OP-Amp with output clamp
Gerhard Mesenich
AD-Converter. The sampling rate is approx. 1MSPS. The input range is 0-4 V.
The OP-Amps will have a dual 5V supply. The gain will be 1-2. The output of
the OP-Amps has somehow to be restricted to 0-5V, so that they do not
overdrive the AD-Converter.
A good choice as a PGA would for instance be the TI THS 7001/7002, which has
such an output clamp feature.
There should also be some simple OP-Amps on the market which have such
internal clamps.
Who makes these limiting OP-Amps?
Which types are available?
(GBWP approx. 5Mhz or better)
Thanks in advance.
Gerhard
rlaury
>
> I need some limiting OP-Amps for driving a fast single supply 12bit
> AD-Converter. The sampling rate is approx. 1MSPS. The input range is 0-4 V.
>
> The OP-Amps will have a dual 5V supply. The gain will be 1-2. The output of
> the OP-Amps has somehow to be restricted to 0-5V, so that they do not
> overdrive the AD-Converter.
Just add a 4.8V zener ACROSS your feedback loop.
A Murphy
I have used Commlinear CLC502 clamping amp in front of high speed ADCs.
They have quite good performance and not too much harmonic distortion close
(0.5V) to the clamp voltage.
Also I have a spec for Analog Devices AD8036/AD8037 that looks like it will
work even better.
Dont use a Zener if you are concerned about distortion since it will
introduce harmonic distortion even when the voltage is well away from the
Zener voltage.
A. Murphy...
John Larkin
wrote:
>I need some limiting OP-Amps for driving a fast single supply 12bit
>AD-Converter. The sampling rate is approx. 1MSPS. The input range is 0-4 V.
>
>The OP-Amps will have a dual 5V supply. The gain will be 1-2. The output of
>the OP-Amps has somehow to be restricted to 0-5V, so that they do not
>overdrive the AD-Converter.
>
>A good choice as a PGA would for instance be the TI THS 7001/7002, which has
>such an output clamp feature.
>
>There should also be some simple OP-Amps on the market which have such
>internal clamps.
>
>Who makes these limiting OP-Amps?
>Which types are available?
>(GBWP approx. 5Mhz or better)
>
>Thanks in advance.
>Gerhard
>
>
Look at the Harris HFA1130; 800 MHz GBW with clamps. Also Analog Devices has a
similar part, maybe 250 MHz, with better DC specs... can't remember the part
number.
John
Hans Dermot Doran
Used their stuff for video signal conversion in the early 90's
Hans
Gerhard Mesenich <mese...@t-online.de> wrote in message
news:851rlk$o2i$1...@news02.btx.dtag.de...
Jerry Avins
>
> I need some limiting OP-Amps for driving a fast single supply 12bit
> AD-Converter. The sampling rate is approx. 1MSPS. The input range is 0-4 V.
>
> The OP-Amps will have a dual 5V supply. The gain will be 1-2. The output of
> the OP-Amps has somehow to be restricted to 0-5V, so that they do not
> overdrive the AD-Converter.
>
> A good choice as a PGA would for instance be the TI THS 7001/7002, which has
> such an output clamp feature.
>
> There should also be some simple OP-Amps on the market which have such
> internal clamps.
>
> Who makes these limiting OP-Amps?
> Which types are available?
> (GBWP approx. 5Mhz or better)
>
> Thanks in advance.
> Gerhard
You can build op-amp bounds with two resistors and a diode for each
bound (upper and lower). If the diode drop is unimportant, you can bound
to zero with a diode alone. You will have an easier time if you can set
the bounds to slightly exceed the range of the converter and use a
converter that itself saturates on out-of-range signals.
Jerry
--
Engineering is the art of making what you want from things you can get.
-----------------------------------------------------------------------
Gerhard Mesenich
Let me share some information after a lot of checking and consideration:
1) Clamping with Zener diodes:
As A.Murphy has already pointed out, Zener diodes are close to useless in
the range below 5V. Just for education I have looked them up in a data
sheet. Small Zeners are classified at a current of 5mA. For instance:
Leakage current for a small 4.7V Zener (1N750) is 1mA! @ 4.3V, 0.1mA @ 3.5V
and still 10uA @ 2.5V. Capacitance is approx. 100-200pF! @ 2V. This is just
another warning to anybody who might be tempted to try.
Things improve a lot at higher voltages. A 9V Zener for instance drops to
less than 0.1uA at approx. 8.5V. Capacitance is still about 50pF@ 5V. With
a lot of caution Zeners can be used as clamping devices in precision
circuits at such higher voltages only.
2) Clamping with signal diodes:
This works very well in most cases. Jerry Avins has pointed out that an
op-amp can be bound to any limit with a diode and a devider at the output of
the op-amp. A 1N4148 has leakage in the nA range. Capacitance is approx. 4
pF. Forward current is approx 100nA @ 0.2V and 1mA @ 0.65V. The feedback
divider needs a relatively low impedance, adds complexity and increases
current consumption. It also introduces nonlinearities if the feedback
impedance is too high.
3) Limiting amplifiers:
Limiting amplifiers introduce a lot more of distortion than I thought they
would. Thanks to Hans and John for the hints.
4) Using a single supply op-amp with a neg. reference:
This is my new clean approach. I am going to use single supply inverting
op-amps. They are inherently self limiting. The op-amps will be connected to
ground on the noninverting input and the output can be centered at mid
supply or any other convenient level by introducing a negative offset
current fom a precision negative reference of -2.5V. (another inverting
op-amp) The gain can easily be set by changing the input resistor. The
inputs can reliably be protected by 2 antiparallel signal diodes to ground.
The negative supply must supply a very low current only, which is constant
(approx. 1mA per amp in my case) and can easily be generated by a cheap
charge pump. (MAX 1697)
PS
I have posted some rather complicated questions on 12.25.99 in comp.dsp
regarding the interface to a TI 6211 DSK. Does nobody dare trying to answer
at least some? -:)
Gerhard
Gerhard Mesenich <mese...@t-online.de> schrieb in im Newsbeitrag:
851rlk$o2i$1...@news02.btx.dtag.de...
Jerry Avins
>
> Thanks to all for the answers.
>
>
> 4) Using a single supply op-amp with a neg. reference:
>
> This is my new clean approach. I am going to use single supply inverting
> op-amps. They are inherently self limiting. The op-amps will be connected to
> ground on the noninverting input and the output can be centered at mid
> supply or any other convenient level by introducing a negative offset
> current fom a precision negative reference of -2.5V. (another inverting
> op-amp) The gain can easily be set by changing the input resistor. The
> inputs can reliably be protected by 2 antiparallel signal diodes to ground.
> The negative supply must supply a very low current only, which is constant
> (approx. 1mA per amp in my case) and can easily be generated by a cheap
> charge pump. (MAX 1697)
>
...
Be aware that most op-amps exhibit significant delays coming out of
saturation. A comparator is essentially an op-amp in which linearity and
stability under feedback have been traded away for fast recovery time.
Crackpot
>Be aware that most op-amps exhibit significant delays coming out of
>saturation. A comparator is essentially an op-amp in which linearity and
>stability under feedback have been traded away for fast recovery time.
cool. I think I'll model a distortion algorithm after that!
--
. . . . . . . . ... . . . . . .
"La la la laaa laaa laaa " | Niente Totalera V.
La la la laaa laaa laaa." -Stereolab | shi...@gweep.net
Art Du Rea: NoGargbageIn:NoGarbageOut
>Let me share some information after a lot of checking and consideration:
>
>2) Clamping with signal diodes:
>
>This works very well in most cases. Jerry Avins has pointed out that an
>op-amp can be bound to any limit with a diode and a devider at the output of
>the op-amp. A 1N4148 has leakage in the nA range. Capacitance is approx. 4
>pF. Forward current is approx 100nA @ 0.2V and 1mA @ 0.65V. The feedback
>divider needs a relatively low impedance, adds complexity and increases
>current consumption. It also introduces nonlinearities if the feedback
>impedance is too high.
For whatever it's worth, I've got some notes from Jim Williams (of National
and Linear Technology fame) about EXTREMELY low leakage clamp diode device
substitutions in the picoamp range. If anyone is interested, I'll look
them up.
All the best, and ENJOY!
Art Du Rea
Knoxville, Tennessee, USA
(Remove NOGI:NOGO for direct e-mail)
Gerhard Mesenich
Regards,
Gerhard
Art Du Rea: NoGargbageIn:NoGarbageOut <A-Du-R...@EsperNOGO.Com> schrieb
in im Newsbeitrag: 387d223c.2177833@news...
Jerry Avins
Jerry
--
Engineering is the art of making what you want from things you can get.
-----------------------------------------------------------------------
Art Du Rea: NoGargbageIn:NoGarbageOut
just now back on the air ...
Regarding low leakage clamp diodes ... here is some early '95 personal
correspondence with Jim Williams ... I'm presuming he is still at Linear
Technology ...
Dear Mr. Williams:
I need a commonly available very low leakage diode for several 5 V
ultra-low power circuit application. I have been useing the 1N456 family
(25 nanoamps) but they are getting hard to find through distributors. I
remember somewhere in your distant past writings that a standard (?)
transistor in a diode connection (collector shorted to base?) makes an
extremely good low leakage diode, but that begs the question "how do I
quantify the leakage for a design review"? Do I just search for a good
I(ebo) spec, such as the MPS8097 (20 nanoamp @ 6 V - $0.22)? What is the
purpose or effect of tying the collector to the base, and how do I quantify
that effect? What commonly available device would you use for this
application?
Jim's notes in the margins an on the back ...
(If collector floats in a high noise environment there is gain from it)
"Yes, 1N456 is (was!) good because is it NOT gold doped."
(From Linear Technology Ap Note 43, pg. 47, Fig E2):
Clamp type Forward Drop Leakage @ 25C (15 V reverse)
(Diode) 1N4148 ~ 0.6 V ~10e-9 A
(Diode) HP5082-2810 ~ 0.4 V ~ 10e-7 A
(NPN) 2N2222 ~0.6 V ~ 10e-11 A
15 v use base-collector
at 5 V use base-emit for lowest leakage
(JFET) 2N4393 ~ 0.6 V ~ 10e-11 A
drain-source shorted
"You will NEVER find a JFET that won't do the quoted #'s, & that's why
every DVM mfgr. uses these things for low leakage clamp diodes. In a real
hard-nosed design review "everybody does it" ain't good enough, of course,
& Siliconix sells (or at least they used to) "low leakage diodes" (PAD2, 5,
10, etc.) fully specified & guaranteed. Of course, all they are is
2N4393's connected as diodes, & sold at enormous profit. Take your choice!
I have enclosed the Siliconix data sheet."
I was able to make a final buy of a couple of hundred 1N456A diodes and
that was enough to see me through to the end of the project, but I have
always enjoyed Jim's "down to Earth" ingenuity and creativity (not to
mention his prolific writing) in making simple practical circuits do
amazing things.
I know this isn't much, but brings back great memories.
mikea...@gmail.com
> I need some limiting OP-Amps for driving a fast single supply 12bit
> AD-Converter. The sampling rate is approx. 1MSPS. The input range is 0-4 V.
>
> The OP-Amps will have a dual 5V supply. The gain will be 1-2. The output of
> the OP-Amps has somehow to be restricted to 0-5V, so that they do not
> overdrive the AD-Converter.
>
> A good choice as a PGA would for instance be the TI THS 7001/7002, which has
> such an output clamp feature.
>
> There should also be some simple OP-Amps on the market which have such
> internal clamps.
>
> Who makes these limiting OP-Amps?
> Which types are available?
> (GBWP approx. 5Mhz or better)
>
> Thanks in advance.
> Gerhard