Printable View
According to this article https://db-electronics.ca/2017/07/05...etro-consoles/ 3 volts memory are nocive to retro consoles.
Now accoring to this video:
https://www.youtube.com/watch?v=IjC0qQREV1w
and this picture:
Attachment 14039
Tec Toy is using single voltage 3v memory (29LV1600) and isn't using level shifters.
Could this be potentially harmful to original retro consoles or all this talk is just bullshit?
Considering just how bullshit the hardware of their most recent clone is? I'm not surprised that they also screwed up here.
Actually, does their clone even use 5V?
AFAIK this new megadrive is based on a RED KID 2500.
I've found some more details here:
According to a guy who was selling a Chrono Trigger repro without level shifters, it's all bullshit.HTML Code:https://assemblergames.com/threads/two-fans-have-reportedly-fixed-the-atgames-genesis-audio-issues.66488/
More or less translated from portuguese:
"if you put this, this protection, nothing's going to be altered.
There are people using repro games with this flash for about 10 years and never got trouble.
This is bullshit. Maybe this is applicable in complex circuits, not where there's only 1 flash and nothing more.
You'll only spend money doing another board - really, you don't even need the regulator, you can put it straight.
I have stories of guys using it for 2 years without regulator the same board, or even with only 2 diodes.
Some electronic theories are applicable in complex circuits only, not on simple circuits wich uses the flash's read pattern and nothing more.
The burn option on a thing like a router doesn't apply in this case. The burn tolerancy would be 3.3 volts."
Is someone developing replacement parts for old consoles with FPGAs?
Because I strongly think we'll soon need it here in Brazil!!
Regarding voltage level shifting, I'm not very acknowledged to discuss it, but the last argument from Krikzz may have been not responded yet by René, since he brought the whole discussion. I don't know if René didn't see Krikzz answer yet tough.
http://krikzz.com/forum/index.php?to...55046#msg55046Quote:
Original topic about this question were accidentally removed, but sense of this topic it two words sound like that: Rene's sensational "research" were based on wrong numbers, by some reason he decide that CMOS device current limits is 5.2mA per pin, but JEDEC specification clearly says that CMOS device current rating is 20mA. Old everdrive carts without level translators has around 7mA load per pin, and this is notably less than max acceptable 20mA, as we can see.
I say it to Rene in comments to his article, but he just ignored it. Anyone can read my comment in his article here: https://db-electronics.ca/2017/07/13...3-3v-mismatch/
Edit: fixed link in quote.
You can use 3.3v ics on pcbs by using a simple 1 cent diode to drop 5v signal to 4.3v which will be under the safe range for the flash IC.
All the rumors of level shifters etc are propaganda from what I would suspect failed businesses in the retro gaming scene.
No more true than growing hair in the palm of your hands from fapping.
Level shifters or bust!
I have huge doubts about this subject, here is why:
I've bought a 3.3v Flash repro of SunsetRiders for my SNES.
It was played by me and the kids for 8~12h a day for 6 months.
I them decided to play Super Turrican, and got heavy backgound flicking.
Them, played TazMania, wich crashed and was unplayable.
I noticed some moves on Super Mario World also were not as smooth as they were at first.
The first stage of Donkey Kong 2 were also showing some artifacts.
Them, after testing these games on a friend's SNES and figuring out they were working normally, I ran a test cart and got a DIV 16/8 test error.
When I asked about it in other forum I was told that this problem can be strongly related to the use of this kind of cartridge.
I really don't want this to happen with my other retro systems, but since original old games are having their prices skyrocketing, these repros seem like a good alternative.
I tried to look up about connecting 5V and 3.3V last night (turns out this is an extremely common issue with microcontrollers) and... um, it depends. There are so many factors to account for that it's hard to tell at a glance whether what a board does is good enough or not. Some chips are more tolerant to getting fed input above Vcc (this particular one in fact seems to be right on the limit? those resistors probably lower voltage enough to put it back into safe range - the resistors are on the opposite side of the board, just to make it clear). Also it seems level shifters are a must for high speeds, while resistors can be OK for slower frequencies (Mega Drive falls squarely on the latter group), and apparently a lot of people indeed prefer the latter where possible. Even the distance to the components can matter, due to noise.
So huh yeah, it depends (on the particular components used). What I can say for sure is that wiring 5V directly to 3.3V (which has happened) is definitely a good way to fry the chip on the cartridge =P (some of the console revisions use 5V CMOS so that can exacerbate the issue). But yeah, I'd be more worried about the cartridge than the console, since the latter is the one oversupplying.
I'm not particularly an expert though, so in case anybody with more knowledge wants to check further, this is the chip they used:
http://www.macronix.com/en-us/produc...%20NOR%20Flash
Also they sell gigabit parallel Flash o_o who wants to outdo the Neo Geo?
The gigabit NOR flash is the series used in the N64, I believe - so 128mb N64 carts could be made with those.
Quote:
Originally Posted by Fisher
Just clean up your pins, should be fine.
If there was a voltage problem with your snes, first the fuse would have popped, and it would not turn on unless you would have bridged it.
SNES CPUs are unusually delicate (especially the earlier "launch" consoles) so it's definitely possible that the bootleg game could've damaged something. Note that floppy-based copiers from the 90s shouldn't have this problem, as they natively run at 5v.Quote:
Originally Posted by pikointeractive
From that link:
I swear, broken mode-7 seems to be a surprisingly common way of failure.Quote:
Some games worked, others had glitchy graphics (SMW looked fine, F-Zero had track issues)
Bad mode7 is a relatively common failure, I even have one such machine where mode7 stopped working (and eventually it stopped working completely).
The 20mA per pin mentioned eralier doesn't really apply to anything but bus buffers, normal memory chips and whatnot are not capable of 20mA per pin from my experience, CMOS or not. Most chips are designed only to work at few tens of mA power dissipation at max, and now you're gonna turn that number waaaaay higher by loading all pins excessively getting to several hundred mA for the device in worst case. Current limits will help here, but that excess current still must be supplied by the voltage regulator(s) and stuff like MD2 with one regulator with a puny heatsink gets very hot indeed.
I can't find any information about the maximum current rating of the I/O pins, where this 20mA limit came from?
I agree about the level translators, they are the best way to go for sure. This will increase the price of the device, but this is not a problem, I'm almost sure everyone prefer to spend more and get the best device possible.
Datasheets of various chips will tell how much a pin can be loaded and how big is the maximum power dissipation of the entire device.
Translation chips are not gonna be more expensive than current limit resistors, it takes far more time to lay down all those resistors than the 3 conversion chips and time is money (though if there's machines doing the laying of parts the savings in time are not gonna be very significant).
I thought the problem here is how to route all the traces (wich are not so much), put a resistor in series on each line is far more simple in this case.Quote:
Originally Posted by TmEE
The most useful ones are pretty much 1:1 passthroughs, a row of pins matching what happens on other end, one could even see them as really wide/tall resistors as far as functionality is concerned. The few power pins in the middle aren't going to be any sort of headache.
You'd think so, but everybody wants to skimp out the very least cent from the board (which is also why I'm not convinced that "memory is cheap" when people go that far). It's also a stupid way to optimize it since most of the cost of a cartridge goes in assembling (the runs are small enough that it's cheaper to hire people than to use machines and you can't go too low there) and there's the box, manual, organizing to ship them...Quote:
Originally Posted by majinga
Also for what's worth it, when I decided to look about this topic separately I found out that the vast majority of people did indeed suggest resistors (although the exact circuits and values changed wildly each time, nobody seemed to agree on anything there) and to consider level shifters a last resort "if you don't feel safe" (sigh). I wonder how much does that skew views.