Hi,
Will this sram work in a nes cart?
http://elcodis.com/parts/5811505/EM51M2 ... #datasheetI'm asking because in the data sheet says:
Output level: 3.3V
The NES is an NMOS device, and have voltage thresholds very similar to even-older TTL logic. 3.3V for true should be fine.
(Putting a 3.3V SRAM with 5V tolerant I/O and a 3.3V regulator inside a package seems odd to me, but maybe that's just how the economies of scale have worked out)
A better alternative would be to replace the power supply in the NES to output 3.3V and run the system at this voltage, I've heard of someone doing it and that it fully works. Running it at 5V is a waste of energy.
However, if your goal is to build a cart which is compatible with all unmodified NES then it's a no-no.
Regulators are cheap and the SRAM chip will dissipate so little power that the regulator also will dissipate almost nothing so I think there is no problem.
You don't even need a regulator, with a Zener diode and an AOP you can make one (since +5V is already regulated).
Bregalad wrote:
A better alternative would be to replace the power supply in the NES to output 3.3V and run the system at this voltage, I've heard of someone doing it and that it fully works. Running it at 5V is a waste of energy.
That was me. Yes it works, but all the colors are darker. It locked up around 3.1v IIRC, so you're pretty close to the voltage corner at 3.3v. I wouldn't rely on it aside from experimentation or temporarily running with non 5v tolerant logic as I was doing.
If it fails at 3.1V, then 3.5V is probably large enough to be considered safe, and is probably still in tolerance of 3.3V chips.
By changing the current flow in the amplification circuit past the PPU you could probably restore the video signal so that it's not darker than usual.
However, even if the CPU and PPU are safe to run slighly above 3V, at which voltage the system will fail depends mostly on the SRAM chips which depends on the cart and of the SRAM chips inside the console, which may vary very greatly so that's a problem.
Is there an easy way to tell if a LVTTL chip will be 5V tolerant? And what does 5V tolerant mean exactly? I've seen a lot of discussions about parts with maximum tolerances on their I/O pins specified at 4.6 volts being "5V tolerant" because of how the pins are connected to ground inside the chip, causing feedback on the I/O channel rather than exceeding the forward voltage on a die transistor. Is this really "5V tolerant", or is this "it won't melt in a 5V system"? Seems to me like it's the latter.
Don't undervolt the NES. Commercial ROMs aren't fast enough when you drop the voltage. There's also no reason to; the amount of power when plugged in is small, and even this 1.5V drop of Vcc isn't enough to bring down power consumption to "reasonable to run on batteries"
The datasheet will tell you whether any give chip is 5V tolerant; they'll explicitly say "5V tolerant I/O" somewhere. Sometimes they limit it to specific pins.
Most ICs have ESD protection diodes from ground to all their outputs and from all their outputs to Vcc. For a 3.3V chip, this means that there's a practical maximum voltage you can apply of about 4.0V before you start powering the chip through its I/O lines, which is bad.
Otherwise, as far as I know, the only way you have real maximum logic voltages are things like oxide thickness on the gates of the MOSFETs. (Or maybe too much power dissipated, when working with BJTs)
Quote:
Is there an easy way to tell if a LVTTL chip will be 5V tolerant?
You should take a look at the datasheet of the component.
Quote:
And what does 5V tolerant mean exactly?
It means exactly that : The device operates at arround 3V, but is tolerant to having 5V to one of the input pins, and will consider it a valid logical '1' instead of frying the chip.
It does NOT mean in any case that the chip is tolerant to be supplied in 5V. This could result in destruction of the chip.
Right, so parts that say "5V tolerant I/O" are designed for 5V logic, those that just don't melt when given 5V I/O signals are not. Got it.
Some parts also aren't described as 5V tolerant, but can be if there are clamping diodes internal to the chip. Add a series resistor to limit current for each input, and you have a voltage divider. AFAIK, datasheets don't describe this usually. But for various parts there are application notes, I know at least this is done with certain Xilinx FPGAs, and Parallax Propeller.
BTW the original question, if the answer wasn't clear, yes, that SRAM should work on NES.
qbradq wrote:
And what does 5V tolerant mean exactly?
5V-tolerant means that 5V logic signals are within the device's maximum ratings, but that it generally uses 3V logic. So if its specs say it can tolerate -0.5V to +5.5V, then it's 5V-tolerant. A maximum rating of 4.6V is not 5V logic-tolerant. A device seeming to work after you apply voltage outside its maximum ratings isn't any indication of 5V tolerance. For example, BJT transistors may seem to work after exceeding the base's reverse voltage (-5V or so), but their hFE will be permanently degraded.
lidnariq wrote:
Don't undervolt the NES. Commercial ROMs aren't fast enough when you drop the voltage. There's also no reason too; the amount of power when plugged in is small, and even this 1.5V drop of Vcc isn't enough to bring down power consumption to "reasonable to run on batteries"
Besides, the 7805 regulator probably dissipates as much as the rest of the system.
Quote:
Besides, the 7805 regulator probably dissipates as much as the rest of the system.
The math is simple, the input voltage is redressed 9VAC so sqrt(2) * 9 ~= 12.7 V (I count redressing diodes as part of the regulator),
so the regulator dissipates from 12.7V to 5V and the rest of the system from 5V to 0V.
So the regulator will always dissipate 1.54 times what the rest of the system dissipates, no matter how much that is.
Quote:
Some parts also aren't described as 5V tolerant, but can be if there are clamping diodes internal to the chip. Add a series resistor to limit current for each input, and you have a voltage divider
The problem I see is that each time a input is to logic '1', there will be a significant current flowing through your inputs and the diodes, and this could potentially be a problem. Be sure to use very large resistors if you want to go this way.
I used another type of sram.