Graphene Semiconductor/Transistor Technology

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I can't wait for seeing a 100GHz computer, for audio apps though never think, but it's really intresting to have a new sound around... maybe in a couple of years we would be sayin... 100% silicon singal path like a vintage sounding something, maybe it will be really linear for those who look to wires with gain for pres, maybe we reach lower noises, maybe it sucks for audio, maybe really nice sounding could be get from it... a lots of maybes, who knows! I've born in 1990, it's the first big change in electronics I've ever seen, really excited about it!

JS
 
this is interesting, a friend of mine was explaing graphene to me about a year ago, it really did sound too good to be true, the only draw back was making it semi conduct, so now it seems they have done it, so laptop procesors running at super high speeds with ultra low power draw due to near perfect effiiency are now within our grasp.

I'll find out if my mate was involved in this research.
 
Its just neatly ordered carbon atoms really, can anyone tell the difference between the sound of a metal film resistor and a carbon film?  Carbon comp resistors change in proportion to the voltage across them apparently, but that just gives slight compression, I wouldn't hold my breath for some new sound, maybe low noise will be the most likely outcome.
best
DaveP
 
It's one thing to develop the process, it's another thing to have the libraries for the transistors etc that allow for low noise.

(my understanding is that they are two separate things... one is a chemistry/physical limitation, the other is a development tools issue)
Normally IC designers get a set of components they can use from the "process development" folks.

With most new processes, silicon component focus on low voltage, high speed transistors, and small resistors/caps... not on low noise components.

to get the matching required on resistors etc, we usually like to use things like laser trim etc... but even that requires a process qualification.
 
JohnRoberts said:
Silicon carbide is the actual semiconductor junction. Making devices capable of switching up at tera Hertz is not for linear analog use.

JR

That's why I'm more axious about seeing computers with this tecnology, maybe with this much processing capability we could simulate every single transistor and any part of some rig and use plugins with this, so instead of simulating the response of the whole equipment we simulate the parts and get a much closer simulation, also in this case we could use ideal caps or transformers in there or whatever imposible to archive real component.

JS
 
joaquins said:
JohnRoberts said:
Silicon carbide is the actual semiconductor junction. Making devices capable of switching up at tera Hertz is not for linear analog use.

JR

That's why I'm more axious about seeing computers with this tecnology, maybe with this much processing capability we could simulate every single transistor and any part of some rig and use plugins with this, so instead of simulating the response of the whole equipment we simulate the parts and get a much closer simulation, also in this case we could use ideal caps or transformers in there or whatever imposible to archive real component.

JS

Depending on what you are trying to sim or model.  DSP can already execute audio filters using unit delay math, instead of RxC time constants, completely free of any real world non-ideal component behavior. 

If you are trying to model non-ideal behavior it's more about the models than processor ticks (imo).

Higher computer power might be useful for finite element analysis/simulation of real acoustic instruments with expanded resolution detail,,, or not. 

JR
 
So we're really looking at advancements here for things like thyristors and speeding up logic circuits? Interesting.

The idea of completely simulating things based on their reactances would be a ton of fun. We could have component libraries of parts that look like the original parts.. Like pulling out an Ohmite Carbon Comp resistor and connecting it to a simulated JJ 12ax7..

I must admit though, at that point I would definitely miss my involvement with building (or trying to). It would be like turning a knob in a plug in with your mouse. Imagine turning the real knob. Getting a package full of parts for a board I had etched and waiting is a great feeling. I wouldn't miss drilling holes though! Not unless I was amazing at doing the chassis metalwork from the ground up.. Dave.
 
JohnRoberts said:
joaquins said:
JohnRoberts said:
Silicon carbide is the actual semiconductor junction. Making devices capable of switching up at tera Hertz is not for linear analog use.

JR

That's why I'm more axious about seeing computers with this tecnology, maybe with this much processing capability we could simulate every single transistor and any part of some rig and use plugins with this, so instead of simulating the response of the whole equipment we simulate the parts and get a much closer simulation, also in this case we could use ideal caps or transformers in there or whatever imposible to archive real component.

JS

Depending on what you are trying to sim or model.  DSP can already execute audio filters using unit delay math, instead of RxC time constants, completely free of any real world non-ideal component behavior. 

If you are trying to model non-ideal behavior it's more about the models than processor ticks (imo).

Higher computer power might be useful for finite element analysis/simulation of real acoustic instruments with expanded resolution detail,,, or not. 

JR

I was talking about physical modeling instead of the conventional DSP digital modeling, I'm thinking something like spice, where you draw the schematic, select the componenst (linear and not linear) from a library, a library that has real models and ideal models, in case we want to avoid one kind of distortion, so we could select from a crappy ceramic or electrolytic cap, a grater NPO or polypropylene one or an ideal one, even extrapolate an NPO of 470µF or whatever, the same with opamps, transistors, resistors, etc...

I see 2 big advantages to this, the first one that I'd mention is (maybe) grater plugins, that sim real analog gear even better than they are now, the second one and really exciting is that you could listen some design in the paper (computer) and trim there the design befor going to real world... would be a great tool for developing new analog gear and save the developer a lot of money in burning tubes and time before getting what he want.

I still love analog stuffs, I can't get a tape delay plugin that sounds even closer to my real basic unit, neither can't make sound a snare of a band I'm mixing right now without my 2150 VCA symetrix comp...



Hey Dave, if your real concern is about touching the knob a MIDI/USB controller is the solution, you still could build the real stuff, build acoustical treatment for the room, a DSP (I don't thing would be necesary having a 100GHz procesor in the computer, but probably really good to live performing) or whatever, but if you want to mod some of your builds you could listen a simulation before taking the iron instead of reading a couple of numbers from spice...

JS
 
JohnRoberts said:
I never messed with it but IIRC Peavey made a tube amp sim where you could mess around with the internals of the virtual amp.

JR

wow cool! I hope I can find that and it still works with modern operating systems
 
Back
Top