Forums

Sega Master System / Mark III / Game Gear
SG-1000 / SC-3000 / SF-7000 / OMV
Home - Forums - Games - Scans - Maps - Cheats - Credits
Music - Videos - Development - Hacks - Translations - Homebrew

View topic - Minimal slot 2 cartridge design

Reply to topic Goto page 1, 2  Next
Author Message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Minimal slot 2 cartridge design
Post Posted: Sun Mar 13, 2022 8:05 pm
Due to the current situation with the semiconductor shortage, I want to design a SMS cartridge using only discrete logic. The idea is that the cart is used for homebrew, so after talking to some devs, these are the specs they need:

* 32 KiB fixed from the beginning of the ROM
* Switchable 16 KiB on slot 2
* Slot 2 bank register mapped on the full $E000 ~ $FFFF range (should not be a problem as long as the homebrew ROM does not use the RAM mirror).
* Up to 512 KiB of ROM

With these in mind, I started reading documentation and I have not all the details ironed out (read here for more details), but I have created an initial draft of the cart schematics. You can download it from the attachment.

The design uses only 3 logic chips (74XX) plus the flash chip, but I am not 100% sure if it will work. These are my main concerns (please look at the schematic to ease understanding the points):

1. Is it necessary to use the CE# signal for the ROMCE#? If affirmative, will it work just switching MREQ# for CE# in the address decoding logic?
2. I have assumed that the EXMx signals (that I have named M0-7#, M8-B# and MC-F#) are generated using only address lines, and do not involve MREQ# or any other signal (RD#, WR# or whatever). Is this true?

In addition to these, I have to think about what to do with the flash WE# pin, but that's not currently a concern.

So, any suggestion or correction is welcome!
sms_sl2map_c.pdf (64.54 KB)
Schematics draft

  View user's profile Send private message
  • Joined: 05 Sep 2013
  • Posts: 3763
  • Location: Stockholm, Sweden
Reply with quote
Post Posted: Mon Mar 14, 2022 10:18 am
since you're using an 8 bit latch anyway you could theoretically support up to 4 MiB too... maybe a different revision?
  View user's profile Send private message Visit poster's website
  • Joined: 24 Mar 2021
  • Posts: 120
Reply with quote
Post Posted: Mon Mar 14, 2022 5:06 pm
I see you've invented something that's almost the Codemasters mapper. (They just used a PAL instead)

doragasu wrote
1. Is it necessary to use the CE# signal for the ROMCE#? If affirmative, will it work just switching MREQ# for CE# in the address decoding logic?
On machines with the menu, -CE is the only one that lets you know which slot is enabled - you have to include it in your decoding logic. (That's why the standard 28-pin mapper-and-ROM chip pinout connects to only /CE and /WR)

Quote
2. I have assumed that the EXMx signals (that I have named M0-7#, M8-B# and MC-F#) are generated using only address lines, and do not involve MREQ# or any other signal (RD#, WR# or whatever). Is this true?
No, they do already factor in /MREQ.

Quote
In addition to these, I have to think about what to do with the flash WE# pin, but that's not currently a concern.
There's no obvious reason to prohibit self-programming...
  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Mon Mar 14, 2022 7:33 pm
Thanks for the replies!

@sverx: The problem is finding a 5V flash chip bigger than 512 KiB. I know no one that is currently in production.

@lidnariq: I suspected I had to use CE#. But since it is MREQ# basically with the bit of the enable register, I suppose it is just a matter of swapping both signals in the address decoding of my design.

And about the EXMx signals, if the have also MREQ#, I have to review the decoding in my design, I suspect the register write enable might not be correct.

Again, thanks for the info!
  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Mon Mar 14, 2022 8:56 pm
I have updated the schematic, and after taking into account that EXMx# signals are ORed with MREQ#, and replacing MREQ# with CE# in the address decoding, I have reviewed the equations and I still think it should work.

I have attached the schematic with these changes.
sms_sl2map_c.pdf (64.67 KB)
Cart schematic

  View user's profile Send private message
  • Joined: 28 Sep 1999
  • Posts: 1197
Reply with quote
Post Posted: Mon Mar 14, 2022 9:52 pm
You could save some logic by replacing 74HC573 with 74HC574 and use the output of U1B directly without having to invert it. Here the data bus would get latched by the rising edge of the write strobe the '139 generates. Freeing up U2A might allow some other design choices too.

Does the logic need to be 74HCT instead of 74HC for compatibility? The Z80 is NMOS and will have similar drive strength as 74LS-series parts parts as I recall. HCT is compatible with that whereas HC expects CMOS only. I've only used HCT parts with console projects so I can't say how well HC works, maybe it's fine.

If you're going to allow flashing the chip in-circuit you may want additional decoupling near U3 as the current draw during programming cycles can be quite high and cause a dip in the supply voltage. The flash chip data sheet will have the worst cases listed for programming current. If it's just programmed in an external programmer then no problem.
  View user's profile Send private message Visit poster's website
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Tue Mar 15, 2022 6:12 pm
Those are nice suggestions, thanks!

Didn't think about using a 574, but you're right, if I need another nand gate that will come handy!

I have used in the past HC devices with TTL chips and had 0 problems, but you are absolutely right I should use HCT chips to be on the safe side.

About the flash decoupling, I wanted to use just one of the 100 nF capacitors in the schematic, I will review the datasheet to se if it is enogh.

Thanks!!!
  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Sun Mar 26, 2023 10:22 am
Took forever to get the time slot for these to be built ant tested. But they work great! I have tested homebrew games like Gotris and Moggy Master, and also commercial games like Castle of Illusion and Master of Darkness, the cart works flawlessly.

I named the thing "GhettoMapper", as a reference to "GhettoRepair" (repair things with whatever you have lying around). I had to lower resistors from 47k to 4k7 for the cart to work, but other than that the initial design worked perfect!

Thanks a lot for the help, could not make it work on the first try without all the advice I got here!

The cart design is Open Hardware, so I will make a little revision (correcting resistor values and removing R1) and publish here the link to the repository in case anyone is interested.

  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Sun Mar 26, 2023 11:36 am
Awesome stuff, well done! These look fantastic and really appreciate you open sourcing these - I'm sure will come in very handy for the community!
  View user's profile Send private message Visit poster's website
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Sun Mar 26, 2023 3:54 pm
Revision A (that should be the final one) ready. This is the project repository:

https://gitlab.com/doragasu/sms-sl2map
  View user's profile Send private message
  • Joined: 24 Mar 2023
  • Posts: 15
Reply with quote
Post Posted: Mon Mar 27, 2023 7:14 pm
Thank's for your hard work! I'll echo Will in thanking you for opensourcing these as well!
  View user's profile Send private message
  • Joined: 14 Oct 2008
  • Posts: 508
Reply with quote
Post Posted: Tue Mar 28, 2023 2:49 pm
doragasu wrote

I named the thing "GhettoMapper", as a reference to "GhettoRepair" (repair things with whatever you have lying around). I had to lower resistors from 47k to 4k7 for the cart to work, but other than that the initial design worked perfect!

I don't think I'd recommend using "Ghetto" in the name.
That word is a reference to neighborhoods of people living in poverty. Or maybe ethnic-based. Regardless, it's a derogatory word.
  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Tue Mar 28, 2023 7:59 pm
I wanted to give that "poverty" meaning to the board, but I don't want to offend anyone, if the term is pejorative, I'll change the name, thanks for the suggestion. And if anyone is offended, I'm really sorry. I'm Spanish and there are many terms I might apply wrong.

Time to search a new name then.
  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Tue Mar 28, 2023 9:19 pm
A few ideas:

- FeatherMapper - because it's lighter
- StreetMapper - maybe has similar connotations to what you wanted, but less derogatory, also a pun on streetmaps.
- GafferMapper - as in the phrase "held together by Gaffer Tape
- HackerMapper - pretty obvious. It kind of rhymes
  View user's profile Send private message Visit poster's website
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Wed Mar 29, 2023 6:32 am
Thanks for the suggestions. What about "No-frills Mapper" or "Frugal Mapper", are they OK?
  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Wed Mar 29, 2023 7:11 am
Yes I should think so!

edit: of the two I think I like FrugalMapper better
  View user's profile Send private message Visit poster's website
  • Joined: 04 Jul 2010
  • Posts: 539
  • Location: Angers, France
Reply with quote
Post Posted: Wed Mar 29, 2023 10:40 am
no$Mapper, lol

PS. just a little hint, you should add a note about the mapper.
It will not work on MD and Mark III as MC-F does not exist on these systems.

You can also remove D5-D7 from the schematic, as they're not used (limited to 512k, but you maybe can make it compatible/another version with old 8/16bits ceramic chip like 27C160 (2048k))
  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Wed Mar 29, 2023 11:36 am
ichigobankai wrote
You can also remove D5-D7 from the schematic, as they're not used (limited to 512k, but you maybe can make it compatible/another version with old 8/16bits ceramic chip like 27C160 (2048k))

Those inputs will still need something driving them and would need to be tied to a rail if not driven from the data bus, feels like no gain in removing them from the bus, plus there are some test points connected to the outputs.

@doragasu just wondering if there's space on the PCB for two more SST devices as then you could make it expandable to take more banks?

I've spent a little time skimming the schematic, and possibly am being a little slow today, but could you explain what the deal with A14 on the SST is and what the logic around A14, A15 and Q0 is doing?
  View user's profile Send private message Visit poster's website
  • Joined: 23 Jan 2010
  • Posts: 417
Reply with quote
Post Posted: Wed Mar 29, 2023 11:39 am
DRM - doragasu repair mapper.
Doramapper
Gasumapper
  View user's profile Send private message
  • Joined: 04 Jul 2010
  • Posts: 539
  • Location: Angers, France
Reply with quote
Post Posted: Wed Mar 29, 2023 11:50 am
willbritton wrote
would need to be tied to a rail

i don't think i'll have to detail this, but yes should be tied to gnd or vcc. nothing complicated.

willbritton wrote
what the logic around A14, A15 and Q0 is doing?

its forcing A14 to be enable when reading $4000-$7FFF and using what the latch Q0 have when reading $8000-$BFFF
  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Wed Mar 29, 2023 12:12 pm
Ah yeah, okay...I was being slow haha!
  View user's profile Send private message Visit poster's website
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Wed Mar 29, 2023 12:31 pm
It's just something just has my spidey sense tingling and I can't quite put my finger on it.

I think that you might be able to replace all 3 of those nand gates with a single diode and one of your resistors to form a wired logic and gate between A14 and Q0; and with Charles's suggestion of using a 574 instead of a 573 you'd drop the last nand.

Not sure whether swapping a cheap 14 pin dip for a diode is much of a gain but in the spirit of frugality haha!
  View user's profile Send private message Visit poster's website
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Wed Mar 29, 2023 6:57 pm
Changed the name to FrugalMapper. It's already in the repository. Thanks a lot for the suggestions, I really liked some of them, but will stick to this name.

ichigobankai wrote
PS. just a little hint, you should add a note about the mapper.
It will not work on MD and Mark III as MC-F does not exist on these systems.


I want to add a readme file explaining the mapper capabilities. I'll also add this. Thanks a lot for bringing it up as I didn't know it (I am new to SMS homebrew). This brings also up a question: is also #M8-B missing on these systems? If it's only #MC-F, I might be able to make the cart compatible by implementing the suggestion from Charles MacDonald and using the extra NAND for A15 and A14. But if also #M8-B is missing, it will need more logic.

ichigobankai wrote
You can also remove D5-D7 from the schematic, as they're not used (limited to 512k, but you maybe can make it compatible/another version with old 8/16bits ceramic chip like 27C160 (2048k))


My original design, attached to the 5th post (and the one in the PCBs from the photograph above) did just that, but when reviewing the PCB to remove the resistor on #WE, I noticed routing these tracks takes no effort (as they already went through that zone in the same order), so I changed it and added the test points, just in case I want to experiment adding something later (as this has 0 impact on board size, cost or performance, just a tiny -almost negligible- increase in power consumption).

willbritton wrote
@doragasu just wondering if there's space on the PCB for two more SST devices as then you could make it expandable to take more banks?


It can be done either making the PCB bigger (thus a little bit more expensive) or assembling components also on the bottom (making assembly more expensive). It would also require a bit more logic (a 2 to 4 decoder basically). Is there homebrew bigger than 512 KiB? In the past I designed some monster carts (like my MOJO-NES MKIII, that allows 16 MiB or 128 Megabits) and I discovered the simpler and dumber my carts are, the greater the probability of projects finding them useful are. So I do not feel inclined to add more chips. Could consider using a bigger flash chip if I can find one compatible with 5V. By the way, with the name of the cart I also intended to separate it from my other cart designs that are more "current tech", like MegaWiFi, the aforementioned MOJO-NES MKIII or ROMBA (a cart for Amstrad CPC 8-bit computers that includes a -never seen before for that computer- tiny DMA engine).

willbritton wrote
I think that you might be able to replace all 3 of those nand gates with a single diode and one of your resistors to form a wired logic and gate between A14 and Q0; and with Charles's suggestion of using a 574 instead of a 573 you'd drop the last nand.


Maybe I am not fully understanding your idea, but I'm not sure that would work. First because you need to take A15 also into account, and second because that kind of wired AND gates require open collector/drain outputs (unless you add more diodes).
20230329_201822.png (652.24 KB)
20230329_201822.png

  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Wed Mar 29, 2023 7:38 pm
doragasu wrote
willbritton wrote
I think that you might be able to replace all 3 of those nand gates with a single diode and one of your resistors to form a wired logic and gate between A14 and Q0; and with Charles's suggestion of using a 574 instead of a 573 you'd drop the last nand.


Maybe I am not fully understanding your idea, but I'm not sure that would work. First because you need to take A15 also into account, and second because that kind of wired AND gates require open collector/drain outputs (unless you add more diodes).

Your "open collector" and your A15 I think are both provided by Q0 coming from the register: Q0 will be high-Z if A15 is low, in which case you want A14 to feed through directly, and if A15 is high then Q0 is actively driven in which case it takes precedent over A14 so by pulling Q0 UP to +5v through a resistor and adding a diode to A14 I think you implement this logic. I'll make a sketch and send later.
  View user's profile Send private message Visit poster's website
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Wed Mar 29, 2023 7:54 pm
Oh, that's interesting!

BTW I have added the README file to the repository, detailing the cart capabilities, and with some more info (like how to burn ROMs to the Flash chip, saving games, etc.).
  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Wed Mar 29, 2023 8:47 pm
Last edited by willbritton on Wed Mar 29, 2023 9:04 pm; edited 1 time in total
Sorry, I posted a sketch briefly and then realised I'd made a mistake so deleted it.

Here, second time's a charm...

It's actually a wired or gate rather than an and, which is even simpler because you can keep the pull-down resistor as it currently is.

You get a diode drop when A14 is high, but there's a huge margin above TTL's VIH to accommodate that so not a problem at all if you're using HCT.

Please let me know if I made a (nother!) mistake...
notepad-or.jpg (45.78 KB)
notepad-or.jpg

  View user's profile Send private message Visit poster's website
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Wed Mar 29, 2023 9:00 pm
Also I just thought of another alternative if you're a bit sniffy about diodes - see attached!

The 74HCT1G126 is a 5 pin SMT device and dirt cheap. Luckily it has an active high gate input vs. the active low on your 573/574, so the two devices are never driving at the same time.

You also need 1 fewer resistor with this one.

(ignore the diode drop footnote I forgot to erase from the previous sketch - that no longer applies either)
notepad-buffer.jpg (44.92 KB)
notepad-buffer.jpg

  View user's profile Send private message Visit poster's website
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Thu Mar 30, 2023 7:09 am
Oh, that's clever and I think it should work! Definitely replacing the 74HCT00 for a diode should be a win, I will have to give it a try, thanks!
  View user's profile Send private message
  • Joined: 14 Aug 2000
  • Posts: 740
  • Location: Adelaide, Australia
Reply with quote
Minimal slot 2 cartridge design
Post Posted: Thu Mar 30, 2023 11:05 am
Is that electrolytic capacitor too tall for the board to fit inside a cartridge case?

Looks like I just missed the naming comp'. I was going to suggest the following:

Doragasaurus - the dinosaur mapper.

SlapperMapper, or the WillSmithSlapperMapper - it’s a sign of our times, with the chip shortage and all.

The Summer of ’74 Mapper - self-explanatory.
  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Thu Mar 30, 2023 11:35 am
asynchronous wrote
Is that electrolytic capacitor too tall for the board to fit inside a cartridge case?

Yeah I was surprised to see an electrolytic there too and had assumed it was a placeholder in the 3D model, but now I see from the schematic that it's specced as 100uF - seems pretty huge for this board, 10uF would surely be more than enough for a bulk cap here. But in any case you can get MLCC caps that big so no need for electrolytic I'd say.

No comment on the name suggestions async, other than it's a small insight into the way your mind works.
  View user's profile Send private message Visit poster's website
  • Joined: 14 Aug 2000
  • Posts: 740
  • Location: Adelaide, Australia
Reply with quote
Minimal slot 2 cartridge design
Post Posted: Thu Mar 30, 2023 1:59 pm
They're all meant to be light-hearted and tongue-in-cheek.
  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Thu Mar 30, 2023 5:13 pm
Fun suggestions for the name, but competition is closed :-)

The capacitor fits without a problem, I measured with a caliper the spacing before designing the case to make sure it fits, and I have enclosed some boards inside a case without a problem. That said, I think the board will work just fine without it, just with the decoupling for each chip. But the footprint is there just in case. I'd like to do some tests with code to write to flash from the console and with the capacitor removed, to see if everything is stable. If it works, maybe I'll mark the cap as DNP.

About MLCC, you have to be careful with values greater than 1uF. They usually are made using poor dielectrics (X5R or even worse like Y5V) and even worse than that, they tend to have huge variations with voltage, making the actual capacitance much lower than you expect, specially for the lower footprint ones. If you have a look to attached graph, for example a 5uF 0805 X5R capacitor loses 50% of its capacitance when powered at 5V. You take into account other effects (like variation with temperature) and things can get even worse. For 10uF and higher, I always try to use electrolytic capacitors (and of course avoid tantalum ones).
sLFzU.gif (19.05 KB)
sLFzU.gif

  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Thu Mar 30, 2023 5:19 pm
Some photographs, one of the cart programmer (with slots to program Megadrive and SMS carts), one of an enclosed cart (including the electrolytic capacitor) mounted in the programmer, and the programmer in the 3D printed enclosure I designed.
prog.jpeg (536.08 KB)
prog.jpeg
prog_with_cart.jpeg (369.11 KB)
prog_with_cart.jpeg
prog-cad.png (185.51 KB)
prog-cad.png
prog-cad-photo.jpeg (466.57 KB)
prog-cad-photo.jpeg

  View user's profile Send private message
  • Joined: 24 Mar 2023
  • Posts: 15
Reply with quote
Post Posted: Thu Mar 30, 2023 6:49 pm
But what about an X7R @ 10uF though? Presumably, the schematic can be used for a game gear cart just as well, but with that cap it's a non-starter.
  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Thu Mar 30, 2023 7:18 pm
Sure, you can use them. I didn't say you can't (I use them from time to time), just that you have to be careful. Expect a 10 uF 0805 X7R cap to drop its value to about 7 uF (plus tolerance, plus temperature along the usage range) when 5V powered and compensate with a bigger one if you need.

Anyway, as I wrote, I strongly suspect the cart will run just fine even without that capacitor (just with the decouplings for each chip). Will need to test to confirm.
  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Thu Mar 30, 2023 7:47 pm
Yeah for bulk tolerance is fairly moot as you say you might not even need one. My default position is to leave redundant board layout for things you might not need just to give the option e.g. you could have pads for the electrolytic and for a smaller ceramic in parallel and give people the option.

P.S. I'd never skip a bulk capacitor entirely - it's just "belt and braces" ... okay you might get away without one but for the sake of a single device you might as well play it safe.
  View user's profile Send private message Visit poster's website
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Sat Apr 01, 2023 3:55 pm
Today I grabbed the soldering iron to modify a board. Tested the "diode trick" and also removed C1. I did not completely removed the NAND chip because I need one gate until I swap the 573 for a 574, but bypassed the output of the 3 NAND block, added a 1N4148 diode I had lying around, modified the needed connections, and... success!

I have completed a full "Castle of Illusion" run without a problem. So I'll do a revision removing the NAND chip, adding the diode and I'll also change the C1 footprint for a 0805 cap and mark it as DNP (I think it's not necessary but whoever makes the boards can mount a 0805 cap in case of doubt). For the diode I'll use a schottky one (maybe a S14) that should be faster and have a bit less forward voltage. I have attached the modified schematic in case anyone wants to have a look.
sms_sl2map.pdf (357.32 KB)

  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Sat Apr 01, 2023 4:27 pm
Great stuff!!

Good call on C1, 0805 is easy enough for anyone to install themselves if they need to.

Doubt you'll really need the speed of schottky just because the address lines in the Z80 are asserted first in the read cycle, well ahead of ~MREQ and ~RD so you have "ages" to stabilise the address bus, but hey - can't hurt to have it! The important thing is the layout.

One more thing that had been running through my mind - I personally found it much more difficult (around 6-12 months ago anyway) to source 573/574 devices than 373/374 devices which are effectively the same but with different pinouts so you might just want to check stock levels before you commit to that device. Possibly it was just part of the chip shortage, but I wasn't sure whether there were any plans to discontinue the 573/574 lines with some manufacturers in favour of the 373/374s. Worth satisfying yourself anyway, if you haven't already.
  View user's profile Send private message Visit poster's website
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Sat Apr 01, 2023 6:43 pm
I have checked the 574 vs 374 (20-TSSOP version) on both Digikey and LCSC. On Digikey there are about 2500 374s and a ton of 574s (more than 100.000). On LCSC there are 980 374s and 1834 574s. So I think I'll stick to the 574. But thanks for the suggestion, better checking in advance than finding the problem later.
  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Sat Apr 01, 2023 6:56 pm
Routing done, and release created on the repository.

  View user's profile Send private message
  • Joined: 24 Mar 2023
  • Posts: 15
Reply with quote
Post Posted: Mon Apr 03, 2023 6:03 pm
Regarding #MC-F being missing on MD and Mark III:

https://www.smspower.org/maxim/Documents/Pinouts

Looks like M8-B still exists on them.
  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Mon Apr 03, 2023 6:24 pm
On that, could you connect ~CE to pin 15 of your 139 and A14 and and A15 to pins 13 and 14?

It would mean that the mapper page selection only worked if the cart slot was active but I think that's okay isn't it? Then you wouldn't need ~MC-F.

EDIT: okay you need ~MEMRQ somewhere don't you. Hmm...
  View user's profile Send private message Visit poster's website
  • Joined: 24 Mar 2023
  • Posts: 15
Reply with quote
Post Posted: Mon Apr 03, 2023 8:20 pm
I also found this: https://members.optusnet.com.au/eviltim/smsmap/

From an old SMS thread: https://www.smspower.org/forums/9821-CPLDPagingICMapper

That details a CPLD mapper, with schematics captures. I've tried busting open the CPLD project files, but my home internet is awful and chokes on large files, like the Quartus Lite installer. Still, the schematic captures might be useful here?
  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Tue Apr 04, 2023 6:18 am
willbritton wrote
On that, could you connect ~CE to pin 15 of your 139 and A14 and and A15 to pins 13 and 14?

It would mean that the mapper page selection only worked if the cart slot was active but I think that's okay isn't it? Then you wouldn't need ~MC-F.

EDIT: okay you need ~MEMRQ somewhere don't you. Hmm...


#CE already involves #MREQ, so I might get away with that. Will have to draw the solder iron again when I have some time.
  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Tue Apr 04, 2023 7:34 am
MossDragoon wrote
That details a CPLD mapper, with schematics captures. I've tried busting open the CPLD project files, but my home internet is awful and chokes on large files, like the Quartus Lite installer. Still, the schematic captures might be useful here?


For me they are not that useful, what I find most useful is good information about the signals in the cart port.

Creating mappers using CPLDs from the mapper description is relatively easy, I made several in VHDL for my MOJO-NES cart (and the flasher automagically loaded the proper bitfile to the CPLD when you burned a ROM by parsing the iNES header). Did also a mapper with a PCM module and a tiny DMA engine for my ROMBA cartridge (for Amstrad CPC 8-bit computers). The problem is that as of today, CPLDs are difficult to find at decent prices. I used LAMXO256C/LCMXO256C for those designs and several years ago you could buy one at Digikey for less than 2€ a piece. Today their price is 11,91€...
  View user's profile Send private message
  • Joined: 24 Mar 2023
  • Posts: 15
Reply with quote
Post Posted: Tue Apr 04, 2023 8:05 am
Not even with the schematic captures that show the discrete logic involved? Sorry!
  View user's profile Send private message
  • Joined: 06 Mar 2022
  • Posts: 598
  • Location: London, UK
Reply with quote
Post Posted: Tue Apr 04, 2023 9:33 am
doragasu wrote
#CE already involves #MREQ, so I might get away with that. Will have to draw the solder iron again when I have some time.

Good point, I always forget ~CE has ~MEMRQ in it.
All the same, I looked at this a little more (was in the pub last time I wrote so not thinking super clearly!) and I can't find a way to get it to work without ~MC-F using only one IC. I think you could pull it off with a couple more diodes but I don't think either a single 139, 138 or quad OR can do it: the 139 because you need a way to OR some outputs (that's where the diodes could help), the 138 because you don't have enough inputs and the quad OR because you'd need to invert A13 somehow.

Shame this, it feel so close but sadly not I think.
  View user's profile Send private message Visit poster's website
  • Joined: 04 Jul 2010
  • Posts: 539
  • Location: Angers, France
Reply with quote
Post Posted: Tue Apr 04, 2023 10:14 am
Signals i'm using in my SLOT 2 (only) implantation into a 22v10.

  View user's profile Send private message
  • Joined: 03 Mar 2022
  • Posts: 55
Reply with quote
Post Posted: Tue Apr 04, 2023 5:54 pm
willbritton wrote
doragasu wrote
#CE already involves #MREQ, so I might get away with that. Will have to draw the solder iron again when I have some time.

Good point, I always forget ~CE has ~MEMRQ in it.
All the same, I looked at this a little more (was in the pub last time I wrote so not thinking super clearly!) and I can't find a way to get it to work without ~MC-F using only one IC. I think you could pull it off with a couple more diodes but I don't think either a single 139, 138 or quad OR can do it: the 139 because you need a way to OR some outputs (that's where the diodes could help), the 138 because you don't have enough inputs and the quad OR because you'd need to invert A13 somehow.

Shame this, it feel so close but sadly not I think.


It's a pity, as you say it feels really close. I was also trying to play with the 574, if you feed any of the unused data pins with a '1', and put a pulldown at the output, you can get the negated version of #M8-B (we could call it M8-B without the '#'), but unfortunately that is not useful either :(.

Anyway, I suppose it's not the end of the world, for SMS this design is OK, and for the others, a design with one more chip is not that bad.
  View user's profile Send private message
  • Joined: 25 Jul 2007
  • Posts: 716
  • Location: Melbourne, Australia
Reply with quote
Post Posted: Wed Apr 05, 2023 10:51 am
I was looking at getting some of these made today, seems the SST39SF040A is in short supply, is there an alternate flash that could be used?
  View user's profile Send private message
Reply to topic Goto page 1, 2  Next



Back to the top of this page

Back to SMS Power!