Update: I heard from the Lava RSC folks and they fixed the sparkle issue with Rev E PPUs with a firmware update! So I guess this is more to do with a timing issue with the Rev E PPU versus any level conversion issue.
The VOUT pin always outputs composite video, but the mod intercepts and modifies the palette data being sent to the PPU so that it produces a simplified video signal. The video signal is then fed to several comparators, which it used to identify whether a particular pixel is background or foreground, and thus the value of the hidden EXT4. As the fpga sniffs the original palette data, it is able to reconstruct the intended signal. The speckle issues is seen on all of these mods (hi-def NES/NESRGB/LavaRGB) and is supposedly due to writes to register $2000 which can corrupt the data on the EXT pins. Some PPU versions don't have this bug, and the problem can sometimes be fixed by reading the data at a different points in the video signal.
Thanks for clarification. Amazingly, just moments ago I got an email from the Lava RSC team and they have released an updated firmware to fix the Rev E sparkle issue! I guess as you said, they are just sampling the data in a way to work around the Rev E sparple issue. Good stuff!
Femicom reminds me a lot on my childhood, so I liked it very much, Adrian. YT algorithms are hard to understand. Sometimes I do a video, which I think should be interesting and then nobody watches is. Another day I release something, what I think is not quite as exciting and it goes up like a rocket. I gave up to understand how it works and just make whatever I like :)
I watched it specifically on that _other_ platform, that I can't say by name without tripping this korp oration's sensor ship AI. The big N also has less power on that _other_ platform.
Super interesting how you notice this tiny little thing and are able to trace it back to the depths of how electronic actually work. This is great stuff don't let the views make you think otherwise.
Hi Adrian, wonderful video as usual. Sorry to hear it performed badly, I liked it! The LVC series logic levels are the same as TTL, TI calls it “3.3V TTL”. Google “ti lvc levels” and you’ll find some of TI’s guides on the subject with a graphic of the levels. The SN74LVC541A is supposed to have 3.3V VCC on the input side, that’s why the data sheet does not list 5V VCC. But if you look at the 3.3V rows of the recommended operating conditions you’ll find TTL levels. The input pins are also 5V tolerant, which is what makes the LVC series useful to interface 5V TTL logic with 3.3V logic. As for the PPU issue, my theory is that the timing of FPGA sampling the signals is calibrated for the G version, and the the E version has slightly different timings for those EXT outputs. Upscalers like OSSC have a “phase” setting to adjust this when in exact pixel sampling mode.
It's too bad these videos weren't popular. I love the Nintendo videos. I had this console when I was young and played it more than any other game console. Even after we had a SNES I still played the NES because it was just a fun console with very well thought out games. Not leaning on its sound and graphics, but on fun gameplay.
Sorry, ive been busy and just now got around to watching your main channel video. I wouldn't say no one is interested in it, its the holidays now and people are busy too. I pulled the trigger on one of these today -you convinced me!
I dunno, I quite enjoyed watching the famiclone vids, just because "the algorithm" disagrees doesn't mean it's not worth going into, sometimes you've just got to do what you love doing, numbers be damned... :)
I"m happy anytime you do an NES repair or hacking video (though worthwhile hacks are a bit few.) I've been using the NESRGB in my Twin Famicom for almost 15 years now, and no glitches or sparkles. The only unusual behaviour I've seen is when I bump the carts or if demos start up and don't initialize all of the PPU Palette - the NESRGB doesn't know what colours to show and displays fine grey noise.
from that video clip, it seems that they are capturing the EXT pin data a bit early or late, causing the sparkles on the leading edge of colour transitions. The timing of the different revs of PPU are slightly different. The RGB mod is using the ext pins to digitally capture the pixel data. The "shimmering" is normal, and is composite dot crawl characteristic of the PPU (its dot crawl is a bit interesting, since every other frame rendered is short 1 clock to give a more favorable pattern)
Nice. I wonder how many orders they ended up with based on the video? At least AliExpress doesn't generally raise prices when orders come in unlike most eBay sellers of similar stuff
Re: the bad main channel video performance... I never saw it in my feed or notifications. So it MAY not have been the topic that was the issue but standard TH-cam BS instead.....
The PPU's may have marginally different timing characteristics which are acceptable in the real NES, but the FPGA code wasn't developed to anticipate that flavor of delay. I mean, FPGA implementations have to account for timing envelopes and gate delays and such, and they may have assumed too much based on the PPU's they had on hand during development. Dang, does that make sense? FPGA implementation's timing characteristics not accounting for every variation of PPU?
Enjoyed the other video. Sort of fun to see somewhat of the current state of modding and what retro people are doing to it new after seeing all KINDS of hackjobs and strange stuff in the 90's and 2000s. I still have my unadulterated, perfect condition, non-yellowing, non-scratched or cracked NES from circa 1991.
Again, it’s definitely not “bad.” When you pick the wrong phase in Hi-Def NES you see the same thing with any PPU. The whole reason Hi-Def NES has that manual setting is because it isn’t something that can be automatically set to eliminate the pixels flickering on every PPU variant. Obviously, this Kit triggers the same phenomenon. Double Dragon is the game that GameTechUS and Kevtris use to demonstrate it and, indeed, the game I use after every Hi-Def NES install so I can get the phase set right. If this RGB kit doesn’t have a way to set the phase then you’ll just have to try other CPU/PPU combinations, unfortunately. :(
I love atari nes and commodore content, but like others I generally just enjoy watching your trouble shooting methods and techniques. Some of the more unique computers you work on are also quite fun to watch.
I like these sorts of videos. There are some actual gems hidden in the sea of cheap tech. If it's good it's cool to learn about. Even if it sucks it's fun to watch someone take it apart and say why it sucks. I hope you'll consider taking another jab or two at these sorts of things in the future.
What a bummer Adrian. I was super exited about this series (I know it failed? on the main channel) but I guess that's why I became a member, maybe we can get a quick video on the consoles on Patreon if you have something exciting come up! Sorry it didn't do well on the main channel, I guess the SNES craze has passed.. too bad, I LOVED my Nintendo and Super Nintendo back in the day. I use an emulator today, but it's not the same.. I'm probably going to grab one of these new ones.. why not? SO much fun!! (I had a good childhood, awesome parents, we were not wealthy by any means, but mom and dad took such wonderful care of my brother and I, and we started with the Atari 2600, had the Nintendo, a Super, then went to PS1, after that I bought my own PS2 and we had an N65, then of course a PS3 ahhaha but we NEVER got into Sega OR Xbox. Huh..
There were several revisions of the board. Notice the cpu-xx. You had 08 and 10. The only differences were the way the cic chip was handled. Some nes's had rev H. They put a diode or some other component to prevent unlicensed games like chiller that used a stun circuit bypassing the cic chip. Also cpu-03 to cpu 5 shows a purple blinking screen when there is no cart vs a cpu-06 and above showing a flashing gray screen.
I've really enjoyed this pair of videos about the NES clone. Maybe it will be a slow burner, because in theory NES videos should do well, given the nostalgia they hold for many people.
Back in 80's on arcade systems as more CMOS came out repairs used to show sparkle like this where we had to change the buffer resistor/ packs to fix. A real pain as you change IC only to get complaints with the change taking longer than the first fix!
I enjoyed the video on the main channel and this one. I find Famiclones interesting. I have an AV Famicom that I modded with a NESRGB and loaded up Castlevania III to see if I could see the sparkles on it. It looked fine on my RGB modded console. I'm not sure if it matters or not but I was running it through an OSSC.
15:29 leaving aside that the colour palette is incorrect in the Famiclon (custom palette?), the white edge on the big squareish bricks is jagged in a different way in revs E and G, not to mention tha, although it may be just a trick of the mind, the G colours seem darker..
I'm sad to hear about the original video. I'm not at all interested in the NES but for some reason it was the most enjoyable video of yours to me for quite a while, and I was literally smiling all the way through it!!
Fascinating discussion of the different 'tolerances' on TTL vs CMOS. Even 'digital binary computers' are subject to the very ANALOG physics of electricity and the infinite possibilities of randomness.
Adrian, I rather enjoyed the main channel Lava video and would love to see more modding of vintage consoles. I'm honestly surprised it wasn't more popular since it's very much along the same lines as many of your other videos: Working with older computer components with possible troubleshooting and modding. I'd personally blame the algorithm or GOOG's AI idiocy for categorizing it as a gaming video rather than a computer video. Maybe if you referred to the Famicom CPU as a '6502 Clone' it'd fall down on the near side of the fence.
That's a bummer with how custom palettes are swapped. I have a feeling they aren't going to just keep providing new firmware to us indefinitely.The hardware is awesome for the price and I hope the community can come up with a better solution for the palette problem.
Adrian, the clone chips are less available near as I can tell. While there are loads of old Famicoms and NES consoles still available at the moment. The whole Famicoms are generally less expensive than buying the CPU or PPU independently. I assume, from it being the case in all the clones I've opened including old models, that UMC cores rapidly moved to a combined CPU/PPU core blobbed on boards rather than using a discrete packaged chip. Incidentally we've never found a modern packaged NoC for sale, just the dies.
I went through the datasheets of the chips used for the voltage level translation and I don't see any issue. The SN74LVC4245A is bidirectional and specifically designed to be placed between a TTL-5V domain and a CMOS-3V3 domain. Such translation requires a bit of care because CMOS voltage levels scale with VCC and are symmetrical. For side A with VCCA = 5V VIH,min (minimum voltage, which, when present on the input, the chip guarantees to recognise as HIGH, over the specified operating temperature range) is 2V (so not a typical CMOS input). VOH,min for TTL-5V is typically 2.4V so the margin is 0.4V and sufficient. With SN74LVC541A things work slightly differently because the voltage levels of the input scale like CMOS but the power supply is 3.3V so VIH,min = 2V and the margin is the same. One need to make sure that the inputs of SN74LVC541A can stand a voltage above its supply voltage. The datasheet specifies an absolute VI,MAX = 6.5V so TTL logic levels seems safe.
About the custom firmware -- maybe try and diff the custom one with the latest stock one and see what changed. You might be able to find the palette data and substitute in your own.
So I was looking at the website that has PPU variants you used in the previous video on this, and the interesting thing is the PPU you had in the famiclone when you bought it apparently has DMA issues natively. Which may be causing the sparkles you're seeing.
that 72 pin connector for the cartridges, my nes came with a new one installed and the cartridges are very tight, gotta wiggle the cart as you pull it out.
I was wondering if the Rev E PPU (or maybe the sample you received) could have been more susceptible to outside noise than earlier/later PPUs. After all, high-speed electronics weren’t a thing back in the 80s, but this Famiclone has an FPGA on it which I presume is running at a higher check speed than the rest of the chips…
I'm wondering if the source of all the GPU's and PPU's in the Famicon were sourced from older nintendo machines in bbulk or something to save costs? I mean since the 'E' versions seem to work find in a "real" NES...
On the TTL versus CMOS issues. If I recall correctly, NPN and PNP transistors are better at pulling the voltage one way but not the other. I believe NPN is better at pulling to ground and PNP is better at pulling high. So that's likely why TTL has a bigger margin for the logical one, the chips being used aren't as good at producing that logical one. CMOS on the other hand uses both types of transistors. They basically contain two copies of the same circuit, one using NPN transistors to handle the low states and one using PNP transistors to handle the high states. So that makes them better at producing logical one states and therefore you don't need as much leeway.
I suspect that the video sparkle is down to two chips having edge of spec delays. Possibly factory rejects. either one alone works, not two together...
I think it’s ok to have videos that are a bit different. Maybe they don’t get so many views at first, but over time they may attract other viewers. I’m not into Nintendo, but it was still very interesting.
I didn't even get served a famiclone video. I do typically save videos that are over an hour to watch until later. IDK if TH-camrs realize people do that.
Intereresting problem, It'd be cool to see if its one that you could change with something like a line filtering capacitor but it almost seems like a voltage could be too low / high somewhere. I notice the color is a lot more saturated on the rev e, I assume they're made to be used place other revisions in terms of requirements but maybe the tolerances for lower / higher voltage varies from one revision to the next.
Please don't get fooled by the algorithm. It's not that people don't like the content, it's just TH-cam not showing it to people. If you want to estimate how much viewers like it, you can really only go off the views/likes ratio.
Does this actually have a straight clone of Tim's NESRGB board on it? If so, there are custom firmware palette files out there already (FireBrandX has a ton).
it's interesting that the fpga had the glitch exclusively on the vertical edges there's some minor glitch here i'm sure that's interacting with how the fpga draws the display it's conflating clock and signal somehow i would guess the glitches are happening in long vertical runs which means on that edge every time a horizontal line of pixels gets to that specific pixel for long series horizontal lines it must be originating in the ppu
Why the 🦆 did TH-cam translated the title in French in my feed? I mean, that's my language, so it happens from time to time, but that's always on the home page, and never on channels I'm subscribed to. even weird, the title completely changed, and become "Mon PPU Nintendo brillant est-il mauvais ?" So "Is my shiny Nintendo PPU is bad?" "Brillant" makes no sense here, "étincelant" would have been more correct, but still wouldn't make any freaking sense. Seriously, TH-cam, go home, you're drunk.
It didn't do too bad when I look at the views? Personally I really enjoy old console videos, as much as I do old pc videos. Thanks for giving the NES some love too. And by the way you seem to enjoy these games systems quite a bit, so isn't that the most important thing?
Hey Adrian, I went through a stage where I was modding NES consoles with NES RGB. I came across a bug which literally nobody else had ever had, it wasn't exactly sparkles but it was some strange graphical corruption similar to when the cart isn't in properly. Turns out it was an incompatibility with a certain type of VRAM, I think it was Sony. I think on motherboards with Toshiba I didn't have issues. I know this isn't the same problem but you didn't consider different types of VRAM/WRAM, and I know that can definitely effect the NES RGB. And your famiclone obviously has a completely 3rd party type.
Yeah, I was just looking at that after he mentioned it. OTOH, all four of those are well below average for the channel (only the IIc floppy drive and Plexus compilation videos are lower in the past year or so as I type), for whatever that means 🤷♂ [Plus, he'll have more data on actual watch time and, of course, revenue; which may swing it further.]
@Adrian'sDigitalBasement][ This and the origtinal video may not have done well, but you're still getting good viewership. I'm a diehard retro gamer and the work on this Famiclone was very fascinating to me. It's motivated me to take a look into getting one for myself. I really want the NES version though. So Thank You for doing these video, please do more in the future, especially the SNES stuff. You'll get more viewship with something of that topic. The SNES is very hot right now..
I liked it but I can see where clone nes would get less interest. Also slightly different content might bring some new folks to the channel even if it performs lower it might grow the channel overall
@@adriansdigitalbasement2 I wonder if it was more a timing issue. People might have been recovering from the Thanksgiving weekend, and that was a bigger hit to the viewership than the content of the video itself.
A plausible explanation, I think. Also, more chips in line will add latency. One (modern) chip extra should cause very little extra latency, but there is some.
![Adrian's Digital Basement ][](http://yt3.ggpht.com/ytc/AIdro_k7CrQRe71qN6c0Ua7A1P_I8KkxjqLa70u0zYeCvdvnKw=s48-c-k-c0x00ffffff-no-rj)
Update: I heard from the Lava RSC folks and they fixed the sparkle issue with Rev E PPUs with a firmware update! So I guess this is more to do with a timing issue with the Rev E PPU versus any level conversion issue.
The VOUT pin always outputs composite video, but the mod intercepts and modifies the palette data being sent to the PPU so that it produces a simplified video signal. The video signal is then fed to several comparators, which it used to identify whether a particular pixel is background or foreground, and thus the value of the hidden EXT4. As the fpga sniffs the original palette data, it is able to reconstruct the intended signal. The speckle issues is seen on all of these mods (hi-def NES/NESRGB/LavaRGB) and is supposedly due to writes to register $2000 which can corrupt the data on the EXT pins. Some PPU versions don't have this bug, and the problem can sometimes be fixed by reading the data at a different points in the video signal.
Thanks for clarification. Amazingly, just moments ago I got an email from the Lava RSC team and they have released an updated firmware to fix the Rev E sparkle issue! I guess as you said, they are just sampling the data in a way to work around the Rev E sparple issue. Good stuff!
Femicom reminds me a lot on my childhood, so I liked it very much, Adrian. YT algorithms are hard to understand. Sometimes I do a video, which I think should be interesting and then nobody watches is. Another day I release something, what I think is not quite as exciting and it goes up like a rocket. I gave up to understand how it works and just make whatever I like :)
I enjoyed the video on the main channel, too bad the viewers didn't like it or the algorithm didn't suggest it 🙁
Probably Nintendo suppression... (I enjoyed it too)
Same
I agree with you. 👍
I watched it specifically on that _other_ platform, that I can't say by name without tripping this korp oration's sensor ship AI.
The big N also has less power on that _other_ platform.
Me too.
I was 200% interested in that topic. Maybe its a TH-cam thing
Super interesting how you notice this tiny little thing and are able to trace it back to the depths of how electronic actually work. This is great stuff don't let the views make you think otherwise.
Hi Adrian, wonderful video as usual. Sorry to hear it performed badly, I liked it!
The LVC series logic levels are the same as TTL, TI calls it “3.3V TTL”. Google “ti lvc levels” and you’ll find some of TI’s guides on the subject with a graphic of the levels. The SN74LVC541A is supposed to have 3.3V VCC on the input side, that’s why the data sheet does not list 5V VCC. But if you look at the 3.3V rows of the recommended operating conditions you’ll find TTL levels. The input pins are also 5V tolerant, which is what makes the LVC series useful to interface 5V TTL logic with 3.3V logic.
As for the PPU issue, my theory is that the timing of FPGA sampling the signals is calibrated for the G version, and the the E version has slightly different timings for those EXT outputs. Upscalers like OSSC have a “phase” setting to adjust this when in exact pixel sampling mode.
It's too bad these videos weren't popular. I love the Nintendo videos. I had this console when I was young and played it more than any other game console. Even after we had a SNES I still played the NES because it was just a fun console with very well thought out games. Not leaning on its sound and graphics, but on fun gameplay.
That's too bad on the main channel video - I really enjoyed it! Sorry about that. Keep up the great work!
Sorry, ive been busy and just now got around to watching your main channel video. I wouldn't say no one is interested in it, its the holidays now and people are busy too.
I pulled the trigger on one of these today -you convinced me!
I ordered one myself immediately after seeing the video on the first channel.
Damn. I really enjoyed the main channel’s video!
i enjoy the videos about the famicom, don’t understand why others wouldn’t!
Definitely understand your decision not to make more videos on it. I enjoyed the main channel video and this one a lot though. Thanks for sharing!
I dunno, I quite enjoyed watching the famiclone vids, just because "the algorithm" disagrees doesn't mean it's not worth going into, sometimes you've just got to do what you love doing, numbers be damned... :)
I"m happy anytime you do an NES repair or hacking video (though worthwhile hacks are a bit few.) I've been using the NESRGB in my Twin Famicom for almost 15 years now, and no glitches or sparkles. The only unusual behaviour I've seen is when I bump the carts or if demos start up and don't initialize all of the PPU Palette - the NESRGB doesn't know what colours to show and displays fine grey noise.
from that video clip, it seems that they are capturing the EXT pin data a bit early or late, causing the sparkles on the leading edge of colour transitions. The timing of the different revs of PPU are slightly different. The RGB mod is using the ext pins to digitally capture the pixel data. The "shimmering" is normal, and is composite dot crawl characteristic of the PPU (its dot crawl is a bit interesting, since every other frame rendered is short 1 clock to give a more favorable pattern)
I learn so much from your videos on troubleshooting and repairs no matter what device you are working on. Thank you for all you do.
When I was a kid all the NES's sparkled. So honestly even if it's "defective" it's not defective. It's normal.
I ordered one the day you posted the main video, I've seen those sparkles on my original NES though.
Nice. I wonder how many orders they ended up with based on the video? At least AliExpress doesn't generally raise prices when orders come in unlike most eBay sellers of similar stuff
It's too bad the video didn't do well. I rather enjoyed it.
Same! I love the NES and retro console content.
Same
Same. TBH it's the first video of his I've watched in a while. 😅
Re: the bad main channel video performance... I never saw it in my feed or notifications. So it MAY not have been the topic that was the issue but standard TH-cam BS instead.....
Your videos are an automatic thumbs up. Love the content, Adrian!
Adrian, if you made a video watching paint dry i'd watch it and be happy. love your videos always. i always learn a great deal from you.
Same here
Weird that the video did poorly. I'm watching a lot of retro channels that main these subjects. I loved it anyway.
The PPU's may have marginally different timing characteristics which are acceptable in the real NES, but the FPGA code wasn't developed to anticipate that flavor of delay. I mean, FPGA implementations have to account for timing envelopes and gate delays and such, and they may have assumed too much based on the PPU's they had on hand during development. Dang, does that make sense? FPGA implementation's timing characteristics not accounting for every variation of PPU?
Thank you for following this up Adrian 🥂
Enjoyed the other video. Sort of fun to see somewhat of the current state of modding and what retro people are doing to it new after seeing all KINDS of hackjobs and strange stuff in the 90's and 2000s. I still have my unadulterated, perfect condition, non-yellowing, non-scratched or cracked NES from circa 1991.
For me that was one of the best of your videos, shame people didn't like it.
I almost missed this video because the thumbnail looked slightly different to your other videos. Great that the PPU works fine.
Again, it’s definitely not “bad.” When you pick the wrong phase in Hi-Def NES you see the same thing with any PPU. The whole reason Hi-Def NES has that manual setting is because it isn’t something that can be automatically set to eliminate the pixels flickering on every PPU variant. Obviously, this Kit triggers the same phenomenon.
Double Dragon is the game that GameTechUS and Kevtris use to demonstrate it and, indeed, the game I use after every Hi-Def NES install so I can get the phase set right.
If this RGB kit doesn’t have a way to set the phase then you’ll just have to try other CPU/PPU combinations, unfortunately. :(
I love atari nes and commodore content, but like others I generally just enjoy watching your trouble shooting methods and techniques. Some of the more unique computers you work on are also quite fun to watch.
I like these sorts of videos.
There are some actual gems hidden in the sea of cheap tech. If it's good it's cool to learn about. Even if it sucks it's fun to watch someone take it apart and say why it sucks.
I hope you'll consider taking another jab or two at these sorts of things in the future.
Great Video Adrian !
I don't even like the NES but the Clone video was pretty good and I watched the whole thing
What a bummer Adrian. I was super exited about this series (I know it failed? on the main channel) but I guess that's why I became a member, maybe we can get a quick video on the consoles on Patreon if you have something exciting come up! Sorry it didn't do well on the main channel, I guess the SNES craze has passed.. too bad, I LOVED my Nintendo and Super Nintendo back in the day. I use an emulator today, but it's not the same.. I'm probably going to grab one of these new ones.. why not? SO much fun!! (I had a good childhood, awesome parents, we were not wealthy by any means, but mom and dad took such wonderful care of my brother and I, and we started with the Atari 2600, had the Nintendo, a Super, then went to PS1, after that I bought my own PS2 and we had an N65, then of course a PS3 ahhaha but we NEVER got into Sega OR Xbox. Huh..
There were several revisions of the board. Notice the cpu-xx. You had 08 and 10. The only differences were the way the cic chip was handled. Some nes's had rev H. They put a diode or some other component to prevent unlicensed games like chiller that used a stun circuit bypassing the cic chip. Also cpu-03 to cpu 5 shows a purple blinking screen when there is no cart vs a cpu-06 and above showing a flashing gray screen.
I've really enjoyed this pair of videos about the NES clone. Maybe it will be a slow burner, because in theory NES videos should do well, given the nostalgia they hold for many people.
This is the earliest I've been able to see one one of his new uploads! 2 mins ago, woot!
Love the Nintendo videos. Probably the algorithm screwing it up due to keywords. Hope you do more though.
Back in 80's on arcade systems as more CMOS came out repairs used to show sparkle like this where we had to change the buffer resistor/ packs to fix. A real pain as you change IC only to get complaints with the change taking longer than the first fix!
I enjoyed the video on the main channel and this one. I find Famiclones interesting. I have an AV Famicom that I modded with a NESRGB and loaded up Castlevania III to see if I could see the sparkles on it. It looked fine on my RGB modded console. I'm not sure if it matters or not but I was running it through an OSSC.
15:29 leaving aside that the colour palette is incorrect in the Famiclon (custom palette?), the white edge on the big squareish bricks is jagged in a different way in revs E and G, not to mention tha, although it may be just a trick of the mind, the G colours seem darker..
I'm sad to hear about the original video. I'm not at all interested in the NES but for some reason it was the most enjoyable video of yours to me for quite a while, and I was literally smiling all the way through it!!
I enjoy all your videos. When the microwave pop-corn is popping the wife knows another digital basement episode is about to be played.
Thanks for sharing how the Retro Nintendo is working
Fascinating discussion of the different 'tolerances' on TTL vs CMOS. Even 'digital binary computers' are subject to the very ANALOG physics of electricity and the infinite possibilities of randomness.
Adrian, I rather enjoyed the main channel Lava video and would love to see more modding of vintage consoles. I'm honestly surprised it wasn't more popular since it's very much along the same lines as many of your other videos: Working with older computer components with possible troubleshooting and modding. I'd personally blame the algorithm or GOOG's AI idiocy for categorizing it as a gaming video rather than a computer video. Maybe if you referred to the Famicom CPU as a '6502 Clone' it'd fall down on the near side of the fence.
round hole = machine or machined pin sockets / pins. Love these.
ADB - The like goes on before the intro begins. 😊😊😊
That's a bummer with how custom palettes are swapped. I have a feeling they aren't going to just keep providing new firmware to us indefinitely.The hardware is awesome for the price and I hope the community can come up with a better solution for the palette problem.
Sending feedback to Lava and getting custom palettes back!? Digital Basement roadtrip to China to work on Famiclone rev 2 coming up? :D
Adrian, the clone chips are less available near as I can tell. While there are loads of old Famicoms and NES consoles still available at the moment. The whole Famicoms are generally less expensive than buying the CPU or PPU independently.
I assume, from it being the case in all the clones I've opened including old models, that UMC cores rapidly moved to a combined CPU/PPU core blobbed on boards rather than using a discrete packaged chip. Incidentally we've never found a modern packaged NoC for sale, just the dies.
I went through the datasheets of the chips used for the voltage level translation and I don't see any issue.
The SN74LVC4245A is bidirectional and specifically designed to be placed between a TTL-5V domain and a CMOS-3V3 domain. Such translation requires a bit of care because CMOS voltage levels scale with VCC and are symmetrical. For side A with VCCA = 5V VIH,min (minimum voltage, which, when present on the input, the chip guarantees to recognise as HIGH, over the specified operating temperature range) is 2V (so not a typical CMOS input). VOH,min for TTL-5V is typically 2.4V so the margin is 0.4V and sufficient.
With SN74LVC541A things work slightly differently because the voltage levels of the input scale like CMOS but the power supply is 3.3V so VIH,min = 2V and the margin is the same. One need to make sure that the inputs of SN74LVC541A can stand a voltage above its supply voltage. The datasheet specifies an absolute VI,MAX = 6.5V so TTL logic levels seems safe.
I can't help but wonder, if people assumed it was just an Android emulation box or a NOAC clone, and skipped the main video because of that.
About the custom firmware -- maybe try and diff the custom one with the latest stock one and see what changed. You might be able to find the palette data and substitute in your own.
So I was looking at the website that has PPU variants you used in the previous video on this, and the interesting thing is the PPU you had in the famiclone when you bought it apparently has DMA issues natively. Which may be causing the sparkles you're seeing.
that 72 pin connector for the cartridges, my nes came with a new one installed and the cartridges are very tight, gotta wiggle the cart as you pull it out.
I wonder what are actually (seeing visually) outputting the EXT pins
I enjoyed that video!
let's see those EXT signals on the scope. you should be able to easily see if the levels or timings are different between the 2 different PPUs.
I was wondering if the Rev E PPU (or maybe the sample you received) could have been more susceptible to outside noise than earlier/later PPUs. After all, high-speed electronics weren’t a thing back in the 80s, but this Famiclone has an FPGA on it which I presume is running at a higher check speed than the rest of the chips…
Thanks for the follow up
I'm wondering if the source of all the GPU's and PPU's in the Famicon were sourced from older nintendo machines in bbulk or something to save costs? I mean since the 'E' versions seem to work find in a "real" NES...
On the TTL versus CMOS issues. If I recall correctly, NPN and PNP transistors are better at pulling the voltage one way but not the other. I believe NPN is better at pulling to ground and PNP is better at pulling high. So that's likely why TTL has a bigger margin for the logical one, the chips being used aren't as good at producing that logical one.
CMOS on the other hand uses both types of transistors. They basically contain two copies of the same circuit, one using NPN transistors to handle the low states and one using PNP transistors to handle the high states. So that makes them better at producing logical one states and therefore you don't need as much leeway.
Loved both videos. Dunno about those other yahoos but yeah
I suspect that the video sparkle is down to two chips having edge of spec delays. Possibly factory rejects.
either one alone works, not two together...
Well, another idea, Adrian... maybe you can shoe-horn in one of these consoles while testing a CRT or doing a RGB conversion or what not!
I think it’s ok to have videos that are a bit different. Maybe they don’t get so many views at first, but over time they may attract other viewers. I’m not into Nintendo, but it was still very interesting.
I notice the E variant was sharper than the G variant, perhaps the sharpness is triggering an artifact in the FPGA rendering?
I didn't even get served a famiclone video. I do typically save videos that are over an hour to watch until later. IDK if TH-camrs realize people do that.
Intereresting problem, It'd be cool to see if its one that you could change with something like a line filtering capacitor but it almost seems like a voltage could be too low / high somewhere. I notice the color is a lot more saturated on the rev e, I assume they're made to be used place other revisions in terms of requirements but maybe the tolerances for lower / higher voltage varies from one revision to the next.
lol yeah that cmos vs ttl difference will get you every time
The sockets with round holes are called "Turned Pin"
Please don't get fooled by the algorithm.
It's not that people don't like the content, it's just TH-cam not showing it to people.
If you want to estimate how much viewers like it, you can really only go off the views/likes ratio.
i love nes games, i play them regularly on my powkiddy rgb30 console
Does this actually have a straight clone of Tim's NESRGB board on it? If so, there are custom firmware palette files out there already (FireBrandX has a ton).
it's interesting that the fpga had the glitch exclusively on the vertical edges
there's some minor glitch here i'm sure that's interacting with how the fpga draws the display
it's conflating clock and signal somehow i would guess
the glitches are happening in long vertical runs which means on that edge every time a horizontal line of pixels gets to that specific pixel for long series horizontal lines
it must be originating in the ppu
Sparkle Sparkle Sparkle!
I hope for more SMM videos soon.
Why the 🦆 did TH-cam translated the title in French in my feed? I mean, that's my language, so it happens from time to time, but that's always on the home page, and never on channels I'm subscribed to.
even weird, the title completely changed, and become "Mon PPU Nintendo brillant est-il mauvais ?" So "Is my shiny Nintendo PPU is bad?"
"Brillant" makes no sense here, "étincelant" would have been more correct, but still wouldn't make any freaking sense.
Seriously, TH-cam, go home, you're drunk.
It didn't do too bad when I look at the views? Personally I really enjoy old console videos, as much as I do old pc videos. Thanks for giving the NES some love too. And by the way you seem to enjoy these games systems quite a bit, so isn't that the most important thing?
Huh... strange. I liked it!
Hey Adrian, I went through a stage where I was modding NES consoles with NES RGB. I came across a bug which literally nobody else had ever had, it wasn't exactly sparkles but it was some strange graphical corruption similar to when the cart isn't in properly. Turns out it was an incompatibility with a certain type of VRAM, I think it was Sony. I think on motherboards with Toshiba I didn't have issues. I know this isn't the same problem but you didn't consider different types of VRAM/WRAM, and I know that can definitely effect the NES RGB. And your famiclone obviously has a completely 3rd party type.
Views must have evened out on the main channel as it’s siting with a similar amount as the last 4 videos.
For me, I watch these types of videos in bursts.
Yeah, I was just looking at that after he mentioned it. OTOH, all four of those are well below average for the channel (only the IIc floppy drive and Plexus compilation videos are lower in the past year or so as I type), for whatever that means 🤷♂
[Plus, he'll have more data on actual watch time and, of course, revenue; which may swing it further.]
It's probably not only about view count but also watch time. If people drop it some minutes in that's gonna impact revenue a lot.
Blu Tack = "Smurf Poo". I stole that from someone else, so I won't take credit (blame?), but I thought it was funny.
Dystopian Nintendont wants to know your location...
I liked the video a lot!!! tx!!!
I have one with a bad PPU. It won't play one of my games, and another game has corrupt or incorrect colors.
Ya i watched both videos, both are awesome not sure why the first one didn't do well
*Sparkle Sparkle Sparkle* 😎
If you watch the nintoaster video from vomitsaw's channel you can see exactly how the composite circuit works.
Adrian, are you going to the Vintage Computer Festival SoCal in February 2025?
@Adrian'sDigitalBasement][
This and the origtinal video may not have done well, but you're still getting good viewership. I'm a diehard retro gamer and the work on this Famiclone was very fascinating to me. It's motivated me to take a look into getting one for myself. I really want the NES version though. So Thank You for doing these video, please do more in the future, especially the SNES stuff. You'll get more viewship with something of that topic. The SNES is very hot right now..
It must have done better than the ice maker! 😂
Adrian! There's a few "Barn Find" NES consoles up on eBay right now. Would cleaning and repairing those up make good channel content?
Doh! I just got to the end of this video. I think the answer is a "no"!
Heh probably not based on NES performance on my channels. Probably for more game console oriented channels?
I liked it but I can see where clone nes would get less interest. Also slightly different content might bring some new folks to the channel even if it performs lower it might grow the channel overall
@@adriansdigitalbasement2 I wonder if it was more a timing issue. People might have been recovering from the Thanksgiving weekend, and that was a bigger hit to the viewership than the content of the video itself.
hi here is a like and a comment thanks for the show
I enjoyed this one too!!!!
@26:33 "edge case"😀
Sometimes I doubt your commitment to sparkle motion.
Most of the corrosion on the pins come from spit aerosol from kids (and adults) blowing on the cartridges.
A plausible explanation, I think. Also, more chips in line will add latency. One (modern) chip extra should cause very little extra latency, but there is some.