The F18a DIY alternative with HDMI :)

[retrocanada76]

18 minutes ago, craftsman1234 said:

Ok I guess it is a done deal just not compatible with the Tang/TN-VDP.  Hopefully someday someone else with a Coleco will be able to test it.  

well you know fpga right ? try copying the same cpuclk/gromclk generation from a 100mhz clock as the f18a original does. I am generating a 3.5mhz signal from a PLL. Maybe that's what is different. 

[craftsman1234]

4 minutes ago, retrocanada76 said:

well you know fpga right ? try copying the same cpuclk/gromclk generation from a 100mhz clock as the f18a original does. I am generating a 3.5mhz signal from a PLL. Maybe that's what is different. 

Unfortunately no,   I can build the boards do soldering etc but no experience with any FPGA.  This project was my first use of one.   If my 4yoa would let me I might get to learn it someday.

 

Sorry if that came across negative..  Just been testing like crazy with all the versions.   Still working great with my Nabu's and still an amazing project.

 

 

[craftsman1234]

On 6/13/2023 at 9:52 AM, retrocanada76 said:

well you know fpga right ? try copying the same cpuclk/gromclk generation from a 100mhz clock as the f18a original does. I am generating a 3.5mhz signal from a PLL. Maybe that's what is different. 

Just a question for everyone out there.   Is there anyone that could do what Retrocanada76 is suggesting and send it to me for testing.   Programing a FPGA is pretty much out of reach for me right now.   I would love to see this running fulling on a Colceovision.  

[wierd_w]

Sounds like you would like to just buy a finished board to me.

 

If these work well inside TI units, Arcade Shopper might start stocking them. F18As are a hot commodity.

[retrocanada76]

On 6/15/2023 at 8:46 AM, craftsman1234 said:

Just a question for everyone out there.   Is there anyone that could do what Retrocanada76 is suggesting and send it to me for testing.   Programing a FPGA is pretty much out of reach for me right now.   I would love to see this running fulling on a Colceovision.  

try plugging a USB-C cable to power up the tang. I don't know it looks to me some electrical issue. Do you just have one unit ? It could the voltage level on the interrupt line. This would be the only signal to make the CPU jump out of it's location. Maybe it needs a capacitor there who knows ?

[djones60]

Well I finally got around to finishing mine after several interruptions.  The Nabu sees it, displays things nicely.  Booted up network CP/M.  But then it does something odd at the command prompt.  It displays several ^D and displays those on every key press.  If I do the same thing with the original video chip, everything works fine.  I am sure I've probably done something dumb.  So any ideas where to look for the problem or what to try?

 

Thanks

    David

 

[craftsman1234]

4 hours ago, retrocanada76 said:

try plugging a USB-C cable to power up the tang. I don't know it looks to me some electrical issue. Do you just have one unit ? It could the voltage level on the interrupt line. This would be the only signal to make the CPU jump out of it's location. Maybe it needs a capacitor there who knows ?

I have built 3 of them so far will build 2 more.   I have been testing on the Coleco with two different PCBs and two different Tangs.   Same issue on both.

 

I plugged the Tang to a separate USB power supply and still the same issue.  

 

As to the Capacitor.   I could solder one in.  Do you mean pin 16 on the TMS.    Would that be from 16 to ground?  I'm using the 1 version of the board with the diode hack.

[djones60]

Well I figured out my problem.  Turns out I did not have USR3 - on.  Knew it was probably something dumb I was over looking   Now time to see about making one for my TI-99.

Edited June 18, 2023 by djones60

[retrocanada76]

On 6/16/2023 at 8:07 PM, craftsman1234 said:

I have built 3 of them so far will build 2 more.   I have been testing on the Coleco with two different PCBs and two different Tangs.   Same issue on both.

 

I plugged the Tang to a separate USB power supply and still the same issue.  

 

As to the Capacitor.   I could solder one in.  Do you mean pin 16 on the TMS.    Would that be from 16 to ground?  I'm using the 1 version of the board with the diode hack.

forget capacitor. This game doesn't seem to use interrupts. Does it crash always at the same time ?

 

Can you confirm the version of the game you play works with a real f18a ? there are 2 DK versions: 16KB and 24Kb.

 

 

I can't help you I am too busy with Tang 20K it's taking my life away. Helping out remotely is insanely unproductive and I don't want to buy a coleco

 

 

[Duewester]

1 hour ago, retrocanada76 said:

forget capacitor. This game doesn't seem to use interrupts. Does it crash always at the same time ?

 

Can you confirm the version of the game you play works with a real f18a ? there are 2 DK versions: 16KB and 24Kb.

 

 

I can't help you I am too busy with Tang 20K it's taking my life away. Helping out remotely is insanely unproductive and I don't want to buy a coleco

 

 

Sorry about laughing. Computers are supposed to be making life easier. Not taking it away.

[retrocanada76]

8 hours ago, Duewester said:

Sorry about laughing. Computers are supposed to be making life easier. Not taking it away.

lol, doesn't happen when you try to make new things..

[craftsman1234]

On 6/19/2023 at 3:06 PM, retrocanada76 said:

forget capacitor. This game doesn't seem to use interrupts. Does it crash always at the same time ?

 

Can you confirm the version of the game you play works with a real f18a ? there are 2 DK versions: 16KB and 24Kb.

 

 

I can't help you I am too busy with Tang 20K it's taking my life away. Helping out remotely is insanely unproductive and I don't want to buy a coleco

 

 

The main one I have been testing with is the 24k version.  I have no idea if it works with a real f18a as I don't have one.   The crashing appears to happen on the first level when going to the 4th platform/girder this can be walking up the ladder or once on the 4th girder once on the left side of the screen when jumping over a barrel.   The barrel jump on the 4th level causes the most crashes.  If you get past the 4th girder you can rescue the girl and go to the next level.   Level 2 seems to play just fine as well as 3 from what I've tried.

 

I've tested the 16k version as well and it also crashes but I've not really figured out a pattern to it.

 

 

[retrocanada76]

well isn't the place where the sprites are "hidden" by sort order? You should ask elsewhere if a real F18a works there. Something is non trivial in this game.

 

[Hans23]

After a having done a couple of TI-99 repairs, I'm back to figuring out how to get sound working with HDMI on the TI-99/4A.  So far, I have determined that the only version of the tn-vdp that successfully embeds sound from the MCP3202 is tn_vdp_v2_v9958 - It does not seem to be register compatible with the 9918, though, so the console starts but does not display anything.  The (negative) startup beep is audible, but at a very low level.  The signal that I'm injecting into the ADC looks like this:

 

 

I'm taking the audio signal after it has been filtered:

 

The high frequency noise is not audible on my TV even with the volume maxed out.

 

I believe that the filtering is a good thing, but I'm not sure why the signal is barely audible in the HDMI output.  The peak-to-peak level is at 1.7V.  If I understand the MCP3202 data sheet correctly, the input range and the supply voltage are the same - 3.3V volts in our case.  Would that not mean that a 1.7V signal should be pretty loud, as it is at ~50% of the input range of the ADC?

 

I'm going to try and backport the audio code to the F18A based version of the tn-vdp, but I would really appreciate some help on the analog side of things as I don't really know what to do next there.

Thanks!
Hans

[lesbird]

On 5/26/2023 at 7:26 AM, lesbird said:

Hi, I'm trying to get this to work with a HA-8-3 Color Graphics Board for the Heathkit H8 computer. I had to make an extension PCB to move the F18 up higher so the HDMI cable has room to plug in. The Color Graphics Board has a TMS9918 and AY3-8910 sound chip along with ports for analog/digital joysticks and the H8 has a Z80 CPU board at 2Mhz. If I run static demos, such as a color bar test, everything is perfect but if I try running any action games or demos that have animation the screen just goes blank and nothing shows up. The monitor will display a message about "out of range" which I'm guessing it means the refresh rate. Any ideas what would cause this? I'm wondering if the games/demos are using a graphics mode that is not supported or something. Not sure what could be wrong.

 

I have a NABU as well and it works great on that.

 

If you have any suggestions or things I can try please let me know.

 

Thanks!

 

Les

 

So revisiting this. I was able to get it working. The issue was with the graphics card, perhaps it was old and had some failing parts (although it worked with the original 9918). I had to actually redesign a new one using the original schematic. I made the new card with your Tang Nano adapter built in to it (the LVC245s) so you can just plug the Tang right into the card. I also have an adapter that will take a TMS9918 so you can pull the Tang and plug in the 9918 if desired. See attached pics. I am now able to run all the color graphics software on the old Heathkit including a Warlords game I made back in the day.

 

Anyway, something wonky with the original card so when plugging in the Tang Nano F18 it just didn't like it but it plays well with the new design.

 

Thanks for making this, it works very well and the picture quality is fantastic.

 

Les

 

 

 

[Duewester]

Got my boards in today. Made my first V2 board and programed with 9918 V2 program.

No audio yet but I get a LOT of static on the HDMI monitor - turned it down.

The Tang has a nice demo display with version 1.03 BUT

when I install the tang on my board, I get random partial screens

sometimes it looks like it's almost perfect but most of the time it's random split screens.

If I plug the ribbon cable into the other pin header, I get a garbled screen. If I plug the ribbon cable in backwards I get the boot screen.

I used the V1 program and still get a random split screen but with a wide black bar between screens.

Hmmm... edit while writing this. I am using V1 and after 5 power cycles the screen has come up normal three times in a rom.  My other TI with the original Tang/F18a FAT board boots up just fine and a little quicker.

Other notes

V1 program - no led's lit on tang when installed or after programming

V2 program - 5 bright leds and 1 dim led when installed and after/before programming.

 

It's past my bedtime. I will investigate further tomorrow.

[Duewester]

Good morning Ti-Tanger's.

During the night I thought a lot about this. - I know....

Anyway. Pulled my board and all the chips and jumpers on it (well the caps are still installed). Compared to unpopulated board and didn't get un expected differences.

Installed all new logic chips and test with Tang-1. Same results as yesterday - V2 has split up screen.

Broke out another Tang, tested the boot screen -Good. Programmed V2 onto it and checked screen - nice Tang 1.03 screen. Installed Tang-2 onto board and booted. Came up perfect. Rebooted and back to split screen. 😞

Pulled the AMP chip and all jumpers. You gotta have GROMCLK to get a boot at all and the CPUCLK and GROMCLK will give random split screens. SCANLINE and MAX SPR don't seem to have any impact on bootup.

So, two tangs and different logic chips with same results. Either the Program or the board. I suspect the board but, I don't have the 24 pin sockets to make up a new board. AND yes, there is sufficient clearance to install a Tang if I don't socket the logic chips but, that takes reprogramming the Tang off the table.

I have ordered more socket strips (I thought I had more than I do) and will build new board when they arrive.

 

In the meantime, while cleaning up my workstation I found my shorter ribbon cables. I've been using a long one for testing because it makes it easier to place the Ti-Tang board for easy reach. I swapped one of the short cables and turned on the Ti. Interestingly the Tang did NOT boot with all the led's lit but only the first and last. It still booted up split. I rebooted and the LED;s started counting up and down and finally all settled on but the screen came up looking good. I powered down everything and left it while making this entry but when I just tried it again it was all weird again and the split screen. Sometimes it boots up good repeatedly and then out of the blue goes back to weird.

Something else to note - there is a single LED all by itself and as longs as it is NOT lit, the screen comes up normal.

Gotto go for now. More on the adventures of T-Tang man later.

[Duewester]

Notes before beddy by time.

V2 board, V1 software kinda works until the graphics get intense. Then gets all crazy and freezes TI. I hooked up my 32k/TiPi and it appeared to be working. Tried the sound out and Nada using Pin 7 on 9919 or pin 3 on audio out. Not really worried about sound till I get graphics working.

V2 board, V2 software boots most of the time and even lets me get to my FG99. Select about any game and it most of the time won't finish logging into the game. The leds on the tang start getting all crazy and the static on the monitor gets louder the longer I leave it on. Still have not hooked up the amp chip. 

 

So, I broke out the Frankennine with the Fat board and software on it. Access to TiPi and FG99 without any issues. Games play. So cable and TV are not the problem. Cartridge is not the issue.

I'll build another Fat board and try V1 and V2 software on it before I'll take apart the Frankennine again.

Well it may be a day or three before I get back to this.

Nighty night

[Duewester]

Well, time to build another Version2 board.

I built a Ti-Fat board - even 2nd time around that diode bodge is a PITA to perform - (without all the VGA) and installed Tang1 with Version 1 Software in it. Boots up and runs like a charm - I really suck at Ti Invaders - and I ran the V2 software on the Ti-Fat board. The V2 software still shows all the LEDs on the tang lit but it worked (at least I know it's me and not the Tang that can't play Ti Invaders). My biggest beef with the Ti-Fat board is instability in the 9918a socket due to the stacked up headers needed to clear the crystal next to the 9918a socket.

At least I know now that THIS Ti-99/4a is functional - another component checked off.

 

So, since I'm building another V2 assy, when I finished the 9918a insert I tried it out with two tangs and original V2 Board. Same bad results.

Another component checked off.

Edited June 30, 2023 by Duewester

[Duewester]

Well the good news is, I have two working Ti-Fat boards.

The bad nesw is, I have two V2 Assys with the same indications using versions 1 and 2 of the software.

1. Version 2 software lights up random LEDs on Tang
2. Version 1 software LOOKS normal until, the graphics get intense.
3. The 20 pin connections are picky about which pair to use. They aren't as "Pick and choose" as you might think.

I guess the next step is get on Gowin and recompile the version(s) and see if that helps. Not much hope for that though since version 1 works great on original Ti-Fat board.

 

In the mean time, the wife has returned from her trip and I need to spend time with her.

[retrocanada76]

On 6/30/2023 at 5:15 PM, Duewester said:

Well the good news is, I have two working Ti-Fat boards.

The bad nesw is, I have two V2 Assys with the same indications using versions 1 and 2 of the software.

1. Version 2 software lights up random LEDs on Tang
2. Version 1 software LOOKS normal until, the graphics get intense.
3. The 20 pin connections are picky about which pair to use. They aren't as "Pick and choose" as you might think.

I guess the next step is get on Gowin and recompile the version(s) and see if that helps. Not much hope for that though since version 1 works great on original Ti-Fat board.

 

In the mean time, the wife has returned from her trip and I need to spend time with her.

Hey, been busy with work and shit. How long is your cable on V2 ? It shouldn't be too long or you get all kind of noises.

 

Good news I finished the Tang20K 9958 128K so now I may have time to check the TI99.

 

It could be that the V2 code got some meta-stability issues and causes these issues. To be honest I didn't stressed it out as I was working on the 9958 version.

 

[retrocanada76]

On 6/24/2023 at 11:57 PM, Hans23 said:

After a having done a couple of TI-99 repairs, I'm back to figuring out how to get sound working with HDMI on the TI-99/4A.  So far, I have determined that the only version of the tn-vdp that successfully embeds sound from the MCP3202 is tn_vdp_v2_v9958 - It does not seem to be register compatible with the 9918, though, so the console starts but does not display anything.  The (negative) startup beep is audible, but at a very low level.  The signal that I'm injecting into the ADC looks like this:

 

 

I'm taking the audio signal after it has been filtered:

 

The high frequency noise is not audible on my TV even with the volume maxed out.

 

I believe that the filtering is a good thing, but I'm not sure why the signal is barely audible in the HDMI output.  The peak-to-peak level is at 1.7V.  If I understand the MCP3202 data sheet correctly, the input range and the supply voltage are the same - 3.3V volts in our case.  Would that not mean that a 1.7V signal should be pretty loud, as it is at ~50% of the input range of the ADC?

 

I'm going to try and backport the audio code to the F18A based version of the tn-vdp, but I would really appreciate some help on the analog side of things as I don't really know what to do next there.

Thanks!
Hans

Hans, try putting a voltage divider for the VREF. Look the new V3 board for the Tang Nano 20K i added one there. I will help setting the better voltage ref for the TI

[Hans23]

38 minutes ago, retrocanada76 said:

try putting a voltage divider for the VREF. Look the new V3 board for the Tang Nano 20K i added one there. I will help setting the better voltage ref for the TI

I'm looking at the datasheet and it seems that the minimum Vdd/Vref voltage is 2.7V.  Does that not mean that I can't really turn the reference voltage down to my output level of 1.7V?

[Duewester]

"Hey, been busy with work and shit. How long is your cable on V2 ? It shouldn't be too long or you get all kind of noises."

 

I have been thinking the same thing. I saw some reliability improvement when I went from a roughly 10 inch cable to a 5 inch. I am going to look at cutting one down to 2 inches or so.

[Duewester]

5 hours ago, retrocanada76 said:

Good news I finished the Tang20K 9958 128K so now I may have time to check the TI99.

 

It could be that the V2 code got some meta-stability issues and causes these issues. To be honest I didn't stressed it out as I was working on the 9958 version.

 

Great news about T20k.

 

I was thinking about putting aside some time to compare the two codes but, I would be delving into strange territory for me. I have no idea about FPGA programming but I have watched a few YouTube lessons.