PSE MPS Process Analysis

Size: px

Start display at page:

Download "PSE MPS Process Analysis"

Gyles Blake
5 years ago
Views:

1 PSE MPS Process Analysis Under the IEEE 802.3af specification, power sourcing equipment (PSE) is obligated to detect the disconnection of a powered device (PD) through one or both of two methods involving a maintain power signature (MPS) protocol: DC MPS: Sensitivity to low current (low power) condition AC MPS: Sensitivity to high load impedance condition PSE s are given full latitude to select the method and a significant degree of range on the disconnect criteria. Some key specifications governing DC MPS method are provided in Table 1. DC MPS essentially involves a continuous sampling or monitoring of the load current furnished by the PSE port to the PD. While relatively simple to implement, the DC MPS method has the disadvantage that PD s must either continuously or at least periodically (3 times per sec) draw 10 ma, or.5 Watts of power or in order to stay powered. For this reason, some PSE s favor the AC MPS method. Parameter Description Criteria Imin1 The dc current level at which a PSE decides that the 0 to 5 ma PD has been disconnected. Hence, the invalid signature threshold. Imin2 The dc current level at which a PSE decides that the > 5 to 10 ma PD is still connected. Hence, the valid signature threshold. Tmpdo The required duration of an invalid signature before power is removed from the port. 300 to 400 msec Tmps The minimum time a valid signature must be < 60 msec present in order to declare it valid. Vopen_max The maximum port voltage during signaling following DC disconnect power-down. 30 Volts Table 1: DC MPS Specifications AC MPS requires that the effective PD input impedance be monitored after the application of power. This is accomplished by combining an AC signal with relatively high source impedance on the (nominal) 48V DC power rail, and then monitoring the AC MPS signal amplitude to measure PD AC input impedance. This method has the advantage that the PD may consume very low or zero DC power for extended periods of time and remain powered. The disadvantage is some added noise on the DC power rail and some added complexity in the PSE design. Some key specifications governing AC MPS processing are provided in Table 2 below. Parameter Description Criteria Zac1 The valid signature impedance (must be resistive or < 27 KΩ capacitive) where the PSE must maintain port power. Zac2 The invalid signature impedance (must be resistive or > 1.98 MΩ capacitive) where the PSE must remove port power. V_open The maximum peak-peak amplitude of the AC MPS signal immediately following PD disconnect and prior to powerdown by the PSE port. < 10% of Vport (e.g. < 4.8 Vpp for Vport = 48V) VportMax The maximum voltage on the PSE port during AC MPS < 60 Volts disconnect detection. V_close The maximum peak-peak amplitude of the AC MPS signal < 500 mvpp while the PD is connected to the PSE. (Note: This is equivalent to Vpp specification for frequency < 500 Hz. Fp The frequency of the AC MPS signal. < 500 Hz SR The maximum slew rate of the AC MPS signal. <.1 V/µsec R_rev The effective source impedance of the PSE port when > 45 KΩ performing AC MPS measurements. I_sac Maximum AC current flow while the PD is connected to the < 5 ma PSE port. Tmpdo The time duration of an invalid signature before the PSE 300 to 400 msec removes power from the port. Vopen_1 The peak voltage on the port following an AC disconnect induced power down event. (Note: This is equivalent to Voc.) 30 Volts Table 2: AC MPS Specifications June 5, 2008 Copyright 2007 Tehnologies,Inc. Page 1

2 AC MPS Specification Issues Despite the obvious advantage of supporting low DC power consumption by the PD, the AC MPS method is too vaguely described within in the IEEE 802.3af specification to be 100% foolproof and interoperable method of PD disconnect detection. Some key areas where specification issues exist are: Wide Discretion Range: PSE s have full discretion on the interpretation of signature loads in the range of 27 KΩ to 1.98 MΩ. A PD with a connection signature of 50 KΩ or even 100 KΩ could probably survive a tremendous amount of PSE-PD interoperation testing with no evident problems. PSE s that use higher MPS frequencies and look for 1.98 MΩ signature impedance to remove power may be vulnerable to high impedance shunt conductance paths that remain after a PD is disconnected. Example: High humidity environments. Behavior Given Small DC Loads: Removal of an AC MPS signature load may or may not be sensed as a cause to power down at the PSE depending upon other characteristics of the PD. A PD that initiates a power down by setting current to less than 5 ma and disconnecting it s AC MPS signature load might remain powered in many cases because whatever load is drawing the low DC current will still in all likelihood appear as a valid AC MPS load signature to the PSE. The threshold for cut-off is 1.98 MΩ and there is no lower bound on AC MPS frequency. A PD must effectively draw zero DC current in order to assure power-down across all PSE s regardless of the specified requirements for DC MPS! Lack of a Minimum AC MPS Frequency: Capacitive AC MPS signatures in a PD are only meaningful given a known frequency of the AC MPS signal. Looking at the 802.3af specification for PD s, there is an AC MPS signature requirement consisting of no higher than KΩ resistance in parallel with no less than.05 µf capacitance. This resistance component forces all signatures to be valid regardless of AC MPS frequency, however it therefore precludes purely capacitive AC MPS signatures that would draw no DC power. PD developers may attempt to sidestep the associated DC power drain associated with the MPS resistive load signature. Undefined Waveform: The specification in no way attempts to describe the shape of the AC MPS waveform it could be a continuous (sine wave) or it could be discrete pulses, or anything in between. This too will affect how PD MPS signature loads appear narrow pulses might be more affected by a capacitive signature than a sinusoidal counterpart waveform. AC MPS Signaling After Power-Down: There is no explicit indication of what happens to the AC MPS signal following the power-down event nor any description of how that signal behaves in the presence of detection load signatures. An inference might be made that the AC MPS should be disabled and should not interfere with the detection signaling and detection measurement processes. However actual devices are not all implemented with capability to control AC MPS per port meaning the signal may or may not interfere with other processes. Implication to the PD Interestingly, the IEEE 802.3af specification allows for the combination of AC and DC MPS techniques. PD s are required to present a 26KΩ or smaller AC MPS signature when powered and are also required to maintain the minimum load current requirements of DC MPS method. As a practical reality however, the DC Method effectively dictates what a PD must do to assure it will stay powered since a PD cannot know whether the PSE is using the DC MPS or AC MPS method. A PD must therefore draw the maximum Imin2 current level of 10mA to stay powered regardless of the presence of the AC MPS signature resistance thus meaning that to a PSE using the AC MPS method, the effective signature resistance in that PD will always be satisfied by the 10mA current flow (effectively a 5KΩ signature). The one case where this is not true is a PD that modulates it s steady state DC Load such that it meets the required 20% duty cycle for valid (greater than 10mA) load current. (see Tmps). In this isolated case, the AC MPS PSE could possibly rely upon the presence of the MPS signature load to keep power alive to the PD. DC MPS Measurements The PowerSync Analyzer s PSE Conformance Test Suite offers two tests, mps_dc_valid and mps_dc_pwrdn to assess DC MPS behaviors at the PSE. These tests generate all four parameters described earlier for the DC MPS method. Imin1 (Maximum) Invalid DC Signature Threshold The mps_dc_pwrdn test measures Imin1 by stepping the DC load current from a valid value (> 10 ma) to a sequence of invalid values starting at 0 ma and verifying that a power down occurs within one second of the load step. Imin1 is then reported as the highest value of DC load current that successfully forces the port shutdown. June 5, 2008 Copyright 2007 Tehnologies,Inc. Page 2

Tmpdo Shutdown Timing The mps_dc_pwrdn test measures Tmpdo as the time interval from a DC current load step between a valid current level and a known invalid level until a power-down transition is

3 Tmpdo Shutdown Timing The mps_dc_pwrdn test measures Tmpdo as the time interval from a DC current load step between a valid current level and a known invalid level until a power-down transition is observed. Ideally, the test is constructed using the invalid load and/or a load transient to actively pull charge from the PSE port in order to accurately capture the power-down transition. The test requires accurate timing alignment to the load current step transition. Imin2 (Minimum) Valid DC Signature Threshold The Imin2 measurement within the mps_dc_valid is essentially the inverse of the Imin1 measurement. The test measures Imin2 by Figure 1: Shutdown Timing sequentially launching 500 msec second load transients looking for the highest level of DC load current that does not result in a port power-down event. Port voltage is sampled over the duration of the low-load transients to find any transitions below the Transition minimum powered port voltage level. To Imin1 Tmps (Minimum) Valid DC Signature Timing Tmps represents the minimum time (or duty cycle) over which a valid DC signature (load current level) must be maintained in order to negate (or cancel ) a disconnect power-down event. The mps_dc_valid test measures Tmps using timed DC current load transitions. After powering the PSE port, a transition to zero ma DC load current is followed by a timed valid DC load Figure 2: Tmps Load Waveform current pulse, then a second interval of 0 ma over a 500 msec window. Valid load current is restored approximately 450 msec following the initial transition to zero load current. Port voltage is also measured over this exact same time window. The duration of the of the timed valid DC pulse is successively stepped starting at 100 msec down to 10 msec to recover the first duration where a port shutdown occurs during the 450 msec time window thus suggesting Tmpdo expiration. Tmps becomes the minimum current pulse duration required to keep power alive throughout the full sequence. Another view of Tmps is the minimum duty cycle of valid load current required to keep power alive. In this view, the total period is the combination of smallest valid Tmpdo (300 msec) and largest valid Tmps (60msec) thus creating a 16.7% duty cycle (60 msec / 360 msec). Figure 2 demonstrates a power-down decision where a 60 msec load pulse led to proper decision to maintain power. Figure 3 on the other hand demonstrates a Tmps violation since the PSE killed power given a 75 msec load pulse over 430 msec (> 17 % duty cycle). Unfortunately, the IEEE 802.3af specification is worded poorly in its description of Tmps if read literally, it says that Tmps (< 60 msec) is the minimum time required in order to interpret valid DC current as valid. In fact, from the PSE s perspective, it is the maximum allowed time criteria for interpreting valid current as valid. Figure 3: Tmps Violation Power-Down despite >60 msec load pulse. Vopen_max This is the peak signaling voltage detected in the interval immediately following DC disconnect power-down. This parameter is bound by the maximum open circuit signaling voltage of 30Volts. June 5, 2008 Copyright 2007 Tehnologies,Inc. Page 3

4 AC MPS Measurements The PowerSync Analyzer s PSE Conformance Test Suite provides three tests, mps_ac_pwrdn, mps_ac_vf, and mps_ac_voff to evaluate AC MPS behaviors at the PSE. The parameters and test methods will be presented for mps_ac_vf, mps_ac_pwrdn, then mps_ac_voff. V_open: Peak-to-Peak Amplitude of AC MPS Signal with Open Circuit Load The V_open measurement is derived from a waveform capture timed to the removal of the AC MPS signature load given that there are no other loads that would appear as valid signatures. The peak-to-peak amplitude is assessed as the difference between waveform maximum and minimum values over the 200 msec duration of the sample. Sampling must be finalized before PSE power removal at 300 msec (or later). The measurement method is fully insensitive to waveform shape. Fp: AC MPS Waveform Frequency The same waveform captured for V_open analysis is also analyzed to determine signal periodicity by assessing an effective mid-level voltage, then counting mid-level transitions over the duration of the sample and calculating signal frequency in Hz. Resolution of this measurement is limited to the fact that sample time duration is restricted by the inevitable power-down event. SR: AC MPS Maximum Slew Rate While the waveform shape is not defined within the specification, a maximum slew rate is defined that is equivalent to maximum slew rates elsewhere for detection signaling and power-up edge. Figure 4: AC MPS Power-Down Slew rate requires a time interval measurement in order to resolve to microseconds, especially for small amplitude signals. The mps_ac_vf test carefully evaluates the waveform captured to converge on optimal trigger levels for a subsequent slew rate (risetime) measurement. Isac: AC MPS Detection Current Level This parameter places an upper bound of 5 ma on AC current flow that is associated with the AC MPS detection process. This represents a compliance limit that could only come into effect in the event of a PD that had a relatively low AC input impedance on the order of 100 Ω (e.g. > 80 µf at 20 Hz) and the Figure 5: AC MPS Waveform PSE had a sufficiently stiff source. In reality, since the MPS cutoff signature can range up to 1.98 MΩ, there is no reason for the PSE to furnish this type of AC sourcing capability. The mps_ac_vf test does make a measurement of steady state AC MPS current measured while the PSE is powered and furnishing a small DC load. The measurement is performed by capturing the AC MPS voltage magnitude with an AC peak measurement and then resolving AC current using known AC input impedance characteristics of the PSA test port. The AC load presented by the PSA test port will be between 10 KΩ and 20 KΩ depending upon AC MPS frequency. The mps_ac_pwrdn test returns the following parameters: Tmpdo: Shutdown Timing The mps_ac_pwrdn test measures Tmpdo as the time interval from the removal of a valid AC MPS signature load until the actual power-down event is observed. The test deploys a very short load transient to actively pull charge from the PSE port at the end of the Tmpdo time window order to assure the power-down transition is captured in the event the PSE does not actively discharge the port. Imin1: DC Load Current Required for AC MPS Shutdown This measurement fulfills two purposes. The first is to validate that an effective open circuit load does in fact cause a powerdown from the PSE. The second purpose is to resolve the maximum DC load current configuration that is interpreted by the PSE port as a valid AC MPS signature. This measurement therefore demonstrates AC MPS sensitivity to the characteristics of the PSA test port s active load circuit. This parameter should not be confused with the specified Imin1 parameter developed when testing DC MPS PSE s. The mps_ac_voff test returns two additional parameters: V_open1: The PSE Output Voltage Following Disconnect Decision June 5, 2008 Copyright 2007 Tehnologies,Inc. Page 4

This measurement allows for an initial power-down transition, then recovers peak port voltage in the time interval immediately following this shut-down event.

The test will adapt in the event that an initial voltage transition following the disconnect decision is positive rather than negative (declining) a condition that will be deemed non-compliant.

5 This measurement allows for an initial power-down transition, then recovers peak port voltage in the time interval immediately following this shut-down event. As with open circuit detection signaling, this parameter is limited to 30 volts. The test will adapt in the event that an initial voltage transition following the disconnect decision is positive rather than negative (declining) a condition that will be deemed non-compliant. Vport(max): The Maximum PSE Port Voltage While Determining Disconnect This test assesses the peak voltage measured prior to port power removal by the PSE. Analysis of power-down waveform is analyzed to resolve most likely transition point between disconnect measurement and power-down. Peak voltage in the time interval preceding this transition point is then evaluated. Sample PSE Conformance Test Reports (DC and AC MPS) Figure 6: DC MPS Test Results from PSE Conformance Test Suite Figure 7: AC MPS Test Results from PSE Conformance Test Suite Test Configuration PowerSync Analyzer POWER 10/100BT RS -232 TRIG PSA-1200 PW R PW AU In In PW In In PSE For more information on the Technologies Power over Ethernet test & measurement solutions look us up at: Technologies, Inc East Street Tewksbury, MA sales@sifos.com Publication # June 5, 2008 Copyright 2007 Tehnologies,Inc. Page 5

PSA-3002-SA Comprehensive PoE Service Analysis & Verification

PSA-3002-SA Comprehensive PoE Service Analysis & Verification Comprehensive PoE Service Analysis & Verification Product Overview Key Features Analysis and Troubleshooting for All Standard PoE Installations 13W Services 25.5W Services 25.5W + PoE LLDP Services Fully