Pi Pico[W] Peripheral Expansion Box Side Port Device

[JasonACT]

On 1/6/2024 at 1:53 PM, JasonACT said:

To properly format a FAT32 SD card, for use with a Pi Pico, take these steps:

diskpart
select disk 1
clean
Open "Computer Management" and create a "New Simple Volume..." and follow the usual process to format.

 

EDIT: This is assuming the SD card is disk 1, use "list disk" to find out.

EDIT2: Forgot to mention, these steps ensure you have a "primary" partition rather than "logical" partitions which the Pi Pico won't detect.

 

On 1/9/2024 at 4:54 PM, JasonACT said:

That's what I'm using too, but I've modified mine to display the 1st error detected - and when the PSRAM chip is overclocked, I often see something like you have shown in the image above with the blue screen (error: >F804 written, >F806 read).  I think it's merged an address bit due to the speed.  I don't have that tester with the blue screen though 

 

Attached is a new fail-safe version, it runs at 256MHz (128MHz PSRAM speed).  Do you get less errors with that?

 

Oh, and this is a special build for visual status via the LED on power-up: 1 flash (short or long) = all good, 3 flashes = no SD card, 4 flashes = SD card present but no "Primary FAT32 Partition" found, 5 or more flashes = Everything seems OK, but the .ROM files were not found on the SD card.

PPEB2.ino.uf2.zip 379.12 kB · 9 downloads

@wiloson Looks like you need to "clean" the SD card so you can format it with a "primary" partition.

 

Make sure you "list disk" so you know which one to select before running "clean"

[wiloson]

1 hour ago, JasonACT said:

 

@wiloson Looks like you need to "clean" the SD card so you can format it with a "primary" partition.

 

Make sure you "list disk" so you know which one to select before running "clean"

i seem to be getting 4 flashes saying that it cant see the primary FAT partition.

I used diskpart to clean the SD card and formatted it as default FAT 32 but it still is giving me the same error.   

I have tried a different SD card and tried to resize it to 8gb and formatted it using different software tools and still no luck.

I'll try again tomorrow night.

 

thanks for all the help.

[JasonACT]

2 hours ago, wiloson said:

i seem to be getting 4 flashes saying that it cant see the primary FAT partition.

I used diskpart to clean the SD card and formatted it as default FAT 32 but it still is giving me the same error.   

I have tried a different SD card and tried to resize it to 8gb and formatted it using different software tools and still no luck.

I'll try again tomorrow night.

 

thanks for all the help.

I tried the clean/partition/format steps again on my el-cheapo 8GB SD card and it worked OK.  Used all the defaults.  I usually use a 32GB one though, so I don't think there's any issues with that size.

 

Maybe it's worth putting up a photo of your build, if it's getting read errors on mounting, there may be something about the board we can see.

 

One other thing, I guess is, I tried a 330 Ohm resistor instead of a 470 Ohm and it worked OK for me on the last board I built..  And that one has 2 x 8MB PSRAM chips, and they were the iffy ones - so changing out the resistor to a lower value may get yours going.

[wiloson]

7 hours ago, JasonACT said:

I tried the clean/partition/format steps again on my el-cheapo 8GB SD card and it worked OK.  Used all the defaults.  I usually use a 32GB one though, so I don't think there's any issues with that size.

 

Maybe it's worth putting up a photo of your build, if it's getting read errors on mounting, there may be something about the board we can see.

 

One other thing, I guess is, I tried a 330 Ohm resistor instead of a 470 Ohm and it worked OK for me on the last board I built..  And that one has 2 x 8MB PSRAM chips, and they were the iffy ones - so changing out the resistor to a lower value may get yours going.

I think that the SD card may be set up as GPT and it should probably be MBR.   I'll have to wait until I get home, after work today, before I can verify this and make changes.   If this doesn't help I will change the pull-up resistor to 330 ohms.

I'll also add a photo of the board tonight as well.

 

thanks

[wiloson]

well no luck with reformatting the SD card.  I still get the LED flashing 4 times.     here is an image of the assembled board.

[RickyDean]

9 minutes ago, wiloson said:

well no luck with reformatting the SD card.  I still get the LED flashing 4 times.     here is an image of the assembled board.

 

Are you sure you have good solder joint in the area I circled?

[+chue]

I would be inclined to touch up the pins on the sd card socket as well. 

[Gary from OPA]

1 hour ago, wiloson said:

well no luck with reformatting the SD card.  I still get the LED flashing 4 times.     here is an image of the assembled board.

I would touch up the pi solder joints mainly the ones around pin 11 to 17 and then couple pins on the other side of the pi near the single transistor, not all the pins of the pi looked soldered fully.

[JasonACT]

I can't add any more to those suggestions, thanks everyone, so instead here's a small mod I made to the case:

 

 

An option to shape the top-back to be angled the same as a beige model.  (That's sparkle-gold top and copper bottom - two of the nicest PLA printer filaments I have, the photo doesn't do it justice.)

ZfinalB.zip

[+chue]

Ok, I did the QI mods:

1. Pulled out the resistor pack

2. Pulled out U32 and put in a "shunt" connecting the data lines

 

 

The QI is now able to run carts from the PicoPEB (previously it didn't work):

I am currently running a SAMS burn-in test.

 

I also tested the FinalGrom99 (without the Pico attached) - it doesn't work with Ralph's 1.3 firmware.  It still needs the QI specific firmware to work correctly.

 

Let me know @JasonACT if you would like to see any other testing.

[JasonACT]

That's awesome!  You clearly have more skill with soldering tools than I do, even your Pico PEB board looks better than mine!

 

3 minutes ago, chue said:

I also tested the FinalGrom99 (without the Pico attached) - it doesn't work with Ralph's 1.3 firmware.  It still needs the QI specific firmware to work correctly.

Ha ha, yeah, the new FinalGrom99 firmware tests for >0000 (standard pull-down) and >FFFF (QI pull-up) - the original firmware didn't - but that's safe for the FG99.  And that's why you have to run a modified version of my DSR.ROM - and I didn't want to post a version of that (just the instructions) because someone might try to use it which is bad for an unmodified QI console and my device.

 

Nice work!

[wiloson]

12 hours ago, RickyDean said:

Are you sure you have good solder joint in the area I circled?

The first two pins you circled aren't used and I have checked all the solder joints with a microscope so I am confident that they are good.   

The glare from my lamp and my camera makes all the solder joints look suspicious but I assure you that they are good.   

I will reflow the solder on all the pins on the SD card holder tonight just to be sure, and see if I can verify that the SD card makes good contact with the cardholder when it is inserted.    

Then I will check the communication at the SD card holder with my scope and see if any of the data lines are bad.  

I will then then try swapping the pull up to 330 Ohm resistor instead of a 470 Ohm. 

 

Thanks for all the help

[JasonACT]

14 hours ago, wiloson said:

The first two pins you circled aren't used and I have checked all the solder joints with a microscope so I am confident that they are good.   

The glare from my lamp and my camera makes all the solder joints look suspicious but I assure you that they are good.   

I will reflow the solder on all the pins on the SD card holder tonight just to be sure, and see if I can verify that the SD card makes good contact with the cardholder when it is inserted.    

Then I will check the communication at the SD card holder with my scope and see if any of the data lines are bad.  

I will then then try swapping the pull up to 330 Ohm resistor instead of a 470 Ohm. 

 

Thanks for all the help

I finished work and rushed home to see an update, but since there's none, I'll assume things are still not going well.

 

Some questions about the picture then:

1. Which PSRAM brand/type did you go with?

2. What is the value of the capacitor C1?  (It's physically much larger than I expected to see.)

3. With the SD card ejected, what's the reading you get on R4 with a multimeter?  (If it's nearer 1K Ohm you may get 4 blinks - I remember trying that value with it getting read errors.)

 

FYI - You're using the right firmware - any of the later ones which have the "blink on issues" should be good, doesn't matter which DSR.ROM if we only consider the 4 blinks error code, it's not getting that far to matter.

[wiloson]

On 4/1/2024 at 1:29 AM, JasonACT said:

I tried the clean/partition/format steps again on my el-cheapo 8GB SD card and it worked OK.  Used all the defaults.  I usually use a 32GB one though, so I don't think there's any issues with that size.

 

Maybe it's worth putting up a photo of your build, if it's getting read errors on mounting, there may be something about the board we can see.

 

One other thing, I guess is, I tried a 330 Ohm resistor instead of a 470 Ohm and it worked OK for me on the last board I built..  And that one has 2 x 8MB PSRAM chips, and they were the iffy ones - so changing out the resistor to a lower value may get yours going.

I got the board working last night.     I was working on the original parts list and installed a 10k ohm resistor into R4 instead of a 470 ohm.     I noticed that the clock signal was horrible and I couldn't find any mention of the 470 ohm resistor in the parts list as per your comment.   

After searching the discussion I found your post about the change to r4.

 

 

Thanks, everyone for your help getting this board going.

 

[JasonACT]

Thanks to the curves (lots of g-code) on each corner and long straight print lines in my 3D case model, I was able to (finally) track down the bug in my 3D printer firmware (Sailfish) - so I had to print one more piece to test...  And it was the best print of a large object I've ever had - no weird fat print lines.  So I'm very happy, this new case is silk-gold on beige - but I still really like the silk-copper.  And this picture probably shows I've gone "just too far" with my little project here - but meh...

 

 

Top-Left is the original (with the broken Pico boot button and makeshift PSRAM install) with a different latch style - to be able to print the top piece upside down without support.  Next one to the right, I got the height right, no feet needed, but the angle at the top-front was way off.  Bottom left, the "final" version - good for an original black-siver model I think.  The rest are the Beige version with angled back-top - just in different colours.  The full-copper version doesn't contain a circuit board, but I have one blank v2 version left...  I just only have one real TI to use all these on

[JasonACT]

On 4/7/2024 at 3:02 AM, arcadeshopper said:

This is covered in the FAQ. By the way, the speach synthesizer section.. 

 

Any 32k sidecar should have a external power supply connection. If it doesn't, then it's short-sighted.. 

And that, I do, too!

Isolate the +5V line using some wax paper and blue painter's tape, since I'm not going to cut my circuit board trace, as I don't really find this necessary, but...  It's important to try these things out.

 

 

Power the unit from a Pico specific USB HUB with a 2.5AMP wall wart, and switch "the power switch" to on, so it powers up the Pi Pico first.

 

And there you have it, no extra load on the '99 power supply.  Along with being able to plug in some seriously power hungy gamepad devices like my PS4 controller.  Hmmm, I probably should have mention to use any powered HUB previously, for something power hungry like that.

 

Enjoy.

[JasonACT]

Well, I've now ripped out the wax paper.  My main concern was the load on the '99 power supply, given quite a few comments I've read in other threads about it, and that the '99 can have a GRAM Kracker with 10 x 8KB RAM chips along with its support chips, also with a fully loaded cartridge plugged into it, and a speech synthesizer hanging off the side port all at the same time.  Not to mention, the original 32KB expansion memory modification "inside the console".  None of which I really considered when I put any of this together to get back into to the TI.  But none of those things have any weight here now, with my device.

 

I also always had my PS4 gamepad controller fully charged, so it never really needed the quoted 800mA power it says on the label, it never actually needs to charge when being used via a USB cable.  I'm fairly sure it just works off a small amount of power from the USB port in the fully charged state.  Though with an empty battery, that might be different.

 

So I'm back to printing in more colours, I like the "frosted bronze" PLA filament for the bottom colour...  Starting to match nicely.

 

[+Ksarul]

You're going to need to score some more consoles just to be able to display and use all of this sidecar goodness. . . 

[wiloson]

What is the status of the pcode card?   I tried running the "call pcodeon" command and it just booted into the Pcode colour screen and sat there without any text.

I have the required Pcode.rom and Pcode.grom files in the root of the SD card...

[JasonACT]

4 hours ago, wiloson said:

What is the status of the pcode card?   I tried running the "call pcodeon" command and it just booted into the Pcode colour screen and sat there without any text.

I have the required Pcode.rom and Pcode.grom files in the root of the SD card...

You need to have the PCode disk sector image mounted in DSK1 too.

 

Edit: And you need to leave it to boot, maybe takes a minute or slightly more.

Edit2: Attached my 360KB disk image.

PASCAL.zip

Edited April 12 by JasonACT

[wiloson]

I have put together 5 of the second version(with R4) boards and the only version of the firmware that passes the SAMS test for me is the 252Mhz with 3 NOPs that you published on Feb23.

Is it possible for you to compile two versions of this software one for the 2GB and the other for the 8GB mod?

 

I purchased APS6404L-3SQR DRAM chips, the PI Picos from reputable suppliers and I have tried both of my older consoles (I opened one up and the power supply was good) and only one firmware worked.

I wonder if the date my computers where manufactured has something to do with it.  

[JasonACT]

8 hours ago, wiloson said:

I wonder if the date my computers where manufactured has something to do with it.  

Mine has LTA2483 (24th week, 1983) on the bottom, I was wondering what everyone else who has made a Pico PEB has?  Mine has run every firmware I've posted (of course).

 

Regarding 252MHz, I've tried that speed just now, it doesn't work with anything over 1MB SAMS.  When you go larger than 1MB, additional C code gets enabled by #ifdefs, this pushes the timing out too far and it fails on my TI, so would certainly fail on most.  When I created that firmware it was a last attempt to get it working on a machine I didn't have physical access to.

 

We're going to have to fix this a different way.  Did you run the stand-alone "bench" memory test firmwares and count the blinks?  That will tell if it's really a speed problem with the PSRAM chips or a timing issue with the particular TI.

PPEB2_258MHz_2MB_3NOP.ino.uf2.zip

PPEB2_256MHz_2MB_1NOP.ino.uf2.zip

Edited April 15 by JasonACT
Add slower 2MB SAMS Firmware (258MHz is the slowest working speed) + 256MHz&1NOP

[+chue]

10 minutes ago, JasonACT said:

I was wondering what everyone else who has made a Pico PEB has?

Silver & Black: LTA5282

Beige: LTA3983

QI: LTA4383

[JasonACT]

9 hours ago, wiloson said:

I have put together 5 of the second version(with R4) boards and the only version of the firmware that passes the SAMS test for me is the 252Mhz with 3 NOPs that you published on Feb23.

1 hour ago, JasonACT said:

We're going to have to fix this a different way.  Did you run the stand-alone "bench" memory test firmwares and count the blinks?  That will tell if it's really a speed problem with the PSRAM chips or a timing issue with the particular TI.

(I'll be sleeping/at-work in my timezone when you read this, so...) If you do get a low blink count, try with a higher R4 resistor value or a different SD card to see if you can get the blink count up a bit.  560Ohm may work slightly better in this case, though I did try 1KOhm and it failed for me.

 

1 hour ago, chue said:

Silver & Black: LTA5282

Beige: LTA3983

QI: LTA4383

Thanks chue, those relatively close dates to my unit really make me think there might be more going on, like the TI I have has a slow master clock...

 

On 4/11/2024 at 7:03 AM, Ksarul said:

You're going to need to score some more consoles just to be able to display and use all of this sidecar goodness. . . 

I've entered your mind, and I liked what I saw, so I've got another TI on the way from eBay   It'll be a PAL unit, but I don't think that matters here.

[wiloson]

13 hours ago, chue said:

Silver & Black: LTA5282

Beige: LTA3983

QI: LTA4383

my two computers are silver and black 

LTA2383 and my original system LTA4582.

They don't seem far from what others have as far as manufacturing dates.