hatmoose
Enlightened
After many days of searching I've found some engineering drawings of the TSOP burn-in test socket
Resurrecting the 32/64MBIT flash project and 8/16 bit addressing.
- Thread starter hatmoose
- Start date
Big_P Big_p
Grand Master
Anyone tested these with other eeproms 27c322 should almost be possible as well.
Tailsnic Retroworks
Champion
2x HCT244 are for Data Output bus and 3x LVC244A are for Data Input bus?
https://www.mvs-scans.com/misc/toshiba-64mbit-adapter-master-ack2.zip
Just to be clear, they need to be 74HCT244 and not 74HC244
If I understand this well, 3x LVC244A prevent +5V data input from MVS enter to flash rom decreasing it to +3'3V, and 2x HCT244A increase +3'3V data output from flash rom to +5V to the MVS
Also you play how to send Q-15 signal, if by one line of LVC244A for 8-bit input mode, or by one line of HCT244 for 16-bit output mode
Last edited:
RepairGuy95
Champion
These would work as a substitute for any 64mbit mask rom?
Tailsnic Retroworks
Champion
For 64mbit CAPCOM Sony ZN-2 eproms not, but for MVS / Sega Model 3 64mbit eproms yes
RepairGuy95
Champion
Excellent. Ocean hunter on model 3 looks cool.
Tailsnic Retroworks
Champion
I'm starting to work on a new rev. of my convert PCBs for all sections. If I can make it work for V-section, C-section 32mbit/64mbit mode, I can share it for all.
The 2x HCT244 are for data (flash to mvs) and the 3x LVC244A are for address (mvs to flash).
The 3 jumpers deal with getting the Q-15 going to/through the correct 244.
1 jumper does the /BYTE signal on the flash
1 jumper picks the trace path from the dip socket pin to HCT244 or LVC244
1 jumper connects the trace from the HCT244 or LVC244 to the flash
Tailsnic Retroworks
Champion
correct, the flash should only see 3.3V and the MVS only 5V
Tailsnic Retroworks
Champion
So you can confirm me LVC244 is a +5V to +3.3V level shifter and HCT244 is a +3.3V to +5V level shifter. I have seen in your PCB you establish LVC244 to be active always and HCT244 being active with OE signal state. I supose in my PCBs both HCT244 have to be active LOW because all the data output bus is shared with the 4 flash roms.
Just note this isn't my PCB, I just tweaked it to add support for picking BYTE vs WORD access.
The LVC244 outputs are 3.3V. The HCT244 outputs are 5V.
The LVC244's are always enabled because they only carry the address lines from the MVS to the flash chip. The address lines will just get ignored by the flash chip when its not enabled.
The HCT244's are the data lines going from the flash to the MVS. These must only be enabled when the MVS is asking for data from the flash, otherwise it will cause a data bus conflict.
The LVC244 outputs are 3.3V. The HCT244 outputs are 5V.
The LVC244's are always enabled because they only carry the address lines from the MVS to the flash chip. The address lines will just get ignored by the flash chip when its not enabled.
The HCT244's are the data lines going from the flash to the MVS. These must only be enabled when the MVS is asking for data from the flash, otherwise it will cause a data bus conflict.
Tailsnic Retroworks
Champion
This shouldn't be a problem if always active in the new implementation I'm making, because the use of the Q0-Q15 of all the 4 roms common bus is going to be OE dependent.
Is the same use I have done in the old PCB with SOP44 PCBs, Q0-Q15 always with info dependant of which rom is OE activated. I think is going to work good.
If for example all OE are high, Q0-Q15 won't have any info inside, like original format.
The other solution is to install an extra SMD component for detect when all 4 OE signals are HIGH and deactivate the 2xHCT244. A 74LS21 can do the trick: when all 4x OE signals are HIGH, output is HIGH, in other case LOW. That way we can control 2xHCT244 combined activation
Last edited:
Tailsnic Retroworks
Champion
If you want to protect the flash completely, use 3xLVC244 + 2xHCT244 version
You will want something like this if you have the HCT244's between the flash chips data pins and the MVS. The HCT244's won't care that the flash chips aren't outputting anything and will still output a signal if they are enabled. Having no input on the HCT244 is an undefined state so it may output high or low or just randomly flip flop between high/low.
Tailsnic Retroworks
Champion
Tailsnic Retroworks
Champion
Well, I have news about my universal 4xTSOP48 adapter. I have to order a few units for testing if I have made everything right. I can test all zones and sizes, so no problem for me.
@ack I have made what we said yesterday. Adding a 74LS21 for activating/deactivating HCT244 activity when all 4 roms are disabled, and also 2 new jumpers for playing with Q15 signal:
- If BYTE is LOW (8 bit), Q15 flash is input (MVS Q15 pin signal goes to LVC244 input / LVC244 output goes to Q15 pin of 4xTSOP48 )
- If BYTE is HIGH (+3.3V) (16 bit), Q15 flash is output (4xTSOP48 Q15 pin goes to HCT244 input / HCT244 output goes to MVS Q15 pin)
RESET and WE signals of the 4xTSOP48 are tied to HIGH (+3.3V)
WP/ACC of the 4xTSOP48 is tied to GND
@ack I have made what we said yesterday. Adding a 74LS21 for activating/deactivating HCT244 activity when all 4 roms are disabled, and also 2 new jumpers for playing with Q15 signal:
- If BYTE is LOW (8 bit), Q15 flash is input (MVS Q15 pin signal goes to LVC244 input / LVC244 output goes to Q15 pin of 4xTSOP48 )
- If BYTE is HIGH (+3.3V) (16 bit), Q15 flash is output (4xTSOP48 Q15 pin goes to HCT244 input / HCT244 output goes to MVS Q15 pin)
RESET and WE signals of the 4xTSOP48 are tied to HIGH (+3.3V)
WP/ACC of the 4xTSOP48 is tied to GND
Nice work.
It looks like you have the Q-15 dip pin connected to both 244s. I suspect you might run into issues when you are in 8bit mode. Specifically in that mode Q-15 from the dip will be an address signal coming from the MVS going to one of the LVC244s, but when the HCT244 that is also connected to that trace is enabled its going output a signal and cause a conflict. So I think you need another jumper that picks if the Q-15 dip is connect to the HCT244 or LVC244. I guess another possibility might be for the high byte HCT244 to never be enabled in 8 bit mode?
It looks like you have the Q-15 dip pin connected to both 244s. I suspect you might run into issues when you are in 8bit mode. Specifically in that mode Q-15 from the dip will be an address signal coming from the MVS going to one of the LVC244s, but when the HCT244 that is also connected to that trace is enabled its going output a signal and cause a conflict. So I think you need another jumper that picks if the Q-15 dip is connect to the HCT244 or LVC244. I guess another possibility might be for the high byte HCT244 to never be enabled in 8 bit mode?