Amiga 600 schematics nearly put into Kicad

[SukkoPera]

I think I've lost track a bit, could you please do a short recap of how the current working board compares to a common rev. 1.5, for instance? What did you do 1:1 and what did you change?

[Mick]

Originally Posted by Stedy

Newer video encoders do a much better job of making composite video, hopefully the CXA2075, which does not need the delay line or bandpass filter gives higher quality video. The composite video on the Amiga is quite bad.

Yeah but like others have said if you do a 1:1 board you can always just mod in the newer DAC's (such as CXA1645 also) whilst keeping everything original for those who just want to replace a damaged board. I'd rather have done that if I'm honest but my main goal late last year was just to try to get it working and then I won't feel like I've dangled a carrot in front of everybody for 2 years and then let if drift away from them.

Originally Posted by Stedy

Mick, I'll check those capacitors in the next day or two and confirm the results. I could design a delay line for the video circuit quite easily, for the bandpass filter, I need to know how wide the pass band is. Could not find a datasheet for the Sony recommended part.

Thanks. I found a Toko datasheet for more or less (I believe) the same part which was just a newer variant which had these specs:

3.58MHz BPF (NTSC)
Bandwidth or cutoff freq. = 2.8/4.3MHz (-3dB)
Input / Output impedance = 1kohm

4.43MHz BPF (PAL)
Bandwidth or cutoff freq. = 3.73/5.13MHz (-3dB)
Input / Output impedance = 1kohm

There are tutorials and calculators online for the BPF at least, that's how (I think) I would have been able to do it, no such luck with the delay line though.

I was thinking of adding a 5V pin and using an active delay line but it would be better for original board owners if a passive delay line replacement can be made.

Originally Posted by SukkoPera

I think I've lost track a bit, could you please do a short recap of how the current working board compares to a common rev. 1.5, for instance? What did you do 1:1 and what did you change?

It's more like a rev2B but with U49 added and the video area modified to take a CXA2075, so schematically it's a 2D. I've looked at a 2D and it seems to be just a modified 2B with bigger SOIC footprints crammed in and I'm wondering if C26 is connected to VCC at all on that revision and if so how.

[Mick]

If stedy or someone can design replacements for BPF and DL I will do a 1:1. Gerbers for Svideo are already in the zone.

[Mick]

I did some more digging and found what seems to be the original datasheet. Sony listed the NTSC part for PAL in their CXA1145 datasheet.

Filters for RGB Encoder

H287BSJS-3108HWD
Application = 3.58MHz BPF (NTSC)
Impedence = 1kohm

H287BSJS-4438HWD
Application = 4.43MHz BPF (PAL)
Impedence = 1kohm

https://www.datasheetarchive.com/pdf...fe4723746d6eec

There's also some typical characteristics graphs (fig. 1 & 2).

[Lemaru]

Originally Posted by seb132

In Europe 5x A1200 boards (green) cost 90€ shipped... so I’d say a bit less for A600?

** Edit: What mods would you do?

Nothing probably too major. A few things I would ideally like...

- Better mounting points for any add-ons (or larger holes rather than relying on the small HDD cradle mounting holes)
- DIP64 CPU instead of the PLCC (maybe limited by space but would make it much more expandable).
- Breakout header for the RGB and clock pins from Denise to make connecting the RGBtoHDMI Pi Zero mod easier to fit

Few other thoughts which although easy to add manually to any board, would be nice.

- 2MB ChipRAM (and maybe RTC)
- Clock port
- A header for a switch to swap DF0 & DF1 (so you can use external drive as DF0)

Sure others could add a few more ideas

[SukkoPera]

Originally Posted by Mick

It's more like a rev2B but with U49 added and the video area modified to take a CXA2075, so schematically it's a 2D. I've looked at a 2D and it seems to be just a modified 2B with bigger SOIC footprints crammed in and I'm wondering if C26 is connected to VCC at all on that revision and if so how.

OK, but U49 is unpopulated on my board.

The "video area modified to take a CXA2075" means you added a proper place for the 2.61k resistor? I don't understand why you don't want to add the original circuit for the 1145 as well and just leave it as untested. Someone will test it out sooner or later and report. It's simple enough and does not prevent using a CXA2075, so you could just have both, but I respect your choice. Actually I prefer the 2075 too, my suggestion was just to make a "more generic" board.

I have checked C26 and actually the positive terminal goes to a via that is not connected to anything. It's probably a routing oversight that you could fix.

I can't measure the exact values of C7, C8, C29, C36 as I would have to desolder them to do so. Since they are decoupling caps, their value is not too critical. I'd say just go with 100nF, but anything up to ~330nF could be fine. 2A schematics report 47nF and that sounds good as well.

I can't find any C9 on my board.

C28A, C28B seem to be 10nF.

About the large footprints: I think most SOIC footprints on this board suit both narrow and large SOIC, they have two sets of pads on one side and I'm pretty sure I've seen this on 1.5 boards as well.

Hope this helps.

[Stedy]

Originally Posted by Mick

I did some more digging and found what seems to be the original datasheet. Sony listed the NTSC part for PAL in their CXA1145 datasheet.

Filters for RGB Encoder

H287BSJS-3108HWD
Application = 3.58MHz BPF (NTSC)
Impedence = 1kohm

H287BSJS-4438HWD
Application = 4.43MHz BPF (PAL)
Impedence = 1kohm

https://www.datasheetarchive.com/pdf...fe4723746d6eec

There's also some typical characteristics graphs (fig. 1 & 2).

Thanks, that came in handy. I designed a Bandpass filter today, something to do in my lunch break. I ran it through LTSpice, it appeared to work Ok. The delay line was also designed but the simulator showed an issue, I might need to add in better inductor model or I just got it wrong!

I'll check some of the other capacitors on my rev 2D board this weekend.

[seb132]

@Lemaru: nice. Every new board of any model should now have a connector to pick Denise signal and allow for an adapter PCB to connect, yeah.
All other mods are good stuff but will interest hardware designers only if more than 1 person have them on its board (fixing holes, PLA CPU, etc.
But good stuff!

[Mick]

Originally Posted by SukkoPera

OK, but U49 is unpopulated on my board.

Yeah it is unpopulated on all 2D's but Commodore added it for some reason, it's the only difference between 2B and 2D electrically, most of the difference is in appearance/layout which is why I describe mine as more of a 2B as I used images of a 2B.

Originally Posted by SukkoPera

The "video area modified to take a CXA2075" means you added a proper place for the 2.61k resistor?

Yes and changed some other things slightly to better match what the CXA2075 datasheet calls for.

Originally Posted by SukkoPera

I have checked C26 and actually the positive terminal goes to a via that is not connected to anything. It's probably a routing oversight that you could fix.

I thought so, no need to fix it is connected. I was just looking into changing footprints to match 2D but it seemed impossible to connect the way they did it, 2D looks a bit of a dogs dinner tbh.

Originally Posted by SukkoPera

I can't measure the exact values of C7, C8, C29, C36 as I would have to desolder them to do so. Since they are decoupling caps, their value is not too critical. I'd say just go with 100nF, but anything up to ~330nF could be fine. 2A schematics report 47nF and that sounds good as well.

I can't find any C9 on my board.

C28A, C28B seem to be 10nF.

They should all be 0.01uF according to the Commodore service manual, I just wanted to check because Amigawiki list 0.047uF instead for some reason, I was wondering if it was a last minute change before boards went into production or something. If C28A/B are 0.01uF then I guess Amigawiki just got a bit creative.

Originally Posted by SukkoPera

About the large footprints: I think most SOIC footprints on this board suit both narrow and large SOIC, they have two sets of pads on one side and I'm pretty sure I've seen this on 1.5 boards as well.

Yeah but I meant the main difference between 2B and 2D is that the pads on 2B are about 2mm long, whereas on the 2D they're about 2.5mm. If you look at pin 14 of U15 on a 2D the pad is still the original size because they couldn't route it otherwise. That's part of the reason why I said it looks a bit of a dogs dinner. I think with 2D they just took a 2B, made the pads slightly longer and then fixed any routing they broke but there's also some other weirdness like the floating via you mention and a bunch of extra random via's under Denise as well as U33. I wondered if they were just experimenting with planes or something but I guess not if it's disconnected.

Thanks for the help.

[Mick]

Originally Posted by Stedy

Thanks, that came in handy. I designed a Bandpass filter today, something to do in my lunch break. I ran it through LTSpice, it appeared to work Ok. The delay line was also designed but the simulator showed an issue, I might need to add in better inductor model or I just got it wrong!

I'll check some of the other capacitors on my rev 2D board this weekend.

Nice one.

Regarding capacitors if you can confirm they're 0.01uF I'll be happy.

[Mick]

Originally Posted by SukkoPera

Actually I prefer the 2075 too, my suggestion was just to make a "more generic" board.

I'd like to just do a 1:1 now, I'm sure other variations will follow, you might even make your own. It might seem unrealistic now but I hope that in the long run, with PCB's available, development of replacement chips (whether real or FPGA) and other parts (23-pin d-subs etc) will eventually follow. One hurdle at a time.

[Stedy]

Hi,

C7, C8, C28 were all 10nF. Have not had a chance to trial the delay line or bandpass filter design with real hardware.

Mick, you've done a 1:1 board, that's one of the boards I've been testing, as it has the original RF modulator pinout and no S-video.

[Mick]

Cheers.

It didn't work though and it was my first attempt when I was using mm rather than mil and I autorouted it, not to mention using a ruler and printout instead of GIMP. I've learned loads since then. Looking back I'm a bit embarrassed I sent it to Acill but that's where I was in the learning process, I thought it was great at the time. I can send you the current files with SVideo if you want them but I want to do a really good 1:1 even if just for myself if/when the BPF/DL are available. The schematics will also be relevent to everyone that way.

I think my main regret with the first board is not sticking with the manually routed power and using planes, that's probably what is preventing it from working.

[Stedy]

Mick,

Send me the files for the board you have working, I'll upload them to github and I'll start the support Wiki, I have a few entries to add. From here, other users can take the designs, create forks and make their own versions, if they wish (and have the funds).

The original board, sent to Acill, then myself is still WIP. I know what stops it working but not why. Website work got in the way this weekend, I plan to get back to it soon.

[Mick]

Okay, will you still make BPF/DL though please?

What is it that stops the original board from working?

[Stedy]

Originally Posted by Mick

Okay, will you still make BPF/DL though please?

What is it that stops the original board from working?

I plan to try the BPF at the weekend. Not sure if I could make them retrofit existing PCBs, will see how small I can make it.

The original board is not working due to U26/U27 not passing the data through to the custom chips, which includes the CIAs/serial port. Not sure why yet, but at least I know where to look.

The Github repository is up, I have the original, not-tested version and the S_Video branch, https://github.com/istedman/A600Reborn. I'll add some notes on build/testing to the wiki soon.

[Lemaru]

Great work Mick. The photos on the GitHub look awesome. I think you should add some silkscreen with your name on it, as it was you who done the hard work.

How did you go about manually routing this? Did you trace an image or spend ages measuring everything?

[Mick]

Originally Posted by Lemaru

How did you go about manually routing this? Did you trace an image or spend ages measuring everything?

For placement I measured the same as the first one but I used mils instead of mm which was key to being able to manually route it in a really orderly way, I also went from 1:1 print outs and a ruler to using 1:1 images in Gimp and the measuring tool as it's a lot more accurate that way rather than squinting at a ruler, just find a reference point and measure a location (footprint/via) to the nearest 10mil. The routing is a bit different I didn't measure the length of traces or anything but it was a bit like doing a jigsaw once you have everything in place.

[katarakt]

Yes, amazing work from Mick. Is the Svideo variant fully working? I thought it was mentioned before that there is no serial output or am i wrong?

[Mick]

It should be fully working, everything is connected where it needs to be. I'm not sure where you've seen that about serial.

If I were stedy I'd just delete all of the old stuff to prevent confusion, all of the progression is here in this thread.