My fun fail at loading 6.03 update

PicoCalc
1

PicoCalc Pico 2H — My 6.03 Upgrade Adventure

I thought I would document my experience upgrading a PicoCalc Pico 2H / RP2350 from PicoMite 6.02.01 to 6.03.00, because I ran into a repeatable cold-start problem and eventually reverted to 6.02.01.

The upgrade itself went smoothly. I put the Pico 2H into BOOTSEL mode, Windows mounted it as RP2350, and I dragged over:

PicoMiteRP2350V6.03.00.uf2

The Pico rebooted normally and reported:

PicoMite MMBasic RP2350A V6.03.00

The first boot reformatted the internal A: drive, so the previous configuration, flash slots and library had to be rebuilt.

I restored the normal PicoCalc options, including:

OPTION HEARTBEAT OFF

OPTION CONTINUATION LINES ON

OPTION RTC AUTO ENABLE

OPTION BACKLIGHT LCD 192

One thing worth noting is that my old 6.02 OPTION LIST had shown:

OPTION LIBRARY_FLASH_SIZE 52000

but this is not something that can simply be re-entered as an OPTION command. It produces Invalid option. Saving the library handles this automatically.

I then loaded my main program, AVTEST.BAS, from the SD card and it ran correctly under 6.03.

The library was also successfully restored using:

LIBRARY SAVE

and MEMORY subsequently showed approximately:

Library: 30K used / 270K free

So at this stage everything appeared healthy.

There was also an encouraging improvement: I set the LCD backlight to 192, powered the PicoCalc off for several minutes, and it initially came back up at full brightness. This appeared to have solved the backlight persistence issue I had experienced under 6.02.

Then the trouble started.

After rebuilding the system, a normal power-off followed by power-on would intermittently produce a completely black screen.

There was no AUTORUN involved during some of these failures. It could happen on a plain cold start.

I tested several power arrangements:

  • PicoCalc battery power: sometimes black screen

  • Micro-USB connected directly to the Pico 2H: did not reliably cure it

  • PicoCalc USB-C external power: could bring it up normally

  • Changing batteries: sometimes made it boot again

Initially this made the batteries look suspicious, but subsequent testing convinced me they were not the underlying problem.

I cleared the RP2350 flash and performed a completely fresh installation of 6.03, rebuilt the options, library and program again, and still encountered the same cold-start black-screen behaviour.

I also found reports in the PicoMite 6.03 development discussions of other PicoCalc users experiencing startup/display problems with some 6.03 builds, including reports that earlier release candidates behaved better than the final release on certain hardware.

At that point I decided reliability was more important than continuing to troubleshoot 6.03.

I downloaded the archived:

PicoMite_Firmware V6.02.01.zip

and reflashed:

PicoMiteRP2350V6.02.01.uf2

After returning to 6.02.01, the PicoCalc immediately passed repeated cold-start tests.

I rebuilt the options, restored the library and restored my program. It is again starting reliably.

For startup I now use a very small A:\AUTORUN.BAS launcher which runs my main program from the SD card:

RUN "B:\AV.BAS"

My AV.BAS also sets the LCD backlight to 192 immediately on startup, which provides a practical workaround for the 6.02 backlight behaviour.

Bottom line

On my particular PicoCalc Pico 2H / RP2350:

6.02.01: reliable cold starts

6.03.00: program and library worked correctly once running, but I experienced intermittent black-screen cold starts even after a clean flash and rebuild

I am therefore staying on 6.02.01 for now.

This is only one machine and I am not suggesting that every PicoCalc will behave this way under 6.03. I am posting the details in case anyone else encounters a similar black-screen startup issue after upgrading.

anyone else having this issue and no, it was not battery issues. high quality set fully charged.

and yet when I plugged the power usb c into the calc I would get a yellow charge light.

in any case, all back to normal with the 6.02.01

2

This has been discussed on TBS and seems to be an issue for some PicoCalc but not others. The fix seems to be a capacitor across the power rails

1 Like
3

I pulled the appropriate capacitor discussion.

And reproduced it here

ChatGPT told me to remove the display and the CF card. In fact, the device often turned on successfully when the CF card was removed. Now I was supposed to solder a 22-47 µF capacitor between VSYS and GND on the Pico.
I had a 22 µF electrolytic capacitor on hand. After I soldered it in, my problems were gone! The PicoCalc started up 25 times in a row without any trouble.

I’d already had the feeling that Clockworkpi had skimped on the capacitance on the board. It does indeed seem that the device was built very tightly to spec. Either a capacitor has actually aged, or it has sustained some other kind of damage.

xxxxxxxxxxxxxx

Has anyone tried this and do they have a photo of the capacitor used and the attachment points.

1 Like
4

I think I will just wait for 6.03.01.

1 Like
5

i’m with you, I’ll wait for an update. It is obviously firmware and not hardware related as the 6.02 was extremely stable.

1 Like
6

If it can be fixed by a capacitor I would say it is obviously H/W. Presumably, 6.03.00 just starts up in a slightly different sequence which requires more power earlier. There is another thread on TBS on this that I can’t find but I remember all sorts of S/W tricks were tried to try and fix it without success. So I wouldn’t assume there will be a S/W solution coming. 6.03.00 works fine on my PicoCalc.

1 Like
7

What was the purpose of changing sequence and what was gained? If the code is rolled back what would be the consequences?

1 Like
8

what are the reasons I wanted to go to 6.03 from 6.02 it’s not on the older version the back light does not follow your option command when the picocalc is turned off and back on

I have used the code option backlight LCD 192 and invariably after a power cycle or sitting for a while off the option back light LCD reads 32.

I have placed the command option backlight LCD 192 in a startup file, but thought against that as I keep rewriting flash when that happens. I’ve placed option backlight LCD 192 is the first line of my operating code on my.BAS file.

after researching an AI Help it appears that the option backlight LCD 192 is being ignored and that was one of the reasons 6.03 was created.

That might make sense as a back light display of 192 is bright and make command a lot of startup power causing issues with loading as described on my first post.

602 is extremely stable and I’ve resorted back to that is at the moment I do not want to start adding more electronics to the board. It seems to me that the firmware change does not treat the startup sequence as far as power cycling smoothly.

It may be that the rewrite cannot be adjusted firmware wise and the only option is to adjust the hardware with the capacitor.

I was involved in the HP 41CL beta testing many many years ago and the hardware of the 1979 versions required a nanofarad capacitor attached to two legs of the LCD panel. I have no problem doing that but if there is a software fix versus a hardware fix I prefer the software fix.

Standing by with bated breath!

As far as my program is concerned if you want it right you hold the alt and keyboard brightness key. What the machine is on in the program is run it maintains the brightness.