Sega Master System / Mark III / Game Gear
SG-1000 / SC-3000 / SF-7000 / OMV
Home - Forums - Games - Scans - Maps - Cheats - Credits
Music - Videos - Development - Hacks - Translations - Homebrew
SC-3000 innersPosted: Sat Jun 01, 2013 3:52 pm
Last edited by TmEE on Sat Jun 01, 2013 4:13 pm; edited 1 time in total
Some photos of the inners of the main unit :
Inners of Sega BASIC level III A (PAL) :
PCB of Sega Flipper, with ROM removed :
I will trace out a schematic of the video modulator in not so far future. I'm hoping to get at least S-video output in an easy manner from it.
|Posted: Sat Jun 01, 2013 3:57 pm
|Great stuff. Could you please use the forum attachment system for the photos?
|Posted: Sat Jun 01, 2013 4:13 pm
|Posted: Sat Jun 01, 2013 11:26 pm
|Next step, add them to the dev section :) http://www.smspower.org/Development/HardwareImages-Index
|Posted: Sun Jun 02, 2013 1:06 pm
|I'm not really sure how to do that...
|Posted: Sun Jun 02, 2013 1:38 pm
|Like this: http://www.smspower.org/Development/EuropeanSC-3000834-5227
|Posted: Sun Jun 02, 2013 6:36 pm
Nice job. I haven't seen that metal shielding before on any of the New Zealand SC-3000s I have opened.
That *may* be how it was done originally. My very first SC-3000 keyboard was soldered to the motherboard. But I've also seen a rubber keyboard SC-3000 with sockets for the cable. So I suspect Sega may have used a mixture of techniques.
I must try that. All my SC-3000s suffer from colour fade where the reds gradually fade towards a brownish tinge over 20-60 minutes. If you leave the SC-3000 off for half an hour the reds tend to saturate again.
Before you create a schematic of the 4095 video daughterboard, have a look at this Sega original.
Ah... actually you probably meant the internals of the RF modulator. Yes, that would be interesting.
Take a look at the thread I've been participating in over here
Basically the TMS9929A outputs can be used to drive component video directly, but it is preferable to stick a buffer / amplifier in the way. We just haven't quite figured out how to get the correct balance yet. It is hard to get a nice bright yellow without losing the balance on the other colours. Note that the Yeno SC-3000s have a different video daughterboard and have RGB output which is awesome - I haven't been able to hit the picture quality of that daughterboard yet, although the component output is a lot cleaner than the standard composite. Far less colour bleeding.
The pictures of my current setup are on page 4 of the thread.
I'm using a THS7314 high speed video amp / buffer and 500 ohm trimpot resistors to provide the correct loading to the vdp and to adjust the voltage levels feeding into the THS7314. I think you may have done similar in the past if I recall correctly from another thread you wrote. A plug-in mod is easy for the SC-3000 because you just unplug the existing daughterboard (although you then have to decide how to feed the output out of the case which will involve a little surgery).
I've been a bit busy to do much with that recently, but I'm still working on it. I have an oscilloscope now so I'm just trying to figure out how the specs and voltage levels in the TMS9929A manual relate to the component video specs used by the TVs.
Sorry - didn't mean to hijack the thread :)
BTW - does your rubber keyboard work? I'm just trying out a couple of different repair techniques for those. One using a CircuitWorks CW2305 rubber keypad repair kit and one using a conductive silver pen. I just applied those last night so I'll give that a few more hours and see how well it works. Unfortunately those rubber keyboards were unreliable back in 1983 and after 30 years you may find the rubber pads have totally lost their conductivity.
|Posted: Sun Jun 02, 2013 8:59 pm
I mean the video daughterboard. It modulates color difference signals onta PAL subcarrier and creates the composhit output for AV out and RF modulator.
I saw your post with the schematic and it made my day. I can see the places where I can take Chroma signal for S-video output, I'll experiment with it in nearer future.
I am well aware there's the component output but I have no devices that take Component but I found a neat chip that takes YUV input and makes RGB out of it so that is what I'll eventually put into the machine.
I did not have any problems typing on the keyboard, all keys seemed to work nicely.
The keyboard wires seem to have been soldered on by whoever repaired the machine before me. I'll get sockets in coming days from my ex-work place (an electronics repairshop).
The unit I got is most likely from Finland. They were most definitely not sold in my country (Soviet occupation at that time) but they were sold in Finland and lot of stuff found their way into my country after the occupation. I am more than sure most of my MD and SMS stuff I found locally is from Finland originally.
...and thanks Tom !
|Posted: Sun Jun 16, 2013 9:49 pm
What is the chip id / model number so I can look up the datasheet? I'm still working on the component video mod, but I'm interested in looking at any alternatives (including the S-Video mod you just posted).
Great. The computer probably didn't have a lot of use then. I've just repaired mine with a Circuitworks CW2605 Rubber Keypad Repair kit.
It worked really well. The kit says it has enough to repair 50 keys, but I did two full SC-3000 keyboards with it and I probably had enough silver epoxy for a third keyboard. Just use a *very* thin coat on each key, and remember you only have a max of about 2-3 days to use the product once you mix part A and part B. I tried leaving it 24 hours after mixing before doing the second keyboard and the mixture was still usable, although definitely a bit firmer than the day before.
Note - I also tried doing one of the keys using a conductive silver pen. That seemed to work quite well too.
In any case, all the keys work brilliantly now. You just need a light touch to activate them.
For the rubber keyboard, you can just cut up a standard machined pin IC socket as the keyboard wires are quite stiff and strong. The hard keyboard ribbon connector is a bit trickier to work with.
Can anybody tell me what you call the type of socket in the photo below (labeled as CN7)? I've had trouble searching for a replacement because I don't know how to describe it. This is for the ribbon keyboard connector for the SC-3000H (with hard keyboard). It has a standard 2.54mm pin spacing, but it doesn't have any hole seperators because it is designed for a ribbon. And it is a single-sided socket too.
I can't really find anything like that when looking for SIP or DIP connectors, and any searches for RIBBON or EDGE connectors usually ends up with something that has very different dimensions.
|Posted: Mon Jun 17, 2013 5:04 pm
The chip is CXA1229
As for the connector I do not know how they are called. I have seen them in big music devices.
I ended up soldering in a single row pin header on the mainboard and matching IDC connectors from the keyboard side.
Posted: Mon Jun 17, 2013 6:19 pm
Last edited by honestbob on Mon Jun 17, 2013 6:46 pm; edited 2 times in total
Thanks for that. The datasheet makes interesting reading - looks like a very useful chip. I see there are two versions of the chip for different B-Y and R-Y colour balances (CXA1219 and CXA1229).
However I suspect that neither of those will be quite right for the TMS9929A. The standard problem that everyone building component output from a TMS9929A seems to run into is that the display is oversaturated with blue. If you try to correct that with a simple voltage divider, then although you can reduce the blue level of the output, errors are introduced and it is difficult to get your colour balance right. Either your yellows are washed out, or your white goes too blue / yellow regardless of how you scale the Y / B-Y / R-Y.
I *think* the solution there is to do proper sync seperation on the incoming Y channel then do DC restoration on all three channels using the sync blanking level as the reference voltage (because B-Y and R-Y don't have a reference voltage built into the signal) and THEN scale your B-Y and R-Y signals. If you scale the signals without doing the DC restoration yourself then you effectively shift the zero level detected by the TV on the B-Y and R-Y input, which is what gives an uncorrectable color balance. That's my best guess at the moment, anyway :)
I suspect that you may have a similar issue here with the RGB conversion in those CXA1219 and CXA1229 chips if the TMS9929A B-Y and R-Y output is not close to the correct ratios out of the box. ie. any attempt you make to scale the input to the CXA1229 (without sync seperation and DC restoration first) may result in uncorrectable errors in color balance.
In any case, the TMS9929A Y output is 1.4v p-p (Vwhite 3.34v, Vblack 2.36v, Vsync 1.92v), and the B-Y and R-Y are approx 1v p-p between Vblack and Vwhite), if that is helpful. (CXA1229 datasheet requires Y 1Vp-p, B-Y 1.27vp-p, R-Y 1vp-p)
I'll try to get one of each of those chips and try them out for interest.
If you do try that chip and it doesn't work as you hope for RGB color balance, then this application note is very useful for working with the TMS9929A
Of course, you may be happy with your current S-Video mod :)
|Posted: Mon Jun 17, 2013 6:43 pm
|All comes down to getting you levels and DC offsets right. There's nothing that cannot be overcame.