Sega Saturn Sophia SH-2 171-6692D board reproduction project

Hi All, I'm underway with reproducing the 171-6692D SH-2 boards that may be found in the following Sega Saturn development systems:

Sophia Programming Box
Sophia Graphics Box
Sega Saturn Address Checker (the big one)
Mars (possibly, uncertain as I've never actually seen inside one)

[GALLERY=media, 2845]FullSizeRender (13) by Shane Battye posted Sep 30, 2018 at 8:00 AM[/GALLERY]
[GALLERY=media, 2844]FullSizeRender (14) by Shane Battye posted Sep 30, 2018 at 8:00 AM[/GALLERY]​

I'm using a local PCB engineer with experience in board layout and assembly. The following are steps he's outlined which are intended to take place over the coming month or two (delayed due to production of the interface Hirose connectors - see below).

1. Depopulate sample PCB and prep for re-assembly (note initially we are attempting to buzz it out just after removing the SH-2 chip and then re-test after replacing with a 'new' SH-2 chip).
2. CAD
   • Set up mechanicals
   • Component data entry
   • SCH entry/check against sample PCB
   • PCB layout (layout for through hole connectors)/to SCH and check to sample PCB
   • Generate output docs
3. Engineer design review
4. Fabrication/assembly (I am supplying all components - see below). Initially intend on producing 10 boards as I'm unaware of the true demand of this part, personally I only need 3 and a fellow member has requested 2, which leaves 5 spares.

COMPONENTS
SH-2 chips
I've gone ahead and purchased 20 pcs from China; these match the part numbers for the original SH-2 chips used in the Saturn (HD6417095 S F28, F28 designating 28mhz).

These are at least form and fit equivalents but we will test functionality after replacing the original SH-2 chip on the donor board with one of these.

[GALLERY=media, 2842]IMG_9849 (1) by Shane Battye posted Sep 30, 2018 at 7:58 AM[/GALLERY]
[GALLERY=media, 2843]IMG_9847 (1) by Shane Battye posted Sep 30, 2018 at 7:58 AM[/GALLERY]​

80p connectors
This was more troublesome to source. The original part (HIF7-80PD-1.27SV) was discontinued some while ago. Hirose still advertise a functional equivalent HIF7CA-80PA-1.27DSA(71), however this has a 4 month lead-time on production (it's a non-stocked part, engineering diagram attached to this post as PDF).

Another issue being this is a through hole part where the original was surface mount, hence some redesign of the original board layout will be required (side by side shown in photo below, the new part is black).

[GALLERY=media, 2846]IMG_0233 by Shane Battye posted Sep 30, 2018 at 8:22 AM[/GALLERY]
​

4p connector
This is on the top-side of the board and connects to the power and ground supply in case of the Sophia. These are made by JAE and apparently still available.

[GALLERY=media, 2847]IMG_0883 by Shane Battye posted Sep 30, 2018 at 8:24 AM[/GALLERY]
​

Remaining parts
The extra top-side connector is not normally used in the Sophia and I have not attempted to source these yet, they may be added later. All residual components (resistors etc) are generally easily obtainable.

So that's where I'm up to. Depopulating the donor board should commence in the next week or two and I'll keep updates.

If you require an SH-2 board for one of your systems or just because you're curious then shoot me a DM.

Oh and just to summarise the reason behind this project; the Sophia Programming Box will play any Saturn game if it contains 2x SH-2 boards. The problem is they shipped with only one (allowing the developer to either purchase a POD, EVA board or SH-2 board for the empty slot). Over time, some systems had their single SH-2 board taken to use in another box, leaving Sophia’s either with no SH-2s or possibly just the original one that it shipped with. These Sophia’s often act as expensive paperweights and so the aim is to bring them back to life; if they aren’t used for development, they can play some games and demos.

Regards,
Shane

 

Awesome project .

My Sophia doesn’t have any so I’ll take 2

 

Update; one of the ten boards has been requested by a fellow member. There are now four remaining unallocated boards from the original planned production of ten.

 

I am very happy to hear this news, my Sophia also missing this parts, I need one, please! When you ready please notice me. Thank you!

 

Update; JAE PCB Header IL-G, 4 Way, 2.5mm Pitch, 1 Row, Straight, Through Hole have arrived and match original header at CN4

[GALLERY=media, 2865]0727EB0E-89A9-4324-AD82-863D92C1A804 by Shane Battye posted Oct 11, 2018 at 8:12 AM[/GALLERY]

[GALLERY=media, 2864]159AFB6A-6303-4FB4-8981-768EE484FC65 by Shane Battye posted Oct 11, 2018 at 8:12 AM[/GALLERY]

 

Project update:

• SCH entry completed
• PCB layout progressing approx. 80% routing completed.

Cleanup, generate outputs and final review to follow.

When the final review is completed, one of the new SH-2 processors will be put back on the donor board PCB to check compatibility.

Once the final layout is confirmed we can move to the next phase and arrange for PCBs to be fabricated and a sample produced.

Note there are 3 unallocated boards still available from the original planned batch of 10 units; DM me if you are interested in purchasing one.

Regards,
Shane

 

Project update (a big one):

I can happily confirm an off the shelf SH-2 (HD6417095 S F28) functions identically to an ‘OEM’ that originally shipped on the Sophia PCBs or Saturn; no internal security lockout etc.

Just tested on the Sophia donor board (that had the original SH-2 stripped for tracing the PCB layout); this is one of the recently purchased new/old stock SH-2 chips I acquired from a supplier in China (there’s literally hundreds of suppliers still with these chips).

Note this is the HD6417095 - which differs from HD6417604 (another SH7604 also produced by Hitachi) by at least pin 90 (not connected in the latter).

I ensured as much as possible compatibility by cross matching with the labelling on the original SH-2 on this particular board, note the S suffix and F28 (28mhz specification); I’m unsure the meaning of XY0707 but was able to match that with this order of new parts too (this doesn’t seem significant as other SH-2 boards of mine have different numbers here such as XY0011).

So, this is a big relief and the first hurdle out of the way before we could commence fabricating the new Sophia SH-2 boards.

Cheers, Shane

 

Project update:

MISSION ACCOMPLISHED

I have just received all 10x initial production run of the reproduction SH-2 boards. They are all reportedly tested and working although I will personally run additional tests over the coming days to confirm.

I’m really happy with the finished result. From the outset my intention was not to ‘duplicate’ the appearance of the original board as such but aim for functionality - to bring back to life Sophia systems or other development machines that were missing this component. Given the discontinuation of the original Hirose connectors, exact visual appearance replication was always going to be impossible.

The blue PCB was a conscious decision and looks great IMO. I will at some point start a new thread specifically detailing the design considerations for this board. Also if you think there is any further demand for more boards shoot me a DM or find me on Twitter to discuss.

If you have purchased a board/boards from me please confirm your shipping address via DM. I expect to start shipping over the next week or two. If you wish to purchase a board there are two unallocated and available from this run.

Regards,
Shane

 

wow , nice !

 

Very cool!

 

It works perfectly !!! Thank you very very much.
It s a japan sophia. And i tested daytona usa it works

 

This is fantastic! Thanks so much for sharing. I’m so glad we can now bring these Sophias back to life

 

silly question could you have used fgpa in stead of sh2?

 

Yep could’ve. The SH-2 is now open source and there was a project a couple years back called j-core that was compatible with it. However for this board given the original SH-2s are still in abundant supply in China (just search in chip supplier for the Saturn’s HD6417095S F28), it made sense to use the original type.