f*****0
发帖数: 489 | 1 here is how "current feedback" works.
R1 is your load, R2 is the sampling resistor that converts the current going
through R1, I1, into a voltage signal to be fed back the inverting end of
the opamp through R6/R3.
so it follows that
I1*R2*R6/(R6+R3)=Vin.
or I1=(1+R3/R6)*Vin/R2.
I typically choose R2 to be < 1/10 of R1, and R3/R6 will depend on what
opamps you use and what input impedance you need.
if you use a regular voltage feedback opamp, I would chose R3=10k.
if no DC blocking capacitor is u |
|
f*****0
发帖数: 489 | 2 if you want to offset the 0.7v drop you can reference the output to a
negative Vf - that can be done with a resistor, a diode and a negative rail. |
|
f*****0
发帖数: 489 | 3 it is hard to find an opamp good for 10mhz, :).
the high voltage isn't an issue: you can use two resistors to attenuate it
down to a reasonable level. let's say 50vp.
then use a 1n4148 (with a resitor in serial) half wave rectification into a
small mica, ceramic or polyester capacitor (anything bigger than 1n will do)
. you will not be able to find a big ceramic or mica capacitor but since you
are running at 10mhz, you don't need a big capacitor to filter out the half
waves.
the whole thing shou |
|
f*****0
发帖数: 489 | 4
output
hooked
the pease article is for simplified schematic. yes, your design would work
but having a resistor (R9) there lowered the gain of the super linear pair (
Q7/Q9) (thus improves the stability) and also limits the amount of current
that can go through Q9.
correct on both accounts. |
|
f*****0
发帖数: 489 | 5 let me just answer one of your questions and then you will get a flavor of
how far off you are:
"3. 如果我们有一个constant current source, 那它肯定不应该随温度，power
source 变化而变化，对吧。总之，就应该总是constant, 对吧？"
what is "constant" is relative. a typical constant current source for
examples uses a series of diodes or zener or a LED to drop a "constant"
voltage off a resistor, in a fashion very similar to the widlar bandgap
topology.
in such a design, you consider the forward voltage off the diodes / zener /
led to be co |
|
f*****0
发帖数: 489 | 6 in general, you will never get 2x of the power from paralleling 2 amps.
there will always be some losses in there (you are likely to use two load
balancing resistors for example). |
|
H***F
发帖数: 2501 | 7 any information about the companies? |
|
|
G******y
发帖数: 86 | 9 我觉得opamp一般都是designed to drive capacitive load，要是去drive heavy
resistive load（small resistor）的话，首先直流工作点就被破坏了，再加上output
resistance降低-》Gain降低。
可是不管resistive还是capacitive load的稳定性都要具体分析吧？大的cap load也未
必一定稳定吧 |
|
g******u
发帖数: 3060 | 10 很简单，不需要编程。
去digikey上找一个现成的panel mount，带LCD的电压计，然后找个塑料盒开个洞，按
进去就可以。这样手持这个盒子就可以测量，然后数字会显示在面板上。我用过C-Ton
industries的，还可以，38刀。
这种电压计的电源是9V电池。你可以再用一个9v电池，接一个resistor divider和你的
变阻器，调整一下变阻器两端电压，直到输出接近你的要求值。
虽然这种电压计比较精确，你只要少接两根线，就是一个整数的输出，就做好了。 |
|
|
c*******l
发帖数: 4801 | 12 低频是resistor divider,
高频是直通 |
|
c******s
发帖数: 197 | 13 闭环的hall effect sensor （例如cln-50)应该是电流输出比较可靠，比如电路周围有
大电流流过，而在sensor附近直接用sample resistor转化成电压再进行传输（25cm左右
的ribbon cable）应该比较容易受到干扰。做了一个电路(PCB),发现电流传输更加容易
受到干扰， 不make sense，牛人帮我开导开导，谢谢 |
|
h**g
发帖数: 94 | 14 1) Transistor small signal model, gm, ft
2) Simple amplifer Gm, Rout: common source, common gate, source follower
3) Current mirror, Layout matching
4) Cascode amplifer, Cascode current mirror
5) Differential pair: gain, bandwidth, swing
6) Opamp with miller compensation
7) Opamp with feedback: resistor, capacitor
8) Bode plot, phase margin, pole shifting
9) Transistor noise model
10) Basic equations for diode and BJT
I took many phone/on-site interviews recently. All these questions were asked |
|
g****2
发帖数: 9 | 15 I had a PCB layout (.max) generated by old OrCAD program. Now I imported
this .max layout and opeded it in Cadenece PCB Editor. I want to create
gerber files based on this layout and some changes to the circuit.
My question is can I just do it in PCB Editor without touching the schematic
using Design Entry CIS?
More specifically:
1. How can I delete a connection line and add a new one between another two
components/nodes in PCB Editor?
2. How can I add a new component, say resistor, in the layou |
|
W***n
发帖数: 11530 | 16
For DISCRETE resistors, except for very special apps, the afore-mentioned
three parameters (resistance, power rating and tolerance) should be adequate.
If you go to VLSI level, it's a totally differnet ballgame.
If you go to nano level, even more parameters'll pop up. |
|
g******u
发帖数: 3060 | 17 A simple PID loop can be built just using some resistors, caps and opamps.
So if you have built some small scale prototypes with a know delay before (
if linear), you can tune them very easily.
For thermostat control, i think it is just a input-compare and amplify-PID-
delay-feedback loop, shouldn't be too hard to model in Matlab. There should
be some references online as well.
调. |
|
w****a
发帖数: 11 | 18 先谢谢大家！
The horizontal scale is 50ns/div. I'm using TI OPA656NB, the decoupling
caps are 6.8uF, 0.1uF and 200pF. Circuit is very simple, just the OpAmp and
decoupling caps and feedback resistor. |
|
s*******y
发帖数: 4173 | 19 Do you ever think about the huge area of inductor in chip. Meanwhile, it is
hard to get accurate value of capacitor, resistor, inductor. |
|
s*******y
发帖数: 4173 | 20 Do you ever think about the huge area of inductor in chip. Meanwhile, it is
hard to get accurate value of capacitor, resistor, inductor. |
|
h********t
发帖数: 555 | 21 biquad is a structure which can be realized in many different ways such as
gm-c, mosfet-c, switched capacitors, resistor+c+opamp, etc. Biquad is a
closed loop system. if you want to use gm-c in open loop,the gm will change
with temperature, process. if you don't care frequency response,gain
accuracy and linearity, feel free to do that. I don't see how that can
save power. |
|
le
发帖数: 190 | 22 if you don't care about power gain, why not just put a 50 ohm series
resistor |
|
b****e
发帖数: 8 | 23 I guess he was asking about the 50 Ohm resistor at 77K |
|
a*t
发帖数: 1309 | 24 Fry's
I saw they sell breadboard, resistor, capacitor,..etc at local store at
least. |
|
f*****0
发帖数: 489 | 25
light (a
most TIA.
I am still unsure as to why you need a transimpedance amplifier.
I believe we call that a "resistor". |
|
f*****0
发帖数: 489 | 26 quite a few problems:
1) you don't need a virtual ground. since you are using four batteries, you
can just ground the circuit to the true ground - center of the batteries.
and if you really really want to use the virtural ground for whatever reason
, put a small eletrolytic cap on the resistor and lower their value to no
more than 1k.
2) you absolutely need to have power supply decoupling or your opamp will
likely oscillate. the particular opamp is a pretty fast one (40mhz gbp), so
you may want |
|
n***r
发帖数: 25 | 27 I agree with cosmos. it could be charge pump.
It depends what kind of amplifier. Some amplifiers have charge pump
integrated. Some chopper amplifiers may have switches working at the
frequency as well.
It could also be the switching power supply problem.
Seems it's not oscillation. when circuit oscillates, it's usually not small
signal.
Why connecting 10kohm to vin+? You are adding resistor thermal noise to your
circuit. And it's easier to pickup external interference signal. |
|
f*****0
发帖数: 489 | 28
100
you have to understand the context under which I made the comment about
the gain. if you don't, re-reading it may help.
the purpose of that exercise is to get some idea about its a/c bandwith.
if you look at the left of the R1 resistor, the circuit is an inverting
amplifier, with the input voltage being I*R1, and the gain being -Rf/R1.
that is a typical process to design an I-V converter using opamps.
that has nothing to do with the signal source's output impedance, which
you implified in y |
|
n***r
发帖数: 25 | 29 Let me go over with you again.
;the gain of the amp is 500k/10=50k x, approaching most opamp's open loop
;gain, if not more.
Let's replace the 10 ohm with 0 ohm. Will your inverting amplifier theory
work? will the transimpedance amplifier work? Admit it, you were misled by
the 10ohm resistor then you came up with a wrong AC bandwith analysis.
;at that king of gain, the bandwidth of the opamp is very low, likely <10hz -
;check out the opamp's datasheet.
Wrong! The bandwidth of the Transimpedance |
|
f*****0
发帖数: 489 | 30
that much capacitance will be very difficult to drive at high speed. the
typical approach is to use a high current amplifier, and to use a serial
resistor (50 - 75ohm) to dampen the capacitance.
you may want to look into how ADSL modems solve such problems: though
the load there is more inductance than capacitance.
and they can achieve very high speed. |
|
f*****0
发帖数: 489 | 31
here is a simple two transistor solution. the resistor values are not
critical at all.
see if you can simplify it down to one transistor. |
|
w**i
发帖数: 136 | 32 It is a system, I really dont know what kind of 漂移 trapped me, but I guess
it is related to stability of capacitors and resistors, and maybe the
careless selection of chip manufacturers. |
|
M**********o
发帖数: 203 | 33 Do you guys have some good experiences or tricks of soldering? Thank you
very much.
Do you use magnifier? or desoldering wick? What's the diameter of the solder
? I feel the thinner the better... |
|
f*****0
发帖数: 489 | 34 1206 or 0805 parts are fairly easy to solder: you just tin the pads a little
bit, use a tweezer to push the part onto the pads and use a fine tip to
melt the tin on the pads.
sot23 parts can be done this way too.
for multi-pin chips, I coat the pcb with solder paste (mostly solder but
also some flux), put on the part, and use a hotair gun to melt the paste.
the trick is to use the right amount of paste: if you used too much, you can
always wipe them off with a fine multi-strand copper wire. it d |
|
M**********o
发帖数: 203 | 35 Thank you very much for your reply. I was heard of using hot air soldering
but haven't tried yet. I want to build a 2812 board, the DSP has 176 pins
and LQFP package. So it should be applicable, right. Will the solder paste
make the jumper(connect 2 pins) because the distance between each pin is so
narrow?
By the way, do you have some recommendations for buying a hot air soldering
station? Thank you.
little
can |
|
f*****0
发帖数: 489 | 36
soldering
pins
paste
so
that's why the paste has flux. But it does take practice to get it right.
If you used too much, just use a fine bread of copper wires and press it
against the pins with a hot iron.
soldering
how about a hair dryer? or an oven? |
|
a*****s
发帖数: 6260 | 37 176个脚的玩意？我强烈建议你去找专业人士干这活，业余选手烧糊这种片子简直
是一定的，何况从你问的问题来看连业余都不算。
so
soldering |
|
f*****0
发帖数: 489 | 38 lqfp is very doable and if you use an oven, it doesn't matter if you are
working on a 176 or 48-pin device.
it is the finer pitch or pin-less devices (mssop, or bga/lga, etc.) that re
a pain.
as to failure rate, yeah, it does taking some practice. |
|
M**********o
发帖数: 203 | 39 Thank you for your help.
re |
|
f*****0
发帖数: 489 | 40 it depends on what kind of current goes through the circuit. the simplest
solution would be to put on a current transformer on the resistor you want
to measure its resistance of. measuring the voltage is fairly easy. |
|
h********o
发帖数: 152 | 41 easy, add voltage, and measure temperature on every resistor. |
|
f*****0
发帖数: 489 | 42
he wants to measure two inseparable and unknown resistors.
that's an impossibility. |
|
g******u
发帖数: 3060 | 43 Only clamp meter is the right answer.
You can clamp it to the wire far away from the resistor, thus it won't
change too much of your soil dynamics. |
|
g******u
发帖数: 3060 | 44 有可能gate resistor太小造成gate ringing过大?建议换成22ohm试试.
这个是经过多久坏的? 坏的情况如何?
你用的是什么FET? |
|
g******u
发帖数: 3060 | 45 have you tried 22 ohm as gate resistor yet?
There is another way to find the problem. Don't use the upper FET at all
since it's synchronous FET anyway, won't help much. Use a function generator
to drive the bottom FET, (put an optoisolator/signal transformer or fuse
there) and see if the FET still dies. |
|
t****i
发帖数: 182 | 46 If you need it for I2C type of operations (open drain), you will need it to
be a bi-dir FET type of bus level shifter. Simple resistor based voltage div
ider won't work well.
Usual industry device for this type of application can be found with TI and
Philips (NXP). e.g : TVC3306.
http://focus.ti.com/lit/ml/scyb030/scyb030.pdf
2V
2V |
|
le
发帖数: 190 | 47 Try to understand resistor and capacitor...KCL, KVL, and charge conservation
If you know inductor, you can do RFIC design as well. |
|
ET
发帖数: 10701 | 48 昨晚simulted a 45nm nmosfet using predictive spice model.
attached here. the spectral density also included a 10K resistor's noise
performance tho.
the flicker noise corner frequency is in the range of ~Mhz.. |
|
a****l
发帖数: 8211 | 49 I think I need to choose a resistor for zener so as not to make zener
current too large or too small? |
|
g******u
发帖数: 3060 | 50 nobody reads what I typed....
use a LM317, connect a 62.5ohm resistor between output and adjust pin, it'll
give you constant 20mA at the output.
you only need two parts, 50 cents at most. |
|
