Photodiode case

[Jure Newsgroups]

It's got 16 parts - too many for an ASCII schematic.
I'll see about putting it up on our public ftp site.
BTW, the parts are all inexpensive.

Basically it's the bf862 with a ferrite in the gate,
a source resistor to Vee for running at 5mA, a drain
resistor plus bypass cap to gnd to reduce the drain
voltage to under 8V (see datasheet), a coupling cap
to an emitter-follower stage, which is biased at say
-10V, depending on the PD, an LC filter to quiet the
bias-voltage source, a pos/neg bias-voltage header, a
low-noise EF BJT, with resistor to run at 4mA, etc.,
with back-back collector pmbfJ309 JFETs to limit the
current in the event of a PD-sensor short. Finally
there's another ferrite at the output to damp parasitic
RF oscillations (small resistors, commonly used for this
purpose, aren't allowed because of their Johnson noise).
The bootstrap follower has an 80MHz bandwidth, probably
limited by the ferrites, and is capable of driving the
guard shield in a triple-coax cable, etc. Armed with
a pencil, one should be able to draw the schematic.

WH >> a coupling cap to an emitter-follower stage,

Do I read correctly, AC coupled ?

Jure Z.

 

[Joerg]

Yep, I think we've got you're number, 1 to 5 cents is
OK, but 18 cents is not. Seems a bit overly contrived.

Oh, I'd use the BF862 where cheaper stuff won't cut it. If it weren't
single-sourced. That has just bitten me and my clients too many times.

One of my recent projects, a 2MW sine-wave generator,
used $850 transistors (thankfully I got used ones on
eBay for a fraction of that). Sometimes one has to
bite the bullet and leave all the cheap stuff behind.

The laser diodes on one of my recent projects were in that ballpark as
well. Ok, maybe half but we needed a few dozen. But they are driven by a
three-cent transistor, works nicely 8-D

 

[John Larkin]

It's got 16 parts - too many for an ASCII schematic.
I'll see about putting it up on our public ftp site.

The free pic posting sites are great... supload, imageshack, many
others. No registration required.

http://img135.imageshack.us/my.php?image=twood3.jpg

John

 

[[email protected]]

What do you mean by
embedded? The LMH6624 loses steam quickly with slight capacitive loading
(page 7 in the data sheet).

By 'embedded amp' I mean I am running the current mode amp as the
output stage of the LMH6624. The OPA694 is set up in noninverting
mode, and has a gain of about 2 and takes the load off the output of
the LMH6624. The OPA694 is inserted between the output of the LMH6624
and the feedback loop for the transimpedance amp.

That's a bit puzzling since the LMH6624 is a pretty good performer here.

I'm looking for an input-referred noise in the 0.2pA/RtHz range (or
better) at the photodiode with a gain of about 1.7e6 from the diode to
my ADC.
I have an optical signal that varies from 5nW to 1.3uW

 

[[email protected]]

It's got 16 parts - too many for an ASCII schematic.
I'll see about putting it up on our public ftp site.
BTW, the parts are all inexpensive.

I'd like to see that, please let us know if/when you get the schematic
on the FTP site.
What is the input-referred noise?

And thanks for all the help, I have a lot to think about!

 

[Joerg]

By 'embedded amp' I mean I am running the current mode amp as the
output stage of the LMH6624. The OPA694 is set up in noninverting
mode, and has a gain of about 2 and takes the load off the output of
the LMH6624. The OPA694 is inserted between the output of the LMH6624
and the feedback loop for the transimpedance amp.

So in essence the LMH6624 plus OPA694 would be the TIA? May I ask why
you needed that buffer amp? If it's for amplitude reasons where the LMH
can't swing it fast enough maybe you could run the TIA at a lower gain
(sans OPA) and then follow that with a separate amplifier that isn't
part of the loop. That avoids potential instabilities.

I'm looking for an input-referred noise in the 0.2pA/RtHz range (or
better) at the photodiode with a gain of about 1.7e6 from the diode to
my ADC.
I have an optical signal that varies from 5nW to 1.3uW

Ok, but you don't have to provide all that gain in one single stage.
That makes the BW requirement tough to fulfill.

 

[Joerg]

The free pic posting sites are great... supload, imageshack, many
others. No registration required.

http://img135.imageshack.us/my.php?image=twood3.jpg

Or just use a.b.s.e.

Ok, some ISPs' news servers don't serve up binaries but in this day and
age there ought to be some web site that allows such access. While
showing some ads ... why is it that the ads I see are mostly from some
Italian dating service? It's all firewalled here, I am not Italian, my
IP address should indicate I don't understand that language well, and I
am married.

 

[WhiteDog]

So in essence the LMH6624 plus OPA694 would be the TIA? May I ask why
you needed that buffer amp? If it's for amplitude reasons where the LMH
can't swing it fast enough maybe you could run the TIA at a lower gain
(sans OPA) and then follow that with a separate amplifier that isn't
part of the loop. That avoids potential instabilities.

It was essentially to extend the bandwidth of the circuit.

Ok, but you don't have to provide all that gain in one single stage.
That makes the BW requirement tough to fulfill.

Yep. But for some reason that escapes me right now, the analysis
pointed towards putting as much gain in the first stage as possible to
reduce the overall circuit noise. I do have a voltage gain stage after
this, and a differential driver at the output to drive my ADC
transformer.

 

[Joerg]

It was essentially to extend the bandwidth of the circuit.

It wouldn't help much with the BW except when you must drive huge output
swings where the LMH runs out of slew rate.

Yep. But for some reason that escapes me right now, the analysis
pointed towards putting as much gain in the first stage as possible to
reduce the overall circuit noise. I do have a voltage gain stage after
this, and a differential driver at the output to drive my ADC
transformer.

That doesn't really work that way. The higher the gain, the smaller the
overall bandwidth will be. For example, during my last high-speed PD
design I kept things at 2.5mV/uA, IOW the amp ran at roughly 20dB.
That's really all you need to overcome the noise, there isn't much SNR
to be harvested with higher gains. I'd even have gone lower but this amp
wouldn't have been stable then.

 

[Winfield Hill]

WH >> a coupling cap to an emitter-follower stage,

Do I read correctly, AC coupled ?

Yes. DC to establish the PD's bias and provide a
return path for the low-frequency current. The ac
signal path provides the bootstrap (which is only
important at high frequencies where the opamp has
lost tight control of the summing junction, and
where Cdiode matters) and the return path for the
high-frequency signal current.

 

[Winfield Hill]

The free pic posting sites are great... supload, imageshack, many
others. No registration required.

http://img135.imageshack.us/my.php?image=twood3.jpg

John

How does that work for pdf files?

 

[LVMarc]

Hi,

I am working on a wideband photodiode receiver for a laser ranging
system. I have a bootstrapped photodiode plus transimpedance amplifier
that has a BW of about 75MHz. My signals of interest are in the
3MHz-6MHz region and the 35-50MHz regions. The whole thing (PD plus
amp) is inside of a stainless steel box that provides E-field
shielding (but not much H-field, I fear). I know stainless was not the
best choice for HF, but there were other factors involved. I am trying
to get that changed now.

My question is regarding what to do with the case of the photodiode
(an InGaAs PIN). The case of the photodiode is providing some strange
coupling effects. The PD case pokes out of the stainless enclosure
that contains the PD front end amplifiers, etc. The PD case is NOT DC-
coupled to anything inside itself (the TEC cooler, thermistor, or the
PD substrate itself).

Initially, I had the PD case mounted on the enclsoure itself. The
thinking here was that the PD case plus the enclosure would form a
Faraday shield. This was apparently a fairly bad idea, as the
enclosure apparently capacitively coupled to the PD, and it created a
feedback path from the amp output to the PD itself, and I had in-band
ripples in the response at the higher frequencies.

I tried driving the PD case with the bootstrap. No good.

I currently have the PD case attached to the 'ground' reference near
the transimpedance amp with a length of copper braid. This seems to
have the best results, but still not ideal

Has anyone else had experience with doing this? What is the preferred
way of isolating the PD substrate from the E/H field environment?

Does anyone else have a sore forehead from beating it against a
network analyzer ;-)?

Thanks for any help,

-WhiteDog

you may want to consider differential signal path to subtract thi s
common mode noise, pick-up, ad lave only nice clean signal. That and
manging the BW of the amplifier chain as well!

best Luck

Marc Popek

 

[John Larkin]

How does that work for pdf files?

I think supload and imageshack will host pdf's or zips. Supload claims
unlimited file size, Imageshack is 1.25 M for unregistered users...
something like that. I don't know the mechanics of downloading... try
it!

John

 

[Joerg]

How does that work for pdf files?

Can't you yank out the "under construction" sign and add a link to this
site?

http://rowland.org/sitemap/index.php

Or to this one?

http://rowland.org/labs/electronics/hill.php

Hey, you look a whole lot younger than on the AoE dust cover. How'd you
do that? Can't just be the beard.

 

[Rich Grise]

Can't you yank out the "under construction" sign and add a link to this
site?

http://rowland.org/sitemap/index.php

Or to this one?

http://rowland.org/labs/electronics/hill.php

Hey, you look a whole lot younger than on the AoE dust cover. How'd you
do that? Can't just be the beard.

Photoshop? ;-)

Cheers!
Rich

 

[Jure Newsgroups]

Yes. DC to establish the PD's bias and provide a
return path for the low-frequency current. The ac
signal path provides the bootstrap (which is only
important at high frequencies where the opamp has
lost tight control of the summing junction, and
where Cdiode matters) and the return path for the
high-frequency signal current.

Thanks Win,
then, the AC coupling is just for the bootstapped
detector bias voltage. The "video" path is good to DC.

My detector amplifiers are not bootstrapped.

The last design has a diff pair of JFETS cascoded
with a couple of BJTs. An OpAmp completes the forward path.

Thanks , Jure Z.

 

[Winfield]

http://rowland.org/labs/electronics/hill.php
Hey, you look a whole lot younger than on the AoE dust cover.
How'd you do that? Can't just be the beard.

OK, a hint: Shaving off a beard and shortening
one's hair indeed makes one look much younger.

 

[Winfield]

Photoshop? ;-)

The picture below was taken by somebody and given
to me, after they edited it to haze another person
in the picture. Anyway, there aren't many pictures
I can use when one is needed, so I grabbed it.
http://members.ebay.com/ws/eBayISAPI.dll?ViewUserPage&userid=win-hill_rowland-institute

That's an old photo, a newer one appears in the
Rowland Institute Junior Fellow program brochure.
http://www.rowland.org/rjf/images/rjf_2008.pdf

 

[Uwe Bonnes]

Oh, I'd use the BF862 where cheaper stuff won't cut it. If it weren't
single-sourced. That has just bitten me and my clients too many times.

I wonder when damagers at NXP will EOL the BF862 ?

 

[Uwe Bonnes]

If you ask TI, they'll give you a straight answer: They badly
miscalculated and didn't replace Burr-Brown's automated test
sets far enough in advance to keep up with the fact the old
ones were failing and could no longer be fixed. The new ones
are in place, but the programming is going slowly. They're
bringing back the products several at a time and they hope to
have everything back by next spring. Oh, they do apologize.

November 06 on Electronica 2006, people at TI were quite optimistic to be
back on par with the BB parts in spring 2006. And they apologized there
too...