Current vs. Voltage Behavior of Digital Optical Module Main Boards

Background

The communications cable pair delivers DC power to the DOM main board. A filter circuit on the DOM MB diverts the power to the DC-to-DC converter, where it is converted to a lower voltage/higher-current for the needs of the digital and analog circuitry. A diode across the input, cathode to positive input lead, protects the DOM MB from reverse voltage.

When the input voltage to the DC-DC converter reaches approximately 19V (observed at room temperature, on only one board), the start-up circuitry begins to "bump"as the DC-DC converter's switching power converter tries to start. As the voltage increases, the rate of pulsation increases. The switching converter succeeds in turning on, delivering specified voltages to loads, within tolerance, when the input voltage reaches approximately 55V. (actually, 40 to 55V, depending on the loop-resistance, and input filter capacitance).

This study captures the behavior of a pair of DOMs below the DC-DC converter threshold, and in the range of [-1..0..1]V, where semiconductors begin to conduct. The study was conducted entirely in a lab setting, without side-effects that might be produced by quad transmission line cable. The affects of the cable plant are assumed to be negligeable.

It is reasonable to expect that a normal pair of DOMs, in the Antarctic ice, should behave in a manner very similar to DOM MBs on the bench. Departures from the bench measurements may result from temperature, or the behavior of the wiring plant at the south pole installation (the IceCube experiment). In fact, the alteration of the bench circuit such that it produces curves similar to the ones measured in situ at the south pole will help diagnose cable plant problems. The spice model below hints at possible perturbations

Data collection in the lab

The simplified schematic:

i-v curve excel spreadsheet

The data in the spread-sheet were collected from a pair of DOM main boards in parallel, powered by a precision voltage source (Fluke model 332D) in series with two (nominal) 100 ohm resistors. The resistor on the right takes the place of typical loop resistance for a pair serving DOMs in the ice. The resistor on the left acts as a current limiter, and current measuring shunt. A Fluke model 8842A DMM (1 microvolt least count, 5-1/2 digit) measured the voltage drop across that left-hand resistors.

In the "normal" direction, the turn-on characteristics of circuits in the DC-DC converter contribute the kinks in the i-v curve.

In the "reverse" direction, the slope of the i-v curve above 1V is determined by the resistance in series with the protective diode on the DOM MB. Most of the resistance is external to the board; Filter components on the DOM MB contribute the rest.

Case 1: A typical pair without a "problem" <bench simulation>

The typical case demonstration faithfully follows the schematic diagram above.

Case 2: An atypical pair with a "problem" <bench simulation>

In this case, the typical case bench simulation set-up is augmented with a 1200 ohm resistor in parallel with the cable pair at the DOM main board. Three objects now consume the power delivered through the simulated cable resistance and the measurement shunt. The presence of this additional, current consuming, component significantly alters the i-v measurements.

Note that the current drawn by the resistor adds a linear function to the curves from case1 above. The power drawn by the shunt overpowers the fine structure in the graph contributed by the circuits in the DC-DC converter.

Simplified SPICE model without and with perturbation

This simulation only models the "reverse" polarity I-V curve for a pair of DOMs.
To illustrate the affect of a 1K ohm shunt on the i-v curve for the reverse power connection (i.e. the connection that forward biases the protective diode on the DOM MB), the spice simulation below was constructed. The simulation generally matches the 0-1V graph from case 2 as measured in the lab.

Testing high-current pairs at the south pole in-situ

Equipment:

I-V Data Measurement Procedure:

These steps collect accurate current vs. voltage measurements of the "normal" powered, and "reverse" powered DOMs connected to a transmission line pair in the ice. It is helpful, but not strictly necessary, to first identify the positive and negative connector pins for the pair being tested. The current consumption is high for "reverse" power, limiting data taking to 10V.

  1. Measure the 100 shunt resistor value to 1/10 % or better. Record for use in curent calculations.Ω
  2. Series connect Voltage source, 100 Ω resistor, and cable pair.
  3. Connect volt-meter lead to each side of the 100 Ω resistor so as to measure the voltage across it.
  4. Set power supply to 0.000V out. Turn on power supply. Measure and record voltage across current shunt. This offset voltage may be needed to correct the data during analysis.
  5. Set power supply to 0.100V. Record voltage. Measure and record corresponding voltage across current shunt.
  6. Repeat step 5 for 0.200, 0.300, 0.400, 0.450, 0.500, 0.550, 0.600 0.650, 0.700, 0.800, 1.000, 2.000, 3.000, 4.000, 5.000, 6.000, 7.000 8.000, 9.000, 10.000 Volts. If the current is less than 25mA, proceed to step 7. Otherwise proceed to step 8.
  7. Repeat step 5 for 12.000, 14.000, 16.000, 18.000, and 19.000V.
  8. If you have executed step 6 once with each polarity, proceed to step 9. Otherwise, reverse power supply polarity (swap leads) and repeat step 4 through step 8.
  9. Analyze the data as done in the spreadsheet example. (Page 2 of the spreadsheet may be modified to suit your data), or send us the data.... Plot voltage setting vs. voltage reading divided by resistor value.

A measurement of the i-v characteristics for a known-good pair will provide a useful comparison with the anomalous pair. This additional measurement is STRONGLY recommended.

If the results are more similar to case 2 than case 1, then the value of the reciprocal of the i-v line for the "reverse" power case gives the value of the shunt resistor.

Switcher start-up threshold

The switcher controller in the 40-to-120V DC-DC converter requires at least 19V before it begins to pulsate. At some voltage near 40V, the oscillates at a sufficiently high frequency to produce the power needed by the DOM main board. We use the switching on-set voltage as remote voltage measuring device, since it "sees" the voltage present at the end of the communications pair. If voltage division reduces the voltage at the far end of the pair, then the power supply voltage must be raised to compensate for the voltage division. The degree of compensation hints at the location of anomalous resistance in the circuit. i.e. the anomalous resistance could be near the DOMs, or near the patch panel, or anywhere in between...

For the "normal" power case, increase the power supply output voltage above 19V until current pulses are observed. Reduce voltage somewhat, and approach the threshold voltage slowly from below to discover the value to within 10mV. Record value. It is best to measure the voltage at the connector pins. A hand-held DMM will do.

Repeat for a known-good pair, for comparison. Choose a pair near to, or at the same temperature as the anomalous pair.

Notes:

For an anomalous pair, if the i-v curve below 1V is like case 1 above, the anomalous power is being consumed by one or the other, or both DOMs.

For an anomalous pair, if the i-v curve below 1V is like case 2 above, then leakage resistance exists "in front" of the DC-DC converters on the DOM main boards.

If the behavior is like case 2, the switcher threshold voltage may be displaced above ~19V as a result of voltage division of the cable loop-resistance, and anomalous resistance, if that resistance is near the DOMs. If the anomalous resistance is near the measurement instruments