TomelloSoulMan Posted February 26, 2022 Share Posted February 26, 2022 Hi, I'm kindly asking for some Atari guru help... Some days ago I built myself a pause mod board, following the instructions at https://circuit-board.de/forum/index.php/Thread/6083-Atari-2600-Pause-Schalter/?pageNo=1 I tried to install it on my Atari 2600 jr (PAL), not soldering directly on the TIA and 6502 pins but following instructions at https://vintagegamingandmore.com/pause-installation-guide-jr/ When I powered on my console, all I got is the screen that is seen in the video attached. Seems like sync is lost somehow. And pause was not working. The worst thing is that I undid the mod (redid the link between TIA and 6502 pins 3 and replaced the resistance) and it still shows the rolling picture. Somehow I managed to screw something badly. Upon visual inspection i can't find anything wrong, also tested connectivity and all seems the same as before installing the mod. This console has a 2 transistors AV mod which I also did myself and was working just ok before this. In fact, I'm trying to install this same AV mod board into a clone and it seems to work ok. Well, not really, because i've not figured yet where to connect the video signal in the clone, but some tests gave me a bad picture with no rolling involved at all. ¿Can someone help me? Some clues will be fine, I'm totally lost here and would like to recover my beloved JR VID_20220226_135226.mp4 Quote Link to comment Share on other sites More sharing options...
Thomas Jentzsch Posted February 26, 2022 Share Posted February 26, 2022 Does it look different when you try other the carts? Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted February 26, 2022 Author Share Posted February 26, 2022 It rolls the same with other carts. The one shown in the picture is Realsports Tennis, and you can actually see the title screen and the players. Quote Link to comment Share on other sites More sharing options...
alex_79 Posted February 27, 2022 Share Posted February 27, 2022 The fact that the console still fails after undoing the mod suggests a fail of one IC (the TIA, likely). Did you take some basic anti-static precautions when handling the board and components (e.g. touching a metal part of your PC case to discharge yourself)? Build-up of static electricity is common, especially during the winter because of the dryer air, and you can destroy the chip just by touching the board. Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted February 27, 2022 Author Share Posted February 27, 2022 Woops... No, I didn't take anti-statics measures. Neither I did the other several times I manipulated this console for installing different AV mods (involving soldering cables directly on TIA pins) and other shenanigans. Maybe this time i've just hit the jackpot... I was suspecting this from the start, but I would like to be completely sure that the TIA is fried. ¿Are there any tests I can do for this? Because of the nature of this test, and what I can see on screen, I would suspect some sync fault between the TIA and the 6507, being the TIA unable to put the 6507 in a halt state and thus losing sync. I might well be wrong tho. Quote Link to comment Share on other sites More sharing options...
alex_79 Posted February 28, 2022 Share Posted February 28, 2022 On 2/27/2022 at 11:32 AM, TomelloSoulMan said: Because of the nature of this test, and what I can see on screen, I would suspect some sync fault between the TIA and the 6507, being the TIA unable to put the 6507 in a halt state and thus losing sync. I might well be wrong tho. That seems indeed the issue by looking at the video: the horizontal sync seems to work as the display has definite left and right borders and you can also see the "HMOVE bars" on the left all correctly aligned. The CPU "RDY" pin is normally pulled low by the TIA to halt the CPU until the start of the next scanline (This is obtained in the program by writing to the TIA "WSYNC" register). Without that functionality, the timing of writes to sprite graphics and motion registers that would have occurred on the next scanline, will instead happen immediately, messing up the display. Moreover, without being able to halt the CPU, the code will take less time to execute, resulting in less scanlines than intended and so in loss of vertical sync. There's nothing else between the TIA and CPU "RDY" pin, other than the trace that you was supposed to cut for the pause mode, and the pullup resistor. If you're sure that you restored that circuit correclty (double check all the connections with a multimeter), then the only way to determine the issue is to watch the CPU RDY pin with an oscilloscope or a logic analyzer. If it's stuck "high" (+5V), then the TIA is busted, if instead you can see activity on it, the CPU itself might be the problem. Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted February 28, 2022 Author Share Posted February 28, 2022 12 minutes ago, alex_79 said: There's nothing else between the TIA and CPU "RDY" pin, other than the trace that you was supposed to cut for the pause mode, and the pullup resistor. If you're sure that you restored that circuit correclty (double check all the connections with a multimeter), then the only way to determine the issue is to watch the CPU RDY pin with an oscilloscope or a logic analyzer. If it's stuck "high" (+5V), then the TIA is busted, if instead you can see activity on it, the CPU itself might be the problem. Thanks a lot for the tip! I think I'll do a quick test by soldering "flying" cables between the two RDY pins and the pullup resistor, ignoring the tracks on the board. I swear i've at least triple-checked connectivity and all seemed ok, but it wouldn't be the first time I check right connectivity and then must solder jumper cables in order to make things work. If it still doesn't work, i'll try to reach some friends to lend me an oscilloscope. I'll post results maybe in a few days, cause i'm still involved in trying to find the right spot to get video out from my other console (Kashiwa 3-chip clone). Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted March 3, 2022 Author Share Posted March 3, 2022 Well, I'm crying out loud right now. I tested some jumpers between pin 3 of both ICs, with the pullup resistor in the middle connected to +5V, and it shows the same problem. I don't have an oscilloscope, but I measured voltage on pin 3 in the TIA and it fluctuated between 3.8 and 4.2V (cheap multimeter with a low pooling rate i guess). Then I cut all the jumpers, disconnecting everything, and got same behavior. So I guess it's time to try to find some working ICs in the wild, which will be difficult I guess. I have a 3-chip clone that could be the donor, but since I managed to AV mod it, I prefer not touching it because it's better to have 1 *meh* working console than taking the risk to have 0 working consoles. I don't have a proper unsoldering tool for removing ICs. Would it be feasible to lift some pins in the broken ICs (I'm thinking about Vcc) and installing the new ones just above them? Quote Link to comment Share on other sites More sharing options...
+Stephen Moss Posted March 4, 2022 Share Posted March 4, 2022 22 hours ago, TomelloSoulMan said: I don't have a proper unsoldering tool for removing ICs. Would it be feasible to lift some pins in the broken ICs (I'm thinking about Vcc) and installing the new ones just above them? Desolder pumps are quite cheap and easy to get hold of. I would advise again piggybacking possible good IC on top of possible bad one for two reasons.... 1) The unpowered IC could draw some power through it I/O pins which could affect operation as it can then partially work depending which pins are connected high hand when causing additional problems. 2) If the unpowered device is faulty the fault may result in no change as it effects (and may even damage) the good IC and so the results you get could be misleading. Therefore it is better to remove the suspect device completely and if possible solder and IC socket in its place as then you can simple insert & remove the IC from the socket in future without soldering. Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted March 4, 2022 Author Share Posted March 4, 2022 1 hour ago, Stephen Moss said: Desolder pumps are quite cheap and easy to get hold of. I would advise again piggybacking possible good IC on top of possible bad one for two reasons.... 1) The unpowered IC could draw some power through it I/O pins which could affect operation as it can then partially work depending which pins are connected high hand when causing additional problems. 2) If the unpowered device is faulty the fault may result in no change as it effects (and may even damage) the good IC and so the results you get could be misleading. Therefore it is better to remove the suspect device completely and if possible solder and IC socket in its place as then you can simple insert & remove the IC from the socket in future without soldering. Thank you very much, these are indeed good points. I've ordered two 6507 chips which will arrive some time between now and two months in the future (Shanghai shipping) and I'll do things right. I'll look forward for some good desolder pumps and, if I can't buy one, I'll try to borrow it from some friend. I've never desoldered ICs, should the pump be particularly powerful or would a cheap one be enough? Quote Link to comment Share on other sites More sharing options...
+Stephen Moss Posted March 7, 2022 Share Posted March 7, 2022 On 3/4/2022 at 10:17 AM, TomelloSoulMan said: I'll try to borrow it from some friend. I've never desoldered ICs, should the pump be particularly powerful or would a cheap one be enough? As you mention power I presume you are thinking of a powered de-solder station. As the 2600 boards are not through hold plated you should be able to do it with a standard manual unit like this. You can probably find some videos on line to should who to use them, the common mistake make by those unfamiliar with de-soldering when using these is to keep them too far from the joint so that the solder has already cooled and started to re-solidify when trying to extract it. I recommend holding the pump to one side of the joint at an angle of approximately 45 degrees, apply the tip of your iron to the other side of the joint to melt the solder then as you remove the iron, tip the de-solder pump upright over the joint and press the button to suck up the solder in an almost simultaneous action (practive with the iron off first). Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted March 7, 2022 Author Share Posted March 7, 2022 2 hours ago, Stephen Moss said: As you mention power I presume you are thinking of a powered de-solder station. As the 2600 boards are not through hold plated you should be able to do it with a standard manual unit like this. You can probably find some videos on line to should who to use them, the common mistake make by those unfamiliar with de-soldering when using these is to keep them too far from the joint so that the solder has already cooled and started to re-solidify when trying to extract it. I recommend holding the pump to one side of the joint at an angle of approximately 45 degrees, apply the tip of your iron to the other side of the joint to melt the solder then as you remove the iron, tip the de-solder pump upright over the joint and press the button to suck up the solder in an almost simultaneous action (practive with the iron off first). Thanks a lot for the tip, I have one of those manual de-solderers and now I know I've been using it wrong. Since I ordered some DIP28 sockets and they'll arrive much earlier than the 6507, I'll try to remove the old 6507 following your advice and install one of those sockets in a few days. Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted March 25, 2022 Author Share Posted March 25, 2022 Coming back to this after some time (China shipping takes a while). I've so far replaced all three chips (CPU, TIA and RIOT). TIA was taken from a nice and simple clone that I got almost for free, and I've socketed both the clone (making a nice test board) and the Jr. The clone is the brown board in the attached pictures. First, I replaced the CPU. On power up, I got a black screen most of the time, sometimes with sound (constant pitch) and sometimes with some random patterns as shown in the pictures. This happened too when I replaced the TIA and the RIOT, and strangely, with no CPU at all! We can blame the CPU for the initial no-sync problem, as the clone shows it with the original Jr CPU. At least, the RDY signal is not working. No problem, since I have two spare CPUs that work perfectly. But this new problem is driving me nuts: I'm pretty sure I've done something wrong while replacing the CPU at first. I've checked thoroughly that all the CPU connections go where they're intended, no open tracks and no shorts (tested pin by pin). I'm started to think that another component like some transistor or something can be broken too, but I don't know where else to look at. There's something that is making the CPU not to work properly, but I don't know what... Quote Link to comment Share on other sites More sharing options...
alex_79 Posted March 26, 2022 Share Posted March 26, 2022 Just guessing... Try checking pin 1 of the CPU with a multimeter while the console is ON. If it's LOW (~0V), the culprit might me the CD4050, as one of the gates is part of the CPU reset circuit. If the pin is low, and you don't have a spare CD4050 on hand, you could try clipping the output pin of that gate, as the console might still work without it. That would be pin 4 according to the schematic, but double check on your board with a multimeter: it's the one that shows continuity with the CPU pin 1. This would be a temporary solution until you get a new CD4050: it's a common part still manufactured and it's cheap. Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted March 30, 2022 Author Share Posted March 30, 2022 Ok, thanks for the tip! I was suspecting this could be the culprit... However, I've done the tests and RES line is at a steady 5.02V (4.96V on CD4050 pin 5 iirc). VCC seems ok too. I don't know if there are more tests I can do with a simple multimeter, and I'm afraid I'll need an oscilloscope to check all reference values as stated in the Atari 2600 service manual, which is not easy to follow for me as it's not oriented for Jr models, and component names are different. Does anybody know if there's some service manual adapted for Jr models? Quote Link to comment Share on other sites More sharing options...
alex_79 Posted March 30, 2022 Share Posted March 30, 2022 3 hours ago, TomelloSoulMan said: Ok, thanks for the tip! I was suspecting this could be the culprit... However, I've done the tests and RES line is at a steady 5.02V (4.96V on CD4050 pin 5 iirc). VCC seems ok too. I don't know if there are more tests I can do with a simple multimeter... If you haven't already, try checking continuity between the CPU and the cart port (a cold solder joint in one of the address or data lines will cause the symptoms you're experiencing. Check also the connections between the CPU and the other 2 chips (TIA,RIOT) as per the schematic above. 3 hours ago, TomelloSoulMan said: ... and I'm afraid I'll need an oscilloscope to check all reference values as stated in the Atari 2600 service manual, which is not easy to follow for me as it's not oriented for Jr models, and component names are different. Does anybody know if there's some service manual adapted for Jr models? While Junior service manuals do exist (I've seen pictures of a couple of them in past), no one ever took the time to scan one and put it online. You can find service manuals for 6-switch and 4 switch NTSC models, and 6-switch PAL model: Atari_2600_PAL_Service_Manual.pdf Combined with the schematic posted above, it might still be useful for troubleshooting on the Jr console, as the differences between the various models are minimal. Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted April 1, 2022 Author Share Posted April 1, 2022 On 3/30/2022 at 5:12 PM, alex_79 said: If you haven't already, try checking continuity between the CPU and the cart port (a cold solder joint in one of the address or data lines will cause the symptoms you're experiencing. Ok, I've done that again. This time I've tested with a bare cart connected to the port to make sure all the pins are making contact. Everything seems to be connected where it's intended to be, with no shorts. At this time the only part I'm suspecting of is the clock circuit (see attached image). I'm not able to check the signal values properly with a 0.5s polling time multimeter, but it's a suggested culprit of this kind of failure in the service manual, and as I have some spare 3906 transistors, I guess it's worth the hassle to replace them and test. What I don't have is a 3.54MHz xtal (as stated in the diagram), would a 3.579545 one do the trick? On 3/30/2022 at 5:12 PM, alex_79 said: You can find service manuals for 6-switch and 4 switch NTSC models, and 6-switch PAL model Thanks a lot for the PAL service manual, will take a dive into it right now. Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted April 1, 2022 Author Share Posted April 1, 2022 And now comes when I don't forget the image and actually attach it ? Quote Link to comment Share on other sites More sharing options...
alex_79 Posted April 1, 2022 Share Posted April 1, 2022 (edited) This screenshot makes me believe that the clock is fine. Those vertical bands are generated by the TIA. That display is lacking the vertical synchronization, because the code in the cartridge is not being executed, and that part of the signal is under software control. But there's the horizontal sync (as the bars are aligned) and that means that the TIA is working and is therefore receiving the 3.54MHz clock signal. The TIA provides the clock to the CPU. In the schematic is labeled "O0" (pin 4) Edited April 1, 2022 by alex_79 corrected info on the TIA clock pin Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted April 1, 2022 Author Share Posted April 1, 2022 7 minutes ago, alex_79 said: Those vertical bands are generated by the TIA. That display is lacking the vertical synchronization, because the code in the cartridge is not being executed, and that part of the signal is under software control. But there's the horizontal sync (as the bars are aligned) and that means that the TIA is working and is therefore receiving the 3.54MHz clock signal. This might be out of my understanding, I guessed the horizontal sync was working because of the PALI/PALS signals are present in the PAL TIA, connected to the schematic part where the 4.43MHz xtal resides. I didn't know well if the 3.54 MHz signal was used by the TIA in some way or only carried by it to the CPU and RIOT through O0 and O2, as you say. I have a lot of things to learn yet... I'll do some more tests this evening with the bare cart to make sure it's receiving the 5V signal and everything is ok. It all points to what you previously said, the console behaves as if there was no cart connected at all, but all connections seem ok. I'll post results later, but I'm slowly losing faith in that I can repair this console, which is sad because it's my only original one. At least I can take out the video signal A/D converter circuitry and use it in one of my clones, to make it a proper gaming/streaming platform... Quote Link to comment Share on other sites More sharing options...
alex_79 Posted April 1, 2022 Share Posted April 1, 2022 16 minutes ago, TomelloSoulMan said: I guessed the horizontal sync was working because of the PALI/PALS signals are present in the PAL TIA, connected to the schematic part where the 4.43MHz xtal resides. That's used for the PAL color generation. If you remove that crystal from a PAL 2600, it will still work, but in B&W only. The 3.54MHz clock is what drives the TIA. Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted April 1, 2022 Author Share Posted April 1, 2022 I fixed it! Dang, it was a short in the end! I swear I've tested each single pin with each other before, but didn't detect this. I've done it yet again and BOOM, there was a short between A2 and D2 near the CPU socket. It seems like I scraped the tracks cover while unsoldering, and some solder leftovers did the rest. I can't stress enough how thankful I am for all your help, Alex, and how shameful I am too for such a silly issue. Now I have another problem: reset and game select switches are not working, but it's an easy fix because the jumper is working ok (so, no RIOT related) and the end of the tracks of the plastic band that's inserted into that jumper seem to be very scratched (too much wiggling in this process, I guess). I'll figure out how to make a replacement, with some push buttons or something, but the most important issue is fixed! Yay! 8 hours ago, alex_79 said: That's used for the PAL color generation. If you remove that crystal from a PAL 2600, it will still work, but in B&W only. The 3.54MHz clock is what drives the TIA. Good to know, I'll add that to the heap of things I've learned in this process 1 Quote Link to comment Share on other sites More sharing options...
alex_79 Posted April 2, 2022 Share Posted April 2, 2022 I'm glad you got it working! 10 hours ago, TomelloSoulMan said: I swear I've tested each single pin with each other before, but didn't detect this. I've been there too... 10 hours ago, TomelloSoulMan said: I'll figure out how to make a replacement, with some push buttons or something, Yes, that's an easy fix. If you don't mind about the original look, you can install two normally-open panel mount switches somewhere on the case. There are a few examples in the forums. There are also low profile tactile switches that will fit under the existing plastic buttons. E.g. see here: https://atariage.com/forums/topic/296786-cheap-diy-2600-junior-select-reset-tactile-momentary-fix/ 1 Quote Link to comment Share on other sites More sharing options...
TomelloSoulMan Posted April 2, 2022 Author Share Posted April 2, 2022 3 hours ago, alex_79 said: There are also low profile tactile switches that will fit under the existing plastic buttons. Thanks for the tip and the link, I'm going to look into this as it seems a pretty neat solution. Quote Link to comment Share on other sites More sharing options...
Like us on Facebook
Follow us on Twitter
AtariAge on Instagram
Watch us on YouTube
Chat on Discord
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.