IC based region mod woes

Which leg of the 7805 are you pulling the voltage from for the IC mod? I'm pulling mine from the left-most leg. I haven't measured it, but it was stable enough to power my IC mod in my TG-16. That was a recommendation I got from PCEngineFX.com's forums.

Also, for TG-16s at least, you don't have to ground Pin 29. Grounding Pin 29 is only for PC Engines.

 

Off hand I don't recall which pin but the 7805 has just 3 - input, ground, output. Typically I don't rely 100% on what is stored in my head as it can be faulty even if I don't think it is and thus check with something to make sure I've got the correct pin.

Though if I had to guess? Facing the 7805 I believe I was pulling power from the right most pin which should be an even 5v if the 7805 was working to spec. Still possible I wired something up wrong but it doesn't really matter at this point. I've been toying with the idea Calpis suggested in order to build a better IC based region mod but haven't had the time to sit down and read datasheets.

 

Just use 4x 74HCT/ACT257, you could make a super tiny PCB with 2x SOIC chips / side.

 

Well, since you already have it available and/or installed, switch up the Vin from the right pin to the left-most pin first on the 7805. See if that might make a difference. The 4551s can handle up to 18V, so anything over 5V would probably make it a lot more stable, especially if you have some kind of video mod installed.

If that doesn't work for you, then seek out BlueBMW on the PCEngineFX forums. He has another IC based mod that works better w/ 5V. Probably would be a little more cost efficient than going with the 4 74HC4066 IC mod. That mod, though, wasn't recommended to me at PCEngineFX.

 

Why is it that modders come up with the worst circuits? No analog switches! This is the order of intelligent design:

1) 4x HCT257, SPST switch
2) 4x HCT244, SPDT switch
3) 2x HCT245, SPST switch, 1x HCT00 to decode

 

Would you be kind enough to sketch up a quick circuit diagram for that?

 

that would be nice because MC14551B chip is hard to find.

 

I found them cheaply at Digi-Key.

Not that hard. Heck if you want the two I used (still perfectly good best I can tell) you can have them at what I paid.

 

I ordered mine from futurlec.com...took a while to get, though.

Calpis: You have a working schematic down? I'm looking into an alternative to the 4551's.

 

I couuuuuuld post a schematic but it's really dead simple, possibly won't work as-is, and it's my opinion that if you can't design it you probably can't troubleshoot it; the '257 design adds 3 more loads to the data bus which might be a problem. Even still I'm sure it will perform better than analog mux. In general modding is not my thing so I'm not sure I want to try it myself, maybe if I could sell a few hundred boards.

 

I think people used analog switches because they did not understand how to make a bidirectional buffer such as the 74LS245 work on this circuit.

So IMHO the best solution is the two 74LS245 and 74LS00 with a switch because it is clean and is easier to understand ...

Just mentioning, not using bidirectional bus buffers/switches break Super CD ROM2 and Arcade system cards compatibility ...

 

The problem with '245 is their size, and that since you have to decode the R/W strobes there will be slight glitches and some contention going on, plus 4x more propagation w/ NANDs which is probably pushing things for older games with slower ROMs or the few with address decoding. It probably will work but to me it's still a bit inelegant considering how the '257 will have a negligible effect on the bus apart from some added capacitance.

 

Well if you buffer the address and control lines too with 74LS244(s) it should be not too bad, right ? That's how bus buffering was made on MSX computers and it was not bad, at all... :shrug:

 

Dead on.

I've looked at the datasheets for the above and the original ICs I used. While I have a decent understanding of the latter and the former it isn't enough to (obviously) troubleshoot this thing. I was assuming voltage was an issue but even when running from an external PSU with a voltage under load of 5.2v it still wasn't working. In all likelyhood I probably wired something up wrong or while soldering I fried one/both chips.

I did manage to get my TG16 back up and running...to an extent. Fixed a ground line I severed but failed to notice (derp) and now it can load graphical data uncorrupted from HuCards but doesn't get far enough along to actually get enough usable data to begin execution. Pretty sure the HuCard slot is shot or (more likely) one of the address lines is having trouble reliably getting signals across...due to a poor soldering job and lack of experience with this hardware (which is getting corrected fast, oh how being an autodidact means breaking things to learn how to fix them).

If you did assemble a schematic I'd give it a shot at putting it together. Not of your caliber engineer Calpis but I'm going for a BA in EE and I'd be the best bet to give it a shot with nothing to lose. Though I'd need to get this bloody thing booting games 100% of the time first. Maybe this time I'll socket everything rather than soldering everything to veroboard first. :banghead:

EDIT:
Just looked at the datasheet for the 74LS244 manufactured by Fairchild specifically, seems they have a pretty tight voltage requirement that I'm not sure my TG16 is capable of handling. I haven't measured the new power supply's output in circuit yet but the old one sure as hell wasn't up to the challenge. Tempted to replace the voltage regulator with something a bit...stronger.

 

'244 are perfect since the strobes don't need decoding but you still have to select the proper buffer for each region. You can eliminate any logic though by using a DPST switch (I said SPDT earlier). The '257 circuit works exactly the same as the '244 but since the "switch" is built in so you need only a SPST and also four 16-pin packages is far preferable to four 20-pin. It's not like the '257 is a rare chip or anything, it's probably the most common multiplexer (thus top 10) and more useful to stock than the '157.

Oh and for sure you don't want to use 74LS chips, four will probably draw 80mA(!) and since the PCE is CMOS (even FC is NMOS) it probably won't like the LS Voh. Better use 74HC/AC (with pull-ups) or 74HCT/ACT (which hopefully will float high w/o P-Us) families.

5.2V from the AC adapter???? That is far too low. You need at least 7V into the regulator for it to work right assuming a standard 7805 (I don't have a TG16).

If you see any coherent graphics on-screen some code is running which would suggest all data lines are OK as are all but one or two high address lines (Hucards don't interface with the VDP, just CPU).

Voltage requirement? They need 5V, exactly like every other component in the console.

Engineer/unemployed hobbyist, same difference lol

 

http://www.hardwarebook.info/HuCard

Pins 15-23 are the only ones I've cut, though I don't have the TG16 in front of me to confirm that none of the other pins have had their soldering touched. 18 was completely severed from the rest of the system care of my cutting into the PCB. Based on a profile view it seems clear that the board is just two sided with no layering to speak of. I did reconnect 18 to an independent ground point once I discovered it was disconnected. Guessing there is another issue to resolve, though for all I know the single HuCard I have is shot since I have no other TG16 to test it with.

I recall measuring 5.5v somewhere in circuit when the TG16 was turned on. I believe it was around 7.2-7.5v from the power brick itself and something like 4.5v was coming out of the 7805's business end (output pin) while the system was running. Low admittedly, recapping everything didn't change anything either.

 

7805 will output EXACTLY 5.0, they contain very accurate voltage references internally. You should not be seeing 4.5V. Either the input voltage is too low or the supply is insufficient (which causes the voltage drop). I'm fairly positive TG16 require <850mA so you must have another suitable AC adapter. If you have a 5V supply you could temporarily wire that in directly.

 

OEM adapter outputs 10.5VDC 730mA.

 

Interesting, that is what the second power brick I got has for its specs. The first one was much smaller, non-NEC branded (IIRC) and was speced at 9v 850mA.

Yours wouldn't have the brick in the middle of two sets of wires would it? Unlike the Genesis/SNES/NES bricks where "wall wart" is a more apt description.

 

Yeah, mine has the transfomer in the midde instead of the wall wart type. Your problem may lie there with the power issues since it might be an off-branded adapter...unless it's a Radio Shack one. I've been told that a 9v 850mA Shack adapter works cleaner than most off-branded ones.