Author Topic: Nintendo64 RGB DAC (Read 25925 times)

Barkuti · « **Reply #40 on:** July 26, 2007, 06:48:04 pm »

Yes, just got a wider head with flatter tip, and got those things removed. The oil worked wonders btw. Each screw made a distinctive "clack" sound upon turning.

Wow, the mainboard is sandwiched between a couple metal plates :huh: (Making a bulletproof game console?). I do not see a good reason to put the upper one back in place; also, it's installed below the heatsink and that's not good for heat transfer (Nintendo designers must had been smoking crack while coming up with this).

More updates once I recover some sleep and get time for the mod.

Cheers

viletim · « **Reply #41 on:** July 26, 2007, 11:16:33 pm »

Barkuti,

Sorry about the missing resistors. I actually found a packet of 560 ohm resistors floating around the other day and thought: I know where these are supposed to be...

I have just recently sourced some 280 ohm resistors which would bring the count down to 42. I'll send you some if you like (and have the patience to wait a week for them).

BTW, you definately need an R20 for each DAC.

lameboyadvance · « **Reply #42 on:** July 27, 2007, 02:38:40 pm »

Firstly, excellent work viletim!

I (along with others apparently) have been waiting for a decent RGB mod for the DENC/etc chips.

...With these 280 ohm resistors, how does that change your circuit diagram? All the 560s change to 280, and the ones in parallel (R8-9, R10-11 etc) are combined into a single resistor?

Also, I noticed that you're selling CPLDs for $9 and resisitors/misc for $6. Does this change affect the price?

viletim · « **Reply #43 on:** July 27, 2007, 08:27:35 pm »

lameboyadvance,

The 280 R replaces two 560 Rs in parallel so the circuit would look like:

Code: [Select]

. 0 ---2R--+--2R---+
.          |       |
.          R      ---GND
.          |
. 1 ---2R--+   R=280
.          |  2R=560
.          R
.          |
. 2 ---2R--+
.          | etc...
.~~~~~~~~~~~~~~~~~~~~~~~~
.          |
. 6 ---2R--+----> vid out

Price of the bits remains the same.

Drewman21 · « **Reply #44 on:** July 28, 2007, 01:32:27 am »

I hate to ask as I am interested in this mod but it seems pretty involved. Tim would you post some pictures or a short walk though of what is need to assemble your mod? And the typical dumb question that I have to ask as I have all ready bought the parts for a few of the MMMonkey RGB mods. How much better is your mod over just the plain old amp he came up with? Thanks.

Barkuti · « **Reply #45 on:** July 28, 2007, 01:03:56 pm »

Hi again guys,

Just dropped by to say this mod is certainly painful. A PCB would really help for suckers like me with 1.5 mm tip soldering irons. I've already dug 3 little holes on the socket's plastic :ph34r: (man the socket was a GOOD idea imho).

A little question: Is the mod going to work fine just soldering the RGB outs to the 1, 2 & 4 pins on the multi-out? I mean, it'll be taking csync from the composite signal, everything ok? I just ask this because Moosman was saying the optimal RGB cable uses a 75 Ohm resistor from pin 20 to ground on the SCART, I've measured mine (a cheap one I bought for a GameCube) and it doesn't seems to have it (infinite resistance from pin 20 to shield with disconnected cable). �?

Cheers

viletim · « **Reply #46 on:** July 28, 2007, 03:19:51 pm »

Drewman21,
If you find that the Assembly Procedure isn't detailed enough for you then you should probably give this one a miss.

Barkuti,
That's the way mine's wired up. Works well. The extra 75 ohm resistor in the cable is only required if you intend to actualy watch the composite signal. If you're sure it's not in the cable (better of open up the SCART end and take a look) then you can just solder it into either the cable or the console.

I don't see how a socket would make assembly any easier. I would have thought it would make it even more arkward with it's double row pin arangement.

I did consider makeing a PCB but it would either be big with lots of holes or small and use SMD resistors. I hate drilling holes so that would mean I'd get them produced professionally and have all the problems that go along with that (minimum order, expensive, etc). If I were to produce an all SMD version then people would whinge about the parts being too small and hard to solder. The floaty resistor method is the best - good soldering practice too!

Barkuti · « **Reply #47 on:** July 29, 2007, 02:33:27 pm »

Greetings,

I've made a dirty PCB design for standard size components + socket. Started to draw it with "ExpressPCB" (one of those free proggies companies make to lure customers, first link in google), and later noticed the software assumes you to be drawing the layout viewed from the component side; I did the drawing with the "solder side" view in mind, so it came up inverted and the prog can't make the flip. What a crap. Well, anyways it served the purpose. Take a look at the attached jpg file I made (only export format supported is by the software is DXF -Autocad?-).
Target size is 2.6 x 1.9 inches (around 66 x 48.3 mm), which is just a tad wider than the heatsink's inner width (at 44 mm), and a bit longer (just 6 mm) than the lower opened central section; a couple of pliers bites at the corners solves the problem. :rolleyes:

This is intended to make a single sided copper clad board. You print the drawing and transfer it to the board with carbon copy paper with a ballpen/pencil etc.; just the edges of course. Then with an alcohol based ink permanent marker the tracks are drawn/filled.
What I'm describing is a simple old method my brother teached me 20 years ago when he was studying electronics. The developing was made inmersing the board in "Sal fumante" (a 25% dissolution of HCl in water) and then spurting hydrogen peroxide over it (100-110 vol dissolutions, 30 to 33%), until the only thing remaining were the painted tracks.
For finishing, a good fresh water wash, alcohol to remove the ink, and a quick check for imperfections for fixup. The method may be somewhat tricky with the thin tracks we're using here.

Ok, this is a dirty job (OMG the picture background is black! :wacko: ). If anyone has a better software to make this at least half of the work is done.

Cheers

Barkuti · « **Reply #48 on:** July 29, 2007, 02:39:07 pm »

Mmm forgot to say, the outer column/row resistors should be 280 Ohm, except the ones I marked as "R20".

Barkuti · « **Reply #49 on:** July 30, 2007, 02:07:00 pm »

Hi,

Just finished the mod; tested with success (well, at first I thought I had screwed up something as I was only seeing a black screen :ph34r: , but quickly noticed my expansion pak wasn't back in so the RAMBUS had no termination).

I'm sure I'll enjoy it a lot when plugged to my arcade monitor or a CRT TV, but testing it on my kitchen's LCD TV shows a way too blocky image, also the colours look somewhat grainy while in motion in certain games (for example in Mario Tennis' intro). Yeah, I know, LCD's suck.

Again, awesome job Viletim.

BTW, I was thinking on using some hotglue over the chips' contacts/socket legs/etc. to firmly fix everything in place. Is this a good idea?

Cheers

Drewman21 · « **Reply #50 on:** July 31, 2007, 02:15:10 am »

Very cool. Still wondering how this compares to the MMonkey amp mod for video quality. Anyone have any thoughts?

Tim could you let me know the cost of one of your kits sent to the states with shipping? If you have my address still that would be cool. I have a extra n64 that I could do up with it. Thanks man!

Barkuti · « **Reply #51 on:** August 01, 2007, 09:22:11 pm »

A little update on this topic.
After something more than the usual quick test to check if it works, I've noticed the screen offset isn't quite properly set with the RGB output. No problem with composite, but through Viletim's DAC on my machine the screen is noticeably shifted to the left; so much, that even using the "screen adjustment" option in Smash Bros and moving the rendering area all the way to the right, it doesn't becomes centered.

Of course, this may not be so much of a problem in a monitor, but on TVs it is no doubt since you can't count on making display shift adjustments.

I wonder if may I have done something wrong. Hope Viletim knows the answer.
�Is there any solution or workaround for this?

Cheers

Moosmann · « **Reply #52 on:** August 02, 2007, 05:32:01 am »

It`s normal with every rgb connection.
I sold "RGB Shifter" (in order to adjust RGB to left/right) in the last years ago, but at the moment I do no longer sold these units. Maybe I will make an update in my homepage with the circuit diagramm.

Greetings Markus

Barkuti · « **Reply #53 on:** August 02, 2007, 09:46:44 am »

Mmm, I've read some info on this.
Seems the problem is related to the fact it's taking the sync signal from the composite video which, taking additional time to be generated and/or decoded, produces a delay that translates into a picture shift.

Possible solutions:

a) Take the composite sync signal from the PLCC, and put it into AV port's pin 9. The drawback is we would need to remove the composite video signal from it. Seems quite hard and messy to me, to say the least. We would also lose CVBS output, unless a switch would be installed.

B) Insert some delay circuitry on the RGB lines, giving the exact same delay to all 3 of them. Damn. Has anyone something to add here? Potential quality issues?

Suggestions?

Barkuti · « **Reply #54 on:** August 06, 2007, 12:23:56 pm »

Hi fellas,

Decided to strip out the composite video signal from the output, and put composite sync on its place. Went the hard way and lifted DENC-NUS' pin 20 from the board (bye bye RF encoder). Next, soldered a wire from DAC's pin 27 to pin 9 on the multi-out.
Output's nice and centered now ^_^ . Maybe its me but I'd say the picture seems even crisper now. :huh:

Cheers

viletim · « **Reply #55 on:** August 06, 2007, 09:40:04 pm »

Barkuti,
The !CSYNC will need a resistor in series to drop it down to nornal video level. About 220 ohms sholud do. Yes, I should have mentioned this before... If you've already put everything back together it's probably not worth bothering about.

Barkuti · « **Reply #56 on:** August 07, 2007, 10:20:52 am »

Hi Viletim,

What's the Vpp level of !CSYNC, and what's the csync level SCART TV's are expecting?

And what about monitors? I plan on building a patch cable for interfacing to an arcade monitor with D-sub connector; should the resistor be added in that case? Monitor supports standard VGA signal levels, including composite sync. Capacitors for the RGB lines, as in the SCART cable, are mandatory, right?

I've not blown up anything yet :lol: , I guess those TV circuits are built to withstand this kind of abuse, and probably much more...

Cheers

Barkuti · « **Reply #57 on:** August 07, 2007, 10:49:45 am »

Mmm, just read your MAME cabinet to SCART TV article Viletim (another awesome job btw). So 0.3 Vpp for csync, if !CSYNC is 1 Vpp, best resistor value should be ~180 Ohms, isn't it?

Still don't have a clue on the sync levels and/or impedances for monitors... when applicable.

Cheers

viletim · « **Reply #58 on:** August 07, 2007, 08:48:23 pm »

Barkuti,
!CSYNC is 3.3Vpp with lots of current available (drops to 2.7v at ~25mA). It's a digital logic signal coming from a CMOS chip (the CPLD) powered from a 3.3v rail.

A composite video signal is usualy 1Vpp which is made up of 0.7Vpp of video and 0.3Vpp of sync. If you like to totaly adhere to the standard you can use a 680 resistor in series with !CSYNC to divide it down to 0.3Vpp with the TV's 75 ohm termination resistor. But as there's no video content, there's no harm in increasing the level of this videoless composite video signal up 1Vpp, especially as the CPLD outputs can easily provide the 13mA required. It's good for the signal to noise ratio. I chose 220 for a series resistor which, when loaded with 75R, will give you a healthy .8Vpp sync.

Arcade monitors are a bit different...they always require sync at TTL levels (max .4V = low, min 2.4V = high) which the CPLD output is compatible with.

Barkuti · « **Reply #59 on:** August 13, 2007, 06:40:58 pm »

Hi fellas,

Another little update on this topic. I've tested my N64 on the TV set that gave me colour picture problems with my ultra-dirty VGA-SCART cable (see here). Guess what? Same thing here. Green tinted picture.

In the first case sync signal was a HSYNC + VSYNC (both negative) mix, N64 uses !CSYNC from DAC through the RGB cable's 220�F cap on pin 20. The common point is, TTL signal levels in both cases.
Does this means that TV set doesn't likes that much sync signal? The greenish hue seems consistent over the picture colours, yet not present for black (intensity driven?).
Whatever the hell's this thing triggering inside the TV circuitry... :huh: I, for one, don't have a clue about what it is.

I'll have to drop that signal level and test it out.

Cheers

jero32 · « **Reply #60 on:** March 10, 2009, 04:53:16 am »

Hey i'm new here. Sorry to bump this thread but I wanted to ask some questions.

First: has anyone tried this on a pal unit? (Seems like it should work anyway.)
Two: has anyone tried it with a "real" dac?
And can I use this to convert the c-sync: http://www.tkk.fi/Misc/Electronics/circuits/vga2tv/circuit.html�

jero32 · « **Reply #61 on:** March 12, 2009, 07:01:09 am »

Also what do you guys suggest is the best way to solder to the tiny pins. (I'll need lots of magnification) I have the 16 pin dac.

Jokeri · « **Reply #62 on:** December 23, 2009, 05:25:59 pm »

Sorry to pump up old topic, but just found this forum while searching PAL N64 RGB mod.
viletim: do you still have components for sale?
Picture quality on my PAL N64 is very blurry, and I would like to try your mod.

viletim · « **Reply #63 on:** December 26, 2009, 07:24:42 pm »

Jokeri,
I have enough parts left for two kits. Email me if you'd like one.