Recreating mod chip designed for the Playstation 1.

[Katcher]

"Routing" has two meanings here. Signal routing in software is the one thing. But in the other post I was referring to the board outline being routed by a CNC milling bit, vs V-score, which is making a V shaped groove partway through the PCB material. The cost for tab routing isn't really anything to write home about, but the kerf can add up when the board is so tiny.

Not that it matters if you are just making 10 board for yourself. Go nuts.

The way you did the capacitor is totally fine, but you could also make a more direct connection for the caps, if you wanted the absolute minimum ESR for your decoupling cap. It really doesn't matter; just a note because you asked.

Oh yeah I'm only gonna be making maybe 10-15 of these things since I'll most likely bugger up a few during assembly. Taking what you've said I've come close to what I'm considering the final design. I could probably do with moving the trace from PAD_5 along to the left just a bit but I feel everything else is fine. I could use a second opinion though.

A decouple capacitor should always be wired directly to the power and ground pins of an IC, as short and direct as possible. In this application it is probably not critical but in things that really need good decoupling those long spindley traces between the IC pins and capacitor would wreak havoc.

Mate. There is no way in hell I'm getting the decoupling capacitor that close to those pins. I believe I've already said that. I have however connected it directly to the proper pins this time with the pads connected to those traces. I have no fucking idea if it will work but if it does...yay.

[Mr. Scram]

It seems you could do quite a bit better than that. As people have stated correctly already, rotating the MCU on the board would probably save you loads of headaches. Decoupling pads need to be connected as directly as possible without unnecessary loops, so even if you're insisting on the complicated design it really pays to connect the VSS_B decoupling pad and the VSS_B IC pad together with a direct path. Just straight from one to the other without looping around.

[Katcher]

It seems you could do quite a bit better than that. As people have stated correctly already, rotating the MCU on the board would probably save you loads of headaches. Decoupling pads need to be connected as directly as possible without unnecessary loops, so even if you're insisting on the complicated design it really pays to connect the VSS_B decoupling pad and the VSS_B IC pad together with a direct path. Just straight from one to the other without looping around.

I appreciate you saying I can do better. I know I can. And I do sort of understand why the decoupler needs to be connected as directly as possible just not exactly WHY it needs to be. I haven't the slightest fucking idea what I'm doing and it's great. However. This is a tiny ass 11x11mm board. I've got to factor in the size of the components and the pads and the space in which the fully assembled modchip has to occupy. The thing is only a single layer as well. If this was a much physically larger project then I'd be doing what I can to get the decoupler as close and as directly linked to the VSS and VDD pins as possible. I know that from looking at what I have right now I can make a more direct connection between VSS_8(Not B btw. It's an 8. Numbering the pins/pads to keep track of what goes where.) and the decoupler. But it isn't feasible to make a more direct connection between Pad 1 and VDD_1 with this design. I'd go for a more rectangular design over a square one if I could but due to needing it to have as small a profile as possible, I can't.

[Mr. Scram]

I appreciate you saying I can do better. I know I can. And I do sort of understand why the decoupler needs to be connected as directly as possible just not exactly WHY it needs to be. I haven't the slightest fucking idea what I'm doing and it's great. However. This is a tiny ass 11x11mm board. I've got to factor in the size of the components and the pads and the space in which the fully assembled modchip has to occupy. The thing is only a single layer as well. If this was a much physically larger project then I'd be doing what I can to get the decoupler as close and as directly linked to the VSS and VDD pins as possible. I know that from looking at what I have right now I can make a more direct connection between VSS_8(Not B btw. It's an 8. Numbering the pins/pads to keep track of what goes where.) and the decoupler. But it isn't feasible to make a more direct connection between Pad 1 and VDD_1 with this design. I'd go for a more rectangular design over a square one if I could but due to needing it to have as small a profile as possible, I can't.

I'm not sure why you'd want to give yourself trouble by using the sideways design. This means traces have to go all over the board, probably requiring more space than would otherwise be the case. Can you explain?

[Katcher]

You can see the image right? The traces are running under the IC to connect to each pin once the IC is soldered in place. If I were to rotate the IC 90 degrees to the left I'd then have to elongate the PCB width-wise, I DON'T want to do that because space(Also see my hilariously poor Space Invader shaped PCB a few posts ago. Looks great, probably isn't ideal). Did I also mention I'm trying to make it as COMPACT as possible too? An 11x11mm PCB is as small as I can make it without running into issue. And given how damn tiny that is I don't really think the distance between decoupler and pins is a really big deal. I'm not having trouble with this design at all frankly and it doesn't seem the maker of the original design or those who have bought a fully assembled modchip and actually used it had trouble either. Sorry if I sound a bit arsey here. I get you're giving me advice I can use with future projects but there's just no way I can get the decoupler any closer than it is using this design.

[Katcher]

Okay so here's what I'm considering as the final board design. I'm happy with how it is but I'd still like some suggestions on some quick tweaks I could make that aren't to do with the decoupling capacitor before I generate the Gerber files. I plan on making an alternate design of this board utilising vias but I'd need to know the basics of properly using those first.

[KL27x]

I admit that decoupling is ugly. But the guys saying "it must" "it has to" are technically completely wrong. Either the chip browns out and faults, or it doesn't. And I guarantee this would work without a decoupling cap, so long as the spot you're tapping power and ground isn't measured in feet/meters. You're not driving any inductive loads or switching large transistors, and the PS board probably has 3x as many decoupling caps sprinkled around as it really needs and rock stable voltage levels.

100% A-ok.

As a learning point, you might want to check DRC settings and run a DRC check to make sure you have all the proper clearances between traces/pads. Typical fabs can do about 6-8 thousandths of an inch between traces without any additional cost. (Usually the trace width and space between traces is roughly the same for small signal traces; but it's totally fine to have fatter traces and thinner gaps as long as you keep them far enough apart, as stated).

[Katcher]

I admit that decoupling is ugly. But the guys saying "it must" "it has to" are technically completely wrong. Either the chip browns out and faults, or it doesn't. And I guarantee this would work without a decoupling cap, so long as the spot you're tapping power and ground isn't measured in feet/meters. You're not driving any inductive loads or switching large transistors, and the PS board probably has 3x as many decoupling caps sprinkled around as it really needs and rock stable voltage levels.

100% A-ok.

As a learning point, you might want to check DRC settings and run a DRC check to make sure you have all the proper clearances between traces/pads. Typical fabs can do about 6-8 thousandths of an inch between traces without any additional cost. (Usually the trace width and space between traces is roughly the same for small signal traces; but it's totally fine to have fatter traces and thinner gaps as long as you keep them far enough apart, as stated).

How do I use this DRC check? On EasyEDA I'm assuming it's Design Rule under tools? Also can I use holes in place of vias? EasyEDA keeps giving me bloody errors when I try to make vias smaller than default and it's annoying the shit out of me.

[amyk]

Space. It's going to be attached to the underside of the PU-18 motherboard around where the EJECT button of the console is when looking it it from above. The smaller the better. The initial square design is much easier to compact if needed where as a more rectangular shape resulting from the pins facing the left and right wouldn't be if I want to make the pads as easy to solder to as possible.

Let's step back for a second.

The problem you want to solve is how to attach a modchip to  your PS1.

The reason they're sold with a PCB instead of as just a bare chip is solely so those who have no soldering experience can install them. Does that apply to you?

If not, just glue the IC in place and dead-bug it along with its decoupling cap.

Problem solved.

[KL27x]

Well, check where you are going to make your boards... minimum via size is another "luxury add-on," when you go smaller than the minimum. Same number, here, the minimum hole size is going to be around 6-8 thousandths of an inch to not incur special costs. edit, sorry, probably about double that, around 15 mils minimum.

Yes, Design Rules is what I was referring to.

I'm not familiar with that software, but "holes" are usually not thru-plated, so they are not what you want.

If Amyk and Co had bothered to read any of this thread, they would know that OP is embarking on an educational voyage to produce something for his own singular amusement.   

OP, don't worry, this is just how this forum operates on a regular basis.   

[Mr. Scram]

Well, check where you are going to make your boards... minimum via size is another "luxury add-on," when you go smaller than the minimum. Same number, here, the minimum hole size is going to be around 6-8 thousandths of an inch to not incur special costs.

Yes, Design Rules is what I was referring to.

I'm not familiar with that software, but "holes" are usually not thru-plated, so they are not what you want.

If Amyk and Co had bothered to read any of this thread, they would know that OP is embarking on an educational voyage to produce something for his own singular amusement.   

OP, don't worry, this is just how this forum operates on a regular basis.   

Education isn't supposed to be fun, so correct decoupling it is! 

[Katcher]

Is all good. You've all been a wonderful help. I did some experimenting with Holes(HAH!). Seems like they can be used instead of vias. I'll utilise these in the meantime and get 5 boards made. If it works, hooray. If not I'll deal with large unsightly holes. Take that as you will. I get what Amyk is saying about just dead-bugging it but honestly that is messy. It's 2018, imma treat my tech with respect and hopefully not kill it in the process. That said, once I actually have the PCB's and components I think I'll install a new modchip into my SCPH-102 as well with the correct Hex. That thing was modded nearly two decades ago. Before I do order the components though, should I get a 100nF capacitor rated for 16v or 25v? I remember reading somewhere that voltage supplied to the components is between 3.5v and 8v so I don't want one rated at like 100v. That'd be silly.

[KL27x]

The 12F508 can handle max Vsupply of 5.5V, so it better not be 8V.

I use 10-50V rated decoupling caps. And my main criteria is w/e is cheapest in the desired package and capacitance, esp where it doesn't much matter. There are additional ratings for tolerances and change with temperature, as well, say for example XR7 is some of the "good stuff," regarding temp stability.

[Katcher]

How can I go about testing the voltage that will be supplied to the IC? I can assume whatever it is I'd need to use will need to be in contact to where pads 1 and 8 will be soldered to via wire on the PU-18 board?

[Mr.B]

If it needs to be small, why not use double sided.
Put the components on the bottom.
Ignore pin numbers in picture attached. Mock up done in <30 secs.

[KL27x]

Well, what you're looking for is the nearest 3.3-5V rail you can tap (it depends perhaps on the chip on the mobo you are messing with.. You want to use the same voltage of its Vsupply so the logic levels match), close to where you want to put the chip. Try looking up installation instructions if you have trouble locating a spot, yourself, using say a voltmeter/DMM. But surely you have instructions or else how are you going to install the other connections? 

^Mr.B's gerber, that was what I was suggesting, before. This gives room for a teeny tiny silkscreen, too. 

[Katcher]

If it needs to be small, why not use double sided.
Put the components on the bottom.
Ignore pin numbers in picture attached. Mock up done in <30 secs.

I thought about making a double sided variant after recreating the original modchip but was having difficulty using vias in EasyEDA. Kept giving me errors. You mind if I use this mockup you made for the 5 boards that would be double sided?

Well, what you're looking for is the nearest 3.3-5V rail you can tap (it depends perhaps on the chip on the mobo you are messing with.. You want to use the same voltage of its Vsupply so the logic levels match), close to where you want to put the chip. Try looking up installation instructions if you have trouble locating a spot, yourself, using say a voltmeter/DMM. But surely you have instructions or else how are you going to install the other connections? 

^Mr.B's gerber, that was what I was suggesting, before. This gives room for a teeny tiny silkscreen, too. 

Yeah I got the installation instructions for my board. Looks like I oughta grab a voltmeter before ordering components.

[KL27x]

^Oh, one final.. if you use Mr.B's board design, some of the bottom dollar prototyping special 10 fer's don't do tented vias. And god knows why we have to suffer the term tented/nontented, because it is so damn obfuscating. Tented means covered with soldermask. Untented is nekked copper.

So anyhows, if you try to put a silkscreen over a bunch of untented vias, it will remove the silkscreen.

Also, be sure to view your soldermask layer to make sure if your vias are tented/untented as you intend them to be.

[Katcher]

Is there away to import Mr.B's design into EasyEDA or am I just gonna have to eyeball it and recreate that one too?

[Mr.B]

You mind if I use this mockup you made for the 5 boards that would be double sided?

I simply posted a PNG into the public domain - feel free to use the idea.

Is there away to import Mr.B's design into EasyEDA or am I just gonna have to eyeball it and recreate that one too?

You will have to eyeball it.
I mocked that up in 30 seconds in Eagle and didn't bother to save it. (Would not import into EasyEDA anyway I don't think.)

[Katcher]

You mind if I use this mockup you made for the 5 boards that would be double sided?

I simply posted a PNG into the public domain - feel free to use the idea.

Is there away to import Mr.B's design into EasyEDA or am I just gonna have to eyeball it and recreate that one too?

You will have to eyeball it.
I mocked that up in 30 seconds in Eagle and didn't bother to save it. (Would not import into EasyEDA anyway I don't think.)

EasyEDA can import Eagle fyi. But no worries I'll eyeball it. What's the outline measurements so I can start from there? mm preferably.

[Mr.B]

... What's the outline measurements so I can start from there? mm preferably.

Slightly bigger than the PIC...   
(Sorry, had to use that.)

If you are happy with your original board outline size...
1. just take your original design, rip up all the tracks to wireframe/ratsnest.
2. flip the components on to the bottom layer.
3. re-rout the tracks, using vias, to look something close to my example.

One other side issue to be aware of...
Most cheap PCB manufacturers have a minimum board size.
The company I use has a 20mm by 20mm minimum.
Your board is going to be a lot smaller than 20mm.
You may need to lay up multiples and either have the PCB fab 'V cut' your board or cut them up yourself with a scroll saw.

[Katcher]

I'm using JLCPCB for my PCB's. Uploaded the Gerber files last night before bed and just waiting for approval before I pay. But I think I'll stick with my 11x11mm board initially then maybe tighten things up from there. Now with that out of the way leaves the elepehant in the room...how do I actually program the fucking PIC12F508? And because screw doing it the easy way and buying a programmer, imma fucking build one because it's a learning experience and it sounds fun.

I am a great bloody idiot. During the design of my PCB I had "Chungus Inc." on the bottom silkscreen layer as a joke and forgot I made it invisible to help with laying shit out. I've already submitted the gerber files to JLCPCB and now my printed boards will have "Chungus Inc." slapped on the back.

[amyk]

If you really want to... http://ww1.microchip.com/downloads/en/DeviceDoc/41227E.pdf

[Sudo_apt-get_install_yum]

Now with that out of the way leaves the elepehant in the room...how do I actually program the fucking PIC12F508? And because screw doing it the easy way and buying a programmer, imma fucking build one because it's a learning experience and it sounds fun.

I’ve read through this thread and seen the versions of the PIC12 break out your making, for a first board it alright!
Since you don’t know much about the embedded world ill help you out, DONT MAKE YOUR OWN PROGRAMMER!!!!

Its so much work to do this and its way to complicated for a beginner like you. I dont think you knwo how much work it is to design a programmer. making a break out board is one thing. Designing the board for the programmer, writing the code is a lot of work.

If your experienced with PIC's it wouldn’t take more than a few days but if your inexperienced with electronics design and embedded programming then it would be suicide, it would take so much time and you wouldn’t learn much since the difficulty level is much too high for a beginner.
It better to slowly build your knowledge than immediately jump on the hard stuff not understanding the basics.

Programmer like this is cheap and readily available so it’s just a time waste.

I’m no pro but I have done a fair bit of PCB design and µC work.
Take what I say with a grain of salt