PedroM on Voyage 200 - Link issues - Page 1

Hi

Having applied the previous fixes some weeks ago to PedroM there's been no further data corruption at all when running out of flash.

I've been able to use it reliably with the TI-92+ and (home made) Grey link (9600 serial using a 328p for protocol conversion and USB to serial chip) cables and Black link (4$ real serial bit twiddler built according to instructions on the ticalc site)

However, I just bought a Voyage 200 as the 92+ didn't really have enough space and I can use the 92+ for AMS and the 200 for PedroM

Things are not going as well as I'd hoped re data transfer:

Ti-Connect just will not work at all to send files to the calculator in PedroM, with either cable, strange error from TI-Connect (Unable to communicate with calculator, please check your connection (after it has successfully detected it and hence communicated with it...) then The parameter is incorrect (80070057) from tisendto) and a Timeout on the v200 which I suppose is to be expected if TI-Connect died mid-transfer. (if I'm installing an OS update or in AMS Ti-Connect /does/ work, but it also works on my TI-92+ running AMS or Pedrom sending files in the usual way)

TiLP connects and seems to work fine (but has other limitations like not putting files in the correct folders when sent)

Have I missed something here? If it is a bug is it in TI-Connect or PedroM? I thought the 92+ and v200 were essentially the same calculator just with storage differences but they're not behaving the same... Would buying the official TI USB cable help at all? (I suspect not but if it definitely will perhaps it's worth my paying the inflated eBay prices for one)

Hi,

I suspect the DEVICE_LINK_ID of PedroM to be incorrect for the V200 (It is the same value than the 92+ one = $88) or the MID sent by TIConnect is not recognized (PedroM expects either $88, $89, $08 or $09)
i.e. error in PedroM protocol handling.
What is the calculator identified by ticonnect?
(Changing the calc won't help)

I don't know why TiLP doesn't send the file in the right folder.

Any tip for this problem about TILP ?

On 1) my real 89T running AMS 3.10, with both SilverLink and DirectLink, I've just:
* created a new folder "test" and a new text file "test" in that folder;
* opened TILP to perform ready, dirlist, then drag & drop that text file to the computer;
* deleted the file and folder on the calculator side;
* run dirlist again in TILP;
* transferred the file back to the calculator: the status line indicates that "test\test" was received, and the VAR-Link agrees, of course.

Then, I switched TILP's configuration to TIEmu cable, port #1, and started TIEmu images 2) 92+ PedroM 0.82 20090730 then 3) V200 PedroM 0.82 20091004 , configured to communicate with TIEmu cable, port #2.
On both models, after reconfiguring the model on the TILP side, the aforementioned file was transferred to test\test, as shown by both dirlist and the usual "cd test" followed by "ls" in the PedroM prompt.

=> there's no issue on the TILP side

@Mog Kupo: make sure that when transferring a variable to the calculator, you're not inadvertently forcing the folder. If you want TILP to respect the folder (if any) specified inside the files transferred to the calculator, drop the selection onto the head of the tree, "Variables", instead of dropping it onto "main", one of the files in main, the head of another folder, or one of the files in that folder.

Hi

@PpHd

TI-Connect does identify it as a Voyage 200 correctly (in that that's what it says when detecting the cable and in the corner of the file transfer window), it just fails sending the files
If I attempt to open TI Device Explorer in TI-Connect it says "Voyage 200 detected but couldn't get status"
If I attempt to capture the screen in TI Screen Capture it works fine
If I open TI Device Info it gives the device type as TI-92 Plus (and the total RAM as 128K oddly, all other fields read N/A apart from battery status)

@Lionel Debroux

Ah, I did just double check that but it doesn't matter where I drop it, the file ends up in the current directory I happen to be in on the Voyage 200
If sent to the TI-92+ in AMS/Pedrom via TI-Connect everything ends up in the right place so I assume nothing is wrong with the files
that's not a huge issue though, given I can send them in blocks, it's just not ideal.

--
Adam

I guess I'll need a copy of one of the files which triggers such behaviour
I'd think that this file contains no folder information.

@ Lionel Debroux

ahh, here's a dump of one of them (I say hoping 'code' in this forum means monospaced)

00000000 2A 2A 54 49 39 32 50 2A 01 00 62 69 6E 00 00 00 **TI92P*··bin··· 00000010 00 00 53 69 6E 67 6C 65 20 66 69 6C 65 20 64 61 ··Single file da 00000020 74 65 64 20 4D 6F 6E 20 46 65 62 20 31 30 20 31 ted Mon Feb 10 1 00000030 31 3A 33 35 3A 34 30 20 32 30 01 00 52 00 00 00 1:35:40 20··R··· 00000040 66 72 65 73 65 74 00 00 21 00 00 00 93 00 00 00 freset··!······· 00000050 A5 5A 00 00 00 00 00 39 61 00 00 20 36 38 6B 50 ·Z·····9a·· 68kP 00000060 00 00 00 00 00 32 00 00 00 03 00 00 00 00 00 00 ·····2·········· 00000070 00 00 00 00 00 00 00 00 00 00 4A 78 00 30 66 04 ··········Jx·0f· 00000080 58 8F 4E 75 20 78 00 34 4E D0 4E 42 4E 75 00 00 X·Nu x·4N·NBNu·· 00000090 F3 4E 09 ·N·
Actual file at http://home.kupo.be/media1/ti92/%5bbackup%5d/7%20bin/freset.9xz (mog/kupo for the http login)

It does seem the bin foldername is there at 0x0a though... (plus ti-connect puts it in the right place)

(it was created as were all the files in my backup folder by my using sendcalc in PedroM in the TiEmu emulator which seems to save them with the folder information, perhaps doing it that way has introduced some sort of issues...)

Thanks. But well... on my side, under Linux, again with PedroM 92+ 0.82 20090730 and PedroM V200 0.82 20091004 in TIEmu, and TILP from Git, I still can't reproduce your problem.

1) When I drop the file onto "Variables", it ends up as bin\freset; when I drop it onto "system", it ends up as system\freset; when I drop it onto "main", it ends up as main\freset.
2) I have made two copies of that file, and hex-edited their respective on-calc names to "freset2" and "freset3". When I drop the set of three files onto "Variables", they end up in bin; when I drop them onto another folder, they end up in the folder I requested.

I do acknowledge that you're seeing an issue, and I'm trying to imagine where it's coming from, but I'm failing at that task... The behaviour I can see here is the behaviour I expect to see.


About your program: 4E42, heh
For that purpose, technically, you don't need to make a "kernel"-based program. If you want to use "kernel"-based features later, then sure, go ahead

It is refreshing to read new topics about the ti calculators. thank you. et pas merci pour le début de kernostubnel, mais je suppose que c'est endémique au sujet

Lionel Debroux (./8):
Thanks. But well... on my side, under Linux, again with PedroM 92+ 0.82 20090730 and PedroM V200 0.82 20091004 in TIEmu, and TILP from Git, I still can't reproduce your problem.

1) When I drop the file onto "Variables", it ends up as bin\freset; when I drop it onto "system", it ends up as system\freset; when I drop it onto "main", it ends up as main\freset.
2) I have made two copies of that file, and hex-edited their respective on-calc names to "freset2" and "freset3". When I drop the set of three files onto "Variables", they end up in bin; when I drop them onto another folder, they end up in the folder I requested.

I do acknowledge that you're seeing an issue, and I'm trying to imagine where it's coming from, but I'm failing at that task... The behaviour I can see here is the behaviour I expect to see

Ah well, if it works at your end I've made an error somewhere, I'll try to work out what I'm doing that's different, curious that it works as expected with ti-connect though, I'd best try with standard PedroM to make sure none of my changes are the issue (though the same build behaves on the ti-92+) and check my TiLP is the latest version etc.

Thanks for taking the time to look into it for me

Lionel Debroux (./8):
About your program: 4E42, heh
For that purpose, technically, you don't need to make a "kernel"-based program. If you want to use "kernel"-based features later, then sure, go ahead

Yes, I forgot the _nostub directive, still getting used to as92, well spotted though

squalyl (./9):
It is refreshing to read new topics about the ti calculators. thank you. et pas merci pour le début de kernostubnel, mais je suppose que c'est endémique au sujet

Ah, I came to it a bit late in the day

Was looking for a replacement for my venerable Casio FX-730P I've had since I was 11 or 12, first got a Casio Classpad 330 plus, then discovered not only is the basic on that slower than the ZX81 I started my programming with as a seven year old in 1982 but they refuse to make the SDK available for it to write addins, american friend recommended the TI machines

Looked at the new ones and noticed TI are actively trying to prevent people running code on the hardware they've paid for so the 68K ones seemed the best bet (plus I'm quite familiar with the amiga, have an A1200 here still) so went for those.

Got a TI-92+ as I couldn't find a v200 fora good price with a working screen, which I've modified somewhat (it's got a USB port now which just works as a grey link cable, uses an arduino nano running a protocol conversion sketch wired to the link port), then three weeks later what pops up on my doorstep for almost nothing but a near mint condition v200 so have that too now, lol, using the 92+ for AMS (so I can run GTC) and the voyage 200 with PedroM which I'm going to add a basic to, only seems to be about 8K available to play with before it exceeds 128K but given the library and editing routines are already there something that just runs in place the program you specify should be doable in 68K ASM (which I'll have to learn, was a Z80 and 6502 man myself, other than a very tiny amount on the amiga back int he day that I long ago forgot. been looking at that assembly the past couple of weeks, first CPU I've used with separate address/data registers and no INC/DEC instructions, if anyone can recommend an alternative to the Rodnay Zaks manuals for the 68000 it'd be appreciated)

Wow, nice!

The Classpad 330 and its newer derivative the fx-CP400 stink, indeed...

inc/dec are addq/subq on the 68000; the latter accept immediate arguments from 1 to 8, and support .b, .w and .l sizes.
Even though they're split between data and address registers (a7 is special, stack pointer), you'll probably love the fact that compared to the Z80 and 6502, the 68000 isn't accumulator-based, offers more instructions (multiplication, division, lea, various others), more registers, more (and more powerful) addressing modes, etc.
The 68020 used in the A1200 has more instructions and even more powerful addressing modes, mainly the great "memory indirect" modes, perfect for some use cases.