[Ktechlab-devel] chewin the phat

[Alan Grimes]

yo,

I dropped some code earlier this evening. There is no public repository
yet so if you want it, you'll have to bug me about it.

Right now the main focus is on stabilizing what's there. Because
Ktechlab is an extremely ambitious project which appears to be designed
to support robotics development -- the mechanics part is only in
skelitin form right now. It will need an extremely stable platform.

For the next few days I'll be focusing on getting a better understanding
of the engine, and improving the base classes. Much still needs to be
done to get the pointers under control and considerable refactoring
still needs to be done. Already on the agenda is the move of the LU
based Linear algebra engine to the new math library.

Another thing I need to do is study how each of the data structures is
actually used by each of the classes. Just now I came across code that
was trying to use a list to provide set-like functionality.

Basically I want it to do everything it's doing now correctly, then I
want to get it to do everything in this ancient rat-shack soft-cover
notebook called "getting started in Electronics"... From there we can
start thinking about enhancements.

For one thing, I'd like to see the zig-zag for normal resistances and
the addition of color-banding for factory-spec resistors.

Another problem is that logic gates are currently powered by ZPMs which
have practically infinite power generating capacity... =P There needs to
be some kind of special power node that sits on the schematic and powers
all logic gates on the same page so that it is possible to integrate the
designs with the necessary analog power circuits.

/me just browsed through that book, found a cool counter circuit, and
then implemented a circuit that counts from 0-9 then resets to 0. =)

oh jeez, it's 2AM already. =\

One problem with many of the circuits in the book is that they are light
sensitive. They may use one of any number of light sensitive devices to
do their thing... -- that seems like a function for the mechanics
simulator???? There are also many audio circuits... It would be useful
to be able to pump the output through a safety filter and then to the
sound card...

What I wanted to talk about is an amplifier that could fit on my
miniscule budget.

I want to learn more about how transformers behave when used as signal
devices. -- Eventually thinking about building a 2A3 amp based on the
RC-20. -- each transformer being $100. Since I don't have any money, I
want to prototype the design on a much smaller scale.

The tube I have in mind is one based on a Russian design (the 6N8P) or
6H8[pi] in crylic. It is currently being produced and marketed as the
ECC99 by JJ/tesla. It's normally a pre-amp/driver tube but it is beefy
enough to use a power tube. =P

The output transformer would be a simple Hammond 1609.

I could then drive each channel with a single signal transformer wired
in push-pull, and then use an op-amp to measure the drive current
required to operate the thing.

+100v,   +,- 15v,  -6v and  6.3 OR 12.6v AC or DC (the heaters can be
wired in two different configurations. The two tubes could be wired in
series or paralell for even more configurations and can be run off of AC
at 6.3v, though would probably be too loud at 12v)

I'd need about four different signal voltage supplies, not counting the
heaters. =\ I was hoping to simply rectify the power line and get 150v
for the main part of the amp... I really can't afford a power transformer!!!

That's probably crazy though... even if it worked, it'd be noisy...

-- 
|/-\|/-\|