Rolf wrote:

Thanks for the advice! I'm fairly new to PCB design, so every bit helps a lot. I just add ground planes after routing because EAGLE won't render them on the fly, so it's tricky to organize things with them there.

There's a designated way of adding ground planes (or other planes for that matter.) If I remembered how to do it, I would explain to you here and now. tongue

Rolf wrote:

The socket actually has to be a DMG socket if it's supposed to fit in the case, because of the way the DMG screws together both inside and out.

Hmm, I don't have a DMG open to compare with, but I'll get back to you on this one.

Rolf wrote:

I'll look into the caps soon, I added the same number of extra caps to this board, but I should probably organize them a bit better.

It's good practice to have a big one near the power supply connection, and a couple of small ones spread out across the board, close to the other chips.

Rolf wrote:

I really like your idea of having extra 5V and ground pads for modders, as well as the extra RAM. I hadn't thought of that at all, but yeah, it should be possible to use up to 64KB with all 16 address lines, correct?

Yeah, but then there would be no space for ROM, graphics memory or anything else. The idea is based solely on the fact that there's an unused area of "mirror" RAM, which basically means that if the two top bits are set (as they are for all addresses in the $Cxxx-$Fxxx range) the external RAM is selected. But the third highest bit is simply ignored, which means the same area is repeated. Actually, it's a little more complicated than that, since the highest 512 bytes, which would otherwise be in mirror RAM, are actually used for other things and don't reach the external memory bus.

scienceguy8 wrote:

Will you be releasing the EAGLE files under an open source or non-commercial use license?  On the off chance someone tries to recreate the GameBoy processor with an FPGA, or tries some other complicated mod, having this board for a template would make things much easier.

I'll most likely end up making the files public. I need to read through the different licenses I could use though. Ideally I'd like it to be open but non-commercial unless I know about/approve it. I hope I don't sound too selfish or whatever by saying that. I'd love for this to help get more development going in the community.

i have a stupid question.. why would you make pads for "pro sound modding" when you're designing a pcb? just make the audio lines the way they should be with a quite buffered output... that being said.. it's a DMG!!! it doesn't need to be modified.. lol.

low-gain wrote:

i have a stupid question.. why would you make pads for "pro sound modding" when you're designing a pcb? just make the audio lines the way they should be with a quite buffered output... that being said.. it's a DMG!!! it doesn't need to be modified.. lol.

Well the main reason for that is my complete lack of any formal training with electronics. I'm no engineer, to say the least. tongue However, I think that part works well enough to leave it alone for the time being. Besides, those pads were just thrown in after everything else as a "why the hell not?".

if you are willing to work together, i can give you a rough design of an LCD PCB that would include a pocket screen, maybe a second arduino for RGB control or other things,

i have no experience with designing circuit boards, so i will just be drawing it out by hand and you would have to do the digital version.

like i said before, this is great work, i think if you can loose the bootloader header and maybe fit some extra big SMD caps too, i will try and get that arduino code tested,

low-gain wrote:

i have a stupid question.. why would you make pads for "pro sound modding" when you're designing a pcb? just make the audio lines the way they should be with a quite buffered output...

It works well enough as it is, at least when driving a high impedance, like a mixer/soundcard input/an amp. Don't fix it if aint broke.

low-gain wrote:

that being said.. it's a DMG!!! it doesn't need to be modified.. lol.

Well... I used to believe that, until I heard the difference myself. It's true that DMG doesn't sound as horrible as GBC when unmodded, and that the modification doesn't (and doesn't really need to) improve frequency response. But what it does do is improve the signal:noise ratio, by both bypassing the amplifier, which adds noise, and providing a higher nominal voltage. This comes at the cost of higher impedance though, which does make the output unsuitable for headphones, but, again, works just fine for high-impedance inputs. If you're actually serious and want a clean recording, or less background noise when performing on a big PA, the mod is the way to go. If you're just starting out and hardly know how to use LSDj, it can wait. smile

the DMG's buzz can get really overpowering live, if not pro-sounded

Last edited by OwenMcGarry (Sep 6, 2010 12:11 pm)

if you need any assistance with PCB design, let me know.

since you are in EAGLE, i'm pretty proficient...  learning altium designer made eagle cad a cinch...  well, apples and orange, but altium can be like pulling teeth compared smile

but, yeah, if you need any advice/assistance, please let me know.  i'm happy to help how i can.  including the grounding plane issue.  which, tbh, is going to be an issue the way eagle handles this and the board is laid out. 

this goes to Rolf and Nex in particular, but to anyone else really.  i don't mind helping out with eagle.  its pretty fun honestly. 

also, be careful in eagle because the free version's license has some caveats concerning the making of things to sell.  the free version is meant for hobbyists really, not for commercial application.  last time i checked, they offered a single license thing to get past this...  but i don't know the cost.  altium ended up being substantially cheaper and more powerful for me personally...

i'd try to talk you out of eagle, but its too late wink

edit -- also, many of the eagle libraries are licensed as well, including the sparkfun one the last i checked (try the adafruit one instead).  meaning, you shouldn't use their footprint/silkscreen/etc...  i always make my own footprints in eagle though, except for caps and resistors really, and can help with this also.  or, can walk you through the basic process of doing this.  its pretty horrible the first couple times, but becomes very easy after you figure out what it is you are doing...

this is a good fucking idea all around and would pave the way for some interesting mods i bet - i would definitely buy a pcb if you made a run.  make sure to put a variable clock in there!  would there be a way to get more S-RAM in there?  then we could extend the size of songs we could have on one LSDJ rom?  like instead of 8K S-RAM put like 16K in there or something.  256k and bring every song you ever wrote?

or is the S-RAM just inside the ASIC CPU chip?  it would be badass if someone reverse engineered the CPU chip and bumped the processor power up so you get crazy FM sounds like TG16 on a DMG.  Sorta like what i think Mr. Baldwin is doing with Pulsar.

ok - i think this post belongs in that "ultimate fantasy gear" thread.

Last edited by animalstyle (Sep 6, 2010 2:50 pm)

animalstyle wrote:

this is a good fucking idea all around and would pave the way for some interesting mods i bet - i would definitely buy a pcb if you made a run.  make sure to put a variable clock in there!  would there be a way to get more S-RAM in there?  then we could extend the size of songs we could have on one LSDJ rom?  like instead of 8K S-RAM put like 16K in there or something.  256k and bring every song you ever wrote?

No, the modification I'm suggesting is for the internal work RAM. Has nothing to do with the external save RAM, that lives on the cartridge. However, my LittleFM LSDj file manager, that I like to namedrop, but actually can't seem to finish, does what you want by allowing the cartridge flash to be used as a save memory.

animalstyle wrote:

or is the S-RAM just inside the ASIC CPU chip?

Actually, it's using external RAM chips.

animalstyle wrote:

it would be badass if someone reverse engineered the CPU chip and bumped the processor power up so you get crazy FM sounds like TG16 on a DMG.  Sorta like what i think Mr. Baldwin is doing with Pulsar.

Actually 1: People have already done that and made a FPGA implementation of the CPU. See: http://www.fpgb.org/
Actually 2: That should be possible to do in software with the current CPU and CPU speed, just that LSDj doesn't do it. (Although the HF vibrato is tasty!)

i think a simple power supply upgrade, and a better output amplifier circuit would be all that is needed to make the dmg's output better.
they didn't design the thing with audio quality in mind.. So they just tossed what ever worked in... and since they were only concerned with the digital
side of the circuit.. again, it just wasn't their main focus....

I really think if a newer supply was designed and installed into a dmg you'd see some serious change in how the thing sounds.
I'm sure the technology of DC:DC converters has improved since the 80's.

low-gain wrote:

i think a simple power supply upgrade, and a better output amplifier circuit would be all that is needed to make the dmg's output better.
they didn't design the thing with audio quality in mind.. So they just tossed what ever worked in... and since they were only concerned with the digital
side of the circuit.. again, it just wasn't their main focus....

I really think if a newer supply was designed and installed into a dmg you'd see some serious change in how the thing sounds.
I'm sure the technology of DC:DC converters has improved since the 80's.

I agree with this, and also I would say that the only reason I can think of that Prosound actually does anything is because it bypasses some really crappy audio hardware. Unfortunately, most prosounds I have seen would also not have any shielding or filtering at all, so you'd still pick up noise from the surrounding address lines. Also, with the audio output basically connected directly to the CPU as many prosounds do, you're greatly increasing the risk of ESD damage to the CPU.

I would say the best sound quality on any gameboy, not just the DMG, would come from a combination of replacing the DC:DC converter with a better one, as well as bypassing the cruddy internal filter and amp circuit and instead running through a properly designed filter and op-amp.

agreed. output amp is very important. you really shouldn't just top off the main outputs of the CPU.

low-gain: If you want to eliminate the need for an extra jack, the idea of using a new amplifier does make sense. But the current "Prosound" modification is easy to do, especially on stock consoles, and works well enough.

As long as you use the modification for high impedance inputs only (mixers/soundcards etc) there is absolutely no problem. In fact, it's not much different from the high impedance input of the DMG's own amplifier. smile

The DC/DC converter then... It does produce some ripple, but I don't think it's the main source of noise, at least not when bypassing the DMG's own amplifier. Where this is really a problem, it can probably be solved more easily with more filtering. The main problem with replacing it, though, is that provides -18V for the LCD in a small form factor, which you'd have to make up for.

low-gain wrote:

agreed. output amp is very important. you really shouldn't just top off the main outputs of the CPU.

This is just a scare tactic without much basis in reality. How about a real argument? Unless you're looking to eliminate the need for a dual output, or you're very keen on getting noise-free headphone sound, I don't see the need for a new amplifier.



Arfink: Picking up noise from address lines: Marginal problem. Insofar that this is a problem, it can be improved by tapping ground off of a location near where you tap the signals.

The ESD argument... No, just no. If you want me to elaborate on why this isn't a problem, I could.

Last edited by nitro2k01 (Sep 6, 2010 9:07 pm)

i love this idea, and do you realise that with the heart arduino it makes a bad ass background?

nitro2k01 wrote:
Theta_Frost wrote:

I was thinking of a expanded RAM idea like nitro as well.

No, you didn't. tongue
(Unless you actually knew about the mirrored RAM on $E000-$FDFF)

I was thinking about it, but didn't know how to implement it. wink

Theta_Frost wrote:

An odd feature I would like to see would be the VIN pin on the connector moved to some pads as well... I plan to experiment with this soon.

What are you trying to do? Since the Vin pin has very snug voltage range requirements, and only works if the program running on the GB activates it, I think you're better off injecting your signal into the post-CPU, pre-AMP part of the chain.

This is a good point, it was more rambling as I was spilling out ideas. tongue  However somewhere along the line a cart with an external soundchip would be rad!

How about 2 crystals on board and a switch to go from underclocked to not?  If work on FPGA boy is really ambitious would it be possible to implement an R800 CPU as a sort of accelerator?  Also I'm learning eagle right now too!  I'm designing some GB carts. (Just 64MB as seen on Reiner, etc.)

Last edited by Theta_Frost (Sep 6, 2010 5:31 pm)