current amplifier or current to voltage op-amp

Google: charge amplifier

--

John Larkin         Highland Technology, Inc 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

tector that mechanically vibrates.

it seems I will need a "trans impedance amplifier" to convert the current signal to voltage signal, in order to provide a virtual ground and avoid lo ading the detector, if my understanding is right.

and use a feedback resistor, then connect to the voltage preamp.

y?

That probably depends on your quartz crystal detector.

Its impedance is going to include an appreciable capacitance to ground - yo ur op amp circuit will have to be designed to cope with that capacitance. I f you use just a feed-back resistor to the virtual ground, the parallel cap acitance of the resistor (typically 0.3pF but some surface mount parts have less) may not be enough to keep the circuit stable.

The Burr-Brown data sheet for the rather faster OPA656

includes a useful discussion of it's use as a transimpedance amplifier.

--
Bill Sloman, Sydney

"Bill Sloman"

That probably depends on your quartz crystal detector.

** The only references I can find to "quartz crystal detector" all involve frequency measurement of a crystal oscillator.

So WTF is the dopey OP on about ??

Surely it must be a " piezo electric " sensor.

Be an idea to clear up the ambiguities in his post first - eh ?

Though I know that goes against usenet's " lets all play a guessing game " culture.

.... Phil

Thanks. I think the circuit should be something similar to the front end circuit de sign such as the one by Philip Hobbs

but I am just intimidated to build the circuit myself. If the commercial one such as SRS 570 will do the job, I tent to buy it. In my breadboard test with a ordinary op-amp tl082cp, the noise level is ve ry high.

:

detector that mechanically vibrates.

0, it seems I will need a "trans impedance amplifier" to convert the curren t signal to voltage signal, in order to provide a virtual ground and avoid loading the detector, if my understanding is right.

ut and use a feedback resistor, then connect to the voltage preamp.

ity?

your op amp circuit will have to be designed to cope with that capacitance. If you use just a feed-back resistor to the virtual ground, the parallel c apacitance of the resistor (typically 0.3pF but some surface mount parts ha ve less) may not be enough to keep the circuit stable.

:

detector that mechanically vibrates.

0, it seems I will need a "trans impedance amplifier" to convert the curren t signal to voltage signal, in order to provide a virtual ground and avoid loading the detector, if my understanding is right.

ut and use a feedback resistor, then connect to the voltage preamp.

ity?

your op amp circuit will have to be designed to cope with that capacitance. If you use just a feed-back resistor to the virtual ground, the parallel c apacitance of the resistor (typically 0.3pF but some surface mount parts ha ve less) may not be enough to keep the circuit stable.

Maybe that's not as hard as you think it is. The tuning fork has a very high Q, which will give you massive voltage (and impedance) amplification. I suspect you need a FET-input buffer amp and possibly not very much voltage gain.

** Hang on a mo.

IF the OP is using a clock crystal ( 32.768 Hz ) then resonance is at impedance minimum.

So his notion that current is key is right.

Op-amps with very good noise figures at impedances between a few hundred and a few thousand ohms are easy to get.

Doubt he needs to go for fancy lab pre-amp units.

... Phil

Phil said

e "

I did ask, perhaps more diplomatically ad covertly than you would have done .

And he has given us some more information - though rather less than enough.

ick up sound wave at its resonance frequency.

to do the I-V conversion and then a voltage preamp can be used.

It would be nice if it did, but it seems unlikely.

If your piezo-electric transducer has a mechanical resonance at or near the acoustic frequency you want to pick up, you are going to have to character ise the transducer - drive it electrically at frequencies around the freque ncy of interest form around a factor of ten higher to a factor of ten lower , and map the complex impedance of the transducer - phase shift and amplitu de.

If you do it right - and the transducer hasn't got complex mix of resonant modes in the frequency range of interest - you should be able to set up ind uctance/capacitance/resistance model of the device (mode coupling can - in theory - be handled by using coupled inductors, but in practice, if you nee d to to do that you are in trouble).

The piezo-electric quartz crystals sold (in vast numbers) as frequency refe rences are regularly modelled in this way.

Using a transimpedance amplifier to get a signal out of such a transducer m ight be tricky. The impedance of the transducer becomes part of the feedbac k network around the amplifier, which can make it difficult to keep the amp lifier stable.

A voltage-follower with gain doesn't present the same kinds of problem. The TL082 is relatively noisy - 18nV/per root Hz - where the OPA655 I mentione d is quieter - 7nV/per root Hz - if faster and a lot more expensive.

The OPA657 is a bit quieter again. You can only use it at voltage gains of seven or higher, but it is a lot faster as well as quieter, if just as expe nsive.

--
Bill Sloman, Sydney

design such as the one by Philip Hobbs

very high.

Making you own transimpedance amp is a lot of fun... if you like that sort of thing. But if you've got more money than time, and the srs preamp works then go that way. If you do want to make your own, then you need to know things like the capa citance of you piezo detector, the bandwidth of the signal, and how much cu rrent you are detecting. The TL082 looks to have about 15nV/rtHz of voltag e noise... (strangely enoguh it's often the voltage noise of the opamp that causes the most noise in a TIA.) I like the OPA134 for (audio freq.) TIA' s.

If you wanted you could post a schematic of your TIA and we could comment. (you may just need a bit of capacitance across the feedback R to roll off t he HF gain.)

George H.

te:

Off on a tangent, if I wanted a detector to tell me that the sodium something-or-other street lamp across the street is on, what would be the best detector to use?

(Keeping in mind that it's to the west and the sun tends to be quite _blazing_ around here :) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142   Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

it design such as the one by Philip Hobbs

t.

is very high.

rt of thing. But if you've got more money than time, and the srs preamp wo rks then go that way.

apacitance of you piezo detector, the bandwidth of the signal, and how much current you are detecting. The TL082 looks to have about 15nV/rtHz of vol tage noise... (strangely enoguh it's often the voltage noise of the opamp t hat causes the most noise in a TIA.) I like the OPA134 for (audio freq.) T IA's.

t.

f the HF gain.)

Hmm well you could maybe put an interference filter in front of a photodiod e to just see the sodium lines... (well and the sunlight that falls in the same bandwidth.) Interference filters are spendy though... ~$50-100 Hey maybe the street lights are modulated at 120 Hz? (or some other freque ncy?) You'd still want to somehow try and filter out as much sunlight as p ossible.

George H.

I was referring to the voltage preamp (560), should have made that clear.

If you connect a substantially capacitive transducer to the input of a TIA, you'll get a huge noise peak, and maybe oscillation.

What kind of transducer is it? What's the capacitance? How much bandwidth do you need?

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation

Thanks all for insightful discussion, with my insufficient information.

The tuning fork I am using is standard 12.5 pF and 35 kOhm 32768 Hz quartz crystal tuning fork. The bandwidth I need is at its fundamental frequency 32 kHz.

I find these prgrammable-gain TIA:

The SRS current amplifier 570 spec says the noise at 32 KHz is around 10E-12 A/Hz^-2 with 10E-6 A/V. This is good enough, I can use another SRS560 to amplify the voltage.

Now I really worried is as Spehro Pefhany says, the SRS570 cannot do the job. The noise level after the TIA should be lower than 0.1 uV/Hz^-2.

Also, I noticed that the link to Phil Hobbs's circuit is down. It's chapter 18 of his get-all-done book.

Hi sorry for cross posting from sci.optics. I just summarize what I want to do:

I am trying to use a quartz crystal tuning fork at 32768 Hz, 12.5 pF, 35 kOhm as a sensitive sound detector at its resonance frequency from its piezoelectric signal.

So I need a transimpedance amplifier to convert its current to voltage when it is mechanically driven by the sound.

I read article by Phil Hobbs' get-all-done about the front end design.

Just wondering in my case if the commercial SRS570 current amplifier is good enough? From literature, the noise level I need is under 0.1 uV/Hz^-2 after the TIA. The SRS570 spec says at 32 KHz noise is around 10E-12 A/Hz^-2 with 10E-6 A/V.

If the SRS570 cannot do the job and I have to build the circuit, what TIA I should choose? Like the programmable-gain TIA from Analog-Device?

and last question, would a 6 pF, 35 kOHm tuning fork has smaller noise?

Putting a higher frequency crystal on a TIA by itself is a pretty good way to make an oscillator. Your a tuning fork has a pretty large equivalent resistance, so it might work fine. You should be able to see the Johnson noise of the crystal resistance.

The Johnson noise is going to be about sqrt(35k/60.4 ohms) = 24 nV/sqrt(Hz), so you don't need a terribly low noise amplifier if you have a second filter in cascade.

32 kHz is out of the 1/f noise of most op amps, fortunately. You might try something like an OPA140 or even a TL084, with about a 10 M feedback resistor. You'll want to roll the feedback off at about 60 kHz or so, which requires about 0.4 pF across the 10M resistor. With a through-hole resistor, that's probably pretty close to the board strays, but if you need a bit more, use a piece of solid wire laid next to the resistor, soldered to the output end, and wiggle it around till you get a result you like.

Oh, and use split supplies (+-15V) if you can--it'll make your life easier.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

ISTM that a 60Hz synchronous detector might do.

to do:

kOhm as a sensitive sound detector at its resonance frequency from its piez oelectric signal.

en it is mechanically driven by the sound.

ood enough?

IA.

A/V.

Hmm 0.1 uV/(sqrt Hz)? I'm not quite sure what the TIA gain is, but if it's anything over 1 meg ohm then that blows your noise budget right there. (Unless you're a TIA guru like Phil H who makes TIA's w/o room temperature resistors as feedback.)

Hey speaking of no R's in the feedback path. I've been reading Mark Johnso n's "Photodetection and Measurement" book. And he shows a circuit I've nev er seen before. It uses optical feedback. The output of the TIA opamp dri ves a resistor/ LED, the light of which goes to a second photodiode that is in series with the "signal" photodiode.

George H.

I should choose? Like the programmable-gain TIA from Analog-Device?

The electrical equivalent circuit of the watch crystal is going to include a significant inductance

For actual values, here's an example Spice subcircuit model I googled.

.SUBCKT XT32768K 1 2 CP 1 2 1.5000P RS 1 3 30.000K L 3 4 10.200K C 4 2 2.3129F .ENDS

Plug that into a Spice circuit for a transimpedance amplifier and see what happens.

I've not done it, but I suspect that the results won't be all that pleasing.

--
Bill Sloman