What's with all the shit talking over at Assembler modding and Hacking section?

[Rosser]

They say the guys over here are  "guys over at famicom are a bit clueless" -CombatRocked

That's not very nice 

http://www.assemblergames.com/forums/showthread.php?44350-Famicom-AV-Mod/page2

[2A03]

It's mainly about people who used 80sFREAK's circuit and found that it didn't work right. I tried it myself and as calpis/kyuusaku said in that thread, it wound up having major sync issues with every TV I tried it on and it was fairly dim on top of that.

[80sFREAK]

Rosser, don't bother, cool modders can't tune one stage amplifier.

[lobdale]

The circuit is not the problem, bad modders are!

[kyuusaku]

Rosser, don't bother, cool modders can't tune one stage amplifier.

"Cool modders" actually design circuits, ones based on principles that will work for everyone. Some of the proposed solutions here like 33 uF AC coupling cap are GUARANTEED to fail on standard 75 ohm inputs... There's been nothing posted which surpasses Nintendo's own cheap common-collector output. It's not even possible to since they themselves chose a tuned circuit which at least is verified working by MILLIONS of users.

There are only two options for more robust output and both require more active components:

-PNP common-collector -> level shift -> NPN common-collector with 75 ohm Re doubling as parallel source termination
-PNP input video amp or op amp with internal offset (Av = 2) -> level shift or 220 uF AC coupling cap -> series 75 ohm source termination

Hmm, never seen those posted. So is it so wrong to call 25 pages of "creative" solutions and other tuned voodoo nonsense clueless? How about all the paranoid shielding for jailbars when it's power rail coupled?

[keropi]

It's mainly about people who used 80sFREAK's circuit and found that it didn't work right. I tried it myself and as calpis/kyuusaku said in that thread, it wound up having major sync issues with every TV I tried it on and it was fairly dim on top of that.

And how come the 2 famicoms me and a friend got from 80sFreak directly work perfectly on several differentl tv sets ?
I keep reading things like that and I've come with 2 conclusions:
a. 80sFreak does some kind of voodoo magic and composite works
b. people that complain just make mistakes on the mod and/or can't mod at all

[VinylDracula]

simple answer is this, if you were good at modding you wouldnt need to follow someone elses guide. You cant follow their guide correctly, thats your problem.

[kyuusaku]

And how come the 2 famicoms me and a friend got from 80sFreak directly work perfectly on several differentl tv sets ?
...
people that complain just make mistakes on the mod and/or can't mod at all

Actually, if the mod has the 33 uF AC coupling cap it works because your TV has a non-standard high-impedance input and/or digital signal processing; older and higher end sets tend to have low input impedance and a simple comparator for sync detection, so this affects MANY people.

Perfect is subjective. In my opinion the perfect amp delivers 1.2 Vpp (1.0 for FC video) into 75 ohms with  minimal distortion, 100% DC coupling and 75 ohm series source termination & my definition just happens to be what's used in just about every modern video circuit.

Maybe I'll measure some of the amps posted around here.

[famifan]

hey
where are the video waveform pics taken from oscilloscore?
where i can definitely see that sync level was reduced by that high-pass filter (33uF + 75 Ohm) ?
hard to define what is wrong without them

i'm up to famiclone circuit that is very closer to sharp twin and exactly the first thing that kyuusaku mentioned (two emitter followers with voltage divider between them and pure 75 Ohm output impedance)
Post Merge: March 02, 2013, 10:13:27 pm

Maybe I'll measure some of the amps posted around here.

Spoiler

[close]

is it good?

[80sFREAK]

...

2 keropi, well, yeah, it's kinda wodoo magic, you know, special spells and so on, which i explained in sticky thread. By the way, thank you for bothering with photos 

simple answer is this, if you were good at modding you wouldnt need to follow someone elses guide. You cant follow their guide correctly, thats your problem.

Correct, modder should understand, how it working, not just soldering whatever sits on the table/draw. The problem is, even there is a guide 95% will do their own way, obviously fail and complain.

Actually, if the mod has the 33 uF AC coupling cap it works because your TV has a non-standard high-impedance input and/or digital signal processing; older and higher end sets tend to have low input impedance and a simple comparator for sync detection, so this affects MANY people.

More than that i highly recommend to use ceramic capacitor. The problem is, when your friend 2A03 a.k.a Apolloboy complained about "bad picture", he does not posted any proofs or clear description what's wrong with picture. Ok, i turned on my remote view ability and gave him solution. He still trying to make shit storm. Oh, well, if that's all able to do, i'm off

Perfect is subjective. In my opinion the perfect amp delivers 1.2 Vpp (1.0 for FC video) into 75 ohms with  minimal distortion, 100% DC coupling and 75 ohm series source termination & my definition just happens to be what's used in just about every modern video circuit.

Maybe I'll measure some of the amps posted around here.

I'm affraid you actually have more 50 Ohms sets, just guess  

hey
where are the video waveform pics taken from oscilloscore?
where i can definitely see that sync level was reduced by that high-pass filter (33uF + 75 Ohm) ?
hard to define what is wrong without them

Don't scare them with too much science   Give'em "ready to eat" solution, so they can make money on it straight away or they will call you faggot  

Guys, no need drama, just some people have to accept - they are not perfect (i'm not perfect too   )

[kyuusaku]

hey
where are the video waveform pics taken from oscilloscore?
where i can definitely see that sync level was reduced by that high-pass filter (33uF + 75 Ohm) ?
hard to define what is wrong without them

I'll try to get a scope demonstration. In the meantime this is the concept:

Only considering the 33u C and 75 R, not the TV's clamp capacitor etc, you get -3 dB at frequency of 64 Hz. This means that Vsync's fundamental frequency component, 60 Hz is ~30% attenuated. Sync should have sharp rise and fall times but instead it will look overcompensated like this:

is it good?

If the PPU outputs 2 Vpp then maybe it's good. I think it outputs 1 Vpp though so 0.5 Vpp will be delivered to the TV.. I'll check soon.

Edit: one problem with it is that the 220 & 2.2k point has a non-linear transfer so the signal is compressed and somewhat distorted. Since I think the signal is directly 1 Vpp out of the PPU we can't attenuate it at all to make a passive level shift... It must be active. I think 3 transistors may be the minimum necessary for a linear amp.

butthurt

Maybe because I'm sick of watching mediocrity succeed? Everyone can't have such low standards or nothing would work right.

he does not posted any proofs or clear description what's wrong with picture.

Well considering the problem, it would be sync distortion.

I'm affraid you actually have more 50 Ohms sets, just guess 

AFAIK no NTSC TV has 50 ohm inputs, all baseband inputs are 75 ohm (same for PAL), coax/cable TV is 75 ohm and most antennas are 75 ohm/300 ohm in the stone age.

[80sFREAK]

You missing one thing here. Important thing

I think 3 transistors may be the minimum necessary for a linear amp.

It's very simple to make complicated things, but its very complicated to make things simple.

Everyone can't have such low standards or nothing would work right.

Unfortunately most people tend to use parts, which they have, not they should "Oh, it's just capacitor, i will solder this one".

Well considering the problem, it would be sync distortion.

Ok, so which exactly parts he used?

AFAIK no NTSC TV has 50 ohm inputs, all baseband inputs are 75 ohm (same for PAL), coax/cable TV is 75 ohm and most antennas are 75 ohm/300 ohm in the stone age.

You never know

Well, if picture out of sync and dim, you have to change one value. That's all.

[famifan]

Only considering the 33u C and 75 R, not the TV's clamp capacitor etc, you get -3 dB at frequency of 64 Hz. This means that Vsync's fundamental frequency component, 60 Hz is ~30% attenuated. Sync should have sharp rise and fall times but instead it will look overcompensated like this:

yep, i know it. Due to R and C tolerances cut-off freq. can be safely assumed to 60Hz. Just curious to see the actual waveforms

[keropi]

Actually, if the mod has the 33 uF AC coupling cap it works because your TV has a non-standard high-impedance input and/or digital signal processing; older and higher end sets tend to have low input impedance and a simple comparator for sync detection, so this affects MANY people.

Let me just comment on this... I have personally tested on:
a. LCD 37LH4000 LG
b. LCD LG TV/PC monitor, some cheap model that costed 150e
c. LCD Cheap chinese(?) one from a super market that was 1366x768
d. LCD 40" Sony Bravia something , don't have it anymore
e. CRT Sony Trinitron 14" 2 different models
f. CRT Philips 21"
g. CRT Comet 26"

so yeah, does the 1998 Comet  crt have signal processing? or the 14" ones? and what is the chance that ALL of them crt and lcd alike accept non-standard signals?
Apparently It only affects many people if they screw up their mod. There are people here that got machines from 80sFreak directly, don't you think they would complain at some point?  

edit: added bravia

[kyuusaku]

OK here's the setup:







Just a common-collector amp with 100 ohm Re for very low output impedance. Go 75 ohm if you have the current to spare and you have the ideal 1 transistor amp. (Do use a large AC coupling cap.)


Here's Vsync DC coupled:



Nice edges obviously:



Here's 470 uF coupled:



Also nice edges:



I thought this was 44 uF but it was really 440 nF, shown for example:



Just some line droop:



Holy crap look at Vsync:



Well here's 22 uF:





It doesn't *appear* too attenuated here, but it is (didn't bother measuring) the important thing is that the entire signal actually wobbles around Vsync which could definitely affect separation, particularly with AGC. The 470 uF doesn't show any of this (neither does 220 uF). I took a video of the wobble but I don't have anywhere to upload it at the moment.

Maybe this hasn't been the most effective exercise but at least anyone can see that there is a difference.

Sync wasn't attenuated to -3 dB is since while 60 Hz, it doesn't nearly have a 50% duty cycle, so there's less 60 Hz power.

Note: if source impedance is increased (as in every mod) the effects should be worse since the source will create another time-constant. I'm not going to bother building every amp, too much time spent already.