3
\$\begingroup\$

my problem statement:

i wanted to digitalize 10ns pulses from a photodiode which are in the range of current 100uA to 1mA (these were earlier much larger in range, dealt here, soon I have understood the gravity of the problem statement and changed them with suggestions by other members)

my circuit approach:

this may not achieve the full performance but still can satisfy requirements to a level

schematic

simulate this circuit – Schematic created using CircuitLab

Results:

Input:

TIA requires a current input pulse, so I have created a current source using a voltage generating pulse generator with series 1K resistor, so to generate a current input of 100uA, I have given an input of 100mV from generator

sorry I don't have a generator with sharp rise/fall times, I was feeding 12ns pulse with rise and fall time of 6ns

enter image description here

stage 1 opamp output (LTC6269) and corresponding LVDS output are shown, which have satisfied me initially, but below response is one i see frequently, some kind of repeating reflections or noise are seen close to the pulse

view 1 :

enter image description here

view 2 :

enter image description here

view 3 :

enter image description here

I have initially thought these may be noise, but as they are repeating I have not understood what exactly are they.

soon I have understood that these repetitive noises are present in the function generator output at low levels, but I don't know what caused this, did my 1K series resistor to TIA has caused these?

so I suspected my setup, now I tried to place the actual diode in place of the current source, which have shown no results at all, i have seen noise even with out any light source illuminated, which is undesired,so i removed the diode, when i power the circuit even with out input i get a output as below with a repetetion

enter image description here enter image description here

is it because of improper grounding ?? or any low frequnecy noise ??

please help me in finding the root cause of the problem

EDIT/UPDATE 1 :

the power supply is generated on board, using below setup, the 12V comes from a regulated power supply, LTC6269 would require +/- 2.5 dual supply , so the below is modified by tweaking resistors, LTC6754 requires only +5V and OPA699 would require +/- 5V dual supply.

coming to probe i am using a 500Mhz 10Mohm probe with capacitance 11pF and in scope i have set ac coupling with 1Mohm impedance

enter image description here

i am clue less to find the source of the this periodic noise, primarily i suspected function generator but now i feel its there even no source is present, can an opmap generate such kind of noise ???

EDIT /UPDATE 2 :

output of opamp and ground, both auto scaled show similar noise pattern(green is signal ground), may be due to non isolation of Analog signal ground and power supply ground ?

enter image description here

EDIT UPDATE3 : Results after addition of pi filters at dcdc outputs

with some suggestions of pi filters i have tried to create a CLC filter using components at my desk

L = 10uH and C being 4.7uF, 47uF, 0.1uF and 0.01uF(all 0603 SMD)

i did not get a 1nF but i was able to see noise suppressed to an extent, this set up is bare soldered and checked whether filter output is proper or not, i did not solder this on actual board, instead i took +/-5V from board and checked the filter output

With out CLC

enter image description here

enter image description here

After CLC

enter image description here

\$\endgroup\$
  • \$\begingroup\$ It looks like something is generating noise after the initial pulse, and since it is in synch with the original signal it must be related. When does that comparator switch relative to the initial signal? If it's not at that moment, what else is beyond that that does? \$\endgroup\$ – Trevor_G Apr 19 '17 at 16:41
  • \$\begingroup\$ What kind of power supply is being used to power the op-amps? And what kind of scope probes are you probing with? Using what kind of ground connection? \$\endgroup\$ – The Photon Apr 19 '17 at 16:44
  • \$\begingroup\$ Yup power supply would be my second guess, since rereading your question. \$\endgroup\$ – Trevor_G Apr 19 '17 at 16:46
  • \$\begingroup\$ @ThePhoton i have updated about power supply and probe that are being used \$\endgroup\$ – kakeh Apr 20 '17 at 2:55
  • \$\begingroup\$ You still haven't said how your probe is grounded. \$\endgroup\$ – The Photon Apr 20 '17 at 3:39
2
\$\begingroup\$

Such a healthy aggressive interferer: 200MHz ringing. On VDD.

enter image description here

The OpAmps and Comparators will have 0dB PSRR at 200MHz. The reactance of bypass capacitor: cap internal ESL + PCB vias + PCB traces ~~ 5nH, is j6.3 ohms.

You may need X2Y.com 4-terminal capacitors, ~~ 0.5nH installed.

Try this

schematic

simulate this circuit – Schematic created using CircuitLab

The PI networks were repeated at the Sensor (2 meters away). For dampening, use R = sqrt(L/C) 0.3 Ohms; thus some inductor/bead R is useful.

NOTE: The GND/RTN path is also broken. We want to implement "local batteries" at the low-noise analog electronics.

COMMENTS ARE VERY WELCOME. Is this overkill?

\$\endgroup\$
  • \$\begingroup\$ This answer would be more helpful to OP if it was more clear that C1 & C2 should probably be constructed from multiple parts to handle both the main switching behavior and the ring issue. \$\endgroup\$ – The Photon Apr 27 '17 at 16:13
  • \$\begingroup\$ @ The Photon Thus also inserting beads between the SwitchReg and the triple-pi-filter? (thanks for the suggestion) \$\endgroup\$ – analogsystemsrf Apr 28 '17 at 3:23
  • \$\begingroup\$ i tried this and posted results here electronics.stackexchange.com/questions/302052/… \$\endgroup\$ – kakeh Apr 28 '17 at 6:10
  • \$\begingroup\$ can i feed back the filter output to dcdc ?, more over i will be much happy , if know the math behind choosing the L and C values, considering C being combination of decoupling caps and L is a bead with resistance equivalent to 0.3ohm ?, kindly clarify in the answer \$\endgroup\$ – kakeh Apr 29 '17 at 4:45
  • \$\begingroup\$ No. Implement the SwitchReg independent of the PI filter. The PI filter needs frequency/phase behavior separate from the SwitchReg servo loop. \$\endgroup\$ – analogsystemsrf Apr 30 '17 at 13:19
1
\$\begingroup\$

Your power source is a switching regulator. Each time its internal switch turns off (or maybe on), these high frequency oscillations are generated and they are coupling by some means into your signal path (or possibly just into your scope input).

Eliminating them likely requires a wholesale re-think of your power management plan, and/or changes to your PCB layout.

Edit

output of opamp and ground, both auto scaled show similar noise pattern(green is signal ground),

The fact your scope measures (some of) this ring when probing the ground net (and presumably with the probe's ground pin also connected to the same net in your circuit) indicates that part of what you're measuring is signal coupled (typically magnetically) from your circuit to the probe, not necessarily signal that is present in the output of your op-amp.

You can minimize this effect by probing the circuit using the shortest possible ground connection between the circuit and the scope probe. An old answer of mine shows examples of some different probes, indicating which ones are bad and which ones are better.

Also pay attention to the orientation of the loop formed by the probe and its ground lead; you might be able to turn it a certain way to reduce this effect.

may be due to non isolation of Analog signal ground and power supply ground ?

You cannot isolate the analog circuit ground from the power supply that is supplying the analog circuit. There must be a complete circuit for the power supply to be able to deliver energy to the circuit being powered.

\$\endgroup\$
  • \$\begingroup\$ sorry for the wrong word usage, i meant to say filter between the grounds, not isolation, something like a bead or choke \$\endgroup\$ – kakeh Apr 20 '17 at 5:19
  • \$\begingroup\$ dont know whether my observation is right, i observed the RT of LT8471 is 187k, which results in a switching freq of 500Khz and my output noise also seems of same frequency, can i confirm form these ? \$\endgroup\$ – kakeh Apr 21 '17 at 3:25
  • \$\begingroup\$ It does confirm. But the noise you see is absolutely typical of switching regulator noise. I would have been shocked if you did not find the repetition rate of the noise equal to your regulator's switching frequency. \$\endgroup\$ – The Photon Apr 21 '17 at 4:02
  • 1
    \$\begingroup\$ You CAN break the ground paths, at high frequencies. Using inductors, or ferrite beads, in the GND/RTN. See my answer. Create "local batteries" at the lownoise circuitry. \$\endgroup\$ – analogsystemsrf Apr 27 '17 at 13:41
  • 1
    \$\begingroup\$ @analogsystemsrf, true enough, but be sure OP understands this is a different kind of "isolation" than we normally talk about with power supplies. I think this falls into the regime of "a wholesale re-think of your power management plan" as I said in my answer. \$\endgroup\$ – The Photon Apr 27 '17 at 16:10
0
\$\begingroup\$

Tempting to consider this? The opamps/comparator have no PSRR at 200MHz. This circuit inherently has lots of PSRR; both bases of the diffpair are referred to +5v. To respond to 2nanoSecond Trise pulses, you'll need faster transistors than 2N3904/2N3906.

schematic

simulate this circuit – Schematic created using CircuitLab

Here are some specs for the ONNN Semi NPN MMBT5179 package SOT-23, rated 900MHz---2,000MHz. enter image description here

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.