| Author |
Message |
yanquijiang@gmail.com
Guest
|
 Posted:
Sun Dec 11, 2005 9:35 am Post subject:
what is transit frequency? |
|
|
Hi, all
It just occurs to me: f_t is defined as g_m/c_gs, but as some
text books say, this quantity is dependent on overdrive voltage, so as
we crank up the overdrive voltage, f_t will grow, but when a guy says
he has a device with f_t=10GHz, what does he mean? and under what
overdrive voltage he gets the result? maybe something wrong with my
thought.
Thanks
|
|
| Back to top |
|
 |
Tim Williams
Guest
|
| Posted:
Sun Dec 11, 2005 9:35 am Post subject:
Re: what is transit frequency? |
|
|
Op-amps I presume?
Check the datasheet. It should list conditions.
Tim
--
Deep Fryer: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
<yanquijiang@gmail.com> wrote in message
news:1134273276.904493.19680@g49g2000cwa.googlegroups.com...
| Quote: | Hi, all
It just occurs to me: f_t is defined as g_m/c_gs, but as some
text books say, this quantity is dependent on overdrive voltage, so as
we crank up the overdrive voltage, f_t will grow, but when a guy says
he has a device with f_t=10GHz, what does he mean? and under what
overdrive voltage he gets the result? maybe something wrong with my
thought.
Thanks
|
|
|
| Back to top |
|
|
Pooh Bear
Guest
|
| Posted:
Sun Dec 11, 2005 9:35 am Post subject:
Re: what is transit frequency? |
|
|
Tim Williams wrote:
| Quote: | Op-amps I presume?
|
Transistors.
Graham
|
|
| Back to top |
|
|
Tim Williams
Guest
|
| Posted:
Sun Dec 11, 2005 9:35 am Post subject:
Re: what is transit frequency? |
|
|
"Pooh Bear" <rabbitsfriendsandrelations@hotmail.com> wrote in message
news:439bac96$0$63060$ed2e19e4@ptn-nntp-reader04.plus.net...
| Quote: | Op-amps I presume?
Transistors.
|
Transistors don't have "overdrive voltage", that I am aware of. (You've got
me if these are RF FETs, something I know nothing of..)
Tim
--
Deep Fryer: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms |
|
| Back to top |
|
|
Paul Burridge
Guest
|
| Posted:
Sun Dec 11, 2005 5:35 pm Post subject:
Re: what is transit frequency? |
|
|
On Sun, 11 Dec 2005 01:31:18 -0600, "Tim Williams"
<tmoranwms@charter.net> wrote:
| Quote: | "Pooh Bear" <rabbitsfriendsandrelations@hotmail.com> wrote in message
news:439bac96$0$63060$ed2e19e4@ptn-nntp-reader04.plus.net...
Op-amps I presume?
Transistors.
Transistors don't have "overdrive voltage", that I am aware of. (You've got
me if these are RF FETs, something I know nothing of..)
|
Nope. Never heard of the term, either. Transition frequency is simply
the point where the gain of a device falls to unity.
--
"What is now proved was once only imagin'd" - William Blake |
|
| Back to top |
|
|
Winfield Hill
Guest
|
| Posted:
Sun Dec 11, 2005 5:35 pm Post subject:
Re: what is transit frequency? |
|
|
Ban wrote...
| Quote: |
yanquijiang@gmail.com wrote:
It just occurs to me: f_t is defined as g_m/C_gs, but as some text
books say, this quantity is dependent on overdrive voltage, so as
we crank up the overdrive voltage, f_t will grow, but when a guy says
he has a device with f_t=10GHz, what does he mean? and under what
overdrive voltage he gets the result? maybe something wrong with my
thought.
In the first place f_t is defined as the frequency, where |beta|=1.
You seem to refer to a FET, as you write c_gs.
|
Generally the parameter f_T applies to BJTs, where it's very useful.
But it can also be applied to FETs, and may be useful, if they are
driven from their sources, or by a high impedance on their gates.
| Quote: | Since also this capacitance varies with Id, you cannot make this
assumption.
|
Actually, Cgs varys very little with Id or with gate voltage. So
the calculation for a FET really comes down to its transconductance,
which as expected increases strongly with Id. There's a tendency to
use the highest possible current, Idss for a JFET, or Id(max) for a
MOSFET. But this isn't very useful, because in practise one usually
tries to avoid operating a FET at its maximum current. An exception
would be for low-current JFETs, which are often used at Idss.
| Quote: | Overdrive was once popular in british sports cars, do you mean base
current, which refers to another type of transistors? Are you
operating the device in the linear region? A FET is indeed voltage
driven, but just of the right amount hopefully.
|
Depletion-mode JFET's Idss current is defined with Vgs = 0V. This is
supposedly their maximum, but the gate diode junction can be forward
biased a bit, further increasing their current and transconductance,
and thus their f_T, although such an "overdrive" would be cheating.
:-)
MOSFETs can also be "overdriven," but shouldn't, because they already
have a serious heating problem at their spec'd maximum drain current,
since this is a value usually meant for switched use, with low Vds.
| Quote: | Maybe all the answers to your questions are in your textbook, but
start reading at the beginning, not in the middle, whereall the
things have been already explained.
|
This ahem, subject, is not dealt with much at all in most books.
--
Thanks,
- Win |
|
| Back to top |
|
|
Ban
Guest
|
| Posted:
Sun Dec 11, 2005 5:35 pm Post subject:
Re: what is transit frequency? |
|
|
yanquijiang@gmail.com wrote:
| Quote: | Hi, all
It just occurs to me: f_t is defined as g_m/c_gs, but as some
text books say, this quantity is dependent on overdrive voltage, so as
we crank up the overdrive voltage, f_t will grow, but when a guy says
he has a device with f_t=10GHz, what does he mean? and under what
overdrive voltage he gets the result? maybe something wrong with my
thought.
Thanks
|
In the first place f_t is defined as the frequency, where |beta|=1.
You seem to refer to a FET, as you write c_gs. Since also this capacitance
varies with Id, you cannot make this assumption. There are data sheets for
each device, in which these parameters are specified in graphs or at certain
bias points.
Overdrive was once popular in british sports cars, do you mean base current,
which refers to another type of transistors? Are you operating the device in
the linear region? A FET is indeed voltage driven, but just of the right
amount hopefully.
Maybe all the answers to your questions are in your textbook, but start
reading at the beginning, not in the middle, whereall the things have been
already explained.
--
ciao Ban
Apricale, Italy |
|
| Back to top |
|
|
Winfield Hill
Guest
|
| Posted:
Sun Dec 11, 2005 11:55 pm Post subject:
Re: what is transit frequency? |
|
|
Winfield Hill wrote...
| Quote: | Ban wrote...
yanquijiang@gmail.com wrote:
It just occurs to me: f_t is defined as g_m/C_gs, but as some text
books say, this quantity is dependent on overdrive voltage, so as
we crank up the overdrive voltage, f_t will grow, but when a guy says
he has a device with f_t=10GHz, what does he mean? and under what
overdrive voltage he gets the result? maybe something wrong with my
thought.
In the first place f_t is defined as the frequency, where |beta|=1.
You seem to refer to a FET, as you write c_gs.
Generally the parameter f_T applies to BJTs, where it's very useful.
|
In part because, unlike FETs, C_be changes dramatically with Ic.
| Quote: | But it can also be applied to FETs, and may be useful, if they are
driven from their sources, or by a high impedance on their gates.
|
It's moderately useful as a figure of merit (FOM), although here it
can direct one off in the wrong direction, because often a "better"
FET is merely a larger one, operating at higher currents.
What can be useful is to calculate the FET's f_T at a fixed current,
from the g_m at that current, and the datasheet typical Ciss plot.
For example, say one is designing a low-dropout linear regulator with
a P-channel MOSFET pass element. g_m is proportional to current, so
we'd want to evaluate loop gain at the regulator's lowest and highest
load currents. Unlike BJTs, which have g_m = Id/V_T, very simple,
MOSFETs in the sub-threshold region (where one always works in such
circuits) have g_m proportional to Id / k V_T, where k is a constant
varying from say 3 to 10 for different MOSFETs. So when evaluating
various FETs for the task, we'd want to know k (not in datasheets)
and Ciss (which is), plus package type and the heat-transfer thermal
resistance, of course. One thing such an evaluation reveals is that
good thermal resistance goes along with large die size, which means
high Ciss. <sigh>
| Quote: | This ahem, subject, is not dealt with much at all in most books.
|
Ditto for the above subject matter. Maybe it'll make it into AoE 3.
--
Thanks,
- Win |
|
| Back to top |
|
|
Paul Burridge
Guest
|
| Posted:
Mon Dec 12, 2005 1:35 am Post subject:
Re: what is transit frequency? |
|
|
On 11 Dec 2005 09:55:15 -0800, Winfield Hill
<Winfield_member@newsguy.com> wrote:
| Quote: | Ditto for the above subject matter. Maybe it'll make it into AoE 3.
|
Speaking of which - any time frame as yet for publication?
--
"What is now proved was once only imagin'd" - William Blake |
|
| Back to top |
|
|
Winfield Hill
Guest
|
| Posted:
Mon Dec 12, 2005 7:43 am Post subject:
Re: what is transit frequency? |
|
|
Paul Burridge wrote...
| Quote: |
Winfield Hill wrote:
Ditto for the above subject matter. Maybe it'll make it into AoE 3.
Speaking of which - any time frame as yet for publication?
|
That's what our publisher keeps asking. And a half-dozen emailers
each week. We worked on it most of today. Commiserating about the
grand old TL011 and TL014 current mirrors, R.I.P. all these years
(page 90). Updating recommended FET part numbers, and adding some
other good stuff.
--
Thanks,
- Win |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in