OneROM Fire 28 and the Tandy 1000 #194
Replies: 1 comment 4 replies
|
Very cool - and thanks for posting details for others. Sounds like you got lucky putting in One ROM backwards - if you leave it powered on while backwards for a while it can burn one of more GPIOs pads Ask me how I know .... Out of interest - do you know if that 128KB ROM U51 is a 32 pin ROM? Alternatively could be a 231024/TCS531000 28 pin, but looks more like 32 pin shape here 
|
To be clear, I did not. But I did power it on without any ROM in the socket, and it definitely sings, so that was just an attempt at humor. It would be sad if you've already fried your OneROM at that point :( I'm not sure what Tandy did to avoid the characteristic 'chirp' of the SN76496 that other systems have like the BBC Micro - it must mute the chip nearly immediately.
Looks like a DIP-28 mask ROM. 
Quick question - The ROM on the 1000 sits directly underneath the floppy drive cage, and there is not a lot of clearance. Even if the 2x5 header would fit, you would still have to disassemble everything to change the jumpers, so I didn't bother here. I don't know if this was mentioned anywhere already, but would there be any issue with just leaving a USB-C cable attached to the OneROM and snaking it out some gap in the case? |
|
I’m going to guess that is a 231024/TCS531000 then, as I’ve not come across any other 128KB 28 pin ROM pinouts. Yes, should be fine leaving a USB-C cable attached. Come future firmware it won’t even stop serving when you plug USB in. |
|
I would assume that it's 256Kbit like the mask ROM in the 1000-non-A, but I admit I don't know for 100% certain. It's frustrating that they did not fully update the schematics in the 1000A technical reference 'addendum', they just circled a few things to comment about what was changed. I assume it is 32KB because that was what they were sending out to Tandy 1000 customers wanting to upgrade to v1.01 to use the hard drive controller card, which is what required a PAL swap. 1000A's shipped with v1.01, and I can't think of any reason Tandy would have produced two different mask ROMs, unless they just liked setting money on fire. I was reading this https://github.com/piersfinlayson/one-rom/blob/main/docs/MULTI-ROM-SETS.md and I thought this might be a way to avoid needing to use two OneROMs with the older PAL, but then I see the note that the OneROM 28 does not have X pins. Any alternative? Tying the chip select lines together is a no go without board or socket surgery. |
|
Not that I can immediately think of. In theory, if it’s just 32kB ROMs you could use the spare address pin as an extra CS line. Would need something soldered to the extra pin, and also a firmware mod. I’m not really keen on this (specifically needing hardware mods), so am more likely to try a new 28 layout at some point with the RP2350B which has more GPIOs and then add some X pins. |
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
I have successfully upgraded my Tandy 1000 from BIOS v1.00 to BIOS v1.01 via a OneROM Fire 28.
I thought I'd document some things here, since the Tandy 1000 (the original, not 1000A) ships with two possible ROM configurations, and it can be a bit confusing. (I'm still a bit confused).
The 1000 originally shipped with two 16K EPROMS of the 27128 variety, in sockets U10 and U9. If you are googling how to upgrade your 1000, you may see old information about how this is going to require replacing a PAL chip. Don't worry, OneROM will come to your rescue here.
Address decoding was performed by the original PAL chip in U53 according to the following equations:
ROMSC0* = RFSH* & /MEMR* & A19 & A18 & A17 & A16 & A15 & A14 & A13
ROMSC1* = RFSH* & /MEMR* & A19 & A18 & A17 & A16 & A15 & A14 & /A13
This decodes two 8K windows to each socket.
U9: FC000 - FDFFF
U10: FE000 - FFFFF
The 16K ROM image on both chips was likely identical, with each chip just serving one half of it.
I am not sure what the point of doing this was. I'm guessing that the board was taped out and the PALs burned with the expectation of using something like the 2764, but the fact that they wired A13 to the ROM sockets meant that they may have anticipated using 27128's, which they in fact ended up doing. If anyone has a better possible explanation, do please let us all know.
If you have such a board, you can simply use two Fire 28's as 27128s. Burn the 16K BIOS ROM image to each. Done. No PAL upgrade required!
The other configuration is a single 32K mask ROM. This is the configuration I have. Mine was manufactured by Motorola, it has some un-googleable part number on it, but the equivalent EPROM pinout is the 27256.
In this configuration, you should also have R1 populated, which is a 0-Ohm resistor that simply passes A14 along. You should also have the new PAL - you might see little round sticker on it, even.
With the new PAL we now have the 32K window F8000 - FFFFF decoded to U9. Our BIOS ROM image is still 16K, though, so the first 16K is padded with FF. I am not sure if anything is even decoded for U10 anymore.
You can try 1.02 and 1.03 on a 1000 as well - if you get a crazy looking screen a whole lot of mess and some blinking characters immediately on poweron with 1.02, it seems there is a bad dump out there.
If you hear the 3-voice sound chip scream at you, you probably have the OneROM in backwards. :D
There are some ROM images you might run into that are advertised as "good for Tand-EM" that are 64KB. I don't know how somebody wound up dumping 64KB from their 1000, but if you look at these ROMs in a hex editor, they appear to be just the 32K layout repeated twice. However, instead of FF they have some weird garbage. I suppose they might still work, you can just set the OneROM flasher to 'truncate'.
Tandy 1000A
The service manual supplement for the 1000A shows a single 128Kbit ROM in the BOM, U51, part 8040328. I guess they ended up agreeing with me that everything prior was confusing for no particular benefit.
All reactions