Still working on GTIA V-Gate?

[+mytek]

6 hours ago, Panther said:

I've all the components on hand, so I'll just have to order some PCBs and try out a couple of these.  I can finally dispose of these original version V-Gate PCBs I have but never used.  I haven't used CRTs in ages, so this is a very useful addition.

Unfortunately those first PCBs turned out to be an experiment that didn't quite work as they were intended. It was my first foray into over scan control, and it took a few more experiments and iterations to get me to where I am today. This project brings it all together, V-Gate and better video all in one compact easy to install package

 

And although it still does get impacted by the supply shortages (PIC), at least there is a way around it for the time being.

 

[pseudografx]

Could this be controllable from the U1MB UI (if @flashjazzcat could add support for it)?

[+mytek]

1 hour ago, pseudografx said:

Could this be controllable from the U1MB UI (if @flashjazzcat could add support for it)?

With the right plug-in it's already supported, because the 1088XEL/XLD have a V-Gate selection that works this same way, have a U1MB integrated as part of design, and @flashjazzcat coded one to work for it.

 

[Panther]

13 hours ago, mytek said:

With the right plug-in it's already supported, because the 1088XEL/XLD have a V-Gate selection that works this same way, have a U1MB integrated as part of design, and @flashjazzcat coded one to work for it.

 

I don't think it requires any particular plugin, I believe this is implemented in the default plugin code.

 

[+mytek]

1 hour ago, Panther said:

I don't think it requires any particular plugin, I believe this is implemented in the default plugin code.

Well it very well may be the default now, but before the 1088XEL it didn't exist

 

[olix]

I just finished my first UGV board.

The image quality is really as good as with a UAV. Brilliant with S-VHS and also very good with normal FBAS (composite) signal. (If you use good cables).

Also the V-gate functions works perfect. The Overscan-area ist perfectly hidden.

 

I’m really excited about this device.

 

@mytek:Thank you so much for developping this little peace of hardware. I really love it.

[+mytek]

Wow you got that built very fast @olix 

 

And I'm glad that you like it 

 

[olix]

Yes, but I'm afraid a little too fast.
I must have made a mistake with the second circuit board that I assembled, because it destroyed a GTIA.
Now, it only delivers a brown image without content.
The first board runs without problems.

[+mytek]

3 hours ago, olix said:

Yes, but I'm afraid a little too fast.
I must have made a mistake with the second circuit board that I assembled, because it destroyed a GTIA.
Now, it only delivers a brown image without content.
The first board runs without problems.

Sorry to hear that

 

The only possible way that I can think of that the UGV could take out a GTIA is if there is something causing an output (Luma, Chroma, Sync) to be shorted, or directly connected to +5V. So the best approach would be to get out the continuity checker and try to locate the culprit (bridged trace/pad or a bad chip).

 

[+mytek]

On 7/27/2022 at 8:09 PM, pseudografx said:

Could this be controllable from the U1MB UI (if @flashjazzcat could add support for it)?

I thought I'd revisit this question. As was already established, there is support for enabling/disabling the V-Gate function in the U1MB plug-in. However having experience with that as well as the alternative method of using a physical switch in the form of either the RF Channel switch or via keyboard control aka: 576NUC+, I find I prefer the physical switch. Reason being is that V-Gate is a technology that is often times better when switched on the fly to accommodate the requirements of a given application or game, which is much easier and faster with a simple switch.

 

On 8/8/2022 at 5:05 AM, olix said:

I must have made a mistake with the second circuit board that I assembled, because it destroyed a GTIA.
Now, it only delivers a brown image without content.
The first board runs without problems.

Did you ever discover what the problem was with the 2nd board?

 

[olix]

39 minutes ago, mytek said:

 

Did you ever discover what the problem was with the 2nd board?

 

I checked the second UGV again very thoroughly for errors. First optically and then again electrically. The problem was that most of the important measuring points are covered by the GTIA. So I soldered a test adapter to be able to test the UGV module in operation and still be able to measure. (Pictures are attached)

I didn't find any errors!

 

I first carried out the measurements with the GTIA that was killed - it still creates a picture. (Even if only a big brown screen and the Atari doesn't start either). But the VGATE signal was clean on the PIC and no short circuits could be measured. So actually all good.

 

I then dared and used my last GTIA from the spare parts box: The UGV works. So there was no problem at all.

I may have destroyed the GTIA by plugging and unplugging it many times. Maybe a static discharge.

 

So all good. So now I already have two working modules and I have sent a third one to an ABBUC Club member. He is also very happy and enthusiastic about the picture quality.

[flashjazzcat]

On 7/28/2022 at 7:30 PM, Panther said:

I believe this is implemented in the default plugin code

Correct. It's in the default 1088XEL/XLD plugins (as opposed to the generic default U1MB plugin).

[olix]

I've attached a photo where you can see the working VGATE function and the high quality of the S-Video signal. (The interference in the first images are not visible on the monitor)

[+Stephen]

2 hours ago, mytek said:

Reason being is that V-Gate is a technology that is often times better when switched on the fly to accommodate the requirements of a given application or game, which is much easier and faster with a simple switch.

TIL that the V-Gate is hot switchable.  Interesting.

[+mytek]

@olix I meant to ask you earlier and forgot... what unit is this being used in?

 

 

I'm thinking a PAL 130XE.

 

Was there enough clearance for the Antic, or did it require a stacked socket to raise the UGV board up higher?  If so, were there any issues with keyboard clearance?

 

I noticed you used lower profile right angle headers than what I had spec'ed in the BOM.

[+mytek]

14 hours ago, olix said:

So all good. So now I already have two working modules and I have sent a third one to an ABBUC Club member. He is also very happy and enthusiastic about the picture quality.

I forgot to congratulate you on getting to the bottom of the problem with your 2nd board 👍

 

Also glad that your fellow ABBUC member likes the one that you sent to them.

[olix]

5 hours ago, mytek said:

I'm thinking a PAL 130XE.

 

Was there enough clearance for the Antic, or did it require a stacked socket to raise the UGV board up higher?  If so, were there any issues with keyboard clearance?

 

I noticed you used lower profile right angle headers than what I had spec'ed in the BOM.

 

Yes, this is an 130XE PAL Board (130XE 64x4 C103579 Rev 1)

And i had to stack a second IC-Socket to the UGV, cause the ANTIC is socketed too. But it's still very tight, so i put a little paper on top of the ANTIC connector row.

I have actually never installed the UGV in a closed housing. So I can't tell if it fits under the shielding plate or if it hits the keyboard. But i will try the next days in a 130 XE and a 800 XL case.

Yes, the right angel headers are lower, but this is no problem. 

[+mytek]

5 hours ago, olix said:

So I can't tell if it fits under the shielding plate or if it hits the keyboard.

I'd be surprised if the shielding still fits without cutting out part of it, but unless you plan on using the RF output, a missing RF shield really isn't a problem. However the keyboard does have to fit on its own, so hopefully that's still possible.

 

Now on an 800XL (or 600XL) installation, stacking another socket should not be required with the tall right angle headers I specified in the BOM (not so sure about the ones you used). There are usually some transistors that stick up kinda high towards the left side of GTIA that can interfere, so they will need to be bent over in most cases. And even the color adjustment pot on a 600XL can be a clearance problem if a low profile header was used for J2 (not a problem on an 800XL).

 

BTW, the tall right angle headers don't stick up any higher than the GTIA chip when plugged into the UGV board. In fact GTIA will be the tallest component. So unless there is a problem getting the specified headers (or something like it), there would be no reason to go with a low profile version.

[+mytek]

20 hours ago, olix said:

And i had to stack a second IC-Socket to the UGV, cause the ANTIC is socketed too. But it's still very tight, so i put a little paper on top of the ANTIC connector row.

I've been test fitting a UGV to my 65XE (i don't have a 130XE). So far I've only gotten as far as removing the soldered in GTIA chip, and installing a socket in its place. If I place the UGV pcb on top of that socket, it appears that it will totally clear the Antic chip that it overlaps. So your situation is probably kinda unique with the Antic also being socketed. Based on what I'm seeing it looks pretty certain that when only GTIA is in a socket on any XE, no stacking of an additional socket will be necessary. When I get back into town on Thursday I'll build up another UGV and install it in my now prepped 65XE, and just for grins see if the RF shielding still fits.

 

[+mytek]

Update: UGV Install in 65XE

 

I wanted to mount the Mini-Din Interface board so that it could use the existing RF Jack hole after I removed the modulator. But the existing hole was so low in relation to the Motherboard I had to science the sh!t out of it to make it work.

 

First I sanded off as much of the soldered pins that I felt safe to do. This is one of the flawed prototypes where had to also dremel a pad that had an inadvertent connection to the ground plane, but the procedure would be pretty much the same with a new board. As I figured out what I wanted to do, it evolved into an idea that really wouldn't have required sanding off anything other than the Mini-DIN jack's mounting tabs (live and learn).

 

 

Originally I was going to drill a bunch of holes where the pins stick up, but for safe measure to prevent shorting out anything I cut-out a hole in the Motherboard that would allow for electrical clearance of the pins and traces. I also had to move a capacitor to get it out of the way. I know it's a pretty rough looking hole, but when I install the interface it won't be visible from the top of the Motherboard. If it was going to be seen, I would have spent time with a file and made it pretty

 

EDIT: In retrospect, I think instead of making a square hole, I might have been better off to use a stepped drill bit to make a round hole instead. This would have not only looked nicer, but also would have been easier to execute.

 

Next I lined it up with where the RF Jack hole was positioned, and soldered the interface board to the ground plane, both securing it and also grounding the Mini-DIN case. It became quite apparent after I did this, that it wasn't necessary for me to have removed so much of the solder mask covering the ground plane

 

 

Next I needed to slightly enlarge the RF Jack hole with a stepped drill to better allow for a molded S-Video connector to easily fit through it.

 

 

Here's the overview of all the pieces in place and the wiring. There was actually a good amount of clearance for the UGV board above the Antic chip since it wasn't socketed, so no stacking of additional sockets to add height was necessary.

(click on image to enlarge)

 

In fact I had ample clearance for the RF shield to fit back in place without any modifications. I just noticed the inscription in Sharpie of when I bought this unit from Bruce Carso over at B&C, who personally dropped it off at my vacation home not too far from his shop in the town of Eldorado. Wow 2016!  That was back when you could still buy these as NOS for only $100 in the original pristine box with all the accessories.

 

 

The XE computers have the channel switch built into the modulator, so with that removed I had no switch to disable V-Gate. I guess I could have mounted a mini toggle in the slot that was provided for the switch in the rear panel, but I opted not to and just stuck a label over it instead. V-Gate will always be ON.

 

 

Here's what it all looks like with all the cables plugged in. I routed the UGV's composite output to where it used to feed out of the modulator, so it's now available on the A/V DIN jack in the usual spot. Of course audio is also present, as well as the original Un-Gated Luma and Chroma outputs.

 

EDIT: When I tested the original Non-Gated S-Video coming out of the Atari A/V DIN jack, I could see that the V-Gate was still influencing the image produced, with a different shade of whatever color was being generated in the border appearing as bars on the left and right side of the screen. This apparently happens due to the non-buffered color signal being influenced by the capacitive coupling to the FMS6400 chip on the UGV. So each time the DG419LEDY Analog switch would 'gate' in the color signal, this would connect it to the R/C network feeding the FMS6400 Chroma input which shifted the phase of the color slightly.

 

Bottom line is that only the UGV's video outputs should be used after installation.  With the UGV to be considered as a video driver replacement board intended to take the place of the stock Atari video circuits, and not to be used in parallel with that.

 

 

So I fired it up and all looked good, with composite and audio on the small CRT monitor to the left, and S-Video to the LCD monitor on the right. Except when I started running some other checks I could see that there was a vertical red and blue line running down both sides of the screen on the wide-screen LCD monitor. That could only mean one thing... (to be continued)

 

(click on image to enlarge - you won't see the red & blue vertical bars in this photo, but it'll give you a good idea of how clean the overall video is coming out of the UGV )

 

So guess what I forgot to do?

(click on image to enlarge)

Yep I forgot to remove that unnecessary W2 jumper that Atari installed in this NTSC 65XE (you PAL folks don't need to worry about this).

 

With that jumper removed the vertical lines where V-gate was asserting itself disappeared, and the ACP test program image produced by the UGV in this 65XE looked great 👍

(click on image to enlarge)

 

Based on what I'm seeing in this 65XE install, I would feel confident to say a 130XE would be virtually the same situation. Of course if you don't want to go through the hassle of installing the Mini-DIN Interface board, the UGV's video signals could all be routed out the standard A/V DIN jack a bit more easily. That would be done in the same manner as shown in the UAV installation manual.

 

[+mytek]

Works very well with the WiiStar WS-Z51 AV2HDMI converter via its S-Video input sourced from the UGV, and then feeding the HDMI output to my VIZIO HDTV.

 

Note: This is an actual photo taken of my VIZIO HDTV Screen using my Samsung tablet's camera (click image to get a slightly larger image). Any blooming or lack of sharpness stems from this method of capturing the TV's image in this way - seeing the screen with your own eyes reveals an even better picture.

 

To match up with the converters input requirements I have UGV's dip switch set to 635mv Video Data level, which appears to be right judging by the ACP test.

 

I bought a couple of the WS-Z51's from AliExpress for just a tad over $16 each (LINK), and will be using them when I want to put the Atari on the big screen TV for some game playing.

[+mytek]

Something fun and perhaps silly, but still about the UGV.

 

 

I saw this done years ago by Long Horn Engineering for one of their boards, and always thought it was so cool. Well the UGV really fit the bill nicely, since it was small enough and thin enough to translate to a business card size.

 

So in this video I'm showing how to extract the embedded UGV board from the card itself with a pair of scissors. BTW, the last two cuts I made were optional, since a slight overhang of the GTIA won't be an issue on either of those sides.

[+Stephen]

Half way done with my build of 5 UGVs (the PICs are already flashed):

[cbelcher]

Just put one in a 400 after doing a 48k RAM upgrade.

 

It is TIGHT on the GTIA card - and the motherboard will not seat correctly in the cast metal shield with the UGV in the GTIA socket.

 

So - to make it fit, I took out the GTIA socket and moved 6 resistors and a capacitor to the back side of the board.  

 

Picked up audio at the Joystick port side of R160 with a 150 ohm resistor in series - so I only have POKEY sound... But some of the old mods address that.

 

Ran it all out to a DIN-5

 

Closed up, and have great S-video out. 

[+mytek]

2 hours ago, cbelcher said:

Just put one in a 400 after doing a 48k RAM upgrade.

 

It is TIGHT on the GTIA card - and the motherboard will not seat correctly in the cast metal shield with the UGV in the GTIA socket.

 

So - to make it fit, I took out the GTIA socket and moved 6 resistors and a capacitor to the back side of the board.  

 

Picked up audio at the Joystick port side of R160 with a 150 ohm resistor in series - so I only have POKEY sound... But some of the old mods address that.

 

Ran it all out to a DIN-5

 

Closed up, and have great S-video out. 

Yeah the 400/800 were questionable on the fit, and I had no way to test it myself (only have XL and XE systems).

 

Glad to hear you made it work, and that you are getting good video out of it 👍

 

Is it possible for you to take a PIC of it all together from the front side as well?  EDIT: just a pic of the CPU card with the UGV installed.