TMS9929 (Sega etc.) display direct to modern TV!

[honestbob]

Thanks Tezza

I finally found some time over the Summer break to tidy up the PCB design for the daughterboard and get them manufactured. AllPCB did an awesome job. They are inexpensive, fast, and the online ordering system is excellent. Lots of options for blinging your PCB design out. http://www.allpcb.com

This is my first surface mount project to learn a new skill, and I even had a stencil made up (solder paste tubs still in the post). So I'm trying out a few different soldering techniques for building the boards.

I started off with hand soldering late Friday night. That is bloody slow work with surface mount components, but my technique is improving. I couldn't find the bloody 100nF capacitors I'd ordered. On Saturday morning I figured out that was because they had actually sent me 220nF capacitors. That's probably fine for everything except the RC reset timer on the LM1881 which is used to mark out the colour burst etc. on the back porch (actually I can't remember what it is for, but a 200nF capacitor gives a RC time constant twice what a 100nF capacitor does, so probably not good).

Unfortunately when I turned it on there was a bang, lots of smoke, and the 22uF tantalum capacitor at the bottom left of the board had exploded and shorted out. It also took out the fuse on the SC-3000, and a couple of components on the PCB - one of the AD826 op-amps and the LM1881 sync seperator. Bugger. The incoming voltage is actually a 100Hz wave between about 9v and 11v from the DC power pack. That is what Sega passed to the original board. I'd hoped that would be within the 16v limit of the surface mount capacitor I used, but obviously not. Either that or I reversed the polarity, but I don't think so. It was a bit hard to tell after it went poof.

I cleaned that up, replaced with a 35v electrolytic cap, and started debugging the other broken bits on the PCB for the next few hours. Finally got it working correctly this morning after replacing both the LM1881 and the CD4066.

So - see lots of attached photos

You can see the test pattern and test blocks of text with different colours. The colours look better on-screen than that (the photos are a little washed out). Still, performing pretty much as designed so I'm happy.

Finally, I want to say a massive thank you to Natshaw for kicking this project off, and pointing me in the right direction about a year back with a link where someone else had solved the colour burst issue that was giving bad colour mixes. And to everyone else who has shown an interest. Kevman - pm me and I'll arrange one for you to try out.

I may make a couple of follow up posts with some oscilloscope images when I have time.

Cheers

All PCB Box containing lotsa goodies




A thing of beauty - the Sega SC-3000 Component Video daughterboard PCB, replacement for Sega 4095 PAL PCB




First assembly with the tantalum capacitor at bottom left




BOOM! That didn't work so well...



Fixed it - assembled and working. Had to replace capacitor, one AD826, the LM1881 and CD4066 (and SC-3000 fuse)




Test Patterns! The photos are a bit washed out - it looks better in real life

[ti99user]

Thanks Tezza

I finally found some time over the Summer break to tidy up the PCB design for the daughterboard and get them manufactured. AllPCB did an awesome job. They are inexpensive, fast, and the online ordering system is excellent. Lots of options for blinging your PCB design out. http://www.allpcb.com

Great, can you please provide schematic and PCB, if possible also in source format? Which CAD program used?

Thnx.

[Carcenomy]

Just want to throw it out there... these days there's a bunch of reasonably affordable scandoublers and things hitting the streets... although most can work with component as well, is there a possibility of pulling straight RGB out of the SC3000 as opposed to YPbPr?

Ed: would help if I read more of the thread wouldn't it... the Yenos have native RGB?

[jbourke]

Hi All,

I have a Ti/99-4a (TMS9918a NTSC) connected to a composite -> HDMI converter -> LCD monitor; the rather fuzzy image was initially great from a nostalgia point of view but better quality would be great!

A TMS9928A is underway and I wonder if I then could use one of these:

https://www.amazon.com/Converter-Compon ... B07GPPJJ6W

Just an example, there are lots of similar converters available. If it works it will save the hassle of building one (too many projects on already).

What do you guys think?

Cheers,

John

[ti99user]

I have a Ti/99-4a (TMS9918a NTSC) connected to a composite -> HDMI converter -> LCD monitor; the rather fuzzy image was initially great from a nostalgia point of view but better quality would be great!

A TMS9928A is underway and I wonder if I then could use one of these:

https://www.amazon.com/Converter-Compon ... B07GPPJJ6W

The TMS9918a NTSC version works directly. For the TMS9928A you need some correction of the signals, see for details:
https://hackaday.io/project/13056-tms99 ... nt-adapter

I've made an enhanced version (had no time to build it yet but should be ok):
https://github.com/wiesl/ti99-rgb

ti99user

[Falonn]

Hi everyone. I know I'm a couple years late, but I wanted to share a few results from a very similar project that we just undertook over at the AtariAge forums over the last six months. Unfortunately I didn't find this topic until after the project was complete or it would have saved a lot of effort! You guys did a great job over here. Most of our lessons learned were exactly the same as honestbob's findings, though hopefully I can add one or two new points.

The project started as a kind of survey of all the known approaches to get the 9928A/9929A to output RGB. We knew about the old "Interface to Color Monitors" document from TI. We'd also seen the Hackaday circuit, which ended up being the most helpful/influential. I ended up building five(!) different adapters on breadboards and the final result used at least one idea from each one.

The first discovery was that the TI datasheet for these VDPs has a major error: the 9928A does have a color burst on (only) the B-Y line. My guess is that it's an unintentional bug in the silicon. (Perhaps after removing the red color burst, they thought their job was done and forgot to also remove it from the blue side?) So the NTSC version of the chip has no choice but to handle color bursts the same as the PAL version. And with only one color burst, the "burst averaging" solution proposed a couple pages ago here wouldn't be possible. So, the sample-and-hold idea from the Hackaday circuit became the best bet.

Like honestbob, we also saw lots of noise from the VDP. The filter on the TMS7374 cleaned most of it up. There are still some faint jailbars in solid black areas though, which can usually be cleaned up by tweaking settings on the receiver (OSSC, Framemeister, etc.).

One nice detail we were able to borrow from a project called ColecoRGB was the idea of packing the whole adapter/mod PCB into the same DIP-40 footprint as the VDP so that it can be soldered right onto the underside of the motherboard. This makes it compatible with anything that uses a 9928A/9929A. Granted, it requires some pretty small surface mount components to get everything to fit.

The whole "TMS-RGB" project is completely open. Schematics and Eagle files are on its GitHub page and I recorded a full set of build/install videos on the official site.

Hopefully that's useful information to someone. It'll be nice to have it in the same place as the rest of this discussion for people that might spot this in the future. In any event, nice work again honestbob!