retrosix

November 4

On 10/30/2023 at 12:13 AM, Atari Dogs said:

Do you offer the option of buying the resistors with the CleanComp? I have a light sixer I want to mod.

Just leave a note on your order "please include jailbar resistors" and I will chuck them in free. Next batch I'll include them as standard

October 26

Tim has that mod where it listens and re-creates the RGB signals from the data bus

There is also the composite option that I made

October 22

Pretty sure nobody has seen or fixed this before, considering the issue only seemed to be noticed once installing the latest CleanComp shows up the issue now the rest of the signal is so accurate

I got my hands on a NTSC 2600, REV B C010433 and it had major jailbars.

image.png.a83b8acacc3d0872de0836aaffb2ffdd.png

I proceeded to diagnose and fix it, and the full article is here Jailbars Fix (Atari 2600) - RetroSix Wiki but for good reference so more people can see it and comment, I'm posting the solution here.

I'd love any feedback as well as any of you guys with jailbars to send me photos of the top and bottom of the TIA area, ideally with the TIA removed and/or socket removed if its in the way so I can provide solutions for every version that has this issue. I can diagnose and fix it remotely with just images, then you could confirm it works.

Below is the fix for the model mentioned.

The noise is caused by the chroma signal and phase signals bleeding into the LUMA pins.

The LUM0 pin gets its noise from the COL pin.

LUM0

COL Pin

And the LUM1 pin gets its noise from the Phase 0 pin.

LUM2 gets its noise from the BLK signal if it is connected to the COL pin as it basically re-introduces that COL noise onto LUM2 as well. This results in distinctive jailbars.

Jailbars

Bottom Trace Issues

Let's start with COL pin and how it adds noise to the LUMA channels.

The LUM0 pin on NTSC models picks noise ultimately from the COLOR (pin 9) as it is right next to it on the NTSC chip, whereas on PAL it is physically separated by PAL-I pin.

TIA Pinout

Unfortunately this physical closeness inside the die of the TIA chip cannot be fixed, but its direct effects are minimal (100mV approx).

The main cause of the cross-talk noise comes from the fact this COLOR trace comes out of the PCB itself both on the top and bottom.

The bottom trace adds noise to the LUM1 signal for two reasons.

Bottom Traces

Firstly, you can see how close the COL pin comes to the LUM1 trace. This adds lots of noise to LUM1.

Cutting this trace disconnects the signal from the part of the trace that is close to LUM1, thereby removing the noise.

On top of that, you will notice there is a resistor (on the top side of the board) that this trace originally connected to, sending COL to !BLK pin, which then is physically close to the LUM1 trace which would re-add this noise.

Cutting the trace resolved this issue at the same time by disconnecting !BLK from COL (something you will see is the correct thing to do in a moment.

New COL Pull-up Resistor

By cutting the bottom trace to remove the noise, we also disconnected the pull-up resistor on the COL pin, so we must reinstate that, in a place that does not add noise back to the LUMA pins.

I chose to place the resistor on the top side, bending out the COL pin from the PCB so that the solder pads which are physically close to each other between COL and LUM0 are not joined, and then sending a new resistor directly on the TIA COL pin to 5V.

COL Pull Up Resistor

At this same time I also chose to weaken the pull-up strength of the COL pin by using a 3.3k resistor instead of the factory 1k resistor.

This step is optional, and lowers the amplitude of the signal, thereby further reducing cross-talk, but as I am passing this video signal through a CleanComp which reconstructs the Chroma completely, this reduction in amplitude is no issue at all.

Top Trace Issues

With this fix, we have removed most of the noise from LUM0 and LUM1 pins but not all.

On the top side we need to remove the LUM0 pull-up resistor to access the trace underneath, which is running close to the COL pins solder pad.

If you have removed the COL pin from the package so it isn't touching the PCB at all, this step isn't required.

Strengthen LUMA Pull Up

Now most of the noise is removed, but still present due to the TIA die package itself, the only choice we have left to further remove noise is to strengthen the pull-up resistors on the LUMA pins.

The TIA can sink around 8mA of current so the pull-ups can be as low as 1k safely.

By cutting the trace from LUM0 in the previous step its already missing the pull-up resistor, and for LUM1 you can remove the original pull-up resistor or stack another on top and calculate the parallel resistance. I find it easier to just remove both resistors, and add a new 1k resistor to LUM0 and LUM1 pins to 5V.

I do this here on the CleanComp pins as they are connected to the TIA pins.

LUMA Pull-Ups

If you do not want to strengthen the pull-ups you can just replace them with the same value for LUM0 if you cut the trace on top, otherwise the original pull-ups are already in place and nothing else is needed.

Cleaned Up Signals

With the fixes done, all LUMA signals are down to around 100mV noise, which cannot be cleaned up much more as they are likely from the physical pins being close inside the TIA die.

Cleaned LUMA

The results however speak for themselves.

Fixed LUMA

Pitfall

Restoring COL to CleanComp

Because we lifted the COL pin completely off the PCB, the CleanComp is no longer connected, so remember to solder a wire from the COL on the CleanComp to the TIA.

COL Wire

Now we have COL restored, I have also noticed on the NTSC versions the COL signals is much stronger, so turn down the S-Video potentiometer anti-clockwise all the way until the image is black and white, then turn it back up just enough to bring the color back. This will give you the best visual.

Or use an oscilloscope to tweak the level to exactly 350mV when connected to the receiver.

!BLK Pin Issues

Even though Atari officially recommended to link COL and !BLK together with a resistor, do not do this. It is 100% wrong.

Atari Note

Here is it disconnected.

!BLK Disconnected

And here is it joined.

As part of the entire Chroma circuit, I can tell you it should look like the disconnected image. The CleanComp completely reconstructs this signal, as does the Atari 2600 circuitry (just not as well).

By connecting !BLK you are massively reducing the chroma wave amplitude, causing the TV/receiver to super over-compensate colors, making them very wrong but over-saturated.

Remove this resistor if doing a CleanComp mod for sure.

Other Causes Of Jailbars

If you still get jailbars it is important to know that the visual issue can come from many other places including:

Cheap composite converters
Chroma Voltage Too High (if using CleanComp turn pot to left to reduce level)
Receivers themselves (some TVs have poor quality circuits)
Wires for chroma/luma and composite physically to close to each other
Poor quality wires
Different variants of Atari 2600

All tests done above were using a genuine power supply, a CleanComp V2 (white board), an OSSC Add-on through OSSC to a UGreen HDMI capture card.

I have previously tested the UGreen HDMI capture card doesn't introduce noise on composite or S-Video using the rest pattern output from OSSC to HDMI, and tested the OSSC Composite Add-on board with a known clean Neo Geo AES to prove the entire signal path is clean from noise, before then testing the Atari 2600.

October 22

So confirmed on 2 units so far, that NTSC consoles have the jailbar issue, nothing related to CleanComp, it is just that CleanComp makes the image that clean you can see the issue.

Fix is here Jailbars Fix (Atari 2600) - RetroSix Wiki

I'll probably do a separate post on here for good reference for others to find it.

image.png.d5e0f8eb701bbd406ff652cdd0ead055.png

Before

image.png.ca448cf2e2bb0746f630ecf1d132cfb3.png

October 21

Just a quick update will follow up tomorrow with full write up.

NTSC consoles are inherently faulty and the TIA itself is worse than PAL due to location of LUMA to COL and phase pins, as well as bad routing decisions causing vertical jailbars.

Fixed and will post solution tomorrow and on wiki.

October 8

On 9/28/2023 at 9:46 PM, -^CrossBow^- said:

Could be, but these signals are being taken straight off the TIA chip itself so the rest of the atari board layout doesn't or shouldn't really matter since the TIA pinout is the same across all models. Now, what I could see is that some TIA revisions might cause differences like this. I see some variance with UAVs installed from 2600 to 2600. I had a 7800 once that had bad jail bars but only from composite and only when 2600 games were played. It ended up being the TIA causing them.

If jailbars remain (seems we found it to be the retrotink so far), I suspect it could be noise from the traces running directly under the TIA and the bottom layer of the CleanComp.

Just ruling out retrotink with those that have jailbars and if its not the retrotink we can continue to diagnose. One test would be to lift the CleanComp up slightly so its not flat on the board.

October 8

On 9/20/2023 at 9:27 PM, Blinky said:

I recently ordered one from you. white PCB and just looks like your image. Is it a V2 or else how can I identify which version it is?

I've installed it on one of my 2600JR (PAL version) and gave a great picture but after 5 minutes the screen started flickering and a gray image with interpolated color frequency bars displayed. Maybe this happened because the system heated up but this didn't happen before the mod. Tweaking the little preset resistor on the CleanComp didn't help but tweaking the preset resistor of the 2600 fixed it.

Unfortunately I noticed the TIA behaves a bit strangly (after warming up) as the missle sprites position seems to randomly change so It looks like my 2600 has a dodgy TIA and it doesn't like the increase load (when I turn the preset pot of the 2600 all the way to GND the problem goes away but then the color signal is gone of course) added a 4.7uF cap between GND and 5V but that didn't help either. I'm gonna tinker around a bit more in the weekend and will try to remove some components.

I'm wondering does the CleanComp rely on those pull up resistors on the luminance/color and sound pins or will it work fine with those parts removed? I wanna see if I can get the TIA to behave again after it's warmed up. if not I'll need to get another one for my other JR or light 6-er as I'm not sure I will be able to desolder the CleanComp properly.

Here are some pictures on my Sony Brava LCD TV

The V2 puts zero load on the pins, they go straight to buffers. So should not cause it. I suspect the chroma CADJ voltage on your Atari is warming up and moving, causing shifting then flickering as you first described perhaps.

Literally zero ohm load on pins.

I'd suspect cartridge pins need cleaning or maybe bad/failing TIA for those glitches. Might be worth swapping it out to see

October 8

On 9/28/2023 at 5:33 AM, -^CrossBow^- said:

Hmm..the jailbars are still present on this new version over s-video still. I'd hope that could be cleaned up more? But the overall color output does look improved over the earlier version board that I've got installed into a jr model.

Zero jailbars. But if you are using new RetroTink hardware that causes jailbars.

Here is RetroTink 2X Pro and 5X

Here is using OSSC and regulator monitors that support S-Video also

Could you confirm if you are using retrotink for testing as if so they are poor quality and add noise/jailbars

September 19

17 hours ago, -^CrossBow^- said:

I'm up for testing as this would have been really nice over the past month or so while I was literally doing over a dozen consoles. That said, I'm concerned anytime I see pot trimmers on a board? Because, it means some dialing in is required and that can vary greatly from display to display. So my concern is trying to install and set something like this up for another person, adjust it looking good on my AV setup, but then they complain when they receive it because it looks over saturated or washed out...etc?

What has been the initial feedback on that so far? Because I can tell you that most of my clients that receive services from me are NOT willing to open up their consoles to have to tweak anything.

The pots are pre-configured at factory for perfect voltage levels. The tuning allows you to dial in chroma to your desired level (sharpness vs noise).

So far all feedback on V2 has been 100%. V1 the only main issue we had was a filter sending 5V to the board had poor soldering so no power. That was fixed. Early pre-assembly units (most the testing on here) had a lot of improvements, so AV output wise I'm pretty sure it will give the best possible comp and S-vid available for the console.

I'd love to get more widefield reviews of the latest version and see if there is any improvements left to be had.

September 18

So V2 is finally out and sold quite a few to get some general feedback.

Issues fixed since V1's.

Audio sorted and correctly buffered via its own buffer so no interference or lost audio
Luma sharpness improved / undershoot reduced
Luma / Chroma timing improved alignment
Much better filtering into board and power to stop the "5V not present" issue
S-Video big improvements, reduced cross-talk, and better levels after more devices to test on (OSSC, RetroTink etc...)
Upgraded capacitors to AV spec tolerances

I feel the composite is as good as it will get, and all we really have left is dot crawl. Without TIA intervention we can't change that bit I don't think.

S-Video I am still looking for feedback on, as I feel it should be able to be possibly improved slightly. It looks really sharp (perfect) on Luma so I think we have zero issue there. Chroma I feel perhaps could be tweaked the waveform a little as you get slight slow bleed between sharp edges. Unless thats just S-Video limitations?

Known possible improvements

Remove cross-talk on S-Video with shielded cables or PCB design updates perhaps
Remove slight vertical bars most visible on dark blues. Perhaps actual Atari 2600 board mods to filter TIA circuit

Anyone want to give me some feedback on this version? If so simply place an order at retrosix.co.uk and in your order notes ask to include a free CleanComp V2 and your username from AtariAge and I will chuck them in for free for testing. I'll limit it to about the first 20 people but I think we only had about 20 people in this thread test anyway.

Any other feedback greatly welcome and thanks again for testing and giving feedback on this product, its made it into a really nice and easy, clean mod for the 2600 that I've had a lot of positive feedback from.

May 6

You should be able to easily diagnos the issue now using the diagnostic steps here https://www.retrosix.wiki/power-circuit-atari-lynx-ii#HJuwv

Start from top and go down doing each test

May 6

Full info here https://www.retrosix.wiki/power-circuit-atari-lynx-ii#HJuwv

Check your Schmitt trigger circuit components

April 16

@batarican you let me know wholesale prices I would love to sell this at retrosix, got a nice growing community for 2600 and 7800 now and they would love this.

April 7

Just for an update on this, the *P isn't anything to do with PAL, it specifies if the game uses Paddles

March 18

On 3/14/2023 at 6:12 PM, cwilkson said:

Why are you using a SMPS? I don't see a need and it will only cause you extra grief as you've already pointed out.

But maybe there's something special in your circuit that requires a weird voltage?

For transistors...are you using BJT or FETs? What type? BJTs are probably faster at the cheap end, but storage delay will kill you if you don't compensate for it correctly.

I'm not using transistors like other composite mods, I reconstruct the entire signal. I'm using FETs just for triggering at the start and will use a lower voltage trigger to improve timing.

I'm not using SMPS I'm showing that many (in fact all) third party and common ones sold on market will always be this so it's good to support them.

March 14

32 minutes ago, cwilkson said:

Real quick cinematography advice. Turn your phone 90 degrees!!! And tell all your friends to do it too!

To take advantage of basic human vision wetware. And every computer monitor everywhere.

You'll have a bigger viewable image without wasted pixels at the top and bottom and you won't need to pan and scan.

You can then use a cheap tripod to avoid shaky cam effects. Even propping the phone against a stack of books will work. I said cheap!

TikTok is portrait only I'm afraid

March 14

Yes the unsharp luma is caused by the TIA delaying the Luma 0 and the requirement to then over dampen with a 1nf cap on luma stage. Need a huge 85ns delay circuit for Luma 1 and 2 for next version then can re sharpen luma.

Chrome causing undershoot on sync fixing by faster response transistors on SYNC input stage.

Hopefully those 2 things remove all remaining wave shape issues.

Final one would be supporting low quality SMPS and seeing if we can clean any more of that 14mV noise. It's not visible at all on any monitor or PVM I have tried only capture card but super hard to photo the monitors without it adding the usual effects you see when photographing them.

Check video out here https://www.tiktok.com/@retrosix/video/7210065739225615622?is_from_webapp=1&sender_device=mobile&sender_web_id=7207524830441883141

Photo attached

Pretty sure the noise on captures is the capture card but we will see if others see any noise.

March 13

So the first version v1.0 is out. https://retrosix.co.uk/Atari-2600-CleanComp-p537997297

Lots of improvements and done a full scope of the outputs for those interested (seems lots here are interested in the hardware side).

This is through a cheap capture card, and output on actual monitors shows even nicer, but at least this shows the worse case.

Composite:

S-Video (brightness +10 to give similar visual as composite as the capture card has different base levels of brightness for them)

Now I have these production boards I have also done some further improvements to come out in the next version. So important notes for this release:

1. LUMA0 is still much more delayed than LUMA1/2 now everything is cleaned up (so you will see the black line on back in pitfall). I have also had to over-dampen the luma to compensate the mixing of the LUMA pins together. Plan to fix with a 85ns delay circuit next time

2. Using switch mode power supplies (instead of transformer based) will introduce 2MHz visual noise all over the image. Not sure if this is normal with RF and other mods, but plan to find a way to filter the system/CleanComp next version. So use high quality/original transformer based power supplies for now.

3. Chroma merges into SYNC rise now I don't put additional load on the TIA pins. Doesn't seem to visually effect anything but I will remove the sync tip undershoot next version with faster transistor

To confirm the signals I have used a PicoScope 4444 (the lowest noise differential scope you can get). Internal noise is 9mV from the scope so thats the lowest measurement we can trust.

Measuring the noise its 14mV so there is a tiny bit of noise to still try and clean up, but overall I'm happy with it.

S-Video has had very little testing, but hopefully I will get plenty of feedback from this release to then improve that if needed.

March 13

On 2/22/2023 at 4:07 PM, -^CrossBow^- said:

As I answered this on the FB group posting where you also stated the same thing...

turning the color trimmer just changes the tint I see on the screen and does not change the presence of the jail bars. They are also present on composite but not as pronounced as they are in s-video.

How are you handling the 'stereo' of the TIA? It would normally require you to severe the trace for pin 12 and 13 on the main board since they are tied together. Or, as was done in the past, lifting pins 12 and 13 off the board to isolate them since the connected trace is usually located under the TIA itself or socket if one was installed?

So I'm curious to know how you are separating these from each other on your board in a manner that prevents the pins from still bleeding over to the other without that separation?

Yes for NTSC you would have to sever the trace or lift pins etc.. I'll explain that in the install vid for those that want to try the stereo version

February 22

Some updates on the production unit.

1. Added support for Stereo audio out for NTSC consoles (no auto switching yet, just solder blob)

2. Removed any loading of TIA pins so zero load on anything on TIA

3. Separated circuit completely for Chroma and Luma until final stage (had a small bit of chroma noise (40mV) mixing into Composite and S-Vid V channels due to mixing inside IC)

4. Plated through hole solder now for easier install

5. Larger and more separated cable pads

6. Adjustable chroma level to balance perfectly with your console, TV and cable, as well as tune out noise or saturate to your preference

7. Support for cheap upscalers (images below)

For upscalers the image looked really bad. On normal TVs and CRTs totally fine, but pass through an upscaler or capture card it was much worse. So I spent a lot of time really improving that.

Top of image is pre-production unit (green), bottom is production unit.

And here is pacman on an LCD that shows any issues up the most (I used the harshest monitor I could find that produced the worst visuals to perfect the image as much as possible).

These are both composite, S-Video is even nicer.

Should be 2 weeks out from being in stock. Then further feedback definitely welcome and anything I can improve on.

Will make 7800 version after this I guess.

February 22

On 2/18/2023 at 7:43 AM, -^CrossBow^- said:

I installed my preview version of the Cleancomp into my NTSC 2600jr this evening finally. Results are mixed but I do like how easy it is to install the board. It looks like the actual production versions are a tad wider and have actual vias to solder onto the TIA pins with? That is good because the solder pad only solution on these earlier made ones requires to basically just blob the solder onto the bad and pins to get them to attach well. Here are some pics of my install into my jr. I didn't quite go full no cut but and I decided to just use composite output for what will be a test bench console anyway for now. But I did connect up the s-video temporarily to capture this shot of pitfall. The colors look good and everything is pretty sharp but the output has bad vertical bars across the entire image. Composite has it as well but it isn't quite as pronounced as what s-video shows.

As I reused the RF jack for my composite, I only had to run 2 wires from the output board since ground is already attached to the old jack. The 3.5mm jack has the ground tab soldered to the side of the RCA jack and also only needed the single wire as these preview version boards are mono output only.

Those lines definitely not normal. Can you do a video while also turning the Atari's color wheel? Looks like what happens when you put an NTSC game in a PAL console or vice versa, and the chroma timing is out.

February 8

RetroSix.co.uk. Will be in Retro Consoles then Atari 2600 when out.

Price is in it will be £22 plus taxes

February 7

Did some updates, added option for stereo audio (for NTSC), added tuneable chroma voltages as some monitors prefer different levels or personal preference for saturation, cleaned up the signal even more so works on even really cheap upscalers nicely, removed more noise and over/undershoots etc...

So its in production now for version v1.0 and then any feedback we can improve next version.

Should be on website in 4 weeks approx.

January 7

On 11/21/2022 at 9:20 PM, cwilkson said:

I really like your 2 write ups. So many of these 2600 mods just kind of hack around and shove stuff together until they get something that "works". But it looks like you put real effort into studying the stock circuitry and finding weaknesses and defects and possible solutions. Great job! There are still a couple of things for you to find once you get this pass tested but most of the things you've already tried to address.

Where are you located? Are you trying to make a PAL mod? (I noticed your reference to COL phase shifting on alternate lines.) Also....Pditfall is supposed to be green not blue!

The PAL version especially interests me. Seems like there are a couple of things that would be easier to fix if you had alternating chroma phase available. I dabbled, but never really dived into it.

When you get it done I would be really interested to see how it stacks up against mine. And at which level. I made several back in the day, ranging from "pitiful hack" to Prosumer to VGA output. I also had an HDMI version mostly worked out but I became disillusioned with the hobby and walked away.

EDIT:

PS. Be careful reducing the pull up resistors. The output transistors were designed for about 1.5 mA of sink capability. You can create local hotspots on the die!

Best case...nothing happens.

Realistic case #1...The pulldowns won't go all the way "low". You may have sync issues, or image dependent luma levels. And if you use CMOS buffers they'll run hotter than they should as their inputs approach VCC/2.

Realistic case #2...(less likely, but possible)...Those local hot spots may affect your COL delay chain => unwanted color shift that is image dependent.

Bad case...(unlikely, but still! It's bad!)....The hots spots are bad enough to cause your semiconductor junctions to blur and the transistors will stop working over time. The TIA is dead from this.

Just be careful.

Thanks for the feedback, drop me your email if you want one sending got a few left.

It buffers all signals instantly now without any change to pull ups or anything altered on original console whatsoever, along with quite a lot more things. Still room for improvement also.

I am thinking about the chroma shifting/delay circuit also. Perhaps a variable inductor would be enough to delay the chroma to allow shifting? Or do you have other ideas. Definitely good to get feedback. I feel there is still room for improvements, but good to go to first release.

January 6

Sorry for the delay. Got 20+ of these out to people on our discord and who order from us. So far they all have them working and like the quality which is good.

Personally I'm still not absolutely happy with the quality and feel I can improve, but I also feel at this stage its good to get much wider feedback.

I have 5pcs left so those who feel they can test S-Video and Composite (not just one) and will be critical of the quality, ideally on both CRT and LCD, email me your address to contact@retrosix.co.uk and the first 5 I will post out tomorrow or as soon as I get the addresses.

I will make a production run after CNY once factories re-open and sell this one, as its clearly good enough based on current feedback and has worked on every variation of 2600 from 6, 4, Jr, Pal and NTSC etc... then I will sell v1 still with the desire to get feedback from everyone on improvements.

Also on that note, after 2600 what would you like to see next? 5200, 7800.. something else? I plan to make composite mods for most consoles using this same technique of completel reconstructing the signal, and have already some improvements I can do on this one but all the things learned will go towards improving the next versions

Profiles

Forums

Blogs

Gallery

Events

Store

Community Map

Posts posted by retrosix