It is currently Tue Nov 12, 2019 11:58 am

All times are UTC - 7 hours





Post new topic Reply to topic  [ 120 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6 ... 8  Next
Author Message
PostPosted: Mon Mar 04, 2019 9:21 am 
Offline
User avatar

Joined: Sat Jul 04, 2015 9:58 am
Posts: 977
Location: -29.794229 -55.795374
Ben Boldt wrote:
I am not sure an HC will even work there

I noticed it's just an idea, and I'm willing to try that as soon as possible. :-)
Also in this topic seems that this change makes a Powerpak work fine with an RGB modded NES.

A guy named Esrael Neto did some mods on a NES clone and published here.
Maybe this can give some clues about what to do on other clones and even the NES itself.
Looks like the problem on the clone was interference on the power circuit, with insufficent current arriving on the ICs.


Top
 Profile  
 
PostPosted: Mon Mar 04, 2019 11:23 am 
Offline
User avatar

Joined: Tue Mar 22, 2016 8:27 pm
Posts: 364
Location: Minnesota, USA
If you were to look at the composite video signal with a scope, and not have a cart handy that could accept the 240p video test ROM, I believe that Super Mario Bros. would be a good one to use because the first 16 scanlines or so are always solid blue during the entire demo loop.
Fisher wrote:
Ben Boldt wrote:
I am not sure an HC will even work there

I noticed it's just an idea, and I'm willing to try that as soon as possible. :-)
Also in this topic seems that this change makes a Powerpak work fine with an RGB modded NES.

A guy named Esrael Neto did some mods on a NES clone and published here.
Maybe this can give some clues about what to do on other clones and even the NES itself.
Looks like the problem on the clone was interference on the power circuit, with insufficent current arriving on the ICs.

Okay, sounds good. I ordered some 74HCT373N from China. It will be a while to get here but I will lend a hand if/when needed.


Top
 Profile  
 
PostPosted: Mon Mar 04, 2019 6:28 pm 
Offline
User avatar

Joined: Sat Jul 04, 2015 9:58 am
Posts: 977
Location: -29.794229 -55.795374
Well, I socketed the IC of my clone and got the same results with an LS part, with an HC part and without the latch.
I just couldn't get good enough pictures. :-(
In fact, I can see no jailbars on the pictures I've got. :shock:
Maybe if I try to get the images with a capture card later?

So probably it's not the latch.
Let's see how your tests go.
I'll test other games to see if the HC part does make some games behave different graphically.


Top
 Profile  
 
PostPosted: Mon Mar 04, 2019 7:48 pm 
Offline
User avatar

Joined: Tue Feb 26, 2019 6:36 pm
Posts: 21
Ben Boldt wrote:
If you were to look at the composite video signal with a scope, and not have a cart handy that could accept the 240p video test ROM, I believe that Super Mario Bros. would be a good one to use because the first 16 scanlines or so are always solid blue during the entire demo loop.
Fisher wrote:
Ben Boldt wrote:
I am not sure an HC will even work there

I noticed it's just an idea, and I'm willing to try that as soon as possible. :-)
Also in this topic seems that this change makes a Powerpak work fine with an RGB modded NES.

A guy named Esrael Neto did some mods on a NES clone and published here.
Maybe this can give some clues about what to do on other clones and even the NES itself.
Looks like the problem on the clone was interference on the power circuit, with insufficent current arriving on the ICs.

Okay, sounds good. I ordered some 74HCT373N from China. It will be a while to get here but I will lend a hand if/when needed.



Hey bud,

Digikey is in your backyard. They are in MN. - https://www.digikey.com/product-detail/ ... -ND/277262

I order from them all the time.


Top
 Profile  
 
PostPosted: Mon Mar 04, 2019 7:51 pm 
Offline
User avatar

Joined: Tue Feb 26, 2019 6:36 pm
Posts: 21
Hope to have time this weekend.

I would really like to get this on the o-scope.

I have my top-loader NESRGB modded. But the board will still output the PPU composite video. I'm curious if the jailbars are seen there, too.


Top
 Profile  
 
PostPosted: Mon Mar 04, 2019 9:09 pm 
Offline
User avatar

Joined: Tue Mar 22, 2016 8:27 pm
Posts: 364
Location: Minnesota, USA
Ultron wrote:
Hey bud,

Digikey is in your backyard. They are in MN. - https://www.digikey.com/product-detail/ ... -ND/277262

I order from them all the time.

Yes I know! :) But the shipping still costs more than the parts for stuff like this. :?

Fisher wrote:
Well, I socketed the IC of my clone and got the same results with an LS part, with an HC part and without the latch.
I just couldn't get good enough pictures. :-(
In fact, I can see no jailbars on the pictures I've got. :shock:
Maybe if I try to get the images with a capture card later?

So probably it's not the latch.
Let's see how your tests go.
I'll test other games to see if the HC part does make some games behave different graphically.

Thanks for trying Fisher. Now we know. I will try it if I am able to find a way to precisely measure the jailbars.


Top
 Profile  
 
PostPosted: Sun Apr 07, 2019 1:20 pm 
Offline
User avatar

Joined: Tue Mar 22, 2016 8:27 pm
Posts: 364
Location: Minnesota, USA
I finally got around to working on this today. I got very lucky (in this case) and found a Famicom with very bad jailbars. Here is the AV mod that I did (video only, no cut traces, no cut pins, no copper foil, etc, only the parts shown):

Attachment:
av_mod.JPG
av_mod.JPG [ 2.73 MiB | Viewed 4372 times ]


Here are the jailbars with this mod:

Attachment:
no_cap.JPG
no_cap.JPG [ 2.58 MiB | Viewed 4372 times ]


By adding a 47 uF Tantalum capacitor from pin 22 to GND, it makes an incredible improvement:

Attachment:
with_tantalum.JPG
with_tantalum.JPG [ 3.49 MiB | Viewed 4372 times ]


Code:
                (GND)
PPU.20 -----------+---------------------------------+
 |                |                                 |
 |             | V (c)                              +----O } Composite
 |    PPU.21 --|< PNP (Reused the 2SA937)                  }   Video
 |             | \ (e)                   33uF       +----O }
--- Magic         |        100       Electrolytic   |
--- Capacitor     +-------/\/\/----+------|(--------+
 |                                 |
 |      (/RST)             150     |
PPU.22 -------------------/\/\/----+


I don't have any particular reason to think my values are any better worse than other mods, this is just the way that I have been doing it in the past. I notice now that my AV mod uses PPU pin 22 (/RST) as the pull-up rail. It seems to me that might be a digital signal and not really a 5V rail, that's interesting, I wonder why it is like that?? I should try using a true 5V, and also tweaking the resistors to be like other mods that are out there.

I tried various "magic" capacitors. I notice these capacitors are good, I can't tell any of them better or worse than each other:
  • 47uF 35V Tantalum
  • 15uF 35V Tantalum
  • 6.8uF 35V Tantalum
  • 4.7uF 35V Tantalum

These caps are almost as good as the tantalums above, almost undetectable difference:
  • 10uF 100V Electrolytic (might be a low ESR cap, doubtful but not sure)
  • 3.3uF 35V Tantalum
  • 2.2uF 35V Tantalum

These caps are are definitely NOT AS GOOD as the ones listed above:
  • 1uF 35V Tantalum
  • 100uF 16V Tantalum
  • 1uF 100V electrolytic
  • 100uF 35V electrolytic
  • 22uF 16V X7R ceramic (much to my surprise, I suspected this would be a winner)

Referring to the oscilloscope, the jailbars should be measurable. Triggering on the first scanline, showing just the first few tiles, running Super Mario Bros. 1 demo (solid sky blue):
Attachment:
tek00090.png
tek00090.png [ 37.47 KiB | Viewed 4372 times ]


With tantalum:
Attachment:
tek00089.png
tek00089.png [ 35.8 KiB | Viewed 4372 times ]


I think that if it was perfect, all of the peaks would line up, so I can try experiments and measure the effects now. Also, when viewed in motion with the oscilloscope, the jailbar one has much more variation to it.


Top
 Profile  
 
PostPosted: Sun Apr 07, 2019 2:25 pm 
Offline
User avatar

Joined: Tue Mar 22, 2016 8:27 pm
Posts: 364
Location: Minnesota, USA
Playing with the 33uF cap (this cap is intended to remove DC bias):

Full frame with 33uF 16V electrolytic:
Attachment:
tek00093.png
tek00093.png [ 88.61 KiB | Viewed 4360 times ]


Full frame with 100uF 35V electrolytic:
Attachment:
tek00094.png
tek00094.png [ 88.45 KiB | Viewed 4360 times ]


Blue = directly from PPU, base of transistor. Notice it is all positive.
Yellow = Composite signal. Notice it is centered positive and negative (i.e. DC bias has been removed)

You can see here that the 33uF creates a quite slanty signal, especially the v-blank, whereas the 100uF seems to be a better choice. My monitor shows no visible difference between 33uF and 100uF. No more or less jailbars, no sharper or fuzzier or anything. I tried adding some ceramic and tantalum in parallel with this 100uF to see if there was any effect to the sharpness but found none.

Edit:
Does anyone here know much about composite signals? I thought it was supposed to be a 1 Volt signal and this is only 0.78V: (Showing V-Blank here)
Attachment:
tek00003.png
tek00003.png [ 89.97 KiB | Viewed 4353 times ]


Also, some distortion comparing before-and-after the transistor, not sure what this means (Showing first scanline here):
Attachment:
tek00001.png
tek00001.png [ 31.43 KiB | Viewed 4353 times ]


I am wondering if it might be better to use an op-amp here instead of a transistor.


Top
 Profile  
 
PostPosted: Sun Apr 07, 2019 5:52 pm 
Offline
User avatar

Joined: Tue Feb 26, 2019 6:36 pm
Posts: 21
Nice measurements! Though I don't see jailbars in the video signal measured on the o-scope. I'm not sure how to bring them out.

Pixel frequency should be ~5.368MHz. Which is ~186.29 ns period.

Your composite signal with the 30uF shows field tilt. That capacitor should be at least 200uF. Here's a good tech note on field tilt.

https://pdfserv.maximintegrated.com/en/an/AN3768.pdf

Composite out of the PPU should be 2V. Then when T'ed off the TV connection, should be 1V. Problem here is, the PNP transistor in an emitter follower configuration has a gain of a little less than 1. I'm thinking that's why it's a little low. Though, from the wiki, it looks like it isn't outputting a 2V composite signal.

https://wiki.nesdev.com/w/index.php/NTSC_video

Thanks for doing this!!


Top
 Profile  
 
PostPosted: Sun Apr 07, 2019 6:28 pm 
Online

Joined: Sun Apr 13, 2008 11:12 am
Posts: 8666
Location: Seattle
Ben Boldt wrote:
Does anyone here know much about composite signals? I thought it was supposed to be a 1 Volt signal and this is only 0.78V: (Showing V-Blank here)
I know more than is healthy :p

But, yes, it's supposed to be approximately 1Vpp, from sync tip to full white (and sync tip to fully saturated yellow is 1.3V). The NES is really rather noncompliant though; see my measurements on the wiki.

Quote:
I am wondering if it might be better to use an op-amp here instead of a transistor.
Op-amps have tremendously lower gain-bandwidth. You should be able to use a series of emitter-followers to get a perfectly lovely result.

The NES/FC have just the first PNP-based emitter-follower already on the board; what happens after it enters the RF modulator is a question with multiple answers. One of the RF modulators has a NPN-based emitter-follower on the baseband NTSC output.


Top
 Profile  
 
PostPosted: Sun Apr 07, 2019 7:46 pm 
Offline
User avatar

Joined: Tue Mar 22, 2016 8:27 pm
Posts: 364
Location: Minnesota, USA
lidnariq wrote:
I know more than is healthy :p

Yay!!

lidnariq wrote:
But, yes, it's supposed to be approximately 1Vpp, from sync tip to full white (and sync tip to fully saturated yellow is 1.3V). The NES is really rather noncompliant though; see my measurements on the wiki.

Interesting, so since there is no white or yellow in the SMB1 demo, I should expect this lower voltage? I suppose I should get the 240p Test Suite running on a cart for this.

lidnariq wrote:
Quote:
I am wondering if it might be better to use an op-amp here instead of a transistor.
Op-amps have tremendously lower gain-bandwidth. You should be able to use a series of emitter-followers to get a perfectly lovely result. The NES/FC have just the first PNP-based emitter-follower already on the board; what happens after it enters the RF modulator is a question with multiple answers. One of the RF modulators has a NPN-based emitter-follower on the baseband NTSC output.

I was looking at the MAX4390 op-amp, it seems to be intended for this purpose. Though, I never imagined that a single SOT-23 part could cost several dollars! Wow. Do you have any emitter-follower ideas or an example I could try to build? Should I try to copy the circuit from the RF modulator you linked to? I am sort of clueless where to start. Would love to come up with a solid, semi-simple solution for the DIYers out there.

I guess my original intention was to find out where the jailbars are coming from and how to make them the least visible. Improving the video signal might make them MORE apparent! lol

Anyway, I think I can keep poking at it for now. The bad jailbars at first seem like each 4 pixels. When I put on the "magic cap", the very slight jailbars that remain are wider, seems like 8, possibly suggesting an additional source elsewhere.


Top
 Profile  
 
PostPosted: Sun Apr 07, 2019 8:39 pm 
Online

Joined: Sun Apr 13, 2008 11:12 am
Posts: 8666
Location: Seattle
Jailbars always seem to be from coupling with some of the various digital signals nearby - PPU A13 (4 pixels each on/off) and PPU /RD (1 pixel each on/off) seem to be common culprits. Slower=wider patterns must be something of PPU A0 through PPU A11 - they'll be very heavily dependent on what exactly is being drawn. Depending on context, PPU A12 might look like PPU A13, or be offscreen; PPU AD0-AD7 will change every pixel but won't be predictable.

In the Famicom, PPU /RESET is just tied to +5V, so that's why it seems to work as a bypass point.

Ben Boldt wrote:
Interesting, so since there is no white or yellow in the SMB1 demo, I should expect this lower voltage? I suppose I should get the 240p Test Suite running on a cart for this.
Yeah, you won't see full range of voltage without full range of picture brightness. NTSC yellow is just particularly high chrominance magnitude. NES yellow-ish isn't any different than any other color in the same horizontal row of the palette.

Ben Boldt wrote:
Do you have any emitter-follower ideas or an example I could try to build?
I'd consider trying the Vs. System 3-stage emitter follower schematic. My two NESes haven't shown any jailbars so I haven't pursued this rabbit hole myself.


Top
 Profile  
 
PostPosted: Mon Apr 08, 2019 3:00 pm 
Offline
User avatar

Joined: Tue Mar 22, 2016 8:27 pm
Posts: 364
Location: Minnesota, USA
lidnariq wrote:
Jailbars always seem to be from coupling with some of the various digital signals nearby - PPU A13 (4 pixels each on/off) and PPU /RD (1 pixel each on/off) seem to be common culprits. Slower=wider patterns must be something of PPU A0 through PPU A11 - they'll be very heavily dependent on what exactly is being drawn. Depending on context, PPU A12 might look like PPU A13, or be offscreen; PPU AD0-AD7 will change every pixel but won't be predictable.

In the Famicom, PPU /RESET is just tied to +5V, so that's why it seems to work as a bypass point.

Ben Boldt wrote:
Interesting, so since there is no white or yellow in the SMB1 demo, I should expect this lower voltage? I suppose I should get the 240p Test Suite running on a cart for this.
Yeah, you won't see full range of voltage without full range of picture brightness. NTSC yellow is just particularly high chrominance magnitude. NES yellow-ish isn't any different than any other color in the same horizontal row of the palette.

Ben Boldt wrote:
Do you have any emitter-follower ideas or an example I could try to build?
I'd consider trying the Vs. System 3-stage emitter follower schematic. My two NESes haven't shown any jailbars so I haven't pursued this rabbit hole myself.

Okay, I will keep it in mind. It looks like the Vs. System uses RGB + Sync, are you saying we should consider trying one of the 3 color channels for composite on the Famicom?

I changed to a 220uF electrolytic to reduce the field tilt, here is the result:
Attachment:
tek00006.png
tek00006.png [ 58.77 KiB | Viewed 4215 times ]


Top
 Profile  
 
PostPosted: Mon Apr 08, 2019 4:13 pm 
Online

Joined: Sun Apr 13, 2008 11:12 am
Posts: 8666
Location: Seattle
Ben Boldt wrote:
It looks like the Vs. System uses RGB + Sync, are you saying we should consider trying one of the 3 color channels for composite on the Famicom?
It's where I'd start, I guess? An emitter follower is a voltage buffer, producing the same voltage on the output as the input, plus or minus one diode drop depending on PNP/NPN. Given matched transistors (level of matching discretionary), a PNP and NPN stage will balance out, removing any net DC bias.
Each stage will effectively decrease the output impedance of the previous stage by the β=hFE of the transistor.

There's also the reverse-engineered schematics from Electronix Corp- RF modulator and mainboard - which also show a different PNP+NPN stage from the one I saw in the previous link.


Top
 Profile  
 
PostPosted: Tue Apr 09, 2019 8:01 pm 
Offline
User avatar

Joined: Tue Feb 26, 2019 6:36 pm
Posts: 21
Ben Boldt wrote:
I was looking at the MAX4390 op-amp, it seems to be intended for this purpose. Though, I never imagined that a single SOT-23 part could cost several dollars! Wow. Do you have any emitter-follower ideas or an example I could try to build? Should I try to copy the circuit from the RF modulator you linked to? I am sort of clueless where to start. Would love to come up with a solid, semi-simple solution for the DIYers out there.

I guess my original intention was to find out where the jailbars are coming from and how to make them the least visible. Improving the video signal might make them MORE apparent! lol

Anyway, I think I can keep poking at it for now. The bad jailbars at first seem like each 4 pixels. When I put on the "magic cap", the very slight jailbars that remain are wider, seems like 8, possibly suggesting an additional source elsewhere.



The MAX4389/4390 would work well, but it is pricey. High speed op amps get high in price, and an emitter follower will work fine here.

The AV Famicom and NES include a ferrite bead, FC2 in the video path. This is very often (always) omitted from AV mods. Ferrite beads provide high frequency noise suppression. Problem is, there are no labels on the bead and anything in the NES schematic. It might be worth it to borrow one from a NES or AV Famicom for testing. I attached diagram of the AV Famicom Video Amp (references are arbitrary).

Attachment:
File comment: AV Famicom Video
AV Fami Video.png
AV Fami Video.png [ 41.08 KiB | Viewed 4090 times ]


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 120 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6 ... 8  Next

All times are UTC - 7 hours


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group