Searching \ for '[EE] Analog "Proportional" Amplifier Design Questi' in subject line. ()
Help us get a faster server
FAQ page: www.piclist.com/techref/io/atod.htm?key=analog
Search entire site for: 'Analog "Proportional" Amplifier Design Questi'.

Exact match. Not showing close matches.
'[EE] Analog "Proportional" Amplifier Design Questi'
2005\06\09@154502 by

Friends,

So I have this sticky analog design problem which I don't think should be
that complicated, but I just can't find a good reference for it.  I suspect
the circuit I'm looking for is related to a diff amp with some kind of level
shifting, and I know this list is about microcontrollers, but I thought
maybe someone would know the answer to this question just offhand.

It's especially annoying because I can't think of what the appropriate
name of this circuit would be.

Here goes:

The input to this mystery circuit has three slow-moving positive voltages
(essentially DC) which are all linear and floating.  Voltage A, although
floating, will always be the highest of the three.  Voltage B will always be
the lowest, and Voltage C will move around between the two over time.  I
know somewhere out there is a circuit which takes those three inputs and
gives an output voltage which is proportional to the floating input voltage,
but over a fixed range (like a rail-to-rail supply).
For example: Voltage A is 1.5V, Voltage B is 1.0V, and Voltage C is 1.25V.
If the rails on the circuit are 5V and ground, the output would be 2.5V.  It
would have exactly the same output (2.5V) if A, B, and C were, say, 1.2V,
0.8V, and 1.0V, respectively.

Another way to think about the inputs would be like the Vref inputs on an
A/D converter -- there are positive and negative reference voltages which
are used to set the scale for the input signal.

So my question:  What's this circuit called?  I've had the hardest time
looking it up in H&H or online or anywhere.  It's sort of like a scaling
amplifier and sort of like a level shifting amplifier, but not either one of
those two.  I'd like to find a reference before I go drawing up some sort of
frankenstein.

Yours, and thanks so much in advance,

Todd Bailey

On 6/9/05, Todd Bailey <tm_baileyhotmail.com> wrote:
{Quote hidden}

Mathematically, your desired output voltage is (B-A)*5/(C-A)

Subtracting A from B is easy.   Scaling is harder.  You need a
variable gain device. If you are just going into a ADC, you can do A-B
and 5*(C-A) in op-amps, then feed the A-B into the ADC with 5*(C-A) as
the reference.   This makes the ADC do the scaling.

Regards,
Mark
markrages@gmail
--
You think that it is a secret, but it never has been one.

> So my question:  What's this circuit called?  I've had the hardest time
> looking it up in H&H or online or anywhere.  It's sort of like a scaling
> amplifier and sort of like a level shifting amplifier, but not either one of
> those two.  I'd like to find a reference before I go drawing up some sort of
> frankenstein.

To clarify:

Given Va, Vb, and Vc, with output Vout and single-rail supply Vdd:
Vout = [(Vb-Vc)/(Va-Vc)]*Vdd
where Va > Vb > Vc, right?

Mike H.

Mike,

The equation you suggested is absolutely right -- that's exactly the
function I want, although the circuit doesn't have to be single supply.

Further, to clear things up w/r/t an earlier suggested solution:
Vout isn't necessarily going into an ADC -- it might very well stay analog,
so the output has to be linear.  Seems like it'd be totally doable with a
few op amps -- is that true?

Thanks again,

Todd

-------------------------------------------------

Mike Hord wrote:

To clarify:

Given Va, Vb, and Vc, with output Vout and single-rail supply Vdd:
Vout = [(Vb-Vc)/(Va-Vc)]*Vdd
where Va > Vb > Vc, right?

Mike H.

2005\06\09@183643 by
On 6/9/05, Todd Bailey <tm_baileyhotmail.com> wrote:
> Mike,
>
>   The equation you suggested is absolutely right -- that's exactly the
> function I want, although the circuit doesn't have to be single supply.
>
>   Further, to clear things up w/r/t an earlier suggested solution:
> Vout isn't necessarily going into an ADC -- it might very well stay analog,
> so the output has to be linear.  Seems like it'd be totally doable with a
> few op amps -- is that true?
>
> Thanks again,
>

Your solution will need a variable-gain device.  "Normal"
op-amp-resistor circuits aren't enough.

Look at National's LMH6503 or similiar.

Regards,
Mark
markragesgmail.com
--
You think that it is a secret, but it never has been one.

Try Googling for:

schematic analog multiplier

I didn't look closely at the results, but they looked promising.

Bob Ammerman
RAm Systems

{Original Message removed}
Do-able probably.
You would need to specify what sort of error is permissable before I
said it was easy.
The problem is that the gain needs to be be able to change. This can
be done in a number of ways but often needs temperature compensation
etc. Changing the gain of an amplifier is normally done by either
adjusting the resistors in the feedback network or operating at fixed
gain and having a variable amount of attenuation - again changing the
value of fixed resistors.
To make the gain dynamic, you need an element you can adjust by
changing a voltage or current. It can be done with fets &/or bipolars
but device charecteristic variations mean you get a wide spread in
results & you may need to calibrate it & possibly temperature
compensate it at well.
A LED/LDR pair can be used, again with device variation concerns
You may be able to use an audio volume control or a multiplier chip.
Or convert the signals to logs, add and then antilog.

There are other methods as well, with a bit more info on the accuracy
required we may be able to come up with some more ideas.

Richard P

On 10/06/05, Todd Bailey <tm_baileyhotmail.com> wrote:
{Quote hidden}

Of course op-amps were designed specifically to do these sorts of
things, as analog computing elements.

Look in the National semi linear handbook from about '83 or so.

Log amps, two and four quadrant multipliers, etc.

At 04:08 PM 6/9/2005, Todd Bailey wrote:

>>Given Va, Vb, and Vc, with output Vout and single-rail supply Vdd:
>>Vout = [(Vb-Vc)/(Va-Vc)]*Vdd
>>where Va > Vb > Vc, right?
>
>  The equation you suggested is absolutely right -- that's exactly the
> function I want, although the circuit doesn't have to be single supply.
>
>  Further, to clear things up w/r/t an earlier suggested solution:
>Vout isn't necessarily going into an ADC -- it might very well stay
>analog, so the output has to be linear.  Seems like it'd be totally doable
>with a few op amps -- is that true?

What frequency response?  How much noise in the output can you tolerate?

If I was trying to do this without involving a micro, I'd look at something
like a ramp generator using PWM to drive a SPDT analog switch.  The input
voltages would set the ramp reference and PWM comparitor voltages; the SPDT
analog switch would alternately switch the output between the desired
output rail and gnd.  The output ripple can be quite low if you set the PWM
frequency high enough (KHz or tens of KHz).  You can put an optical barrier
at the PWM stage if you require voltage isolation.

First glance suggests somewhere near 4 op-amps (better yet:
comparitors)  and something like a 4053 triple SPDT analog switch.  That's
assuming that the input signals are always within the common-mode range of
the op-amps.

Va feeds the ramp upper limit comparitor, Vb feeds the ramp lower limit
comparitor (ramp output oscillates between Va & Vb), output of ramp
generator feeds one input of PWM comparitor, Vc feeds other input of PWM
comparitor.  PWM comparitor feeds control input of SPDT analog switch with
NO contact to Vdd, NC contact to Gnd, output comes from armature output via
low pass filter.  Add active filter if you need better ripple filtering and
/ or lower output impedance.

So: a single LM339 quad comparitor with a few passives, a single cd4053
(might as well parallel all 3 switch sections), plus whatever low pass
filter you want.

dwayne

--
Dwayne Reid   <dwaynerplanet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax

Celebrating 21 years of Engineering Innovation (1984 - 2005)
.-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-
`-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'
Do NOT send unsolicited commercial email to this email address.
This message neither grants consent to receive unsolicited
commercial email nor is intended to solicit commercial email.

At 03:28 PM 6/9/2005 -0500, you wrote:
{Quote hidden}

The AD632 or MPY634 will do this with reasonable accuracy with almost
no additional parts. They are NOT cheap.

Can you just use a [PIC], digitize the three voltages, spit out the
result into the PWM, filter and buffer (and enjoy)?

When B approaches A you will have the most difficult case for both the
analog and the digital methods of solving this problem. With the digital
circuit, the resolution will probably bother you first, with the analog it
might be the offsets and linearity.

BTW, using a micro with a 24-bit ADC and precision DAC would probably not
be much more expensive than the analog multiplier and would be really, really
accurate (at very low frequency).

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
speffinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff

The OP mentioned 'slowly changing signals'. This is a hint that the problem
can be addressed by using a time varying signal. See:

http://www.edn.com/archives/1996/101096/21di_01.htm

which does just that using a simple triangle wave generator, a comparator,
an analog switch, and a couple of discretes.

Really rather elegant.

Bob Ammerman|
RAm Systems

>The equation you suggested is absolutely right -- that's exactly the
>function I want, although the circuit doesn't have to be single supply.
>
>  Further, to clear things up w/r/t an earlier suggested solution:
>Vout isn't necessarily going into an ADC -- it might very well stay analog,
>so the output has to be linear.  Seems like it'd be totally doable with a
>few op amps -- is that true?

Use a dsPic ? get a free sample from Microchip to play with, use the ADC to
measure the voltage, and output on a PWM after a variable gain calculation?

Todd Bailey wrote:
{Quote hidden}

I guess this was a very long winded tedious way of saying you want a circuit
that does:

5V * (C - B)
OUT = ------------
A - B

It would really help if you learned to communicate technical concepts
clearly.  Terms like "slow-moving" are meaningless without numbers.  I can't
even imagine why you meant by the input voltages are "linear".  Linear with
respect to what?  Floating, but over what range?  At what impedence?

Anyway, this is not at all easy to do with analog electronics since the
divide in the equation can't be made to go away.  However, this is what
microcontrollers are good at.  Since the signals are "slow" and you didn't
specify accuracy, run the three signals into a PIC, do the math, and control
a PWM output with the result.

What will the final signal be used for?  It would be better to perform the
math in the PIC then use the digital value from there on.

*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com
Todd Bailey wrote:
>> Given Va, Vb, and Vc, with output Vout and single-rail supply Vdd:
>> Vout = [(Vb-Vc)/(Va-Vc)]*Vdd
>> where Va > Vb > Vc, right?
>
> The equation you suggested is absolutely right

Really?  I thought you defined B as the bottom voltage and C as the middle.

*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com
> >> Given Va, Vb, and Vc, with output Vout and single-rail supply Vdd:
> >> Vout = [(Vb-Vc)/(Va-Vc)]*Vdd
> >> where Va > Vb > Vc, right?
> >
> > The equation you suggested is absolutely right
>
> Really?  I thought you defined B as the bottom voltage and C as the middle.

Perhaps he did, but I "re-labeled" to make it a little more logical for me.

At any rate, I'd suggest using a digital solution, too, if you can make it work.
Not knowing right off what the actual ranges are makes it tough, but if they
are correct, using (with my nomenclature) Va as the Vref+, Vc as Vref-, and
DAC of some sort with the end result being a voltage scaled between the
supply rails of the DAC in exactly the manner you're looking for.  No muss,
no fuss.  In fact, even if Va and Vc AREN'T within the acceptable range for
the Vref+ and Vref- on the ADC, you can force them to be with pretty simple
op-amp circuits.

Since you mentioned that these voltages are "slow-changing" (to me that
means Hz or even sub-Hz range), this sort of method should work quite
well.

Mike H.

Todd Bailey wrote:
>>Given Va, Vb, and Vc, with output Vout and single-rail supply Vdd:
>>Vout = [(Vb-Vc)/(Va-Vc)]*Vdd
>>where Va > Vb > Vc, right?
>
>The equation you suggested is absolutely right.

Thanks for all the replies!

Based on the above equation, as Spehro and others suggested, I've been
drawing up a circuit using two diff amps to perform the two subtractions
followed by an analog multiplier (the AD632 -- I'd used the 633 before and
thought THAT was an expensive IC) to divide the results of those two
operations.  I expect the final version will need a lot of trimming.
My next question is:
Is there a way to multiply the result of the division operation by a
constant (Vdd or some other scaling voltage) without another analog
multiplier?

ie:

(Signal - LowReference) via diff amp--------------------|
|---
Division via AD632 ----?Multiplication? --- Vout
(HighReference - LowReference) via diff amp--------|

I suspect everyone who has suggested using an A/D and a PWM is totally
right, but I worry about having 8 or 10 bit precision as a limiting factor.
Someone suggested a uC with a 24-bit A/D.  Can you guys recommend one?

Thanks again for everything,

Todd

At 07:51 AM 6/10/2005 -0500, you wrote:
{Quote hidden}

The closer Va gets to Vc, the lower the effective resolution of the
output will be. If Va is within 0.1V of Vc with a 5V reference, you'll
only have a few bits of resolution in the output with a 10-bit ADC.

Of course things get completely intractable right at Va=Vc (divide by
zero), but they also get ugly for both analog and digital solutions
as Va approaches Vc. The analog is going to be noise and Vos drift
limited, the digital also has the quantization to contend with.

It may be possible to add a bit of complexity (without a multiplier) and
improve the digital solution- for example by digitizing the difference
with a programmable scaling.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
speffinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff

{Quote hidden}

Multiplication by a constant is simply gain, i.e. a classic inverting or non-inverting amplifier configuration.

Regards

Mike

=======================================================================
This e-mail is intended for the person it is addressed to only. The
information contained in it may be confidential and/or protected by
law. If you are not the intended recipient of this message, you must
not make any use of this information, or copy or show it to any
received this e-mail, and return the original to us. Any use,
forwarding, printing or copying of this message is strictly prohibited.
No part of this message can be considered a request for goods or
services.
=======================================================================

Mike--

You think a digital solution is the way to go, too?
Maybe I'm being hard-headed tyring to do this in an analog way, but it
wanted an analog output.

And just to specify, the voltage difference between the Va (Vref+) and Vc
(Vref-) will be pretty small sometimes (100mV or less worst-case) with a
range easily within what the inputs of an ADC could handle (easily scaled to
absolute values of <5v).  And you're correct, "slow moving" in this case
meant sub-audio, mostly less than a few Hz.  Call it 20Hz worst-case.

With that in mind, any more suggestions?

Thanks again,

Todd

From: Mike Hord <mike.hordgmail.com>

> >> Given Va, Vb, and Vc, with output Vout and single-rail supply Vdd:
> >> Vout = [(Vb-Vc)/(Va-Vc)]*Vdd
> >> where Va > Vb > Vc, right?
> >
> > The equation you suggested is absolutely right
>
> Really?  I thought you defined B as the bottom voltage and C as the
middle.

Perhaps he did, but I "re-labeled" to make it a little more logical for me.

At any rate, I'd suggest using a digital solution, too, if you can make it
work.
Not knowing right off what the actual ranges are makes it tough, but if they
are correct, using (with my nomenclature) Va as the Vref+, Vc as Vref-, and
DAC of some sort with the end result being a voltage scaled between the
supply rails of the DAC in exactly the manner you're looking for.  No muss,
no fuss.  In fact, even if Va and Vc AREN'T within the acceptable range for
the Vref+ and Vref- on the ADC, you can force them to be with pretty simple
op-amp circuits.

Since you mentioned that these voltages are "slow-changing" (to me that
means Hz or even sub-Hz range), this sort of method should work quite
well.

Mike H.

Todd Bailey wrote:
> I suspect everyone who has suggested using an A/D and a PWM is totally
> right, but I worry about having 8 or 10 bit precision as a limiting
> factor.

So what is your accuracy spec?  Analog multipliers, log/linear converters,
and the like are not that accurate either, especially at "low" prices.  1
part in 1000 accuracy is not going to be trivial.

So once again, how accurate does it need to be?  What frequency range?  What
common mode range?  What min/max differential range between A and B?

*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com
> I suspect everyone who has suggested using an A/D and a PWM is totally
> right, but I worry about having 8 or 10 bit precision as a limiting factor.
> Someone suggested a uC with a 24-bit A/D.  Can you guys recommend one?

I don't know if there ARE any uC's with built-in 24-bit ADCs.  There was a
good article a few years ago in Circuit Cellar about achieving 24-bit
accuracy in gas chromotography (?IIRC?).

Issue 140- March 2002.  "High Resolution Data Acquisiton Made Easy"
by Brian Millier.  I remember it being a pretty good article about achieving
24-bit accuracy, and it specified an ADC and gave good examples of using
it to measure (you guessed it!) slow-changing voltages with ease and
accuracy.  I don't remember the part number, though; I'll try to remember
to look it up later, if you don't get a deluge of "my favorite ADC is xxxx"
messages.

Mike H.

>-----Original Message-----
>From: piclist-bouncesmit.edu [piclist-bouncesmit.edu]
>Sent: 10 June 2005 14:49
>To: Microcontroller discussion list - Public.
>Subject: Re: [EE] Analog "Proportional" Amplifier Design Question
>
>
>> I suspect everyone who has suggested using an A/D and a PWM
>is totally
>> right, but I worry about having 8 or 10 bit precision as a limiting
>> factor. Someone suggested a uC with a 24-bit A/D.  Can you guys
>> recommend one?
>
>I don't know if there ARE any uC's with built-in 24-bit ADCs.

Silicon Labs C8051F350 and C8051F350 have 24bit sigma delta converters.  Blows most PICs into the weeds in terms of performance as well.

Mike

=======================================================================
This e-mail is intended for the person it is addressed to only. The
information contained in it may be confidential and/or protected by
law. If you are not the intended recipient of this message, you must
not make any use of this information, or copy or show it to any
received this e-mail, and return the original to us. Any use,
forwarding, printing or copying of this message is strictly prohibited.
No part of this message can be considered a request for goods or
services.
=======================================================================

>With that in mind, any more suggestions?

get a couple of sample dsPics from MChip and try to see how good the
solution might be? At least that would give you a feel for how good the

If you need a more accurate PWM, have a look at this design ideas from EDN.
http://a330.g.akamai.net/7/330/2540/20041102222656/http://www.edn.com/contents/images/93004di.pdf

It does it using two 8 bit PWM to make one 16 bit accurate out put, but the
same could be done for higher resolutions if you could get suitable
resistors from 10 bit PWM.

For a 24 ADC you will need an external one, such as those available from
Linear Technology.

> You think a digital solution is the way to go, too?
> Maybe I'm being hard-headed tyring to do this in an analog way, but it
> wanted an analog output.

Nothing says you can't have an analog output- you just need a DAC
stage, as well.  And perhaps you should assess that analog output
need as well- if the data is just going to be digitized later, can you
skip the middleman?

> And just to specify, the voltage difference between the Va (Vref+) and Vc
> (Vref-) will be pretty small sometimes (100mV or less worst-case) with a
> range easily within what the inputs of an ADC could handle (easily scaled to
> absolute values of <5v).  And you're correct, "slow moving" in this case
> meant sub-audio, mostly less than a few Hz.  Call it 20Hz worst-case.

As Spehro suggests elsewhere, can you use a PGA to ensure a larger
difference between the two?  Or a differential ADC?  He's right- a 10-bit
ADC with a 5V range will give you very few counts within a .1V range,
but a (say) 16-bit ADC would give you ~1300 counts in that range.

Mike H.

At 02:58 PM 6/10/2005 +0100, you wrote:

> >-----Original Message-----
> >From: piclist-bouncesmit.edu [piclist-bouncesmit.edu]
> >Sent: 10 June 2005 14:49
> >To: Microcontroller discussion list - Public.
> >Subject: Re: [EE] Analog "Proportional" Amplifier Design Question
> >
> >
> >> I suspect everyone who has suggested using an A/D and a PWM
> >is totally
> >> right, but I worry about having 8 or 10 bit precision as a limiting
> >> factor. Someone suggested a uC with a 24-bit A/D.  Can you guys
> >> recommend one?
> >
> >I don't know if there ARE any uC's with built-in 24-bit ADCs.
>
>Silicon Labs C8051F350 and C8051F350 have 24bit sigma delta
>converters.  Blows most PICs into the weeds in terms of performance as well.
>
>Mike

Also TI and AD. All those are 8051 core. For true overkill I think AD has
some ARM core processors with on-chip 24-bit ADC here or coming RSN.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
speffinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff

Mike Hord <mike.hordgmail.com> wrote:
> > I suspect everyone who has suggested using an A/D and a PWM is totally
> > right, but I worry about having 8 or 10 bit precision as a limiting factor.
> > Someone suggested a uC with a 24-bit A/D.  Can you guys recommend one?
>
> I don't know if there ARE any uC's with built-in 24-bit ADCs.  There was a
> good article a few years ago in Circuit Cellar about achieving 24-bit
> accuracy in gas chromotography (?IIRC?).

Silicon Labs has a new 8051-based part with a built-in 24-bit ADC. Tom
Cantrell talks about it in an upcoming Circuit Cellar (#181).

"The Silicon Labs C8051F350 microcontroller isn't only the '51
MIPS master (100 MIPS peak), but it also leads in the integration
of evermore complete and precise analog subsystems (see Figure 1).
With amazing 24-bit resolution and a full complement of signal
conditioning, the C8051F350 enables what Silicon Labs calls direct
sensor interface technology (DsiT). As the name implies, the idea
is to eliminate the need for the external analog front-end add-ons
(e.g., op-amp and filter) typically required by low-gain sensors
that are often too noisy.
"At the ESC, Cygnal demonstrated a great example of the concept
in action in the form of a digital compass (see Photo 1). The
C8051F350 is directly connected to the various low-level sensors
(e.g., magnetometer, accelerometer, and inclinometer), and all
those MIPS are put to work crunching the nontrivial navigation
algorithms."

See http://www2.silabs.com/tgwWebApp/public/web_content/products/Microcontrollers/en/DSiT.htm

The input mux supports up to 8 fully differential inputs.

I've used other SiLabs microcontrollers in the past.
It's amazing to see 8051 code running at almost 100 MIPS!

At about half the price of the analog multiplier/divider chip, and
giving at least two orders of magnitude greater accuracy, it seems
like a no-brainer.

-- Dave Tweed
Mike Hord wrote:
> Nothing says you can't have an analog output- you just need a DAC
> stage, as well.

Or PWM.  At such low frequencies, PWM will have quite decent resolution.  I
would seriously look into a 30F3013 doing this.  It has 12 bit A/D and can
do 16 bit PWM at this frequency, all in a 28 pin package.

*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com
How good are they (Silabs MCU) in terms of reliability?

We just start to use the C8051F3xx part for the fast
ADC speed. In terms of features, nothing can compare
with them in the 8-bit world. The price is generally
in the higher side though. The 11 pin parts are very
good for the small sensor application (3mmx3mm MLP11
package). PIC is just too big. With Atmel cheaper and
Silabs better, I think Microchip should make some
changes to the pricing, packaging and features.

Xiaofan

-----Original Message-----
From: Dave Tweed [picdtweed.com]
Sent: Friday, June 10, 2005 10:36 PM
To: piclistmit.edu
Subject: Re: [EE] Analog "Proportional" Amplifier Design Question

I've used other SiLabs microcontrollers in the past.
It's amazing to see 8051 code running at almost 100 MIPS!

At about half the price of the analog multiplier/divider chip, and
giving at least two orders of magnitude greater accuracy, it seems
like a no-brainer.

-- Dave Tweed

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...