Photodiode transimpedance amplifier for slowly changing signal

AC

s

nV,

a noise

equal at

ast.

the

bout

se a

nd 1

SNR

ing

ements

r the

. The

an

ctor of

cally,

=EF=BF=BDs a

of

for

use?

urrent

DC

in order to

but

or

the

. e p t

I've got some UDT sensors that are designed (well supposedly optimized) for zero bias. But I know what you mean about a bit of bias reducing the capacitance.

Hey, the book is targeted to sell op amps because the author designs op amps! Much of the book is dedicated to optimal stability.

There do exist photodiodes that work best at or near zero bias, but those aren't the fast ones. Oh, the other thing is that the series resistance is dominated by the undepleted part of the I region...if you crank the bias up right near breakdown, so that the wafer is really really fully depleted, most PIN photodiodes speed up amazingly.

I have no big beef with Graeme--he's no worse than most apps folks, but I've never talked with one who really had a clue about photodiode front ends. It's just that dispelling that disinformation has taken up many hours of free consulting time. I don't mind helping folks at all, but the apps folk and their song and dance make it appreciably harder.

(One time, I had to explain to the inventor of the electron-multiplying CCD why its gain mechanism was so cool--he had the math all wrong and didn't realize it was as good as it was!)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net

I have now ordered a copy of Mark Johnsons book and I also added a copy of your book Phil. Hopefully they will be of help. It?s kind of interesting that the more you learn about a subject the harder and more complex it gets. Initially I thought this task would be pretty simple and straightforward but now I have started to hesitate about that. This is what?s so great about engineering; you learn something new every day.

I?m not really following you about the square in detected electrical power. The sensitivity of photodiodes is measured in A/W which seems as a linear relationship between optical power and detected current. The same relationsship (but oposite) should be valid for the LED side. So an increase in LED current should give a linear increase in photocurrent. The detector voltage is linear to the photocurrent (in a TIA) and therefore the LED current and detector voltage should be linear. Or am I thinking wrong?

Point taken about the Tee. I?ll skip it :)

What about bootstrapping of the photodiode, could that be of any use? I?m thinking of something like in Linears datasheet for LTC6244.

(page 18)

use?

Regarding Lock-Ins I have made some reading and it sure looks interesting. I have some questions and thoughts around it that I gladly discuss but perhaps that?s better off in a new thread. I need to deliver a first ?test setup? at the end of this week and it will obviously not be lock-in based.

Regards

--------------------------------------- Posted through

The OP stated that he only needed five decades of accuracy. And he had to ask -- this is one of those realms where if you have to ask, you can't do it quickly. I'm not saying that electronics _couldn't_ be built that work down to more accuracy than you can get out of an honest

24-bit ADC -- just that if you're aiming at five significant digits, that particular ADC shouldn't be the salient barrier.

The hypothesized electrostatic dipole moment of the electron requires measurements that can discern a signal that's something like 10^28 times smaller than the standing electric field in the experiment.

_That's_ an impressive level of precision.

(And I really appreciated the comments made by one of the researchers. I can't remember the exact wording, but it was along the lines of "even if we get the results we expect, they won't mean anything until someone else gets the same results using entirely different equipment." It's good to witness that level of integrity!)

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

Your book is a bit pricey, 82$ at Amazon

Regards LM

I'd say that's actually on the "cheaper" side for an in-depth technical book these days.

There are some 3rd-party sellers asking ~$42. Some say they're offering the "US student edition," though -- what in the world is that?

Strangely, the 1st edition of Phil's book is available new for $125 or used for $115!

ook

g the

ed

Without actually ordering the "student" edition to see one, my guess is it is not intended for the US market. Often in foreign countries there are legit copies of engineering books in soft cover. You often seen foreign students with these books. There are also Chinese copies, well Taiwan usually that might not be legit. Also in theory, the soft cover isn't supposed to be sold in the US.

Depending where you live, you can find engineering books in used book stores. Very true in the bay area. Perhaps not so likely in East Podunkazona.

Hi Miso,

Hmm, that does sound most likely. I suppose they don't say "international" version since Amazon is probably on to their game, then...

Yeah... I think there was some discussion about just how enforceable that was a year or so ago. I.e., is the "only to be sold in the U.S." restriction just a contractual arrangement between the publisher and whomever they're selling to (e.g., Amazon) and hence some, e.g., regular Joe China dude who bought it at a local bookstore (not directly from the publisher) is perfectly within his rights to sell the Chinese edition back to someone in the U.S., or is there some dusty bit of copyright law that might be applicable?

I'm in southern Oregon, which is pretty close to East Pdunkazona. :-) Portland is ~4 1/2 hours away, and they certainly have good bookstores such as Powell's.

---Joel

Please note that this Thread is NOT about book prices and bookstores!!! Any suggestions and information about photodiode amplifiers for weak, near DC signals are however much welcome.

I repost the latest relevant post in this thread: If you have anything to say about the subject I'm all ears.

of

about

new

The

the

wrong?

I?m

interesting.

Thanks!

--------------------------------------- Posted through

Any thread titled "Photodiode transimpedance amplifier" is going to be about Phils' book :)

This is s.e.d. so you should count yourself lucky it's not all about global warming and healthcare by now!

Well you've had lots of suggestions, some from experts in the field.

But OK, I never saw anyone answer this:

Power is proportional to current squared. So a linear optical power translates to a quadratic electrical power in the detector circuit.

[...]

--

John Devereux

So with a suitable choice of operating conditions, we get squared power out for linear power in. Free energy!

[Duck!] ;-)

Jeroen Belleman

I admit I was not entirely happy with that explanation even while typing it :)

But I suppose it's no different from any other transducer. For a photodiode the energy is supplied by the bias circuit. if operated at zero bias, the response will saturate before "over unity" is achieved.

--

John Devereux

g the

nal"

was

on just

ing

it

n his

re

ch as

I was in Portland last year and spent many of my few free hours in Powell's. Consider yourself lucky to have such a nice used bookstore. (They seem to have disappeared from the rest of the country.)

George H.

!!

near

ctrical

near

. The

re the

rong?

Each photon absorbed creates a electron-hole pair. So energy goes to charge. The time rate of change then says power goes to current. (~0.6 amps/ Watt near the peak sensitivity wavlength) The square law... power to current, also means you can use your photodiode as a mixer. A useful trick for measuring diode laser linewidths.

George H.

You can easily get power gain from a photodiode, just the way you can from an amplifier. There's no violation of energy conservation--the extra power comes from the power supply as always.

Consider a 1 mW laser beam at 900 nm wavelength hitting a photodiode with a 10V bias supply and a 10k ohm load. The responsivity will be about 0.5 A/W or a bit more, so the photocurrent is 500 uA.

The power delivered to the load is 10k*(500uA)**2 = 2.5 mW. An equal amount is dissipated in the PD itself. (It's not unknown to water-cool photodiodes.)

The more usual circumstance is that you get a lot less dc power than optical.

Square-law detection involves a huge dynamic range increase, which is one of the main things that makes optical measurements so challenging.

Re the price of the book: it's over 800 pages long, so $85 is pretty cheap actually. The first edition was $145 for real, and was 10-15% shorter. Wiley is one of the lower cost outfits among first-tier academic publishers--the same thing from Elsevier or Springer would be $250. (I picked a publisher by the very scientific process of walking over to my bookshelf and seeing which house published more of my favourite technical books. Wiley won by a landslide, with Academic and McGraw-Hill also-rans.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net

Only in a resistor, not in an LED.

John

Just following up here...

I went ahead and purchased one of the ~$42 copies. It arrived on Friday, and it looks to me to be a completely "normal" copy -- no pages telling you it's only for sale in Asia, it's the same decent quality paper you'd expect, etc.

So apparently some sellers just have a line on cheap new copies... woo hoo!

(If I'm feeling flush, maybe I'll just sent a check for $40 directly to Phil or to a charity of his choosing... :-) )

The advertising is way more valuable to me than the royalties--social media, 15th Century style. Read it in good health!

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net

, d

hoo!

t

I was expecting you to say "My favorite charity is Phil Hobbs." ;-)

I'm seeing a lot of those foreign textbook in the used bookstores in say the last six months. Singapore and India mostly. I wonder if some foreign engineers are dumping their books and heading home.

Given the budget cuts across the nation, there is oh so much hand wringing about the cost of tuition. Pity our children! Funny how nobody does anything about the cost of books.