Hi From London England

[Pr0f]

Even then you'd have to move A15 to the top right left hand pin (pin 1), which nominally is at +5v, as for whatever reason Sinclair decided no one would fit a single rom. Only a single gate is needed with a 64K chip, and pin 1 and pin 20 and pin 22 are the ones that need to be displaced to connect up correctly. ROMOEH, needs inverting for use by *OE pin (22) on the 512 eprom, and A15 needs to go to pin 1, but it is currently wired to pin 20. Pin 20 can be either tied to ground, and let the *OE pin on the Eprom do the work of selecting, or tie both these pins together. I am not sure what scheme they used on the Minverva carrier from memory. Fairly sure there is a post about it on here somewhere. You can bundle up TK2 with the ROM image and use up the top quarter of the 512 eprom, so you have a full 64K image.

[Pr0f]

viewtopic.php?f=2&t=409&p=3040

like that - I found it.

[Derek_Stewart]

viewtopic.php?f=2&t=409&p=3040

like that - I found it.

Which is the link I referred to on page 4 of this message:
https://www.theqlforum.com/viewtopic.php ... prom#p2740

[tofro]

Even then you'd have to move A15 to the top right left hand pin (pin 1), which nominally is at +5v, as for whatever reason Sinclair decided no one would fit a single rom. Only a single gate is needed with a 64K chip, and pin 1 and pin 20 and pin 22 are the ones that need to be displaced to connect up correctly. ROMOEH, needs inverting for use by *OE pin (22) on the 512 eprom, and A15 needs to go to pin 1, but it is currently wired to pin 20. Pin 20 can be either tied to ground, and let the *OE pin on the Eprom do the work of selecting, or tie both these pins together. I am not sure what scheme they used on the Minverva carrier from memory. Fairly sure there is a post about it on here somewhere. You can bundle up TK2 with the ROM image and use up the top quarter of the 512 eprom, so you have a full 64K image.

I've not looked too deeply into this, but the issue 5 circuit diagram says different (http://www.sinclairql.net/srv/QLSchematicIssue5.gif, box bottom right, option 7):

You can fit a single 23256 into the left socket, then set JU2 and JU3 - Voilá, single ROM solution.

Tobias

[Pr0f]

Even then you'd have to move A15 to the top right left hand pin (pin 1), which nominally is at +5v, as for whatever reason Sinclair decided no one would fit a single rom. Only a single gate is needed with a 64K chip, and pin 1 and pin 20 and pin 22 are the ones that need to be displaced to connect up correctly. ROMOEH, needs inverting for use by *OE pin (22) on the 512 eprom, and A15 needs to go to pin 1, but it is currently wired to pin 20. Pin 20 can be either tied to ground, and let the *OE pin on the Eprom do the work of selecting, or tie both these pins together. I am not sure what scheme they used on the Minverva carrier from memory. Fairly sure there is a post about it on here somewhere. You can bundle up TK2 with the ROM image and use up the top quarter of the 512 eprom, so you have a full 64K image.

I've not looked too deeply into this, but the issue 5 circuit diagram says different (http://www.sinclairql.net/srv/QLSchematicIssue5.gif, box bottom right, option 7):

You can fit a single 23256 into the left socket, then set JU2 and JU3 - Voilá, single ROM solution.

Tobias

The only problem there is that 23256 is only capable of holding a 32K image, and the other issue is that the original poster has 2 x 27256 chips which are negative going OE and CE, unlike the ROM chips which have been masked programmed to accept a positive going OE and potentially CE as well.

[Ruptor]

Which is the link I referred to on page 4 of this message:
https://www.theqlforum.com/viewtopic.php ... prom#p2740

Yes Derek I was not ignoring your suggestion but it is only now that Pr0f has shown me that putting two 32K EPROMs in is harder than one 64K. It doesn't matter to me if the internal ROM area takes the ROM slot space since I was going to try the Tychie code anyway.
Like you say Pr0f invert ROMOEH and connect the two enable signals then put A15 on pin 1 for 27512 chip. Simple when you say it quick.

[Ruptor]

After jumping through hoops I have decided not to modify the board to fit one 64K EPROM because it just looks too complicated track wise. So I am going to make a daughter board where I can add a transistor to invert the ROMOEH line and just wire all the other lines where I want. It is better to pull the chip enable *E line permanently to zero to avoid any timing issues because the data says it has to be set first before the output enable *G. I shall straddle the daughter board across the two ROM sockets so it picks up lines 1 to 14 on IC33 and 15 to 28 on IC34 so it makes the board bigger and makes things easy to wire. I still don't know if a new ROM is going to fix the QL but the board will give me the ability to program EPROMs with test code if normal code doesn't work. If there is any errors in these thoughts please let me know before I cock it up.

[tofro]

Well, you could do worse than looking at the QLSD hardware (http://www.dilwyn.me.uk/qlsd/) schematics.

This uses a 64k EPROM connected to one single ROM socket - It obviously also uses a CPLD for address decoding (which is way too large for your purposes and way too small for soldering at home), but that could be simplified to a GAL, for example. And it doesn't need any mods on the mainboard.

Tobias