lordpil's blog

i actually redrew the opamp because i HATE inv input on top and non inverting on the bottom
then i had to learn how the pin associations work in the subcircuit files
its not so intuitive really!
kk, nite
btw thanks for convo i feel alot better than i did like an hour ago

Posted by renesis at 16:29 | permalink | 0 comments

i thought that was going to be some form of self deprication on your part
welcome to the human condition
so your are concious of it so you have no excuse now
fuck i should really wake up early and take the deadphone amp to work to test
yeah you download the file
you put it in the directory with your simulation file
you do .include NAMEOFFILE.MOD
there is like a '.op' button or something, use that
then you name the BJT same as file
not the ref designator, like Q1, change the value
so like .include 2N3904.MOD
then just use 2N3904 as the BJT value
opamps are same way, you just have to know the right generic opamp file to use

Posted by renesis at 16:24 | permalink | 0 comments

not so much
its good to have but i dont think i would have gotten it just for that
i was a cnc tech student
and i bought it with grant money
so basically free money
if you want a precision machine tool its worth it
haha why?
honestly the amount of shit i get done is pretty sad compared to how much shit ive halfassed or abandoned
i mean i guess its all learning experiences
oh whatevers its only midnight

Posted by renesis at 16:19 | permalink | 0 comments

values of like 10K to 50K are common, any higher and current thru the parts is very low, making them sensitive to EMI
lower values tend to be less noisy
curiously strong, and i never finished them
i did do a curiously strong headphone amp
i want to do the full set, but i want to spend a good amount of time on them and do them up right
i dont think i not blogged it, have to look for the pics
one day ill actually wake up early so i can test it for THD+Noise and freq response at work
those are the homebrew pcb i spent like 4 manic days machining
they werent designed to be homebrew, china just took to long to tell me they were CNC retards
really i waited to long to start
looking for headphone amp...

Posted by renesis at 16:14 | permalink | 0 comments

1/(2*pi*24000*.0000022) 3.0142981646192298441076470335703
24K and 2.2uF is 3Hz, you see shit like this alot
well, for full range amps, for system specific amps matched to drivers and cabs, it could end up being anything really
anyway, night
well, however its setup
if youre going straight into the bases, impedance will be very high, basically a function of your base current
but typically, you take a resistor from the input the ground
on a non inverting input, and this is your imput impedance
on an opamp, it basically *is* your input impedance
on a BJT input, its in parallel with the impedance seen thru the BJT base
usually much higher, as your run the diff input section BJT at low collector current
yes theres an equation with a constant
that i dont quite remember
but basically, you put a resistor there to set up a consistent input impedance

Posted by renesis at 16:09 | permalink | 0 comments

yeah that works
try putting caps on the input to get it to roloff where you want
put like 50R or 100R in series with your source to simulate output impedance, itll affect your filter poles
ive actually fucked up like that, because im using to doing shit after low impedance opamp buffers
so i calced agressive filter poles without taking that into consideration
fucked up response in the audio range until the values were reworked to much more conservative ones
coupling the input?
ac coupling?
or you mean decoupling like adding a LPF
yeah that cap + input impedance determines your first HPF poles
typically values from like .47uF to 22uF are used, standard input impedances are like maybe 10K to 50K
so like, poles from DC to maybe 20Hz are pretty common for full range amps

Posted by renesis at 16:04 | permalink | 0 comments

danielson: tell me what gain its flat at before i go to bed
suspense is killing me or some shit
cause yeah if it does fine in transient, i dont think it would break in ac analysis
theres only like one or two parameters for ac analysis, hard to fuckup
whats yur Y scale?
it is
you have it thru a divider?
there you go
gain is 40db
sounds like class AB to me
you can do V*I in the plotting in ltspice somehow
it accepts math in the ploted parameters
ok cool, its neat it knows to do the scale in watts

Posted by renesis at 15:59 | permalink | 0 comments

but if youre doing 10Hz to 100KHz, a flat line is not very surprising, even if it isnt so realistic
anyway, for audio, you dont want that, you want it to roloff somewhere in the hypersonic spectrum
anyway, i gotta go crash out
if its 0db and flat, it means you did it right, its just maybe not a very realistic simulation
if its like, -80db flat or something, then your got probs
actually since you doing an amp and not a signal level section, it should have positive gain
but 0dB is unity, so it means its passing signal

Posted by renesis at 15:54 | permalink | 0 comments

should be fine
whats the error?
you might be right, but i think itll sim with generic silicon, just using typical parameters
sounds like a circuit prob
most likely your circuit is such that most BJT parameters dont come into play enough to make it fail
is it flat @ 0db?
because that actually means the amp resonds exceptionally well
in fact, if this is ultra simplified and contains no capacitive poles, mathematically the response might be flat well into the MHz range
usually after that, parasitic values will cause the response plot to roll off anyway

Posted by renesis at 15:49 | permalink | 0 comments

it could be something from the oscillating circuit, kinda kicking back into the source
your input impedance and your output impedance form a divider, so like even tho you have a voltage from your source, it doesnt brake math to be producting a voltage on your input, too
<tecan> i thought cheese comes from horses
danielson: we thought that was funny!
i dunno but it makes me laugh
because you know somewhere in the world, they digon horse cheese fureal
naw you prob dont, alot of times its just not funny in context
thats why it ends up in a topic all by itself
why does that page zoom at me funny after its loading
like its tryin to punk me or something

Posted by renesis at 15:43 | permalink | 0 comments

but at high frequency, that 1uS is significant
compared to a 2uS, or 500KHz wave i guess?
its 180deg negative phase
so at the higher frequencies, this phase reversale basically turns your negative feedback into positive feedback
positive feedback causes a boost in the frequency domain, or oscillation
low pass the input is a good idea
100pF caps are pretty good for this, with mosr common audio resistances
the poles are usually out of audio range
however, filtering in the circuit itself can actually make the problem worse
due to the phase reversal thing i mentioned
anyway, your source, prob a sig generator, has an output impedance, its most likely not an ideal source like an opamp
this is actually standard for signal audio output circuits, it protects outputs from shorting and helps fend off ringing from capacitive signal lines
yeah that sounds correct
but this means that the noise may not be coming from your source

Posted by renesis at 15:38 | permalink | 0 comments

like, you maybe dont have a low pass filter pole in there to reject hypersonic signals
you might not hear it, but its likely causing distortion in the amp
the noise can basically overwork the amp
well, im not sure
but since its AB, you guys are prob running feedback back to your input stage
like, you have a differential amplifier, with one side hooked to you input
and the inverting side hooked into your output thru a divider
okay well your input is tied to your output there
okay check this out tho
if there is propegation delay thru your amp
right, say it takes 1uS for your amp to react to input
because of various filter poles and response times of transistors
naturally, you get a negative phase shift
the higher the speed, the more this delay is significant, the more negative phase
because 1uS is no kind of delay is your sound signal is cycling at like 100mS

Posted by renesis at 15:33 | permalink | 0 comments

prob mixed i doubt theyd be teaching you chipamps
hehe i did some analysis on a discrete amp, walked a speaker engineer thru the circuits on a sub amp
er, did some analysis today, i meant to say
so like, a normal amp =)
like a long tail pair for input/feedbck, driving a bias circuit hopefully a Vbe multiplier, and maybe some buffer transistors before the AB section
long tail pair is the dual bjt differential input section
thats not audio yo
its oscillating tho
so basically its working the amp out for no reason
positive feedback
either EMI from the circuit itself, possibly something about how your differential input is setup
you could be picking up some EMI on the input, and its just doing its best to amplify it

Posted by renesis at 15:28 | permalink | 0 comments

*typos, find them
also, LTspice has all the raw spice directives pretty out in the open
like, all the clicky dialogs for setting parameters will actually generate the associated spice command in front of you
and thats how it notates it in your drawing
so its also a good tool to learn standard spice, i dont totally remember but i think even pspice tended to hide that shit
i kinda remember psice being somewhere in between ltspice and multisim in its tendency to obfuscate the raw spice directives
discrete or opamp based or mixed?

Posted by renesis at 15:23 | permalink | 0 comments

22:01 <@danielson> Can LTSpice accept spice libraries freely available online?
is that a joke or what?
(yes its standard spice, duh, else it wouldnt be useful)
theres an ltspice yahoo group
prob the best single source of information on using and building spice libs/mods/subcircuits/whatevers(same shit)
im fucking tired already
this full time job + school thing is challenging
yes theres standard spice models and their proprietary bullshit or other simulation app files
its netlists and values
and honestly, most 'different types' of spice files are the basically the same shit with diff file extensions

Posted by renesis at 15:18 | permalink | 0 comments

m$ exchange server is wasting my life

Posted by renesis at 01:38 | permalink | 0 comments

damn =\
everyone at work uses the china digi micrometers
im the only one that busts out the dial and vernier calipers we have =\

Posted by renesis at 01:09 | permalink | 0 comments


Posted by renesis at 01:01 | permalink | 0 comments

Top | Add to Technorati Favorites

© 2007 lordpil.   XHTML 1.0! CSS! Site design by GNAA  Blog Engine by pbx | MULTI2 | ian hanschen | lolwat