] [Thread Prev
Re: Tubes versus Modeling Caged Death Match
At 11:40 PM 4/27/2003, Mark Sottilaro wrote:
>I said this a while ago and a friend of mine started quoting it in her
>signature file and I thought it might shed some light on this subject:
>"You can't simulate dressing like a clown. It's a do or not do thing.
>If you're wearing a clown suit, you're a clown."
>I hope this helps.
perfect Mark, thanks.
for a reference point, I spend a lot of time in my professional life
simulating electronic circuits. In my case it is usually the output driver
circuits on high-performance IC's driving PCB traces, assorted parasitic
impedances, connectors, and the input receiver on another IC. This is
reasonably comparable in complexity to simulating an audio amplifier. In
work I simulate transistors, but tubes would be a similar chunk of math
when you consider what the electrons are doing.
There are two approaches to this that I use. Either a fast "behavioral"
simulation, that runs quickly but with less detail. (or perhaps good
detail in the operating region I care about.) Or a detailed "transistor
level" simulation that gives a much more accurate result across all
operating regions but takes much longer (using a simulation engine called
spice, for those who care).
I usually simulate about 3 cycles of a given waveform. On the PIII 850MHz
machine on my desk the fast behavioral simulations take about 1 second.
slow, detailed, spice simulations can take as much as an hour and a half,
but they come out looking very much like the real thing. One of the guys
working for me has a P4 1.9GHz, when he's not around I VNC into his pc to
run spice. It speeds up by a factor of 5. Still, 15-20 minutes to get 3
detailed waveform cycles is a long way from real time. My guess is your
tube amp modeler isn't doing that kind of detail....
Kim Flint | Looper's Delight
firstname.lastname@example.org | http://www.loopers-delight.com