Famicom AV/Power Board

[HyldenKing]

Hello
I'm interested in making a Power/AV PCB for HVC 07 Famicom
I drew these schematics based on the HVC 05-06 diagram and Ben Boldt's AV mod posts

https://imgur.com/U2JDUvb.jpg

I'm a newbie to the electronics and can't say if this will work or not, so I have some questions:
1. Is everything correct?
2. Is it okay to add the power switch before the regulator?
3. Which caps would you change?
4. Is it better to solder a tantalum cap to PIN20-22 or add it to the board?

[krzysiobal]

This is the board I have made some time ago:


It is both compatible with
* first revision famicom (with a big metal shielded dc/rf pcb, a'ka 5 pin), which has power switch on the 5V rail
* second revision (with a small dc/rf pcb a'ka 7pin), which has power swtch on 9V rail

(just connect JP1/JP2 differently according to the description)

It does not have the gretz's bridge, just a simple schottky diode placed on the DC input to protect against wrong polarisation.
Video amp is as in most of famiclones.

And the most important thing - there is a place to solder either
* mini jack A+V connector (this way you don't have to drill any holes in your shell)
or
* Audio RCA + Video RCA where originally two switches were placed (you need to enlarge those holes a little)

[aquasnake]

It is recommended to set the power switch on the 5V power rail after 7805 output. And keep the video amplifier part powered forever, which is conducive to prevent the surge at the moment of switching and will not make the video output flash white


For LCD TV, the highlight at the moment of switching will not have any problem, just interference. For CRT TV, the white level of highlight output exceeds the standard, which may lead to the instantaneous high voltage excitation of kinescope filament

[HyldenKing]

Thank you for the response.

krzysiobal
1. Why the large cap is at the input and not the output?
2. Can I use the schottky diode bridge instead of the regulars? What's the difference?
3. I can't find the quality brand ceramics for the regulator's input/output, can I change it to the 470nf film caps?

I've made this on EasyEda , should I connect 5v pins on the right side?

aquasnake
That makes sense. I think I'll add the holes for the second switch

[krzysiobal]

krzysiobal
1. Why the large cap is at the input and not the output?
2. Can I use the schottky diode bridge instead of the regulars? What's the difference?
3. I can't find the quality brand ceramics for the regulator's input/output, can I change it to the 470nf film caps?

1. Why large cap at the output, not at the input?
Cap acts as a charge reservoir for fast current transitions. If some chip needs lets say 100 mA for a fraction of second, this current need to be sourced from somewhere. 7805 is linear voltage regulator, so if you load 100mA at output, it will acquire 100mA from input, so this current will be first driven from the input capacitor. So putting larger cap at output than at input is useless.

So what about putting big cap at the output and at input? Better put many small caps near vcc of every chip, instead of one big.

And finally, cap is similar to voltage source for AC analysis. It means that cap during power-up acts as almost a short circuit for some time (the bigger the cap is, the longer it lasts), so it might trigger some overcurrent protection or even break the voltage regulator.

During power down, the energy stored in cap need to be discharged somewhere.. If you will put bigger cap at output, you will go with a situation that the input cap will be discharged before output one, so the 7805 will have higher voltage at its output that input, this is no good.

And some low drop voltage regulators (which 7805 is not) have expclicitly said not to put big cap on output as it will result in oscillations.

2. Schottky diode has lower voltage drop, giving you higher margin. 7805 to give proper 5V at output needs at least 7.5V at input. So if you have 9V DC power supply, then if it is a transformer one, the voltage at output will still fluctuate between VL and VH, depending on the current consumption (9V is a average voltage).

Puttin schottky diodes on the graetz bridge might not always be good idea as they have low reverse breakdown voltage (but for 9-12V this should not matter)

[HyldenKing]

Quite a lot of info there, thanks!
So I'll change the input cap to 1000uf or lower if it's possible and the output cap to 470uf and use 100nf decoupling caps.

[stefanobaron]

This is the board I have made some time ago:


It is both compatible with
* first revision famicom (with a big metal shielded dc/rf pcb, a'ka 5 pin), which has power switch on the 5V rail
* second revision (with a small dc/rf pcb a'ka 7pin), which has power swtch on 9V rail

(just connect JP1/JP2 differently according to the description)

It does not have the gretz's bridge, just a simple schottky diode placed on the DC input to protect against wrong polarisation.
Video amp is as in most of famiclones.

And the most important thing - there is a place to solder either
* mini jack A+V connector (this way you don't have to drill any holes in your shell)
or
* Audio RCA + Video RCA where originally two switches were placed (you need to enlarge those holes a little)

Hello Krzysiobal!

I've download your awesome Eagle project and I'm trying to update it changing the SMD components by through-hole components.
It's the first time I use Eagle PCB and I'm having hard time to learn the PCB part

Maybe I'll make a manual PCB using the classic pen as it's a simple design.

Thanks again!

[Simnamedjim]

Hi. I've just built this board, thank you for providing it. Is it meant to plug directly into the existing mainboard av output or should I be routing video directly from the ppu pin 21?

[KevinHarte]

Are there jailbars with that board?

[aquasnake]

“putting larger cap at output than at input is useless”
This is nonsense
It is customary for the output capacitance of 7805 to be larger than the input capacitance. A capacitor is a charge storage device, and the 7805 itself has internal resistance. The output capacitor will first tend to save energy and obtain charge near it before absorbing it from the input. However, there is impedance from input to output, and it will not immediately demand it from the input. Only after a certain period of time, the output capacitor would be drived by the front-end.
Generally speaking, the output capacitance of 7805 is greater than the input capacitance. But the output capacitance value should not be too large. If it is too large, it will cause oscillation after 7805 is powered on. It is not satisfied with stable voltage within 6 m2 clocks and cannot provide a good power on environment, resulting in gray screen when starting up
On the contrary, if the input capacitor is too large (1000uF), it will also have problems. It will generate excessive impulse voltage at the moment of shutdown, which will be transmitted to the later stage and cause damage to the cartridge storage chip. Continuously pressing the power button quickly several times can easily lead to such situations. In some clones that are prone to damaging cartridges, I found that a large capacitor (1000uF) was added to the input end

Suggested parameters: Input capacitor 16V/220uF, output capacitor 10V/470-560uF, and add 1 or 2 104 ceramic capacitors to bypass surges

[nathanozlegend]

just got the pcb in production

[kevtris]

“putting larger cap at output than at input is useless”
This is nonsense
It is customary for the output capacitance of 7805 to be larger than the input capacitance. A capacitor is a charge storage device, and the 7805 itself has internal resistance. The output capacitor will first tend to save energy and obtain charge near it before absorbing it from the input. However, there is impedance from input to output, and it will not immediately demand it from the input. Only after a certain period of time, the output capacitor would be drived by the front-end.
Generally speaking, the output capacitance of 7805 is greater than the input capacitance. But the output capacitance value should not be too large. If it is too large, it will cause oscillation after 7805 is powered on. It is not satisfied with stable voltage within 6 m2 clocks and cannot provide a good power on environment, resulting in gray screen when starting up
On the contrary, if the input capacitor is too large (1000uF), it will also have problems. It will generate excessive impulse voltage at the moment of shutdown, which will be transmitted to the later stage and cause damage to the cartridge storage chip. Continuously pressing the power button quickly several times can easily lead to such situations. In some clones that are prone to damaging cartridges, I found that a large capacitor (1000uF) was added to the input end

Suggested parameters: Input capacitor 16V/220uF, output capacitor 10V/470-560uF, and add 1 or 2 104 ceramic capacitors to bypass surges

Having a larger capacitor on the output doesn't help much, and can indeed be bad. In all my designs, I place a 0.1uF capacitor on the output of the regulator, then have bypass caps near every or most every chip. The regulator's job is to regulate, and it is very good at this job. Adding large output capacitors is thus counterproductive, and takes up excess space, and costs money to boot. The regulator is relatively fast, and will handle any medium speed current transients your circuit might pull, and then the local bypass capacitors handle any high speed transients.

Also, if the input is disconnected and pulled down (i.e. via other circuitry, or a short due to the plug being removed from the barrel jack or similar) the energy stored in a large output capacitor can destroy the regulator. If you really want a large output capacitor, you need to place a reverse biased diode across the input to the output (anode to output, cathode to input) to shunt the energy across the regulator. But again this is generally useless and will just waste space/money.

The TI datasheet seems to suggest 0.1uF output capacitors also, and indicates if your output is larger than 10uF, the diode is a good idea.

As for having a large cap on the output than input, this is false. The input capacitor's job (if running off rectified AC) is to act as a reservoir for the pulsing DC, so that the regulator does not "crash" and fall out of regulation, during the zero crossings of the sinewave. If this happens you will get ripple on the output, and having a very large output capacitor will not really help you here. If you are running on regulated DC (i.e. a regulated line lump or similar) the input capacitor can be 10uf or so. I tend to place a 0.1uF on the input of the regulator, then a 10uF to 100uF where power enters the board.

ST's datasheet also suggests 0.1uF on the output.

ETA: I forgot to mention low dropout regulators. The above is for the typical 78xx positive regulators found on the NES, famicom, SNES, etc. Low dropout regulators however are another beast. Be very careful when using these and read the datasheet carefully. The input and output capacitors on them can be very critical and using the wrong ones will result in output voltage oscillations. The LM1117 tends to need 10uF capacitors, and some of them require a tantalum capacitor. Using a ceramic is not advisable. Each regulator is different and it can even differ between brands. The TI xx1117 might be different from the ST xx1117.



TI datasheet:

https://www.ti.com/lit/ds/symlink/lm340.pdf

(page 13, section 8.1.1, 8.2.1, 8.2.1.1 )

[Ben Boldt]

When I used an LM1117-3.3 regulator recently, I could hear a squeaking sound when there was no load on the output. I was wondering where that sound was coming from. When I looked at the 3.3V rail, it had sawtooth ripple about 20kHz.

The input was 6VDC from a bench supply and I had a 0.1uF ceramic cap on the input side, and I had a 10uF ceramic cap on the output side. When I added an extra 10uF electrolytic cap on the output side, the ripple went away.

Are we saying that it would have worked better to put this cap on the input side of the regulator? I thought the bench supply probably already had a large cap there for me.

The purpose of this regulator is so I do not set the voltage too high by accident when I am powering 3.3V stuff.

[kevtris]

Yeah that's the exact thing that can happen. I checked the datasheet and they recommend a 10uF tantalum capacitor on the output, and a 10uF tantalum on the input (I am checking this one:)

https://www.ti.com/lit/ds/symlink/lm1117.pdf

Pages 14/15 talk about the capacitors. The input capacitor in your case is probably too small. The kind of capacitor is relatively critical too. The datasheet says the output capacitor's ESR has to be between 22 and 0.3 ohms. If it is outside this range, it will cause instability and oscillation. Using an aluminum electrolytic is probably OK but it will need to be larger to compensate. A ceramic is probably a bad idea because the ESR is too low (and its capacitance will be much lower than specified too, depending on the dielectric. a Y5V 10uF capacitor might only read 3-4uF when it has half its rated voltage on it).

So I suggest increasing your input capacitance to 10uF to 47uF tantalum and your output capacitor should be a 10uF tantalum. If your regulator is made by someone else it might be a good idea to see what their datasheet says.

In my designs I try to find LDOs that are stable with ceramic capacitors, since I do not like to use tantalums (or aluminum electrolytics) if possible. There are some out there stable with ceramic capacitors.

Annoyingly, these LDOs don't tend to have a very standard pinout. Even though it has 3 leads, different parts will have different pinouts. This makes substituting due to availability, etc. a huge pain. Be careful about that; the LM1117 for example has a different pinout from the 7805. Its tab is VOUT, unlike the 78xx where the tab is ground.

[nathanozlegend]

just about to give this a shot the only thing I'm using is a switching 7805 it don't need heatsinks never gets hot make a video soon been a long time