Project: HDMI/DVI out for STFM

Progress on our FPGA cores.
User avatar
exxos
Site Admin
Site Admin
Posts: 10925
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
YouTube: https://www.youtube.com/user/exxosuk
Contact:

Re: Project: HDMI/DVI out for STFM

Post by exxos » Sat Nov 03, 2018 1:32 pm

Smonson wrote:
Sat Nov 03, 2018 1:14 pm
Yep - 10x. Presumably the load is in the ST side of things. Or some effect that I'm not aware of. I will try it with a smaller series resistor when I get a chance.
It does look like to much resistance there.
https://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxoshost.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.

User avatar
Smonson
Trusted Guru
Trusted Guru
Posts: 491
Joined: Sat Oct 28, 2017 10:21 am
Location: Canberra, Australia
Contact:

Re: Project: HDMI/DVI out for STFM

Post by Smonson » Wed Nov 14, 2018 11:00 am

The new, smaller, shifter socket adapter board (fanout buffer version)... Unfortunately there's a slight overlap on the edge of an electrolytic on the ST that I failed to check, but I can always file it down a bit 😊

Other STs might not have that capacitor there anyway.
IMAG0011.jpg
IMAG0011.jpg (2.24 MiB) Viewed 1124 times

User avatar
exxos
Site Admin
Site Admin
Posts: 10925
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
YouTube: https://www.youtube.com/user/exxosuk
Contact:

Re: Project: HDMI/DVI out for STFM

Post by exxos » Wed Nov 14, 2018 11:25 am

:bravo:
https://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxoshost.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.

User avatar
Icky
Moderator
Moderator
Posts: 2232
Joined: Sun Sep 03, 2017 10:57 am
Location: UK
YouTube: https://tinyurl.com/yals4a3k
Contact:

Re: Project: HDMI/DVI out for STFM

Post by Icky » Wed Nov 14, 2018 8:26 pm

Nice.

Makes it more stable when in the shifter socket too without the overhang.

User avatar
Smonson
Trusted Guru
Trusted Guru
Posts: 491
Joined: Sat Oct 28, 2017 10:21 am
Location: Canberra, Australia
Contact:

Re: Project: HDMI/DVI out for STFM

Post by Smonson » Thu Nov 15, 2018 10:00 am

I just made the schmitt trigger version of the board and used a 0-ohm resistor for the (outgoing) clock - because I don't have anything less than 33ohms on hand - and I have recreated Exxos' symptoms! It's interesting to me because it looks pretty decent on the scope. I'll get a bunch of intermediate values in for more tests... but It's good to know that my ST isn't cursed with incessant reliability.
DS0000.gif
DS0000.gif (11.97 KiB) Viewed 1410 times

User avatar
exxos
Site Admin
Site Admin
Posts: 10925
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
YouTube: https://www.youtube.com/user/exxosuk
Contact:

Re: Project: HDMI/DVI out for STFM

Post by exxos » Thu Nov 15, 2018 10:09 am

Good you replicated my issue!

Problem with waveforms, they vary at random, you have to assume "what you see" can actually be 50% worse (as it usually is)
https://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxoshost.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.

User avatar
Smonson
Trusted Guru
Trusted Guru
Posts: 491
Joined: Sat Oct 28, 2017 10:21 am
Location: Canberra, Australia
Contact:

Re: Project: HDMI/DVI out for STFM

Post by Smonson » Tue Nov 20, 2018 9:14 am

So, the solution was the same capacitor you originally used to fix it, Exxos. I actually decided that the series resistors were working to prevent ringing based on my tests, so I left the capacitor off the final board design. :oops:

Fortunately it's easy to add a ceramic underneath the board. I used 20pF. This is on the 3.3v input side of the clock buffer, the 0-ohm resistor on the output side was not involved after all.

User avatar
exxos
Site Admin
Site Admin
Posts: 10925
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
YouTube: https://www.youtube.com/user/exxosuk
Contact:

Re: Project: HDMI/DVI out for STFM

Post by exxos » Tue Nov 20, 2018 12:10 pm

That "solution" wasn't a good fix.. it caused really bad ringing without a resistor..

Something must be over sensitive to the clock line. resistors will cure ringing and some noise, but on my STE booster, I had a 100R resistor in series with the schmitt buffer, but I also have 22pF on its input.. this is running around 40Mhz, for some reason it doesn't work without the 22pf. I think its just noise which causes glitches in the clock despite a schmitt buffer.. I did try without a schmitt buffer and I don't think I could get the STE to even boot.

That's the problem when you get into MHz ranges. slight capacitance or inductance in the wrong place and you can make or break a circuit. Think I did like 5 revisions of my STE booster before I got it all right :roll:
https://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxoshost.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.

User avatar
Icky
Moderator
Moderator
Posts: 2232
Joined: Sun Sep 03, 2017 10:57 am
Location: UK
YouTube: https://tinyurl.com/yals4a3k
Contact:

Re: Project: HDMI/DVI out for STFM

Post by Icky » Thu Nov 22, 2018 11:01 am

It's actually an interesting problem as I get constant resetting when using the board but have been working with Smonson this morning on running a test program that disables the vblank vector and reads the shifter's resolution registerer over and over and displays the values if they change.

After numerous resetting to get there I finally got the ST to run the program. I must say its the first time the ST has stayed solid for a while using the board for me. I had the program running for around an hour with no resets before turning it off.




This got me thinking as there must be programs that may not use vblank vector (I could be wrong) so I thought I would try out a few random programs from boot. Once I managed to get them through initial resetting that happens just after power on etc. I managed to get Goldrunner and Magic Pockets to load and they have been stable all morning and I have been quietly enjoying the clear graphics and playing the games :)

IMG_3301 (1).png
IMG_3301 (1).png (708.34 KiB) Viewed 1317 times
IMG_3302.png
IMG_3302.png (775.35 KiB) Viewed 1317 times

User avatar
Smonson
Trusted Guru
Trusted Guru
Posts: 491
Joined: Sat Oct 28, 2017 10:21 am
Location: Canberra, Australia
Contact:

Re: Project: HDMI/DVI out for STFM

Post by Smonson » Thu Nov 22, 2018 11:30 am

Awesome, it's great to see someone actually able to use this thing for its intended purpose for a change :)

Icky, you motivated me to write a quick status update here. So, after Troed's discovery that the resetting behaviour comes from the vblank vector, when TOS reads the shifter's resolution register once per frame and checks if it's changed, I started looking for possible reasons why the value could be wrong.

That led me to notice that the 1.5K bus series resistors (on the shifter socket adapter board) may be slowing down the rise and fall times of the bus signals - I tried replacing those resistors with 0-ohm links and that seemed to square up the signals quite a bit, so I asked Troed if he would do the same modification. Unfortunately, no improvement resulted. Nevertheless, those resistors will be reduced to 100 ohms for future boards now that I'm confident that there's no bus contention.

I wrote the test program that Icky mentioned earlier today because I wanted to find out what kind of "different value" TOS is seeing from the shifter's register - and as seen from the video, it's not just a stray bit flip, it's a totally wrong value on certain reads - up to 11 bits are wrong. That seems to point away from bad signal driving behaviour and strongly towards incorrect bus timing. So I'll be investigating in that area now.

What I intend to do is put the real shifter into a test circuit that will do some register reads and measure how long it drives the data bus for.

Post Reply

Return to “FPGA DEVELOPMENT”