RevEng Posted January 9, 2019 Share Posted January 9, 2019 Ah, that's too bad. I was hoping for a less intrusive fix. Thanks for trying! Quote Link to comment Share on other sites More sharing options...
MissCommand Posted January 9, 2019 Share Posted January 9, 2019 (edited) Yeah, I guess it is only a fix for those of us who want a working AtariVox+ and don't care about the Vectrix side of things. Especially since the original AtariVox is not available to purchase anymore. Plus, we did narrow the problem down to the PIC, so hopefully that will help in coming up with a proper fix that will keep the 2600 *and* Vectrix users happy! The DIP switches should be used to configure the PIC, not bypass it (i.e. passing the pulses through in the 2600 mode and keeping Vectrix as is. The bypassing is where the problems come in). I think the fix will be to reroute all the lines into the DIP switch and reprogram the PIC to accept the configuration from it. Just a theory though! Edited January 9, 2019 by MissCommand 3 Quote Link to comment Share on other sites More sharing options...
RevEng Posted January 9, 2019 Share Posted January 9, 2019 For sure! This pic removal fix is big news. (hoping you saw me saying as much, on the previous page) Crossbow made an observation earlier about the PRGE original atarivox units not encountering the glitch. You've confirmed and narrowed that down to the PIC. Well done! Quote Link to comment Share on other sites More sharing options...
+-^CrossBow^- Posted January 9, 2019 Author Share Posted January 9, 2019 (edited) Yeah..and there is obviously too much attached to the PIC as simply lifting pin14 of the PIC (Vdd +5) wasn't enough to make it work still. So even with the PIC effectively dead without power, there is still something with it preventing it from working properly. I was hoping that I could simply lift pin14 and then attach a small wire off that pin to the +5 off the PIC controller connector since it always has +5 available. This way I could put a header connector off the wire from pin14 and just plug it in along with dip settings...but nope...still causes the AVOX+ to reset when the Wiz shows up in Dungeon Stalker. I don't have quite the tools needed to remove the pic completely in my case unless I clip the pins. I was able to lift pin14 by removing the speakjet and socket for it off the PCB allowing me room to wick the solder off the pin that way. Edited January 9, 2019 by -^Cro§Bow^- 2 Quote Link to comment Share on other sites More sharing options...
RevEng Posted January 9, 2019 Share Posted January 9, 2019 I read MissCommands' post to mean the PIC is being back powered from signal on another pin, which is common enough with PICs. Quote Link to comment Share on other sites More sharing options...
+-^CrossBow^- Posted January 9, 2019 Author Share Posted January 9, 2019 Pin14 on the PIC is also connected to the 3 +5 connections on the speakjet. Quote Link to comment Share on other sites More sharing options...
RevEng Posted January 9, 2019 Share Posted January 9, 2019 Is that through one of the dip switches, or is it permanently connected? Quote Link to comment Share on other sites More sharing options...
+-^CrossBow^- Posted January 9, 2019 Author Share Posted January 9, 2019 (edited) The Vdd for the PIC is always connected to the +5. Only the ground seems to be affected with Dip 3 on or off. I'm sure there are other pins connected to the dip of course but my concern was just the voltage and grounds. There is only one pin on the PIC for voltage according to the pinouts for it so again, I had hoped that if I simply killed the +5 to it, that would be enough to render it dead in circuit enough to allow the Avox+ to work in standard Atari mode only. There is still something else in the PIC just being connected that is causing these issues. I'm not familiar enough with the internals of the PIC to suss it out, but simply killing voltage to it isn't enough. According to my meter there is always continuity with the +5 also to the speakjet on pins 11(Reset), 13(Mode Select), and 14(+5). Or at least there is on my AVox+ pcb. Also found it is a bad idea to have switch 2 on in certain instances as it will cause the 7800 instantly go to scrambled lines screen condition like you get with a dirty cart along with a buzz in the speakers. I'm assuming something is being shorted in that case that is backfeeding through the cartridge connector. Edited January 9, 2019 by -^Cro§Bow^- 1 Quote Link to comment Share on other sites More sharing options...
RevEng Posted January 9, 2019 Share Posted January 9, 2019 Got you. Thanks for the correction - I had wrongly assumed that +5v was also switched. I really hesitate to mention this, but if you can remove +5v from the joystick in, and isolate+feed the +5v pin on the speakjet, I'm pretty sure the unit should operate like a old atarivox, minus the eprom saving. (I put together a breadboard equivalent of the original atarivox a while back, using speakjet reference designs. I'm not clear on the detailed role of the PIC in atarivox+, but MissCommand has shown that taking it out of the picture works.) The isolate part probably involves bending the pin to avoid the socket, which is why I hesitate to mention it. That old tin won't take kindly to bending, and could easily break off. So this isn't a suggestion or request, but maybe it will inspire some safer idea. Richard H is on the thread. Hopefully he'll check in after he sees the updates. 1 Quote Link to comment Share on other sites More sharing options...
+-^CrossBow^- Posted January 9, 2019 Author Share Posted January 9, 2019 (edited) I really hesitate to mention this, but if you can remove +5v from the joystick in, and isolate+feed the +5v pin on the speakjet, I'm pretty sure the unit should operate like a old atarivox, minus the eprom saving. (I put together a breadboard equivalent of the original atarivox a while back, using speakjet reference designs. I'm not clear on the detailed role of the PIC in atarivox+, but MissCommand has shown that taking it out of the picture works.) The isolate part probably involves bending the pin to avoid the socket, which is why I hesitate to mention it. That old tin won't take kindly to bending, and could easily break off. So this isn't a suggestion or request, but maybe it will inspire some safer idea. Richard H is on the thread. Hopefully he'll check in after he sees the updates. I mispoke above... I meant Joystick connector not cartridge connector... Could probably just cut the trace on the pcb from the +5 pin on the joystick connector and then wire that to the vin on the speakjet to get it working. But don't the amp ic need that power as well to operate? Again I already lifted the +5 vin to the PIC and that wasn't enough on its own. Other connections from the PIC whether powered up or not still cause the issues to occur. Or at least it does with Dungeon Stalker which, is the primary game I use since it pretty much fails in the exact same place all the time on the 7800s I've tested it with. When I first lifted pin14 from the PIC I was able to play Dungeon Stalker with speech and it actually didn't screw up until about halfway through the second board. And in fact I was excited by this because I thought I had found a semi simple solution for me. But then as I stated after picking up the goblet treasure and it said 'Jackpot' it played a strange beep and then nothing after that. So if anything, killing the voltage to the PIC only delayed the issue in my case. But that was a one time deal as it continued to mess up around the point of the wizard showing up at the end of the first board like normal there after. Edited January 9, 2019 by -^Cro§Bow^- Quote Link to comment Share on other sites More sharing options...
RevEng Posted January 9, 2019 Share Posted January 9, 2019 Yeah, you're right on the op amp - forgot about that, since it wasn't part of my bread-vox. Its also a fair point that the other connections remain to the pic and may be the source of the headache. Your "delayed" result was probably just chance. During all of the testing for fixes, people have had similar promising results on one long section of a play-through, only to re-encounter the issue later on, or in the next attempt. If it's a matter of the PIC interfering, I'm thinking that a particularly long sequence of data bits allows the PIC to fully boot up and start doing vecvox stuff. Not sure if that's high or low bits, and really it's just a guess. Richard would be able to say if that's nonsense or not. Quote Link to comment Share on other sites More sharing options...
Jinks Posted January 10, 2019 Share Posted January 10, 2019 (edited) Ok once again I got my die grinder humming. What pins do I grind off? : Edited January 10, 2019 by Jinks Quote Link to comment Share on other sites More sharing options...
+SmittyB Posted January 10, 2019 Share Posted January 10, 2019 Ok once again I got my die grinder humming. What pins do I grind off? : If you grind off all of them you can then get a child with a cold to make the sounds for you. "Out of awwows... 'tchoo!" (I'll leave a serious answer to the others) 1 Quote Link to comment Share on other sites More sharing options...
Jinks Posted January 10, 2019 Share Posted January 10, 2019 If you grind off all of them you can then get a child with a cold to make the sounds for you. "Out of awwows... 'tchoo!" (I'll leave a serious answer to the others) Just a joke from the concerto cart days. Quote Link to comment Share on other sites More sharing options...
emmanuelf Posted January 11, 2019 Share Posted January 11, 2019 Instead of simply lifting pin14 of the PIC, did you tried to ground pin14 ? It will kill all back powering. But it could be not sufficient to "fix" the problem. If there is some internal leak capable of back powering the chip, the leak could be one source of the speakjet disturbing. It's worth a try. If it does not fix the problem, we should get the complete schematic and with the datasheet and errata of the pic chip find witch pin need proper additional care / signal conditioning under no power condition to not disturb the connected signal. Quote Link to comment Share on other sites More sharing options...
RevEng Posted January 11, 2019 Share Posted January 11, 2019 Crossbow found that it's getting +5v supplied on pin 14, and also on a few other pins. If he grounds pin 14, then there may be a short between 14 and those other pins, or at the very least some very unplanned reverse current flow. Quote Link to comment Share on other sites More sharing options...
+-^CrossBow^- Posted January 11, 2019 Author Share Posted January 11, 2019 I think we would be good in that regard actually. Because it is only connected at the pad. If I run the wire from the pin itself and ground that it shouldn't be a problem. However, you state this AFTER I've already soldered that pin back down again. I'm not sure that little pin will take being lifted again without breaking... Quote Link to comment Share on other sites More sharing options...
KatTech70 Posted January 12, 2019 Share Posted January 12, 2019 Hello All! I’m a friend of MissCommand. I have been reviewing this issue with her. A quick warning. Do Not turn on DIP switches 1 and 2 at the same time. This causes a short between +5V and GND. Bad things could happen. In the current circuit layout, instead of disconnecting power, I recommend disconnecting the data pins. The issue is that the “disabled” PIC is still influencing the serial data, so the data lines (PIC pins 9, 11, 12 and 13) should be disconnected from the circuit. I have been looking at the schematic and don’t see an “easy” fix that keeps the Vectrex modes. 3 Quote Link to comment Share on other sites More sharing options...
MissCommand Posted January 30, 2019 Share Posted January 30, 2019 It's been pretty quiet here, so I thought I would provide an quick update on mine. All has been fine since removing the PIC chip! Been playing lots of AtariVox games including the Wizard of Wor from Champ... and I found another Wizard of Wor hack from Nukey Shay and others using AtariVox... and the Wizard talks continuously! So much so, that in between level changes, he keeps on chatting to empty the buffer. Anyway, it is a somewhat easy fix if you are not concerned about the Vectrex part of the AtariVox+. Quote Link to comment Share on other sites More sharing options...
+SmittyB Posted January 30, 2019 Share Posted January 30, 2019 It's getting very tempting to try cutting the PIC out, but I haven't so far because if I somehow wreck the Vox it would be a costly mistake. I did find that the issue could be triggered repeatedly first time by a specific short phrase, so it may be a particular sequence of bytes or specific timings that set it off. I wish I had noted the data I was using but it had a 700ms pause (03) that caused the issue and if I changed the pause it would work. I was able to use the same pause with a different combination of commands so maybe it was a coincidence. Quote Link to comment Share on other sites More sharing options...
RevEng Posted January 30, 2019 Share Posted January 30, 2019 @MissCommand: thanks for the update. I'm glad that things are working perfectly for you! AtariVox is a great little addition to the console. @SmittyB: Even though MissCommand's results are dramatic, I'd say listen to that little voice. I'd hate for you to kill your vox too. Richard H is looking into the issue presently, to see if something less dramatic can be done. About your specific phrases, it does remind me of some people reporting the Wizard speech as particularly problematic. The thing is, it seemed to be a probability thing; removing the wizard speech just delayed the problem, which would just rear it's head later with other speech that had a lower probability of triggering the issue. Quote Link to comment Share on other sites More sharing options...
+-^CrossBow^- Posted January 30, 2019 Author Share Posted January 30, 2019 Yes when using the avox testing rom I did find that on the 7800 some of the phrases would trigger the reset or overflow on the buffer within just one or two times of executing them quickly. Others didn't seem to trigger it at all or took much more work to do so. The 2600 was even more difficult to trigger, but could still eventually be triggered. I also recall that I had found that triggering the phrases of the test with the joystick would cause the issues much quicker than just pressing the fire button to trigger them. But I might have that backwards..would have to go back in postings to review that. 1 Quote Link to comment Share on other sites More sharing options...
MissCommand Posted January 30, 2019 Share Posted January 30, 2019 (edited) It's getting very tempting to try cutting the PIC out, but I haven't so far because if I somehow wreck the Vox it would be a costly mistake. I did find that the issue could be triggered repeatedly first time by a specific short phrase, so it may be a particular sequence of bytes or specific timings that set it off. I wish I had noted the data I was using but it had a 700ms pause (03) that caused the issue and if I changed the pause it would work. I was able to use the same pause with a different combination of commands so maybe it was a coincidence. I agree with RevEng, I would not cut out the chip either. Removing that specific chip requires a hot air station, KatTech is going to make a post for those interested. As far as the speech, I could get it to fail reliably with "Spell and Speak" by adding the "?" at the end of any sentence. So codes to change the pitch might be at least one thing that triggers the data lines to do weird things. I need to go back and see what that does now that he chip has been removed. I will try that tonight and see what happens and report back. Edited January 30, 2019 by MissCommand 2 Quote Link to comment Share on other sites More sharing options...
KatTech70 Posted January 31, 2019 Share Posted January 31, 2019 I removed the surface mounted PIC chip with a hot air re-work station. Here are my steps: De-soldered the ICSP header. Shielded the adjacent ICs with Kaowool and secured with Kapton tape. Pre-heated board to 180C for two minutes Heated the chip leads to 320C and removed chip with tweezers. Replaced the ICSP header. Quote Link to comment Share on other sites More sharing options...
MissCommand Posted January 31, 2019 Share Posted January 31, 2019 Okay, I did try the Spell and Speak this morning with the "?" at the end of a phrase... and it still puts it in a weird state where it needs to be reset. So apparently, that particular problem has nothing to do with the PIC chip. But otherwise, I have not seen any problems with it using any phrases or any other games using AtariVox (at least that I have encountered yet). 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.