USB SPI Flasher with PIC18F2455 - now with source and binaries

[threesixtyuser]

......
And good work threesixtyuser, the board you build looks really nice, and your drawing too

Thx, I am quite the electronic theoretics nitwit, but fortunatelly I do have a some soldering and drawing experience

Now we seem to have figured out a somewhat "correct" version here's my final drawing of how I should have built it (used 27pF instead of 22pF and 220nF instead of 470nF) with specific components and links to all these components on the dutch farnell site (reckon they use the same article numbers on other international sites).



Flasher:

• PROTOTYPE BOARD
• PIC18F2455-I/SP
• SOCKET IC, DIL, 0.3", 28WAY
• SOCKET, USB, TYPE B WARNING: Faulty pinout in datasheet!
• CRYSTAL, 12M, 18PF CL, HC49/4H
• RESISTOR, 0.25W 5% 10K
• RESISTOR, 0.25W 5% 100R
• CAPACITOR, 470NF +/-10% X7R, 50V unfortunately no 470nF available in X7R with 2,54mm lead spacing
• CAPACITOR, 100NF +/-10% X7R, 50V
• CAPACITOR, 22PF +/-2% LowK, 100V
• HEADER, 1 ROW, R/ANGLE, 2WAY
• JUMPER, 2WAY

Connection:

• WR-BHD 2.54MM MALE 16 PINS RIGHT
• WR-BHD 2.54MM FEM IDC W STRAIN 16 P
• SPECTRA-STRIP RIBBON CABLE, 3C, 12 CORE

JTag

• 1N4148 DIODE, HIGH SPEED, DO-35

Credits go to all who contributed in the result! DarkstarTM, Sandungas, Cory1492, straßenkampf etc. etc.

A more "universal" schematic with explanations is made by Sandungas and can be found here

And this is what the endresult would somewhat look like with the above components:

[maximilian0017]

0.47uF was the biggest I tried, as the cap is just a filter for Vusb and isnt absolutely necessary (at least thats how I understood the datasheet).

The datasheet is a %$&^ to read but it is clearly stated that a 220nF (+-20%) is needed for stable operation.

threesixtyuser: Looks nice, still think the connector is overkill but looks nice  

Edit: 200nF -> 220nF

[threesixtyuser]

thx, I have big fingers, so a big connector gives me more comfort

edit: noticed you changed it btw doesn't datasheet state 220 nF (±20%) and minimal 0.22 μF/typical 0.47 μF as far as I can see (and like you posted earlier)

[SpikedCola]

0.47uF was the biggest I tried, as the cap is just a filter for Vusb and isnt absolutely necessary (at least thats how I understood the datasheet).

The datasheet is a %$&^ to read but it is clearly stated that a 220nF (+-20%) is needed for stable operation.

threesixtyuser: Looks nice, still think the connector is overkill but looks nice  

Edit: 200nF -> 220nF

Mine says 0.47uF typical/recommended, 0.22uF minimim.

[maximilian0017]

thx, I have big fingers, so a big connector gives me more comfort

btw doesn't datasheet state 220 nF (±20%) and minimal 0.22 μF/typical 0.47 μF as far as I can see (and like you posted earlier)

Jup, thanks, typed so fast the keyboard didn't understand  , Corrected

Spikedcola: Yes the datasheet has that on a different page, and an older one has other values too.
(no wonder we can argue about this)
Either one will be fine, would use big one if you used Vusb for a pullup also.

[Rescue Toaster]

When I first assembled mine, I had forgotten the cap on pin 14 completely, and it caused Windows to BSOD, hilariously enough. The standard 0.22uF worked fine for me.

[maximilian0017]

When I first assembled mine, I had forgotten the cap on pin 14 completely, and it caused Windows to BSOD, hilariously enough. The standard 0.22uF worked fine for me.

Thats a long time to not find a bug, Win98 presentation by Bill Gates with USB scanner
http://www.youtube.com/watch?v=RgriTO8UHvs

[SpikedCola]

When I first assembled mine, I had forgotten the cap on pin 14 completely, and it caused Windows to BSOD, hilariously enough. The standard 0.22uF worked fine for me.

Thats a long time to not find a bug, Win98 presentation by Bill Gates with USB scanner
http://www.youtube.com/watch?v=RgriTO8UHvs

Thats an awesome video!

[doveman]

I just wanted to check something before I get started.

I'm building to this design: http://www.xboxhacker.org/index.php?topic=12306.msg96809#msg96809

I'll be using it to flash a large-block (512MB) Jasper.

Should I use the files linked to in this post: http://www.xboxhacker.org/index.php?topic=12306.msg96457#msg96457
or the v3bplus2 ones linked here: http://www.xboxhacker.org/index.php?topic=12306.msg89624#msg89624

[doveman]

Hmm, I seem to be having some problems installing WinPic800 3.64. It completes the installation, but then another process starts named Wp800d~1, which just hangs. Is this trying to install a driver? I'm using XP Pro SP3.

Nevermind, I managed to install the driver manually. Now I'm getting an error when I try and program the PIC.

I've got it configured as shown in these two pictures, which matches the settings in the PDF:





The error I get when I try and program it is shown below:



If I try and erase it, it says "Detected -> Unknown, Device erased: OK" so does that means that everything's wired up right? Is it likely to be either a too long cable or wrong LPT port settings, or does the error indicate something more fundamentally wrong?

[sandungas]

@doveman
V3bplus2 & nandpro 2.0b (older versions are not compatibles with big boobs blocks jaspers)

You must configure it like this (is the default "layout" for art2003)... and you must see this "lights" turned on in the "hardware config" window

If not.... reboot your PC, change your bios settings for LPT & try again
If not... try with another PC

After you see this windows like the image i posted....
Test device... and test hardware (both must show correct info)

Then you can load the .hex
After loading the hex, configure this window

*Change the value for the oscillator you are using

[doveman]

Thanks sandungas,

I'm a bit confused about the lights in Hardware config, 'cos the ART2003-LVP PDF shows that only Data light should be lit and that is the only light that's lit for me. Any idea what would cause Data to be lit and not DataIn and Clock?

I've tried all the different LPT BIOS settings now to no avail and unfortunately my other PC doesn't have an LPT port (damn progress!).

If you haven't got any other ideas, I'll try rebuilding it on stripboard, rather than the breadboard I'm currently using.

[sandungas]

Maybe is your pc power supply that is unable to give the correct volts, try to measure volts in vdd
You can feed the art 2003 externally, in this squematic they are using the usb

*This image is from another forum, not mine

[doveman]

Thanks for the suggestion, but I doubt that's the problem 'cos I've been using this port to flash the NAND successfully.

Edit: Well I've rebuilt it on pad board and it's still not working. In fact my PC booted with the LPT port set to EPP and it didn't work so I changed it to SPP and then my HD got corrupted and now it can't read my local profile and logs me on with a temporary one. Then I disconnected the LPT lead and the next time I tried to boot I got a BSOD saying "Unmountable boot volume"!

So maybe you were right about my PC not being able to spare the power. Just how much does this thing need anyway?

What's the purpose of the diodes? I'm just wondering why it's OK to take 5v direct from USB but we need the diodes when taking the power from the LPT.

[maximilian0017]

Thanks for the suggestion, but I doubt that's the problem 'cos I've been using this port to flash the NAND successfully.

= Something completely different

Edit: Well I've rebuilt it on pad board and it's still not working. In fact my PC booted with the LPT port set to EPP and it didn't work so I changed it to SPP and then my HD got corrupted and now it can't read my local profile and logs me on with a temporary one. Then I disconnected the LPT lead and the next time I tried to boot I got a BSOD saying "Unmountable boot volume"!

So maybe you were right about my PC not being able to spare the power. Just how much does this thing need anyway?

Setting the LPT mode or draining the power of the LPT port will not corrupt your hd, there must be something else, There are a lot of good utility disk's around for windows to repair it.

Did you turn the pc off with the button on the pc case?, or maby moved the pc while it was powered on?, that will be the reason.

What's the purpose of the diodes? I'm just wondering why it's OK to take 5v direct from USB but we need the diodes when taking the power from the LPT.

A diode lets current through in only one way and not the other.
The datapins of the LPT port alternate between 0-5v or 0-3,3v if you were to connect them togheter you would short these pins and probably damage the LPT port.
The usb connector outputs 5v and is always on(unles you create a short or in some cases if the pc is off), so no diode is needed.

Please connect the flasher to the LPT port and measure the voltage across the different pins on the PIC, this should be at least 3,3v, take care to not short any pins.

[doveman]

Setting the LPT mode or draining the power of the LPT port will not corrupt your hd, there must be something else, There are a lot of good utility disk's around for windows to repair it.

Did you turn the pc off with the button on the pc case?, or maby moved the pc while it was powered on?, that will be the reason.

Well, yes and no. After changing the port to SPP and rebooting, that's when it couldn't find my profile (which suggests the HD was already corrupt. I didn't see the boot process before it got to that, so it might have done a disc check) and tried to use a temporary one. Then my screen just went completely black and stayed like that, so I had little choice but to power off manually. The next reboot, I saw that the disc check started and identified loads of corrupted files.

I had this happen to me a few months back actually. When it happens, I can hear the hard drive spinning up and down, like it can't get enough power to stay on. I'm using an EF28 DC-DC board powered by a Dell DA-2 brick and last time it happened I think it was a dodgy graphics card. I had to connect the HD to my other PC to backup what I could and then do a several hour long Western Digital diagnostic on it before it was fixed 

Once I get my PC sorted, what I think I'll do is take 5v from a spare PC PSU I've got and use that to power the programmer, just to avoid the risk of the same thing happening. Strange though, 'cos I was powering a USB HD from the PC the other day and I'm sure that draws more current than the programmer.

A diode lets current through in only one way and not the other.
The datapins of the LPT port alternate between 0-5v or 0-3,3v if you were to connect them togheter you would short these pins and probably damage the LPT port.
The usb connector outputs 5v and is always on(unles you create a short or in some cases if the pc is off), so no diode is needed.

Please connect the flasher to the LPT port and measure the voltage across the different pins on the PIC, this should be at least 3,3v, take care to not short any pins.

Ah, thanks for explaining, that makes sense.

I imagine I'd need to run some software to cause the LPT port datapins to go high. I think there was something with the port driver used for flashing the NAND that did that. Is it possible with WinPIC800? What exactly are you suggesting I measure? Won't there only be voltage on PIC pin 20?

[maximilian0017]

I had this happen to me a few months back actually. When it happens, I can hear the hard drive spinning up and down, like it can't get enough power to stay on. I'm using an EF28 DC-DC board powered by a Dell DA-2 brick and last time it happened I think it was a dodgy graphics card. I had to connect the HD to my other PC to backup what I could and then do a several hour long Western Digital diagnostic on it before it was fixed 

Once I get my PC sorted, what I think I'll do is take 5v from a spare PC PSU I've got and use that to power the programmer, just to avoid the risk of the same thing happening. Strange though, 'cos I was powering a USB HD from the PC the other day and I'm sure that draws more current than the programmer.

The only way it can happen the way you are describing it is if your powersupply is marginal and you make a short in the wiring of the programmer. 9The programme should not use more than 40mA with a chip connected as a normal external 2,5 Inch HDD takes about 450mA during startup and 300mA.
The difference between a light pc with no load apposed to heavy load can be about 3000mA so your pc setup should chrash also if you had real problems supplying current.
It's more likely that a program was still using the registry of your computer and did not release it in time or didn't get to write everything it wanted to write thus corrupting your profile.
Restarting during a disk check is really bad and can easely cause the troubles you are describing.
Maby your hd is dying?

Please repair your pc setup and check all connections of the programmer(correct numering of the LPT port etc(not the wrong way around etc, then without the chip connected(you did use an ic socket didn't you?) measue between the pin 20 and pin 8/19, this should be at least 2.9v or so.
Winpic800 will sort out all the pin toggeling for you.

If you want to use an external powersupply make sure it has exactly 5v across it and connect the ground pin to the lpt ground before doing anything else, an usb cable connected to the same computer is better because the ground is already connected and the 5v is also correct.

Good luck,

[doveman]

Yeah, I didn't think the chips draw could really compare to a USB HD!

The PC was shut down safely before changing the LPT to SPP, so I don't think anything was still using the registry and corrupted the profile. The first time I booted after changing the LPT port, I got the corrupted profile and then the screen went off after logging in to a temporary profile. This is when I had to do a hard shutdown. The disk check happened on the next boot (and found many corrupted files) and I didn't restart during that.

I thought my HD might be dying last time it happened, but it seems to work fine when connected to my other PC and I can access the files (except for the corrupted ones obviously). Doing the Western Digital Quick diagnostic test fails, as it did last time this happened, and I have to do the Extended test (which takes a couple of hours) after which the Quick test works again, so probably something fundamental got corrupted. I'll see if WD think it needs to be returned though.

Anyway, I've re-wired my PC powersupply so that it's drawing more current directly from the 12v AC-DC adapter and less through the DC-DC board, which hopefully will help. I've also removed the PCI modem in case that was dodgy.

Once I get it back up and running I'll do the checks you suggest (yep, thankfully used an IC socket). When you say that WinPic800 will sort out the pin toggling for me, what exactly do I need to do in it to make it send the datapins high, or will it do that as soon as I load it and select the Art2003 profile?

[maximilian0017]

The PC was shut down safely before changing the LPT to SPP, so I don't think anything was still using the registry and corrupted the profile.

When you say that WinPic800 will sort out the pin toggling for me, what exactly do I need to do in it to make it send the datapins high, or will it do that as soon as I load it and select the Art2003 profile?

You werent overclocking were you?, that can have the same effect sometimes.
I'll let the pc part up to you as not to clutter this thread anymore, if you need specific help about that you PM me.

Thought you were talking about the programming not the testing, but if you have all the diodes the right way around and measure the voltage over the supply pins at any time it will probably be all right, just test before and after opening winpic800.

To test the LPT data port you can do this:
1. Open C++ PortIO and select HEX under format (The PortIO package is in the Nandpro20b archive as Port95nt.exe)
2. Adres is the adres of your parralel port (0x378 for lpt1 etc, check under the lpt1 properties of the windows device manager)
3. If you change "value", do-d7 will change to:
0x00  d0-d7 0v
0xFF  d0-d7 3,3v
0x55  d0=3,3v, d1=0v, d2=3,3v, etc
0xAA  d0=0v, d1=3,3v, d2=0v, etc

Cable pinout
http://pinouts.ru/ParallelPorts/ParallelPC_pinout.shtml

[doveman]

You're quite right, this has gone way off-piste I'll just say no, I don't overclock this machine.

Thanks for the tip about PortIO, I'll get round to testing the programmer again once I've recovered from repairing my PC