Phase shift in cabling?
in [Design]
|
From: DigitalPlease on 3 Jun 2010 13:19 I'm working with linear hydraulic actuators which have multiple LVDTs (for the ram and main control valve) and I have a spec for allowed phase shift from primary to secondary for the transformers. I'm trying to make sense of the Labview code I've been given for measuring this parameter and have found an unexplained hardcoded offset value related to this phase measurement. My analog skills are down so I feel like I'm missing something obvious. The setup is: The excitation board is connected to the actuator via a ~25ft cable. Excitation is ~1900Hz at 5V. If I probe either near the excitation board or near the actuator, I get similar measurements for a couple of different parts that are in the 3-7 degrees range, well within the +- 10 degree spec. The problem is, the hardcoded offset is 5 degrees so some parts pass and some parts fail. I assume the cabling can impart some sort of delay thanks to capacitance/inductance of the line and perhaps this is where the offset is coming from, but I'd like to understand if this is correct and how I can characterize this to calibrate my measurement.
From: Tauno Voipio on 3 Jun 2010 14:18 DigitalPlease wrote: > I'm working with linear hydraulic actuators which have multiple LVDTs > (for the ram and main control valve) and I have a spec for allowed > phase shift from primary to secondary for the transformers. > > I'm trying to make sense of the Labview code I've been given for > measuring this parameter and have found an unexplained hardcoded > offset value related to this phase measurement. My analog skills are > down so I feel like I'm missing something obvious. > > The setup is: The excitation board is connected to the actuator via a > ~25ft cable. Excitation is ~1900Hz at 5V. If I probe either near the > excitation board or near the actuator, I get similar measurements for > a couple of different parts that are in the 3-7 degrees range, well > within the +- 10 degree spec. The problem is, the hardcoded offset is > 5 degrees so some parts pass and some parts fail. > > I assume the cabling can impart some sort of delay thanks to > capacitance/inductance of the line and perhaps this is where the > offset is coming from, but I'd like to understand if this is correct > and how I can characterize this to calibrate my measurement. Come on ... The free space delay is roughly 1 ns / ft. In most common cables it will be about 1.5 time as much. For your cable, it is less than 40 ns there and another back. How big phase shift is this at 1900 Hz? -- Tauno Voipio, MSEE (and OH2UG for nearly 50 years) tauno voipio (at) iki fi
From: Vladimir Vassilevsky on 3 Jun 2010 14:29 Tauno Voipio wrote: > DigitalPlease wrote: > >> I'm working with linear hydraulic actuators which have multiple LVDTs >> (for the ram and main control valve) and I have a spec for allowed >> phase shift from primary to secondary for the transformers. >> >> I'm trying to make sense of the Labview code I've been given for >> measuring this parameter and have found an unexplained hardcoded >> offset value related to this phase measurement. My analog skills are >> down so I feel like I'm missing something obvious. >> >> The setup is: The excitation board is connected to the actuator via a >> ~25ft cable. Excitation is ~1900Hz at 5V. If I probe either near the >> excitation board or near the actuator, I get similar measurements for >> a couple of different parts that are in the 3-7 degrees range, well >> within the +- 10 degree spec. The problem is, the hardcoded offset is >> 5 degrees so some parts pass and some parts fail. >> >> I assume the cabling can impart some sort of delay thanks to >> capacitance/inductance of the line and perhaps this is where the >> offset is coming from, but I'd like to understand if this is correct >> and how I can characterize this to calibrate my measurement. > > > Come on ... > > The free space delay is roughly 1 ns / ft. In most > common cables it will be about 1.5 time as much. > For your cable, it is less than 40 ns there and > another back. How big phase shift is this at 1900 Hz? The stray capacitance of a cable can be substantial: for OP's 25 ft it could be in ~nanofarad range. This capacitance combined with impedances and crosstalks could make all kinds of peculiar phase shifts. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
From: Tim Wescott on 3 Jun 2010 14:51 On 06/03/2010 11:18 AM, Tauno Voipio wrote: > DigitalPlease wrote: >> I'm working with linear hydraulic actuators which have multiple LVDTs >> (for the ram and main control valve) and I have a spec for allowed >> phase shift from primary to secondary for the transformers. >> >> I'm trying to make sense of the Labview code I've been given for >> measuring this parameter and have found an unexplained hardcoded >> offset value related to this phase measurement. My analog skills are >> down so I feel like I'm missing something obvious. >> >> The setup is: The excitation board is connected to the actuator via a >> ~25ft cable. Excitation is ~1900Hz at 5V. If I probe either near the >> excitation board or near the actuator, I get similar measurements for >> a couple of different parts that are in the 3-7 degrees range, well >> within the +- 10 degree spec. The problem is, the hardcoded offset is >> 5 degrees so some parts pass and some parts fail. >> >> I assume the cabling can impart some sort of delay thanks to >> capacitance/inductance of the line and perhaps this is where the >> offset is coming from, but I'd like to understand if this is correct >> and how I can characterize this to calibrate my measurement. > > Come on ... > > The free space delay is roughly 1 ns / ft. In most > common cables it will be about 1.5 time as much. > For your cable, it is less than 40 ns there and > another back. How big phase shift is this at 1900 Hz? That logic would apply if the source were impedance matched to the cable. But a typical LVDT has tons of leakage inductance and series resistance in its transformer -- it is, after all, a sensor, not a power conversion device. All this leads to a fairly high source impedance. So a more accurate model is to look at the cable capacitance working against the LVDT source impedance -- do this, and you'll see that the phase shift can get severe. -- Tim Wescott Control system and signal processing consulting www.wescottdesign.com
From: mook johnson on 5 Jun 2010 12:24
"DigitalPlease" <mrperfectionest(a)gmail.com> wrote in message news:5aafaf2e-ad18-419b-8141-54734805c857(a)w3g2000vbd.googlegroups.com... > I'm working with linear hydraulic actuators which have multiple LVDTs > (for the ram and main control valve) and I have a spec for allowed > phase shift from primary to secondary for the transformers. > > I'm trying to make sense of the Labview code I've been given for > measuring this parameter and have found an unexplained hardcoded > offset value related to this phase measurement. My analog skills are > down so I feel like I'm missing something obvious. > > The setup is: The excitation board is connected to the actuator via a > ~25ft cable. Excitation is ~1900Hz at 5V. If I probe either near the > excitation board or near the actuator, I get similar measurements for > a couple of different parts that are in the 3-7 degrees range, well > within the +- 10 degree spec. The problem is, the hardcoded offset is > 5 degrees so some parts pass and some parts fail. > > I assume the cabling can impart some sort of delay thanks to > capacitance/inductance of the line and perhaps this is where the > offset is coming from, but I'd like to understand if this is correct > and how I can characterize this to calibrate my measurement. Could it be that there is some analog filtering before the A2D that causes a 5 degree phase shift at 1.9Khz? |