Model 1 MD PCB hacks topic

For hardware talk only (please avoid ROM dumper stuff)
powerofrecall
Very interested
Posts: 237
Joined: Fri Apr 17, 2009 7:35 pm
Location: USA

Model 1 MD PCB hacks topic

Post by powerofrecall » Wed Oct 29, 2014 3:26 am

I was thinking about this earlier and thought it really deserved its own topic. I've never seen information about this all contained in one place, and I figured people on this board would have a lot of practical background to contribute so why not.

We all know there's a few official board revisions, some pretty rare and some more common. Not including things like the VA7 model 1 board, or the consolidation of custom chips over time, what are some of the more common PCB 'hacks' (jumper wires, additional capacitors/resistors soldered in weird places, etc) that were common on model 1 boards?

The ones I have seen in person are the VA6 jumper wires, the resistor soldered to the /AS leg of the 68k, a disc capacitor soldered down next to the crystal (didn't notice at first but there were no board through-holes for it and it was attached to a via), and some others including another 68k that had a disc capacitor soldered to one of the legs and to what I guess is a ground plane under the CPU. I have a VA3 board that has a capacitor soldered to the bottom near the VRAM, and one unpopulated capacitor space on top.

I know there others, like the little daughterboard on Japanese launch day boards and probably ones I'm not even aware of.

I'm curious why some of these were obviously patched in at a later stage in production. I understand that PCB design doesn't always come out perfect, especially when you're ramping up design of a popular game console.

The one I'm personally most interested in is the VA3 jumper wires hooked to the RAM & to another pad near the VDP, as they obviously anticipated this one--why else would they print on jumper pads instead of just printing the 'fix' on the board? My guess is it has something to do with available RAM parts and their speed as 2 of the wires go to the RAM's /CE but I'd love to know as I'm not super knowledgeable in such things. Especially since I have a board I added some 'new' (well, NOS) RAM to that isn't working correctly right now. :oops:

http://i.imgur.com/GvR7hvH.jpg

The main reason I think this topic is worth having around is that it makes for good service notes as well--later fixes made on later boards might be beneficial to improving older revisions as well. Particularly, as much as I love the model 1 Genesis, it sometimes seems like a really finicky, somewhat shoddy board. It might be component age or EM interference or some other factor, but MD2's seem more 'solid.' (I can't leave a mobile phone anywhere near a model 1 if it is in use, for instance.) Of course, some of this might be having more bugs worked out or more modern fabrication processes.

Hell, if I get bored enough tomorrow I might try 'undoing' the jumper wire fix on one of my other va6 boards just to see what happens. :D

l_oliveira
Very interested
Posts: 53
Joined: Mon Mar 07, 2011 12:58 am

Post by l_oliveira » Sat Mar 14, 2015 2:51 pm

I know this post is old... But you deserve a reply.

The 220 ohm resistor on /AS at the 68000 is put when the console is using a FC1001 VDP (315-5313A) with a Motorola CPU. When the CPU is Hitachi CMOS HD68HC000, like this one:
Image
it is omitted.

The daughterboard on VA1 seems to be something to control the speed of CPU access to VDP registers. Like it keep watching HSYNC pulses and stall the CPU if accesses are flooded. (not 100% sure about that) But for sure it was added after the design was complete.

On VA3 that design was completely absorbed into the bus arbiter IC.

MEGA-CD has three known ASIC versions. MCE1, MCE2 and MCE3.

MCE1 based MEGA-CD units use a patch board on clock oscillator with a 74LS74 IC in it and it has a 48 pin IC on the breaker board which connects the CD unit to the console:
Image

The black/gray wire pair send the clock to the chip on the breaker board.

Charles MacDonald
Very interested
Posts: 292
Joined: Sat Apr 21, 2007 1:14 am

Post by Charles MacDonald » Sat Mar 14, 2015 5:54 pm

All the Genesis consoles I've seen have a few wires going to the two work RAM chips. Any reason why that is?

It seems intentional because even as the motherboards revisions change those wires remain. It doesn't seem to be a routing issue that required a few flying leads to fix.

Given that the wires required some manual assembly you'd think it would have been more cost efficient in terms of parts and production to remove them. So there must be a reason for them?

l_oliveira
Very interested
Posts: 53
Joined: Mon Mar 07, 2011 12:58 am

Post by l_oliveira » Sat Mar 14, 2015 6:33 pm

I made these pictures from my VA6 Japanese MD1:

Image
The green wire links VDP pin 119 (/RAS0) to the /CS pin on both of the PSRAM chips working as the CPU work ram.

Yellow and blue on the other side are there for the UDS/LDS feature of the 68000. I believe the gate array+bus arbiter (on this case a 315-5433 IC, an NEC custom made for SEGA) has logic in it to deliver refresh pulses through these wires while the PSRAM is not in use by the 68000. (on this particular PSRAM model the refresh signal is delivered on /OE pin while the device has /CE high)


Image
As you can see, it has a plain MC68000 CPU and YM7101 VDP (1st version 315-5313) so there's no patches under the MC68000 chip.

Now, the only thing I could think of that would make them use these wire links is signaling problems.
Funny that VA3 has no wires like that, while using the exact same PSRAM chips.
Image
Picture shamelessly "stolen" from google (that guy is a SEGA-16 member I believe)

Jorge Nuno
Very interested
Posts: 374
Joined: Mon Jun 11, 2007 3:09 am
Location: Azeitão, PT

Post by Jorge Nuno » Fri Mar 20, 2015 4:24 am

l_oliveira, your board looks like a VA5 to me, even though it appears to have a 315-5433.. A VA6 has slightly different copper.

I noticed in the CPU swap/mods that I did that /AS has a very high undershoot, the 220 Ohm resistor definitely helps smoothing that out.

The PSRAM wires coming from the IO chip are useless on the VA6, but (AFAIK) needed on the VA5. BTW the VA6.8 still has these wires. All 3 require the long wire coming from the VDP, looks like they couldn't route it.
Either way, those wires are the 68k ram OE signals (/NOE /EOE) which also deal with refresh, yes.

The lil board on the original MD release is to make EDClk from the 53MHz and HSync, later it migrates to a little PAL chip (which also deals with a bit of SMS/IA14) and gets integrated into the 315-5364.

On the other hand look how clean is an early MD1, a VA2:

Image

l_oliveira
Very interested
Posts: 53
Joined: Mon Mar 07, 2011 12:58 am

Post by l_oliveira » Fri Mar 20, 2015 1:54 pm

MD1 VA5 and VA6 have the exact same PCB.

What make a VA5 (exactly as it's written on the PCB) is the I/O + bus arbiter chip be 315-5402. The slightly different copper you mention is a Asian model which has built in RF modulator. That makes it look very close to the "EXPORT" (aka European/USA) models.

Contrary to what you say, removing the wires prevent the console from working/booting.

And the only difference from 315-5402 to 315-5433 is that on 315-5402 the TMSS is not enforced when region jumper is set at "DOMESTIC". Meaning a Japanese console with 315-5402 I/O + bus arbiter chip does not display TMSS.

I don't think VA5s actually exist outside Japan. I did own one many years ago. I don't know what happened to it though.

Jorge Nuno
Very interested
Posts: 374
Joined: Mon Jun 11, 2007 3:09 am
Location: Azeitão, PT

Post by Jorge Nuno » Fri Mar 20, 2015 3:51 pm

I challenge you (or anyone) to remove them on a USA/EU VA6 (the EOE/NOE wires)
Last edited by Jorge Nuno on Fri Mar 20, 2015 3:56 pm, edited 1 time in total.

l_oliveira
Very interested
Posts: 53
Joined: Mon Mar 07, 2011 12:58 am

Post by l_oliveira » Fri Mar 20, 2015 3:54 pm

Jorge Nuno wrote:I challenge you to remove them on a USA/EU VA6 (the EOE/NOE wires)
That's interesting. But why keep the wires anyway ?

Jorge Nuno
Very interested
Posts: 374
Joined: Mon Jun 11, 2007 3:09 am
Location: Azeitão, PT

Post by Jorge Nuno » Fri Mar 20, 2015 4:28 pm

Maybe they forgot about them? I mean the VA6.8 *still* has them and the markings of 5402:VA5 | 5433: VA6, except it says V6.8 on IC BD M5.. text

l_oliveira
Very interested
Posts: 53
Joined: Mon Mar 07, 2011 12:58 am

Post by l_oliveira » Fri Mar 20, 2015 4:35 pm

Jorge Nuno wrote:Maybe they forgot about them? I mean the VA6.8 *still* has them and the markings of 5402:VA5 | 5433: VA6, except it says V6.8 on IC BD M5.. text
My Japanese MD1 says "IC BD M5" on the board, no VAxx.

The VAxx is near the I/O + bus arbiter chip.

Here's a picture of someone else's VA5:

Image

Jorge Nuno
Very interested
Posts: 374
Joined: Mon Jun 11, 2007 3:09 am
Location: Azeitão, PT

Post by Jorge Nuno » Fri Mar 20, 2015 4:44 pm

VA6.8:

Image

l_oliveira
Very interested
Posts: 53
Joined: Mon Mar 07, 2011 12:58 am

Post by l_oliveira » Fri Mar 20, 2015 4:48 pm

So that VA6.8 was released when they were almost ready to switch into the chipset with VDP+Bus Arbiter+I/O chips on the 208 pin SMD thing ?

I did not know about that type. PAL only perhaps ?

Jorge Nuno
Very interested
Posts: 374
Joined: Mon Jun 11, 2007 3:09 am
Location: Azeitão, PT

Post by Jorge Nuno » Fri Mar 20, 2015 4:52 pm

It's quite rare yes, I had one and yep it was PAL. No idea if it ever came in NTSC

l_oliveira
Very interested
Posts: 53
Joined: Mon Mar 07, 2011 12:58 am

Post by l_oliveira » Fri Mar 20, 2015 5:01 pm

Jorge Nuno wrote:It's quite rare yes, I had one and yep it was PAL. No idea if it ever came in NTSC
I've never seen it. And I've seen units from all over the world. Of course PAL stuff is rare here. I bet Asian stuff is quite scarce in Portugal. Is that true ?

Jorge Nuno
Very interested
Posts: 374
Joined: Mon Jun 11, 2007 3:09 am
Location: Azeitão, PT

Post by Jorge Nuno » Fri Mar 20, 2015 5:04 pm

Somewhat. There was something weird back then in the old times where we got asian games with megakeys, and I think asian MD systems too, but I don't remeber details :|

Post Reply