. I get similar results when probing directly at the module board with the TULIP uncabled. :-(

I see messages like these on the HPIL trace. The variety changes at power off and power on, plus when I execute DIR. No F1 data I fear.

> 600 IDY
< 600 IDY
> 490 IFC
< 490 IFC
> 500 RFC
< 500 RFC
> 600 IDY
> 490 IFC
> 623 IDY
> 623 IDY
> 623 IDY

Does this make sense? Is FI asserted all the time from the 41CX? This seems odd.

rocket.scientist · (This post was last modified: 10-21-2024 05:21 PM by rocket.scientist.)

(10-21-2024 02:34 PM)MeindertKuipers Wrote: I have messed up something. I had this issue in an earlier firmware version and thought it was fixed. It might be a corner case due to timing of the FI output. I had some changes in my VS Code and GitHub configuration and things went wrong there.
A new firmware file is now on my GitHub. Please observe the following:
- follow the instructions in the documentation to go to BOOTSEL mode and upgrade firmware
- firmware now runs at 125 MHz (was 150 MHZ)
- FI output is fixed, and HP-IL now works
- FI input is broken, so you will see all flags active in the tracer, which is not true of course, to be fixed
- I have removed the sources as this is for a non-working configuration
- make certain that JP2 is closed, this enables driving FI

Maybe my 41CX is okay. I get the same behavior with my 41CL, meaning both say "TRANSMIT ERR".

Does the new version say "0.01.02 BETA - build time: 2024/09/17_19:56:54"?

MeindertKuipers · 10-21-2024, 05:23 PM

No worries, FI is an active low signal, meaning it is mostly high when not active. The 'noise' is possibly the activity on the FI line from HP-IL. When the HP41 is powered down the FI line will go down.
If you see the messages then the HP-IL emulation works, but no device is connected, and it means that there is FI activity, but you will not see this in the tracer, JP2 does not matter in this case. This is a bug I need to check out.
When you do a DIR now do you see NO DRIVE ot TRANSMIT ERR?? I am pretty sure you willsee NO DRIVE.
It is now time to connect pyILPer and check what happens. Also enable the PILBox tracing with the tracer pilbox command to verify if you get the 'serial' frames.

MeindertKuipers · 10-21-2024, 05:40 PM

(10-21-2024 05:16 PM)rocket.scientist Wrote: Does the new version say "0.01.02 BETA - build time: 2024/09/17_19:56:54"?

My version says "VERSION 00.01.03 BETA - build time: 2024/09/17_19:56:54"
And runs at 125 MHz

Sorry about all the confusion

Craig Bladow · (This post was last modified: 10-21-2024 08:13 PM by Craig Bladow.)

I'm seeing the same issue with my TULIP4041 completed this weekend.
I've updated the firmware to "VERSION 00.01.03 BETA - build time: 2024/09/17_19:56:54".

I verified continuity from the 41c module F1 connector pin to pin 3 of U5 and from pin 2 of U5 to pin 4 of the Pico 2. Power and ground connections to U5 check out.

Oscilloscope shows F1 going low for about 10 uS, I only see this once in 500ms.

A little section of trace output which may or may not be helpful.

Code:

      244  77CC  1  0E4  0.4800100000.4.2C  ...  R000  C0  FI0123456789ABCD  SELP 3                                                                                                                                

      245  77CD  0  001  0.4800100000.4.2C  0E4  R000  C0  FI0123456789ABCD  001  reg 3=00                                                                                                                         

      246  77CE  1  024  0.4800100000.4.2C  001  R000  C0  FI0123456789ABCD  SELP 0                                                                                                                                

      247  77CF  0  005  0.4800100000.4.2C  024  R000  C0  FI0123456789ABCD  005  reg 0=01                                                                                                                         

      248  77D0  1  024  0.4800100000.4.2C  005  R000  C0  FI0123456789ABCD  SELP 0                           Reg 01* 01* 00  00  01  01  01  00  00                                                               

      249  77D1  0  3C1  0.4800100000.4.2C  024  R000  C0  FI0123456789ABCD  3C1  reg 0=F0                                                                                                                         

      250  77D2  1  064  0.4800100000.4.2C  3C1  R000  C0  FI0123456789ABCD  SELP 1                           Reg F0* 01  00  00  01  01  01  00  00                                                               

      251  77D3  0  305  0.4800100000.4.2C  064  R000  C0  FI0123456789ABCD  305  reg 1=C1                                                                                                                         

      252  77D4  1  130  0.4800100000.4.2C  305  R000  C0  FI0123456789ABCD  LDI                              Reg F0  01  00  00  01  01  01  00  C1*                                                              

      253  77D5  0  01D  0.4800100000.4.2C  ...  R000  C0  FI0123456789ABCD  01D                                                                                                                                   

      254  77D6  1  106  0.4800100000.4.2C  ...  R000  C0  FI0123456789ABCD  A=C S&X                                                                                                                               

      255  77D7  1  130  0.4800100000.0.1D  ...  R000  C0  FI0123456789ABCD  LDI                                                                                                                                   

      256  77D8  0  023  0.4800100000.0.1D  ...  R000  C0  FI0123456789ABCD  023                                                                                                                                   

      257  77D9  1  280  0.4800100000.0.1D  ...  R000  C0  FI0123456789ABCD  HPIL=C 2                                                                                                                              

      258  77DA  1  12C  0.4800100000.0.23  280  R000  C0  FI0123456789ABCD  ?FI 8 ?FRAV         IL> 623 IDY  Reg F0  00* 00  00  01  01  01  00  C1                                                               

      259  77DB  1  037  0.4800100000.0.23  ...  R000  C0  FI0123456789ABCD  JC 6

rocket.scientist · 10-21-2024, 09:03 PM

(10-21-2024 05:40 PM)MeindertKuipers Wrote:
(10-21-2024 05:16 PM)rocket.scientist Wrote: Does the new version say "0.01.02 BETA - build time: 2024/09/17_19:56:54"?
My version says "VERSION 00.01.03 BETA - build time: 2024/09/17_19:56:54"
And runs at 125 MHz

Sorry about all the confusion

Most of the confusion is mine. Thanks for being adaptable.

I can't get the latest UF2 to load (the one in the root page of the TULIP github dated earlier today). The PICO does not reboot when I copy it to the virtual drive. Is this by chance for a 2350? I am still on a 2040 until my 2350 arrives.

Sorry to be so needy right now.

MeindertKuipers · 10-22-2024, 06:38 AM

I am away from my calculators for a few days, will look into this later this week.
And indeed, the newly published version is for the Pico2 only!

rocket.scientist · (This post was last modified: 10-23-2024 06:07 PM by rocket.scientist.)

Update thus far. I am using a stock HP-41CX.

1) Native HPIL with HP 82162A printer works perfectly. Print functions work as expected. DIR returns "NO DRIVE".
2) Native HPIL with HP 82161A drive works perfectly. DIR returns list of files on cassette, or "NO MEDIA" without a cassette inserted.
3) PIL-Box with virtual HP 82161A drive and HP 82162A printer works perfectly for both print and drive commands.

At this point I know my 41CX and HPIL circuit works as expected. Now, on to the TULIP.

4) RP2 received and installed into TULIP-devboard. Firmware 00.01.03 installation went perfectly, and confirmed in com0.
5) JP2 is installed, as is U5. Continuity for FI confirmed from the HP native module connector, through the adapter, cable and J4, to pin 3 of U5. Continuity also confirmed from pin 2 of U5, through J2 to pin 4 of the PICO 2.
6) ran plug hpil, confirmed enabled with plug status, then ran tracer hpil.
6) pyilper reconfigured to connect to PILBOX serial port on TULIP, and restarted.
7) Executing DIR on the 41CX again returns TRANSMIT ERR. Ugh. :-(

I am still digging and will post any new revelations here.

MeindertKuipers · 10-24-2024, 12:06 PM

A new firmware file has just been uploaded to my GitHub, version 00.01.04, with the following changes:
- fixed FI input (was fixed in my development version, but forgot to put it in the release Sad

)
- double checked HP-IL, is working
- as a bonus the trace buffer is increased to 5000 samples, there is enough memory on the Pico2

This version is ONLY for the Pico2!

With FI input now working you can close JP2 (if you have U5 installed) and verify if you can see the FI line changing when there is HP-IL activity. In case you have a CX or a C/CV with TIME module, start the clock with [SHIFT] [ON] and have a look at the trace window. You should see the trace starting every second with flags C and D set every now and then.

For troubleshooting please activate HP-IL tracing with the tracer hpil and tracer pilbox commands and share the results. I have tested this firmware version on 2 different TULIP boards and on my 41CX and a 41CL. When testing I also saw that the system poweron commands sometimes fails. Will be investigated (added as an issue to my GitHub).

Please let me know how things go

rocket.scientist · 10-24-2024, 04:17 PM

In the HP-41C/CV/CX Service Manual schematic, the FI line comes straight from the CPU itself. Interestingly my CX keeps FI at about battery voltage even when the calculator is off. It drops to ground only when I remove the battery.

Does this make sense? Can someone with a TULIP measure the voltage for FI right at the module adapter? I'd love to know if this is just a pull-up that is always powered.

MeindertKuipers · 10-24-2024, 05:18 PM

(10-24-2024 04:17 PM)rocket.scientist Wrote: In the HP-41C/CV/CX Service Manual schematic, the FI line comes straight from the CPU itself. Interestingly my CX keeps FI at about battery voltage even when the calculator is off. It drops to ground only when I remove the battery.

Does this make sense? Can someone with a TULIP measure the voltage for FI right at the module adapter? I'd love to know if this is just a pull-up that is always powered.

Confirmed. FI is high even when the calc is off. Measured on my HP41CX.
In the CPU Detailed Description document it is mentioned that the FI line is pulled high by the CPU when open.

Craig Bladow · (This post was last modified: 10-24-2024 06:13 PM by Craig Bladow.)

Installed "VERSION 00.01.04 BETA - build time: 2024/10/24_12:01:05 " and it is working!

Note that it does not work if a terminal window is open and connected to the ITF_HPIL (/dev/ttyACM2) serial port. In that case there is a "TRANSMIT ERR". On page 41 of the TULIP4041 documentation under "HP41 interface test" an instruction should be added to exit the ITF_HPIL terminal window if it is open at the beginning of this step. I'm wondering if this has contributed to the problem all along.

@rocket.scientist: I measured F1 on the module adapter with TULIP4041 disconnected and it is pulled up to V+ with the calculator on or off.

MeindertKuipers · 10-24-2024, 06:41 PM

(10-24-2024 06:13 PM)Craig Bladow Wrote: Installed "VERSION 00.01.04 BETA - build time: 2024/10/24_12:01:05 " and it is working!

Note that it does not work if a terminal window is open and connected to the ITF_HPIL (/dev/ttyACM2) serial port. In that case there is a "TRANSMIT ERR". On page 41 of the TULIP4041 documentation under "HP41 interface test" an instruction should be added to exit the ITF_HPIL terminal window if it is open at the beginning of this step. I'm wondering if this has contributed to the problem all along.

Good to hear it now works.
I will add a check. Currently if nothing is connected to the HP-IL interface the loop is simply closed and there is no transmit error. If something is connected I assume it is always the virtual IL device. There is no check if a virtual device is connected. I will add that and improve the documentation. Good point.

rocket.scientist · 10-24-2024, 06:43 PM

(10-24-2024 12:06 PM)MeindertKuipers Wrote: A new firmware file has just been uploaded to my GitHub, version 00.01.04, with the following changes:
- fixed FI input (was fixed in my development version, but forgot to put it in the release )
- double checked HP-IL, is working
- as a bonus the trace buffer is increased to 5000 samples, there is enough memory on the Pico2

This version is ONLY for the Pico2!

With FI input now working you can close JP2 (if you have U5 installed) and verify if you can see the FI line changing when there is HP-IL activity. In case you have a CX or a C/CV with TIME module, start the clock with [SHIFT] [ON] and have a look at the trace window. You should see the trace starting every second with flags C and D set every now and then.

For troubleshooting please activate HP-IL tracing with the tracer hpil and tracer pilbox commands and share the results. I have tested this firmware version on 2 different TULIP boards and on my 41CX and a 41CL. When testing I also saw that the system poweron commands sometimes fails. Will be investigated (added as an issue to my GitHub).

Please let me know how things go

Things are changing. I am now seeing activity on FI when I execute DIR or PRA, or power on/off the 41C. I see some activity in the HPIL tracer in these instances. Sadly the 41C still says TRANSMIT ERR when I execute PRA to the virtual HPIL Printer.

Here are logs for the various virtual comm ports. Maybe there is a gem hidden in there that a more knowledgeable person can spot. This TULIP is working perfectly in every other regard. I am at a loss for what is causing this. The tracer.log file is huge. Sorry about that.

Thank your for your assistance and persistence.

rocket.scientist · (This post was last modified: 10-24-2024 06:50 PM by rocket.scientist.)

(10-24-2024 06:13 PM)Craig Bladow Wrote: Installed "VERSION 00.01.04 BETA - build time: 2024/10/24_12:01:05 " and it is working!

Note that it does not work if a terminal window is open and connected to the ITF_HPIL (/dev/ttyACM2) serial port. In that case there is a "TRANSMIT ERR". On page 41 of the TULIP4041 documentation under "HP41 interface test" an instruction should be added to exit the ITF_HPIL terminal window if it is open at the beginning of this step. I'm wondering if this has contributed to the problem all along.

@rocket.scientist: I measured F1 on the module adapter with TULIP4041 disconnected and it is pulled up to V+ with the calculator on or off.

Thanks for that confirmation on FI. That was what I was hoping to hear. Let me investigate your ACM2 disclosure. I too wonder if that is my problem all along.

edit: I just checked my last log captures. HPIL port is closed. Drat!

TULIP> system cdc
CDC Port 1 [console] connected? yes
CDC Port 2 [tracer] connected? yes
CDC Port 3 [HP-IL] connected? no
CDC Port 4 [IL scope] connected? yes
CDC Port 5 [printer] connected? no

MeindertKuipers · 10-25-2024, 08:03 AM

Summarizing this issue:
A terminal emulator is used to verify connections and find out the COM port number for the various interfaces. The HP-IL virtual interface simulates a closed loop when there is no virtual device connected. This results in the NO DRIVE message when doing a DIR.
If the connected device is a terminal emulator the loop will appear open, sine the emulator does not pass the HP-IL packets. This results in the TRANSMIT ERR.
Only a virtual HP-IL device (like ILPer or pyILPer) will process the HP-IL packets in the correct way and the TULIP then acts as an HP-IL module with PILBox attached.

MeindertKuipers · 10-26-2024, 01:12 PM

I have made some changes in the handling of the serial ports. First of all I now also report (in the console) when a serial ports disconnects, and the connection changes of the PILBox, so it is now clear if a virtual HP-IL device is connected.
When there is a serial connection to the HP-IL port, but no HP-IL device behind it, the firmware still reports this as an open loop (TRANMIT ERR). The console issues a warning when this is the case and I wanted to be able to check for an open loop. The user will know that something is wrong instead of seeing only a closed loop and search for a solution in the wrong direction.

As a bonus I have added the sample counter to the HP-IL Scope window (but it cannot be switched off).
As I am working on a few other minor items the firmware is not yet published.