large bandwidth amplifier (help needed) / Electromagnetic Radiation / Forums

Forums

4hv.org :: Forums :: Electromagnetic Radiation

« Previous topic | Next topic »

large bandwidth amplifier (help needed)

1 2 3 next

Move Thread

LAN_403

hboy007

Thu Jan 14 2010, 10:16PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Hi there,

I hope this thread fits in "RF energy" better than in "general electronics".

I have read some application notes about photodiode amplifiers (which is what I in fact need) but still I find it difficult to start with this subject. Most of the circuits are either slow or use rare and expensive operational amplifiers. Cheap video opamps are readily available but have insanely low input impedances.
That's why I tried to throw a discrete photodiode amplifier for photoconductive mode together. I assume that photovoltaic mode is too slow because I have rather large diodes (the current noise will be a nightmare, too).

So here is what I have in mind:
I will build an AC amplifier. A laser diode will be pulsed at 1 MHz with 40-50% duty cycle. I bought a 14bit,10MSps ADCs that will be made to amplify and convert the signal. Afterwards, the data is transferred to my computer via ethernet. both on and off phases will be sampled and the measured value will be the difference of the values, corrected by the response function of the photodiode.

My Question would be:
What else, other than my first, basic approach can I do to amplify the photodiode signal? I have already added a red band-pass filter to the optical assembly. (Scattered) light is being collected by a large numerical aperture photographic lens (50mm f/1.4)). The photodiode area is 5 mm^2 and it can be biased with 10V in reverse ( I gathered that it is good practice to increase the voltage until avalanche effects take over, then reduce it slightly ).

Would it be a better idea to use a well-suited opamp? It can't be that difficult to design an
amplifier with 0.1-50 MHz bandwidth, can it? I selected a high transition frequency bipolar transistor with low parasitic capacity and the jFET has to handle a relatively large current to keep the stochastic noise down, that is contributed by the resistors.

The emitter ciruit performs rather well despite parasitic capacities, but I assume I'd be better off with a base circuit. However, I have to repeat the theory first.

Thanks in advance for your replies!

hboy

Steve Conner

Thu Jan 14 2010, 11:05PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

[quote]
Would it be a better idea to use a well-suited opamp?
[/quote]

Yes, you can get op-amps with 3GHz gain-bandwidth product for a few dollars now. You'll probably need a JFET input with low bias current though, and those tend to be slower. The super-fast op-amps (like the cheap video ones you mention) all seem to draw too much bias current.

At my last job we made photodiode amps that went up to 40MBit/s using some fast op-amp, I forget the part number.

wrote ...

It can't be that difficult to design an
amplifier with 0.1-50 MHz bandwidth, can it?

Oh yes it can!

Two commonly used discrete amplifier circuits are the "series-shunt feedback pair" and the "Ring of three".

Mattski

Thu Jan 14 2010, 11:54PM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

I'm not an expert, but I seem to recall a resistive feedback transimpedance amplifier being somewhat useful for being the first stage in an optical receiver. There is a book that covers all the necessary topics in quite a bit of detail.

hboy007

Fri Jan 15 2010, 11:43AM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Thank you, Steve! As I have stumbled upon a 433MHz receiver unit that uses a standard CMOS inverter array it might be a nice approach to abuse some logic ICs first

They have interesting transfer functions and are really cheap.
At least that is what came to my mind when you mentioned the "ring of three", I hope I got that right.

I'll have a look at the shunt series feedback pair ciruits, too.

ps. Mattski: nice book, but unfortunately I can't find it at the university library and I can't bring myself to buy it to find out whether I like it :/

Proud Mary

Fri Jan 15 2010, 12:11PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

LMP7721 has GBW 17 and can be bought as 8-pin SOIC on ebay for a few dollars.

I have dealt with this vendor and he seems honest and reliable.

Steve Conner

Fri Jan 15 2010, 12:35PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

A GBW of 17MHz is far too slow for a 50MHz transimpedance amp, though.

also see

page 11

Note that photodiode capacitance puts an ultimate limit on the bandwidth, no matter how fast the op-amp is. Large diodes simply won't go fast, end of story.

Proud Mary

Fri Jan 15 2010, 12:56PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I must have misunderstood 007's question then, Steve, as I assumed from his remark that "It can't be that difficult to design an
amplifier with 0.1-50 MHz bandwidth, can it?" that he was looking for a photodiode-friendly op amp which would function between 0.1 and 50MHz - which would include LMP7721's 17MHz offering.

As an afterthought, Hobbs's essay on photodiode front ends demonstrates novel solutions to the problem of photodiode capacitance:

Steve Conner

Fri Jan 15 2010, 01:22PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I assumed he meant an amp that would amplify all frequencies between 0.1 and 50MHz more or less equally.

It's like the 0-60 time of a car, I could get a great result if I were allowed to choose any speed in that range

Proud Mary

Fri Jan 15 2010, 01:25PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

I added as an afterthought, while Steve was preparing his answer a reference to an illminating essay by Hobbs and ways and means of dealing with photodiode capacitance issues:

hboy007

Sat Jan 16 2010, 05:13PM

Registered Member #1667 Joined: Sat Aug 30 2008, 09:57PM
Location:
Posts: 374

Hi there,

I found a nice fet video amplifier that seemed to be promising at first. It showed quite nice results until i added an output stage. So what happened? It couldn't drive a 5pF load... (BF245 has 4pF feedback capacitance and 1pF input capacitance)
So I added some offset voltage to the input (is it bad to run 1ÂµA trough the gate?) and a cap parallel to the source resistors.
The amplification is > 500x from 20 kHz up to 2 MHz, the -3dB point is then reached at ~4 MHz. That is an order of magnitude behind the desired 50 MHz :(

So... is there something wrong or am I just not getting anywhere with this topology?

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob