1. Noise In Photodiode Applications
By Art Kay and Bryan Zhao ( 赵 伟 )
June 1, 2011

2. Contents Noise Review Photodiode Review Photodiode Noise Theory
Bandwidth and Stability
OPA827 Hand Calculation
OPA827 Tina Spice Analysis
OPA827 Measurement Example

3. Noise
Review

4. Intrinsic Noise Error Source Generated by circuit itself (not pickup)
Calculate, Simulate, and Measure

5. Time Domain – White noise
normal distribution
This slide shows the time domain waveform for broadband noise. The distribution for broadband noise is also shown; note that it is Gaussian.

6. What is Spectral Density? (Noise per unit frequency)

7. Convert Spectral Density to RMS Convert RMS to Peak-to-Peak

8. Why Photodiode Noise? Noise is a key parameter in photodiode design
Wide bandwidth (integrate more noise)
Low signal levels (noise more critical)
Photodiode amplifier noise is more complex
Parasitic capacitance and sensor capacitance
Poles and zeros
Gain peaking

9. What can we do with the photodiode knowledge?
The competition is in trouble!
Our new friends from National. TI

10. Photodiode
Review

11. Photodiode Basics Introduction Photodiode type
Photodiodes convert light into current or voltage.
Photodiode type
PN photodiode – more wavelength selective
PIN photodiode – wide spectral range (less selective)
APD (Avalanche photodiode) – sensitive to low light, fast

12. Basic Photodiode Physics
Principle of Operation:
. - The P-layer material at the active surface and the N material at the substrate form a PN junction which operates as a photoelectric converter.
- When light strikes a photodiode, the electron within the crystal structure becomes stimulated. If the light energy is greater than the band gap energy Eg, the electrons are pulled up into the conduction band, leaving holes in their place in the valence band.
- These electron-hole pairs occur throughout the P-layer, depletion layer and N-layer materials. In the depletion layer the electric field accelerates these electrons toward the N-layer and the holes toward the P-layer.
- This results in a positive charge in the P-layer and a negative charge in the N-layer. If an external circuit is connected between the P- and N-layers, electrons will flow away from the N-layer, and holes will flow away from the P-layer toward the opposite respective electrodes.

13. Photodiode Basics Figure1.3 Photodiode Equivalent Circuit
Figure1.4 Current VS. Voltage Characteristics

14. Photodiode Basic
Figure1.4 Photodiode Equivalent Circuit
Use left equivalent circuit, the output current is given as :
Figure1.5 Current VS. Voltage Characteristics
The open circuit voltage Voc is the output voltage when Io equals 0. Thus Voc becomes:

15. Photodiode and Control Source TINA model
Light exciting source:
1) Use VG1 and VG2 voltage sources to simulate light power wave.
2) Use R1 and C1 shape light signal
3) The Voltage Control Current Source (VCCS1) simulates photodiode sensitivity.
Photodiode Equivalent Circuit:
1), Current Source Id simulates Dark current
2), Diode is a ideal diode
3), Cd and Rd simulate photodiode's junction capacitor and dark Resistance.
4), Rs is series resistor, which is far smaller than Rd.

16. Photodiode
Noise Theory

17. Photo-Diode Amp Noise Model

18. Photodiode noise
The lower limits of light detection for photodiodes are determined by the noise characteristics of the device. The photodiode noise in is the sum of the thermal noise (or Johnson noise) ij of a resistor which approximates the shunt resistance and the shot noise isD and isL resulting from the dark current and the photocurrent.
If IL >> 0.026/Rsh or IL >> ID, the shot noise becomes predominant. The lower limit of light detection for a photodiode is usually expressed as the intensity of incident light required to generate a current equal to the noise current . Essentially this is the noise equivalent power (NEP).

19. Noise Gain Simplify the model to compute Noise Gain
Gain seen by the noise voltage source.

20. Noise Gain

21. Noise Gain

22. Noise Gain

23. Simulating Noise gain and noise bandwidth
Break the loop to measure Aol, 1/B, and I to V Gain

24. Voltage Noise eni , eno and Eno
Region 1
Region 2
Region 3
Region 4
Region 5
Eno4
Eno5
Eno3
Eno1
Eno2

25. Voltage Noise eni , eno and Eno
Region 1 noise:
Region 2 noise:
Region 3 noise:
Region 4 noise:
Region 5 noise:
Total voltage noise:

26. Voltage Noise eni , eno and Eno
Log scale
R.1
R.4
R.2
R.3
R.5
Eno^2
Linear scale
eno
Linear scale
Two left figures obviously present:
although 1/f noise density is very large in noise density, but noise gain and bandwidth make region (3,4,5) contribute to output noise (Eno) most.
Noise gain effect region is controlled by Ci,Cf,Rf,fc.
Compensation design and op amp selection is critical for noise characters.

27. Resistor Noise and Current Noise
Current noise and resistor noise are limited by the transimpedance (I-V gain) bandwidth
Poles Kn 1
1.57 2
1.22 3
1.16

28. Bandwidth and
Stability

29. Parasitic Capacitance Limits the Bandwidth
Max bandwidth with Min Cf
Low Cf may be unstable
Wide BW increases noise
As shown Cf=Cs (stray cap)
Higher feedback resistor value potentially make parasitic capacitance bypass dominate the photodiode amplifier’s response and set the circuit bandwidth.
Good construction parasitic limit Cs around 0.5pF.

30. Feedback Capacitance Required for Stability
Noise Gain is key to stability
Also called 1/Beta (in stability analysis)
ROC = Rate of Closure
ROC = (Aol slope) – (1/Beta slope)
Unstable when ROC > 20dB/decade

31. Feedback Capacitance Required for Stability
Applying a Step Input shows instability at output

32. Choosing a Minimum Cf for Stability

33. OPA827
Hand Calculation

34. Noise Model for Simple Transimpedance Amp

35. Example Photodiode: PDB-C158
Unfortunately Cj is not specified at Vr=0V.
We called the manufacturer for this info Cj=70pF for Vr=0V

36. Calculate Diode Current Noise

37. OPA827 Noise Hand Calculation (key numbers)

38. OPA827 Noise Hand Calculation

39. Poles and Zeros in Noise Gain Curve

40. 1/f (flicker) Noise Corner

41. Output Noise from OPA Noise Voltage

42. Thermal (Resistor) Noise at Output

43. Current Noise to Voltage Noise at Output

44. The Final Total Noise

45. Reducing Noise (Higher Cf = Lower BW & Noise)

46. Flux Capacitor Advantage Disadvantage ADS1118 Reduces noise to zero
Available on E-Bay
Disadvantage
Requires 1.21 gigawatts
Size
Quarter

47. OPA827 Tina Spice
Analysis

48. OPA827 – test the model
Low Noise : 4nV/√Hz at 1kHz
Low Offset Voltage: 150µV max
JFET Input: IB= 15pA typ
Wide Bandwidth: 22MHz
The noise performance is the same as datasheet.

49. Simulated Spectral Density and Total Noise
Output Noise Density
Calculated
(rms)
Simulated
116.7uV
109.1uV
Total Output Noise

50. OPA827
Measurement
Example

51. Validating Test Equipment Capability

52. Tektronix DPO 4034 Oscilloscope
STDEV: 48uV (same as RMS)
P-P: *STDEV=319uV
40s P-P: 320uV
Set DC couple, 20MHz bandwidth limit
Short input channel to measure noise floor

53. Agilent 4395A Spectrum Analyzer
Frequency Range: 10Hz~500MHz
Noise floor: 10nV/rtHz
Input Impedance: 50Ω
Reference voltage
Noise Floor: 10nV/rtHz

54. The Noise Floors are Not Good Enough

55. Solution: Use A Post Amp What amp do we Choose?

56. Use OPA847 as post amplifier
Wideband, Ultra-Low Noise, Voltage Feedback, Operational Amplifier
At the gain of 150, the bandwidth is 26MHz
0.85nV/Hz Input Voltage Noise
2.5pA/Hz Input Current Noise
±100uV Input Offset Voltage (Typical))
575uVrms post amplifier noise in 20MHz bandwidth.

57. Post Amplifier Relatively small error!
Gain=150
Vn827 = 109.4uV rms
Vnpost = 16.03mV rms
Vnpost/Gain = 16.03mV/150 = 106.8uV rms

58. Test the Noise Floor

59. Test The Noise Floor – Post Amp Noise Scope
Simulated
(rms)
Measured
575uV
518uV
STDEV: 518uV
P-P: *STDEV=3.4mV
40s P-P: mV

60. Hardware Connections PDB-C158-ND photodiode
70pF junction capacitance at Vr=0 V
100dB I-V gain
4pF compensation capacitor
±5V power supply.

61. Shield the Circuit if Possible
Power Supply
Vcc,Vss,GND
Input & Output BNC connectors
Shield

62. Divide by Gain for OPA827 Output Noise

63. Measured Total Output Noise (DPO 4034 Scope)
OPA847 Measured at Scope:
STDEV: mV
P-P: *STDEV=143mV
40s P-P: 170mV
At OPA827 (DUT) Output:
Divide by gain
Vn827 = 21.7mV/150 = 144.6uV
Calculated
(rms)
Simulated
Measured
116.7uV
109.1uV
144.6uV

64. Measured Spectral Density 4395A Spectrum Analyzer
Measurement 4395A
Linear Scale
Agilent 4395A Spectrum Analyzer test 1Hz~20MHz span, 3uV/div, REF=24uV.
The tested noise density curve shape is the same as simulation.
Tina Spice
Linear Scale
Tina Spice
Log Scale
Linear Scale

65. OPA827-Noise Density 4395A Spectrum Analyzer

66. Before Questions, I must ask….
Thanks
Bryan Zhao ( 赵 伟 )
Matt Hann
Collin Wells, Peter Semig, Curtis Mayberry
References
Jerald Graeme <Photodiode Amplifiers>
Art Kay < Op-Amp Noise Calculation and Measurement >
HAMAMATSU <Photodiode Technical Information>
Tim Green <Operational amp stability>
Before Questions, I must ask….

67. Do I move over to the dark side?
You will move to a new topic my young apprentice
More noise presentations have we must

68. Thank you
for your Time.
Contact Info:
(phone)
(cell)