Trace latency on the N64

ShockSlayer · **Posted:** Wed Jun 16, 2010 11:18 pm

Does anyone know for sure it is exists or not?

Electric_rain and I were chatting last night and we couldn't come up with much, but I did find a few articles and junk:

Clock speed of n64's ram: 500MHz

http://www.physicsforums.com/showthread.php?t=206271

http://blogs.mentor.com/hyperblog/blog/tag/trace-delay/

http://www.technick.net/public/code/cp_ ... microstrip

I figure Ram would be the most sensitive out of everything, so I was working off of there.

If anyone can offer any input that could help then please let me know.

We did throw around a few things, from what I can tell having shorter traces would not cause problems.

SS

ShockSlayer · **Posted:** Thu Jun 17, 2010 6:58 am

http://www.youtube.com/watch?v=CCizUqujueQ

That video had no real use until now.

Maybe I'm just retarded.

Maybe.

Beta and I say this video proves there is no latency on the chips.

SS

Basement_Modder · **Posted:** Thu Jun 17, 2010 7:06 am

I say connect the exp. pak directly to the RCP and cut off the whole section of mobo that once had the RAM chips. (and rewire some flax.)

And no, latency is not an issue on the N64 RAM because they designed the system to potentially have RAM upgrades, thus extending traces.

Bibin · **Joined:** Tue Jun 02, 2009 5:43 pm **Posts:** 1469

Given the fairly slow nature of n64 RAM and that of other systems of that age, it shouldn't matter much until you get into lengths greater than like 2-3 inches or so - the main problem there is signal attenuation and data noise, though.

Basement_Modder · **Posted:** Thu Jun 17, 2010 7:19 am

We're discussing making it shorter, not longer.

Also, 2-3" would not work. Unless you mean 2-3" from the RCP and not from the original solder points.

XCVG · **Posted:** Thu Jun 17, 2010 2:54 pm

Didn't Bac hit a clock trace and screw up his board? If that's true, methinks that would be the only hypersensitive trace.

bentomo · **Posted:** Thu Jun 17, 2010 3:22 pm

This would be something to ask kibble about from the benheck forums, if he's around that is. He should know if there's any latency.

Basement_Modder · **Posted:** Thu Jun 17, 2010 5:31 pm

XCVG wrote:

Didn't Bac hit a clock trace and screw up his board? If that's true, methinks that would be the only hypersensitive trace.

That was all flax. He just fried his board and said that's what did it.

jleemero · **Joined:** Sun Jun 07, 2009 1:17 pm **Posts:** 1936

I am the closest thing to an Electrical Engineer on ModRetro.
The latency caused by traces on the N64's PCB are negligible the vast majority of the time.
There is no reason for RAM to be any more "sensitive" to this latency than anything else.
99% of the failures and difficulties people have had in the past regarding this matter were NOT related to Trace Latency.

NOW, the semi-bad news.
The way the CPU and RCP talk is, in my opinion, a bit bizarre.
Understand, that the latency caused by traces really ought not effect even this.
But, variance in length could.
Between these two, pretty much the fastest communication that takes place outside of the CPU or RCP occurs.
But listen, if you are halfway decent at soldering, this should NOT be a problem.
Just be careful and kind to it.

Basement_Modder · **Posted:** Thu Jun 17, 2010 6:28 pm

Imma be a CE or EE does that count?

Trace latency on the N64

Who is online