HB-F1XV black screen

Page 1/5
| 2 | 3 | 4 | 5

By fallen1

Resident (39)

fallen1's picture

22-10-2019, 15:18

My previous thread disappeared, so a short summary: Got a HB-F1XV that had leaking damage + traces cut + repairs, non woking A/V out, HIC board recapped. It had a noise on video and some sort of cold solder joint that was cutting off RGB out. After resoldering and trying to wobble the HIC again, my msx went black Crazy

Current status:

  • Caps-lock not responding
  • No out RGB on VDP, but sync is coming thru
  • Tried memory cart, but no difference. Memory measured ok
  • Voltages are ok (+5,-12,+12)
  • CPU clock ok
  • CPU /WAIT ok
  • ROM has problems with data lines
  • 2 drams have data, 2 not

I have put the ROM into socket allowing me to hook up more channels:
Addresses are coming in
Output 1 is almost constantly high (verified by swapping channels)

I assume that the ROM is shot. If anyone can burn me the ROM, I'd happy to buy it :)
I found HN27C101AG to be pin compatible with the 831000 MASK ROM.
Is there a guide how to assemble the hex file for burning an EPROM for the F1XV?

Got following images:
Filename Filesize CRC
f1xvbios.rom 32768 2C7ED27B
f1xvdisk.rom 16384 04E4E533
f1xvext.rom 16384 B8BA44D3
f1xvfirm.rom 1048576 77BE583F
f1xvkdr.rom 32768 B4FC574D
f1xvkfn.rom 262144 7016DFD0
f1xvmus.rom 16384 5C32EB29

Login or register to post comments

By fallen1

Resident (39)

fallen1's picture

23-10-2019, 00:23

The f1xvfirm.rom looks like full ROM - plays ok in BlueMSX, so just need to get it burned

By Pentarou

Master (223)

Pentarou's picture

23-10-2019, 02:44

Sorry, can't help you with the ROM layout, but f1xvfirm.rom is 8 Megabit, so I doubt it will fit into an 1 Megabit mask ROM.
Besides, that file doesn't look like part of the BIOS, but something related to the Kanji driver. More info in this post:
https://www.msx.org/forum/msx-talk/hardware/sony-hb-f1xdj-internal-software

Does your HB-F1XV have an 831000? All the PCB pictures I've seen (the one on the wiki too) use LH531000.
Pinout should be the same:

  531000/831000           27c301/27c1000
 
                       Vpp  1 +-v-+ 32 Vcc
                       OE   2 |   | 31 PGM/
A15  1 +-v-+ 28 Vcc    A15  3 |   | 30 nc
A12  2 |   | 27 A14    A12  4 |   | 29 A14
A7   3 |   | 26 A13    A7   5 |   | 28 A13
A6   4 |   | 25 A8     A6   6 |   | 27 A8
A5   5 |   | 24 A9     A5   7 |   | 26 A9
A4   6 |   | 23 A11    A4   8 |   | 25 A11
A3   7 |   | 22 A16    A3   9 |   | 24 A16
A2   8 |   | 21 A10    A2  10 |   | 23 A10
A1   9 |   | 20 CE/    A1  11 |   | 22 CE/
A0  10 |   | 19 D7     A0  12 |   | 21 D7
D0  11 |   | 18 D6     D0  13 |   | 20 D6
D1  12 |   | 17 D5     D1  14 |   | 19 D5
D2  13 |   | 16 D4     D2  15 |   | 18 D4
GND 14 +---+ 15 D3     GND 16 +---+ 17 D3
Quote:

I found HN27C101AG to be pin compatible with the 831000 MASK ROM.

According to my device programmer library, HN27C101AG is a 32 pin device so it's not exactly the same, but the PCB seems to have extra holes, so you can probably replace the ROM with a generic 32 pin 1 Megabit non-JEDEC device.

By fallen1

Resident (39)

fallen1's picture

23-10-2019, 10:09

Pentarou wrote:

Sorry, can't help you with the ROM layout, but f1xvfirm.rom is 8 Megabit, so I doubt it will fit into an 1 Megabit mask ROM.
Besides, that file doesn't look like part of the BIOS, but something related to the Kanji driver. More info in this post:
https://www.msx.org/forum/msx-talk/hardware/sony-hb-f1xdj-internal-software

Ok I assumed that there is some (bit->byte->bit) conversion taking place but that would be just silly. Correct, the 1Mbyte image is firmware from the daughterboard.
But you brought me to a different idea when reading about the extra ROM - to get the ROM config from BlueMSX directly for this machine - and BINGO (well, it looks very promising):

0 0 0 0 24 "Machines/MSX2+ - Sony HB-F1XV/f1xvkfn.rom" ""
0 0 0 0 33 "" ""
0 0 0 0 84 "" ""
0 0 0 0 32 "" ""
0 0 0 4 20 "Machines/MSX2+ - Sony HB-F1XV/f1xvbios.rom" ""
0 3 0 6 40 "Machines/MSX2+ - Sony HB-F1XV/f1xvfirm.rom" ""
3 0 0 8 22 "" ""
3 1 0 2 20 "Machines/MSX2+ - Sony HB-F1XV/f1xvext.rom" ""
3 1 2 4 42 "Machines/MSX2+ - Sony HB-F1XV/f1xvkdr.rom" ""
3 2 2 4 65 "Machines/MSX2+ - Sony HB-F1XV/f1xvdisk.rom" ""
3 3 2 2 78 "Machines/MSX2+ - Sony HB-F1XV/f1xvmus.rom" ""

Now just to figure out how to use this info - need to dig in the BlueMSX docs

Pentarou wrote:

Does your HB-F1XV have an 831000? All the PCB pictures I've seen (the one on the wiki too) use LH531000.

Yes, I saw that and wondered too, but it's 831000-20 so I guess it's a different revision of the board.

Pentarou wrote:

According to my device programmer library, HN27C101AG is a 32 pin device so it's not exactly the same, but the PCB seems to have extra holes, so you can probably replace the ROM with a generic 32 pin 1 Megabit non-JEDEC device.

Right, there are 32 holes + they are already wired to hold also 32pin EPROM it seems (pins 30,31,32 are all pulled high, /OE wired to /CE). If I understand correctly, JEDEC device would mean swapping /OE and A16 via an adapter (as I didn't find any non-JEDEC EEPROM with required properties, this may be a solution too)

Wanted to avoid getting a programmer as I don't usually deal with (E)EPROMs, but as I don't have 1:1 dump of the ROM, some testing/revisions will be necessary, so I may need to get one at the end.

Thank you Pentarou for your input!

By fallen1

Resident (39)

fallen1's picture

23-10-2019, 11:48

Grabbed this directly from BlueMSX UI

The ROM in question has 128kB, looking at the F1XV slot map, not sure what all I need to put there? Is it just the 32kB Main-ROM at 0000h in Slot 0-0 (and rest empty)?

By Pentarou

Master (223)

Pentarou's picture

23-10-2019, 13:21

Unfortunately blueMSX uses splitted ROM files.
Maybe a dump of the original ROM, even if it has stuck bits, can be useful to recognize the ROM layout?

You absolutely need these:
Main ROM/f1xvbios.rom = 32Kb
Kanji driver/f1xvkdr.rom??? = 32kb
Sub ROM/f1vext.rom = 16Kb

The rest are optional, i.e. The machine will still boot and give you the basic prompt without them:
Disk ROM/f1xvdisk.rom = 16kb
MSX Music/f1xvmus.rom = 16kb

By fallen1

Resident (39)

fallen1's picture

23-10-2019, 14:32

Thanks for your hints Pentarou!

I had that idea of dumping what I could get (getting shift registers today so I can dump it using arduino - have no programmer), but thought it would be no good as all bytes would change (but I understand now that I could compare it bit wise and to find the overlap).

You brought me again to another idea by mentioning the stuck bit - I pulled the ROM out and connected the socket's O1-O3 data pins to logic analyzer and the O1 is constantly high! Without the chip inside! So the problem is elsewhere - something is pulling the 1st bit on data bus high Crazy

I wonder can I measure which device does that without desoldering & socketing everything?

By fallen1

Resident (39)

fallen1's picture

23-10-2019, 22:41

Desoldered RAM as one of the chips had data line with very low resistance - but the D0 signal remained high (tried also with megaram but no difference).

I thought there may be a resistor busted - there are 3 resistor arrays, but all seem to be fine ( so all pins are showing the same resistance).

I got an idea that in order to see what is pulling the D0 high was to desolder all pins that connect to that bus using desoldering station: I was able to disconnect VDP, (socketed) FDC controller, buffering IC located near slots, printer buffer, Z80, the large black ROM card, (socketed) ROM, those 2 removed RAM ICs but D0 is still high. What remains is S-1985 and the other SMD IC at the bottom (not touching those...). Measured resistance, but on both ICs have D0 and D1 equal values.

Will check again tomorrow whether I missed any pins, but after that, I'm running out of ideas (don't want to lift SMD legs)

By fallen1

Resident (39)

fallen1's picture

23-10-2019, 23:30

This is with RAM + FDC + ROM fully disconnected (and D0 pins disconnected for ICs from previous post):

1(yellow): CE
2(cyan): D0
4(blue):D1

By Pentarou

Master (223)

Pentarou's picture

23-10-2019, 23:55

I took a look at the F1-XD schematics and saw that there are pull-up resistors next to the Z80.
If the F1XV has the same resistor networks then it's normal that the signal is high if no device is driving it low.
What happens if you connect D0 to GND? Does some of the few remaining ICs become hot?

Maybe I'm wrong in thinking it, but I believe you are trying to find a fault in the wrong places...
If you are sure that the machine died when you "wiggled" the HIC you should be looking at that, I don't understand how a module with only analog signals can cause problems to the address bus.

Edit after scope picture: So D0 is not high but it seems to follow some other signal?

By fallen1

Resident (39)

fallen1's picture

24-10-2019, 01:17

I'm not sure about the cause anymore - the last thing I did was wiggle the board and it went black. I must have it shorted somehow (but no idea how as there is everywhere 5V and closest is the VDP). Sure thing is that the RAM chip connected to D0 was shorted at that pin, so something has happened there

I was looking into the schematics 1000 times and never saw the pull-up there Sad So it means that CPU is not driving that line down, doesn't it? So I was definitely searching at a wrong place Shocked!

Connecting D0 to GND does nothing with those 2 large SMDs (rest is disconnected + checked and none gets hot).

I can remove the HIC just to simplify the equation - like you said, it's just analog and drives nothing relevant at the moment (actually, I have no idea what is under that board - it was desoldered/repaired by the previous owner mitigating leaking damage, destroying one trace in process, so if there is a printed trace under it, it may be the cause).

Is there are reason why would the Z80 leave D0 high after reset? Also the voltage at CPU-D0 is 5V, according to the schematics, it should be around 3.3V if I read it correctly - moreover it was 5.88V at the ROM-D0 pin, but Z80 max input is 7.7V so should be fine, it's just strange.

I have no device where to test the Z80 so I will have to build something

Thanks a lot Pentarou!

EDIT: DO has approx. 5v with 1.52V peak-to-peak. have not seen it on the logic probe of course, but yes - follows (inverted) CE it seems

Page 1/5
| 2 | 3 | 4 | 5