The dangerous of 3.3v in flashcart console cartridges.

[dragon]

I found this article in the web. I am sure the hardware guys like It.

https://db-electronics.ca/2017/07/05/the-dangers-of-3-3v-flash-in-retro-consoles/


Big winner sd2snes . Bigger  loss Ness chinese cartridges.

[Bryce]

A nice description of the problem. I was aware of the Everdrive issues, but I don't own any. I modded my NeoGeo 161in1 cartridge to remove this problem before I ever plugged it in.
You'll be pleased to hear that there are no CPC 3.3V hardware (that I am aware of) that has this issue. The C4CPC for example connects to the CPC via a Xilinx XC95144XL which has 5V tolerant inputs and a 74LS245 Buffer which is also a 5V part.

So you can sleep well knowing that your Amstrad isn't about to die (at least not due to some 3.3V hardware being connected to it.

Bryce.

[freemac]

In fact a 5v TTL old components does speak at 3.3v (you can measure it). And 3.3v components does speak at 3.3v (by speaking, I want to say all wires except the Vcc wire)


If the component is very hot, you have a problem... If your resistor is hot (one used to pass from 5v to 3.3v), you will have a problem.


You normaly can test first your interface using another 5v source, not connecting it yet to CPC...


It's better to have a "common source", for sample a 3.3v build from 5v from CPC, instead of a "external 3.3v". In case of using an external Vcc source (a battery...), normaly GND are linked each others (common ground). External Vcc is a problem because you power on one ... and then... the other one so you -in this case- need to add protections (somes littles capacitors)

[Bryce]

If the component is very hot, you have a problem... If your resistor is hot (one used to pass from 5v to 3.3v), you will have a problem.

If the 100R resistor is getting hot you really do have serious problems. As the author of the article showed, the voltage drop across the resistor is 1.25V between 5V and 3.3V (including the voltage drop across the diode). This is 12.5mA which would mean the resistor is dissapating 15.6mW. As most SMD resistors are rated at 1/8W (125mW) there is no way the resistor could ever get hot.

Bryce.

[gerald]

You'll be pleased to hear that there are no CPC 3.3V hardware (that I am aware of) that has this issue. The C4CPC for example connects to the CPC via a Xilinx XC95144XL which has 5V tolerant inputs and a 74LS245 Buffer which is also a 5V part.

So you can sleep well knowing that your Amstrad isn't about to die (at least not due to some 3.3V hardware being connected to it.

Bryce.

C4CPC is a 3.3V hardware that interface nicely with the 5V CPC 
As Bryce said the pld used is 5V tolerant, and all signal from the CPC are going to it.
The output (data and acid output) are converted to 5V. This conversion is only needed because the 3.3V output where not high enough to be interpreted as 1 by the CPC.
Now a bit of warning regarding the 5V. While the PLD inputs are 5V tolerant and connected to IO and not straight power, the upper margin is small.
Using a power supply that is higher than the 5V + 10% may damage the PLD input, hence the C4CPC.
So be sure to use proper power supply, mainly when using switched ones. While the voltage reported by a multimeter may looks OK, you may have ripples that will be more than the max 5.5V with light load.

BTW, why is that thread in the German section 

[Gryzor]

I read about this while on vacation. A real shame and a pity, because I, erm, drive my MegaDrive with an EverDrive... now I'm very hesitant

[||C|-|E||]

I drive mine with original cartridges  but the issue is interesting because I was thinking about buying an Everdride X7, and they are very expensive 

[Gryzor]

If I recall correctly the 5 model is ok, too, and is significantly cheaper?

[||C|-|E||]

Yup, but it is a pity, because I was hoping to play things like Paprium, that is already 80MB, and I was looking forward to the save states. The other option is to mod it to solve the problem... but paying so much for something delicate that you need to mod is a bit crappy