Delicon Posted July 28, 2006 Share Posted July 28, 2006 I have my TIA A/V outputs (LUM, COL, SYNC, AUD) running into a CPLD. The CPLD doesnt do anything to them, just sends them right out the other side into the VCS. All the other TIA signals just go into the VCS, no CPLD. Any ideas why this doesnt work. I thought maybe the VCS draws more current than the CPLD can handle for A/V. So I stuck a transistor on each line to boost the current, still nothing. Assuming all my wiring is correct (been over it a hundred times), any other reason why a CPLD cant buffer the TIA A/V signals? Vern Quote Link to comment Share on other sites More sharing options...
supercat Posted July 28, 2006 Share Posted July 28, 2006 I have my TIA A/V outputs (LUM, COL, SYNC, AUD) running into a CPLD. The CPLD doesnt do anything to them, just sends them right out the other side into the VCS. All the other TIA signals just go into the VCS, no CPLD. The luminance and sync outputs and are open-collector digital. The audio outputs are open-collector analog. I think the color output is open-collector digital, but I'm not sure. You need to configure the CPLD so that the output enable for each output is gated appropriately. For example: lum0out = lum0in lum0out.oe = /lum0out You'll need pullups on the video pins feeding into the CPLD. The analog can't be fed into the CPLD, except at full volume, unless you use some sort of PWM arrangement. For example, if you had two sets of four CPLD outputs serving as a high-speed counters, they were configured so as to only output when high (float when low), and they had binary-weighted resistors attached to the audio pins, then if the resistor values were chosen suitably you could get a PWM signal into the CPLD that would indicate the audio volume. Note that it would probably be better to use five bits of counter and arrange things so that all 1's on the PWM was strong enough to prevent even maximum volume from pulling down the pin. This would ensure that the full dynamic range of the TIA was captured. I'm not sure what you're doing, but if you have enough pins, you could also include in the CPLD circuitry to detect writes to AUDV0 and AUDV1. That would require A0-A5, A12, R/W, phi2, and D0-D3. Thirteen pins. What are you doing this for? Quote Link to comment Share on other sites More sharing options...
Delicon Posted July 28, 2006 Author Share Posted July 28, 2006 (edited) The luminance and sync outputs and are open-collector digital. The audio outputs are open-collector analog. I think the color output is open-collector digital, but I'm not sure. I am an idiot, thanks for pointing that out. I completely missed where they blew up the open collector block on the schematics. Its a little unfortunate that the audio is analog, I thought the numbers were metal sizes used in the transistors. It seems that they are resistor values for the built in resistor ladder. My CPLD cant reproduce that internally, so I will just hold off on audio until later. You need to configure the CPLD so that the output enable for each output is gated appropriately. For example: lum0out = lum0in lum0out.oe = /lum0out I dont understand this. Once the pull resistors are in place, I just need to make sure the value on lum0out is the same as lum0in. How does the OE come in. I float lum0out when lum0in = '1' and drive '0' when lum0in = '0', is that what your saying? You'll need pullups on the video pins feeding into the CPLD. The analog can't be fed into the CPLD, except at full volume, unless you use some sort of PWM arrangement. For example, if you had two sets of four CPLD outputs serving as a high-speed counters, they were configured so as to only output when high (float when low), and they had binary-weighted resistors attached to the audio pins, then if the resistor values were chosen suitably you could get a PWM signal into the CPLD that would indicate the audio volume. Note that it would probably be better to use five bits of counter and arrange things so that all 1's on the PWM was strong enough to prevent even maximum volume from pulling down the pin. This would ensure that the full dynamic range of the TIA was captured. I am holding off on audio for now. Maybe later once feasibility becomes more clear. What are you doing this for? I am setting up my testbench for a TIA replacement. I have the real TIA running in parallel to my implementation. I can use the CPLD to multiplex my design with the real TIA. I figure it will be easier to debug if I can see the real TIA outputs at the same time as my outputs in real time. Thanks, you helped a ton. Nobody get your hopes up about this, its really just a dream. Vern Edited July 28, 2006 by Delicon Quote Link to comment Share on other sites More sharing options...
supercat Posted July 28, 2006 Share Posted July 28, 2006 I dont understand this. Once the pull resistors are in place, I just need to make sure the value on lum0out is the same as lum0in. How does the OE come in. I float lum0out when lum0in = '1' and drive '0' when lum0in = '0', is that what your saying? Precisely. Quote Link to comment Share on other sites More sharing options...
Delicon Posted July 28, 2006 Author Share Posted July 28, 2006 (edited) Precisely. I just got so excited about the open collector stuff that I didnt process your statement, I immediately started forming equations of my own. Mine was: lum0out <= 'Z' when lum0in = '1' else '0'; The two didnt match exactly, so I didnt understand. Had I taken one second, I would have seen that the solutions are identical. Thanks again, Vern Edited July 28, 2006 by Delicon Quote Link to comment Share on other sites More sharing options...
Delicon Posted July 28, 2006 Author Share Posted July 28, 2006 I made the changes and everything works fine, except audio. Thats for a later time. Bonus, because of the open collectors I can run the CPLD at 3.3V. Vern Quote Link to comment Share on other sites More sharing options...
djmips Posted July 28, 2006 Share Posted July 28, 2006 Offtopic-> sorry <- but do either of you hardware types know what is the best circuit for adding a cheap hand built composite output to a VCS? I've read over all of the different modifications and I've never been satisified that I understand the pros and cons and it doesn't seem that most of the people doing these circuits really know what they're doing... at least I don't know what their doing. Quote Link to comment Share on other sites More sharing options...
Delicon Posted July 28, 2006 Author Share Posted July 28, 2006 Offtopic-> sorry <- but do either of you hardware types know what is the best circuit for adding a cheap hand built composite output to a VCS? I've read over all of the different modifications and I've never been satisified that I understand the pros and cons and it doesn't seem that most of the people doing these circuits really know what they're doing... at least I don't know what their doing. I am mainly digital and software. When I play with analog, if it doesnt smoke or smell, its a success story. I wish I knew a decent mod too. I do all my development on an RF system, and the noise can be too much to take sometimes. Vern Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 30, 2006 Share Posted July 30, 2006 I thought the numbers were metal sizes used in the transistors. It seems that they are resistor values for the built in resistor ladder. My CPLD cant reproduce that internally, so I will just hold off on audio until later. Vern, that's exactly right. Those are sizes for the transistors. Each audio channel has 4 transistors connected in parallel, open drain (open drain -> nMOS version of open collector). Because the devices act as current sources, and they're ratiod in 2:1 fashion, you get a nice "linearly" stepped current source when you add them in parallel. Feed that through a resistor and you get a stepped voltage, referenced to VCC in this case. Cheap D2A converter. If you want to buffer things, you could reconvert to digital, feed that through your logic and then re-reconvert to analog. I don't know if there's much point for the audio signals though.... -Chris Quote Link to comment Share on other sites More sharing options...
cwilkson Posted July 30, 2006 Share Posted July 30, 2006 Offtopic-> sorry <- but do either of you hardware types know what is the best circuit for adding a cheap hand built composite output to a VCS? I've read over all of the different modifications and I've never been satisified that I understand the pros and cons and it doesn't seem that most of the people doing these circuits really know what they're doing... at least I don't know what their doing. Yeah, I can help. What's your budget? That will probably guide you more than anything else assuming you have decent assembly skills. Basic advice on a couple of mods, based on quality level: (note 1: these are my opinions, but technically speaking, they're pretty accurate.) (note 2: I haven't seen the 8bitdomain board perform, so I'm not considering it here). 1) avoid the Heckendorn mods like the plague. They don't produce correct video and you have to be lucky for it to work with your TV/monitor. 2) the CD4050 mod isn't bad. Easy to build, ok quality. Still not spec video, but "close enough". Nice middle of the road mod, but there are multiple versions, so be careful which one you use. 3) the CyberTech boards are by far the best. It's the only design that produces correctly engineered video. But it's not public domain because it's harder to build, so you'd have to buy one, and they're (temporarily) hard to get a hold of. Working to fix that. RSN. If you have specific questions, I'd be happy to answer them And once I know your budget, I might have better answers... -Chris Quote Link to comment Share on other sites More sharing options...
supercat Posted July 30, 2006 Share Posted July 30, 2006 Basic advice on a couple of mods, based on quality level: One thing I've wondered about as a mod would be cutting out the audio modulator and feeding the audio into an amp. My understanding of such things (warning: possibly 100% wrong) is the 2600 uses a shortcut to inject audio into its RF signal which kinda-sorta works, but isn't really "right". Consequently, this circuit adds interference to the video and doesn't work all that well for audio; on manually-fine-tunable TV sets, the "fine-tuning" adjustment needed for a good picture often differs from that required for good sound. Quote Link to comment Share on other sites More sharing options...
A.J. Franzman Posted July 31, 2006 Share Posted July 31, 2006 One thing I've wondered about as a mod would be cutting out the audio modulator and feeding the audio into an amp. My understanding of such things (warning: possibly 100% wrong) is the 2600 uses a shortcut to inject audio into its RF signal which kinda-sorta works, but isn't really "right". Consequently, this circuit adds interference to the video and doesn't work all that well for audio; on manually-fine-tunable TV sets, the "fine-tuning" adjustment needed for a good picture often differs from that required for good sound. The difference in "fine tuning" positions would be due to a mis-adjusted variable inductor in the console, not any inherent flaw in the audio mixer circuit (though I agree that the circuit is a bit of a bad hack). Lacking an oscilloscope and a frequency counter that goes high enough, I find it very touchy to get both the video and audio outputs optimized at the same time, but perseverence always pays off eventually. Quote Link to comment Share on other sites More sharing options...
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.