Author Topic: TIA Ground Plane: Better to just remove it?  (Read 1136 times)

0 Members and 1 Guest are viewing this topic.

Offline LoveLaikaTopic starter

  • Frequent Contributor
  • **
  • Posts: 561
  • Country: us
TIA Ground Plane: Better to just remove it?
« on: May 11, 2021, 02:56:09 pm »
I'm working on a design for a transimpedance amplifier on a 2-layer board, with both layers being the ground plane. I may be overthinking this, but according to the article below, while a ground plane is best when not broken, one exception is for high-speed op-amps. They recommend removing the ground plane from the op-amp's input/output pads. My signal pulse has a frequency of 1-2 MHz (and from what I can tell, that's not considered high frequency, which is 3-30 MHz looking at radio frequency), but does that just refer to the op-amp itself, or is it better to surround the entire traces for input/output? Or, perhaps it is unnecessary?

My schematic and top copper plane with the cut-out are shown below. My no-fill zone to remove the copper plane encompasses the op-amp and its input/output connections.

Despite the conventional practice of removing the copper, the the surface area that is cut out seems to be quite big (for a 2 inch by 2 inch board).


https://www.analog.com/en/analog-dialogue/articles/high-speed-printed-circuit-board-layout.html
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19488
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: TIA Ground Plane: Better to just remove it?
« Reply #1 on: May 11, 2021, 03:14:10 pm »
A few points, not complete answers...

Your signal's fundamental frequency isn't that important. The edge rate is more important, as is the frequency/phase response of your circuit.

The importance, or otherwise, of parasitic capacitance and inductance can be understood in terms of the poles and zeroes they insert into your circuit's transfer function. That kind of ting has been well described elsewhere, and I couldn't do any better.

One starting point w.r.t. TIAs would be Phil Hobb's works. He pushes TIAs to their limits as part of his day job. See his postings on sci.electronics.design and has website https://hobbs-eo.com/Building_ElectroOptical_Systems

There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline awallin

  • Frequent Contributor
  • **
  • Posts: 694
Re: TIA Ground Plane: Better to just remove it?
« Reply #2 on: May 11, 2021, 03:52:46 pm »
For 1-2 MHz that design should work OK. The op-amp has maybe 3pF of input capacitance anyway and you may need a larger C1 cap for stability.
There's usually a reverse bias circuit close to the photodiode, to reduce the photodiode capacitance. Hanging the photodiode off a long wire (as maybe suggested by your PCB?) is not a good idea.

more ground vias where you do have 2 gnd-zones don't hurt.
BNC or SMA output from the board would be a suggestion, close to the output-impedance resistor.

If you want to go high speed the board-capacitance matters more. And also capacitive loading of the op-amp. I've used BUF602 as an output buffer after the op-amp (some designs there: https://github.com/aewallin/One-Inch-Photodetector)
 
The following users thanked this post: LoveLaika

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16611
  • Country: us
  • DavidH
Re: TIA Ground Plane: Better to just remove it?
« Reply #3 on: May 11, 2021, 04:59:04 pm »
They recommend removing the ground plane from the op-amp's input/output pads. My signal pulse has a frequency of 1-2 MHz (and from what I can tell, that's not considered high frequency, which is 3-30 MHz looking at radio frequency), but does that just refer to the op-amp itself, or is it better to surround the entire traces for input/output? Or, perhaps it is unnecessary?

They are referring to the bandwidth of the circuit and not the signal.  Parasitic capacitance at the inverting input combined with the impedance of the feedback network creates phase lag which lowers the phase margin.  A compensating capacitance can be added across the feedback resistor however this lowers the bandwidth so for maximum bandwidth, the ground plane needs to be cut away from the feedback network and inverting input.
 
The following users thanked this post: LoveLaika

Offline LoveLaikaTopic starter

  • Frequent Contributor
  • **
  • Posts: 561
  • Country: us
Re: TIA Ground Plane: Better to just remove it?
« Reply #4 on: May 11, 2021, 09:01:34 pm »
Oh, believe me, I got plenty of ground vias all over the board. Thanks for the suggestions. Got them all taken care of, though I'm definitely not hanging my photodiode off-board via wire. Otherwise, this whole thing's a bust.
 

Offline LoveLaikaTopic starter

  • Frequent Contributor
  • **
  • Posts: 561
  • Country: us
Re: TIA Ground Plane: Better to just remove it?
« Reply #5 on: May 11, 2021, 09:02:49 pm »
Thanks. I guess I'll have to do this for every design from now on if I can. This seems like a good practice.
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: TIA Ground Plane: Better to just remove it?
« Reply #6 on: May 11, 2021, 10:00:30 pm »
The main place needing the capacitance as low as possible is the -Ve input, which is all of the CAT1 track including the other side of the board under that track.
Are you sure that gain is correct, 50R to 17R seems very low.
.  That took much longer than I thought it would.
 

Offline LoveLaikaTopic starter

  • Frequent Contributor
  • **
  • Posts: 561
  • Country: us
Re: TIA Ground Plane: Better to just remove it?
« Reply #7 on: May 11, 2021, 10:10:01 pm »
That 50-ohm value is just a place holder. The actual values will vary.
I made sure to remove the copper along the CAT 1 track on both sides.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf