lordpil's blog

so for 1.2A, R2 = 0.7V / 1.2A, or 0.58 ohms
i just see underscores
some of your funk ass characters just show up in my terminal as underscores
maybe utf8 issues, shrug
dont paste shit
R1 drops the input voltage minus two diode drops
B is beta?
so whatever r1 is, the current above 70mA goes into q2

Posted by renesis at 13:18 | permalink | 0 comments

_unreal_: R2 sets current across Q2 Vbe, which is a fixed diode drop
thats going to set current through Q1
base current of Q1 will be set current plus base current of Q2 divided by Vbe
select R1 as base current of Q1 and collector current of Q2, which sets Q2 base current through Q2 beta

Posted by renesis at 13:12 | permalink | 0 comments

kickit2: put a resistor from the opamp side of the capacitor to ground
try that, and then try connecting the Vin+ to ground through same value resistor as the feedback resistor
bad grounding or power?
omg say already
sounds like bad ground
anyway, cool

Posted by renesis at 13:03 | permalink | 0 comments

i wonder if thats a bias current thing

Posted by renesis at 12:58 | permalink | 0 comments

then buy some line drivers and use resistors
shrug, doubt it unless youre running some polarized caps backwards
maybe me

Posted by renesis at 12:53 | permalink | 0 comments

the two sides are usually loaded with a current mirror on the collector side and some sort of shared load on the emitter side
its a Vbe current source
ive built those
why not use the LM317 circuit?
you just have to calc the currents with betas
kickit2: its normal
kickit2: they sell 1% resistors in 5% values
but its a bit goofy because theyll be in between typical 1% increments in a lot of cases
is this for production or personals?
if for personals, cant you just by some h bridge IC?

Posted by renesis at 12:47 | permalink | 0 comments

shrug, is how it works with 5% values
anyway, look up the input voltage offset for the amp, multiply it by gain, and thats about what you should be seeing at the output
my guess is like 15 to 50 mV unless you spent $$$ on the amp
i just mean e24 values, not tolerance so much
330k not even supposed to exist in e96 1% values =)
_unreal_: almost every long tail pair ever has a current mirror
parts are usually a little better than spec
_unreal_: differential input circuits

Posted by renesis at 12:42 | permalink | 0 comments

also star ground back to power supply ground if you can
100mV is a pretty big ground lift for a little circuit, tho
do you have more?
put like, 3 each in parallel
you were maybe just pushing the amp too hard
yeah, drop the resistance of each to 1/3
or just use something smaller, 10k and 33k was pretty reasonable

Posted by renesis at 12:37 | permalink | 0 comments

maybe bad grounding?
is the Vin+ grounded directly or through a resistor?
also scope the voltage rails make sure they arent oscillating along with the circuit somehow
i guess very high input bias current could also offset the output
you can put a resistor same value as your input and feedback resistors in parallel for quick input bias current compensation
if theyre equal that should work
if the transformer has high output impedance, i wouldnt just expect that to mess with gain, not offset
what are they now/
should be fine, maybe a little low

Posted by renesis at 12:32 | permalink | 0 comments

connect the inverting amp to ground and see if you get what you think you should
this is without offset divider?
do you get output offset with no current through the transformer primary?
if you have a really, really shitty opamp it could just be input offset voltage *3
is the no signal at ground or is there offset?
but with signal there is DC bias?
it would need to be like 64mV, which is bad
but i guess not impossible

Posted by renesis at 12:26 | permalink | 0 comments

are you seeing this on a scope?
how are you setting the offset?
what is the transformer referencing?
right you have one side connected to the input of the inverting amp
what is the other side connected to?
the divider?
is there a problem with your period key?
its like half an ellipses, it makes you seem like an ass

Posted by renesis at 12:21 | permalink | 0 comments

need more information
in general, no the frequency shouldnt offset the dc bias if everything is ideal
a CT? center tap?

Posted by renesis at 12:16 | permalink | 0 comments

cheater, blackmoon is upset at flyback because of tradition
this is like trying to reprogram a 74 series inverter

Posted by renesis at 03:00 | permalink | 0 comments

flyback: you already linked this here

Posted by renesis at 01:44 | permalink | 0 comments

i think maybe V_02 to do the loop

Posted by renesis at 00:46 | permalink | 0 comments

depends on if doing standard or electron flow i guess
standard flow, its just I_2 * 20R
the + and - indicated are correct polarity on the 10R and 20R assuming the current is positive and this is standard flow
jero32: np

Posted by renesis at 00:35 | permalink | 0 comments

voltage across the 20R resistor?
you need to label the nodes
jero32: can be configured for various servo controllers, too
oh i see that in the corner
V_20 is referencing the bottom of that resistor, V_02 is referencing the top

Posted by renesis at 00:29 | permalink | 0 comments

thats done in the controller
gcode just specifies some basic machine configuration, like units, absolute versus incremental movements, and then everything is basically coordinates and speeds
go to this XYZ location at this feedrate
efsnable: usually yes, voltage at point A referencing point B
so Vbe is voltage at the base referencing the emitter
jero32: yes, GRBL
i would look into linuxcnc stuff
its pretty standard, very developped
i dunno its kind of basic, honestly ive never used it ive just heard odd things
linuxcnc is much more typical for a machine controller
but you would set up step incriments and maximum accelerations and speeds in the controller configuration
jero32: gcode is basically gerber files
cnc gcode is rs274, gerbers are rs274x, its a subset for photoplotters

Posted by renesis at 00:24 | 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