Clean 5v protected supply for retro console mainboard remake

[L10N37]

Hi guys,

I've watched quite a few of these EEV videos on youtube. Good stuff. Decided this was the place for assistance.

I have used original schematics to "clone" an old sega mainboard just for project as I find it enjoyable to muck around with kicad and learn more about electronics.
I'm basically ready to measure out some mounting holes and add the power circuit but I don't want something as simple as - input cap - regulator - output cap type thing

I want reverse polarity protection for one thing. The original board is centre negative.. mine will be centre positive (I use aftermarket supplies and a little section of cable to swap polarity anyway)
Also the factory DC packs are rated at 9 but do just off 14v with no load.

I found a crowbar schematic with all the bells and whistles and made some adjustments

original:
http://www.electronicecircuits.com/electronic-circuits/7805-5v-1a-regulated-power-supply-with-overvoltage-protection-circuit

mine:


This was considered overboard by other people, it is also a little heavy on components...

2 of the original supplies are here:


The one diode in series seems to do the trick for reverse polarity protection.
Nothing gets fried when using the wrong adapter on the original boards, including the regulator.

A lot of sites are saying the diode between the input/output of regulator is not necessary, this is just for reverse discharge protection from caps?

Also concerned about noise as the originals seem to have ferrite beads (discontinued: FBA04VA600VB-00)
I was looking to substitute for: BL01RN1A1D2B

BUT, is this really necessary??
I've also read having the ferrite bead on the ground connection is not a good thing either?

Having a super clean supply is a priority, as the original boards are known as noisy, although mine is for RGB output only and doesn't contain the extra video encoding circuitry and clock line which would attribute greatly to the noisiness of the board.

So do I keep it simple, copy the original schematic across, with ferrite beads? I wanted to use an LM2940CT in replacement of the 7805 as it's a cleaner supply and its LDO so I can run this down to 6v input if i wanted to.. less heat.

BUT then i found the MIC2940 which has built in reverse polarity protection.

One thing that confuses me is does this protect every IC on the board, or just the regulator from being fried? Seen as how the original boards can handle reverse polarity and nothing is hurt surely I'm going a little overboard??

I know i could just smash on any simple, cheap as chips 5v supply and it would be ok.. but I want to ensure the output never goes over ~5v and fries my IC's, and I want super clean! The proprietary IC's are hard to find and aren't cheap yet alone the rest of the components.

Cheers

[L10N37]

I think i'm stressing too much over something that should be simple
This should suffice

[Ultron81]

The diode across the input and output is for protection in the case if you plug a power source into the output. In a case like this, it is not needed.

A 7805 is fine, and having a diode on the input would protect it from reverse polarity. Just be aware of the heat dissipation (use a heat sink, calculate the size you need). You can install a fuse on the input if you want some over current protection, though the 7805 has current-limiting circuitry which would protect it (in most cases).

Pay attention in the data sheets for minimum capacitance needed on input and output. You can always go higher on the output, but too high can cause issues as well. I don’t think the ferrite beads are needed, those are usually for reducing RF emissions (you aren’t making a product that needs to be checked out by the FCC).

I would use small value ceramic caps instead of tantalum. Make the larger output cap electrolytic. You don’t need a large cap on the input, as there would already be one in the adapter.

Think of the output cap as a reservoir. If a large current draw is abruptly needed (as it is in CPU logic circuits), the output cap would help maintain the voltage as the load changes. The larger the cap, the better it is at holding the output voltage at 5V. According to the MIC2940 data sheet, it requires 10uF or greater (no limit) other than the fact it needs to have an ESR of 5 ohms or less, and a resonant frequency above 500kHz. As capacitance increases, the resonant frequency drops, so keep that in mind. Also, it says you should use a .22uF cap on the input, not 0.1uF.

Data sheets hold a lot of information, don’t overlook them.

If you are designing a board, the ceramic caps should be as close as possible to the regulator.

[L10N37]

https://www.smspower.org/forums/18067-NewMasterSystem2Mainboard#112166"
this is the original thread/ blog of my board which is just a recreation using readily available schematics.

Thanks for the information, I originally adjusted the input cap to suit an LM2940 going off the data sheet, I also read about ceramic verse tant caps and from what I gathered, tantalum for heat properties and ceramic for better ESR?
I will make some adjustments and I've been advised a diode is still required to protect the input capacitors

[L10N37]

Also I was a bit confused with all the different versions of the PSU providing power to the same IC's. You can see they use different value caps across each version of board even though it's all using a 7805

[Ultron81]

The ceramics are better for high frequency transients. Hence why you always see them near ICs. Use them instead of tantalums.

As far as the size of electrolytics, looks like originally they had 100uF on the outputs. I would suggest go with a 16V or 25V 1000uF and call it a day. That should be good enough. Again, you don't need one on the input side of the regulator.

I've done a lot of work with consoles, designing some RGB circuits. Was in the process of redesigning a switchmode power supply and AV adapter board for the Famicom. Just havent worked on it in a while (life gets in the way). So if you need suggestions on other things, let me know.

I saw you have your CSync going through a resistor divider. The CSync input on monitors most always have 75 ohm termination (unless it is a TTL Sync only, which are rare). I would suggest making sure the output from your divider is ~ 1.4V, then put a 75 ohm resistor in series, so when it is connected to the monitor, it will half that to 0.7V (which is an acceptable CSync level, usually anything below 1V is OK). Also, be aware that RGB lines are 75 ohm terminated as well (all video lines actually), and you should have 220uF DC blocking caps in series as well.

Just looked, Master System outputs TTL CSync. So either use a resistor divider to reduce it to 1.4V and with 75 ohm in series, redesign divider to output 1.4V, or just take it, throw a 470 ohm resistor in series between the pin and the output. Also, make sure the cable you are using does not have any resistance on the CSync line (this is common). I'd also throw a 220uF cap in series as well.

[TimFox]

I would recommend adding a rectifier diode between the input and the output of the regulator (as in your first drawing), reverse-biased in normal operation, in case the load capacitance at the output holds the voltage higher than the input when the input jack is disconnected.  Having the output higher than the input for 7805s can damage the device, and a well-chosen diode has no effect on normal operation.  I have normally used a 1N400x device.