Just added some more data. Free actuator displacement, free force transducer reading. I had to use a frame with sampling rate set to 1024 Hz so I went up to only 20 Hz cycling, but it looks interesting.
The PVS002 Sine Sweep Analysis (link is below) has been expanded to include some of your more recent plots. The analysis is converging on a repeating pattern of the sine sweep plot, then a representative error plot and lastly a representative loads (dynamometer and system) plot.
It will be interesting at some point to get data for a tension-tension sine sweep. If that also shows the three regions mentioned at the beginning of the analysis then that may turn out to be a great way to simplify the overall process for users. Let's discuss at your convenience.
The sine sweep shows a curious phenomenon. The error amplitude is nearly independent of frequency. This suggests the errors are not inertially induced. Starting a post to explore that.
Just at a glance, I'm not seeing a strong correlation between frequency and error across the data I'm seeing so far. Unless it's at frequencies below 5Hz.
All data here was collected on a single test frame, a Landmark 370.10, 20-kip capacity.
Crosshead height was 29.5 inches.
Grips were MTS 647.10 wedge type, weight of each grip body is ~ 66 lbs.
Data was collected using MTS Basic Testware on an MTS Flextest 40.
Data channels are synchronized.
A single "HCF" type round dynamometer was used, steel, 1/2" diameter reduced section, 3.1 inches grip to grip, so fairly stiff.
Sample rate, test frequencies and applied force are as indicated.
"Noise" data recorded with dynamometer and force transducer unclamped without cycling, showing 5 seconds and also showing zoomed in to the first 0.03 seconds.
Shown is the difference "error" between force transducer and dynamometer for cycling as well as "noise".
Did not observe appreciable difference in noise between using very low dither (0.01 - 0.25 v 400-528 Hz) and no dither. Further data will be collected without dither.
My intention next is to "fill in the gaps" on what I have here, plus show fully reversed vs tension or compression only, at these and a few more frequencies and endlevels.
Unfortunately, I might not be able to attend the meeting today.
Here's some pertinent info on the data for the group, though.
All data here was collected on a single test frame, a Landmark 370.10, 20-kip capacity.
Crosshead height was 29.5 inches.
Grips were MTS 647.10 wedge type, weight of each grip body is ~ 66 lbs.
Data was collected using MTS Basic Testware on an MTS Flextest 40.
Data channels are synchronized.
A single "HCF" type round dynamometer was used, steel, 1/2" diameter reduced section, 3.1 inches grip to grip, so fairly stiff.
Sample rate, test frequencies and applied force are as indicated.
"Noise" data recorded with dynamometer and force transducer unclamped without cycling, showing 5 seconds and also showing zoomed in to the first 0.03 seconds.
Shown is the difference "error" between force transducer and dynamometer for cycling as well as "noise".
Did not observe appreciable difference in noise between using very low dither (0.01 - 0.25 v 400-528 Hz) and no dither. Further data will be collected without dither.
My intention next is to "fill in the gaps" on what I have here, plus show fully reversed vs tension or compression only, at these and a few more frequencies and endlevels.
I really like what you have done and where you are headed with this. This information more than covers what I was looking for. Pete and I can take it from here for today. Thanks!
Just added some more data. Free actuator displacement, free force transducer reading. I had to use a frame with sampling rate set to 1024 Hz so I went up to only 20 Hz cycling, but it looks interesting.
The PVS002 Sine Sweep Analysis (link is below) has been expanded to include some of your more recent plots. The analysis is converging on a repeating pattern of the sine sweep plot, then a representative error plot and lastly a representative loads (dynamometer and system) plot.
It will be interesting at some point to get data for a tension-tension sine sweep. If that also shows the three regions mentioned at the beginning of the analysis then that may turn out to be a great way to simplify the overall process for users. Let's discuss at your convenience.
********
https://open.substack.com/pub/albrightj/p/pvs002-sine-sweep-analysis
The sine sweep shows a curious phenomenon. The error amplitude is nearly independent of frequency. This suggests the errors are not inertially induced. Starting a post to explore that.
https://open.substack.com/pub/albrightj/p/pvs002-sine-sweep-analysis
Just at a glance, I'm not seeing a strong correlation between frequency and error across the data I'm seeing so far. Unless it's at frequencies below 5Hz.
I added some more data showing a summary of Max Error as % of Range vs. Range at 3 load ratios, with frequencies labelled.
I added quite a few more tests today.
5, 10, and 30 Hz at +/- 2000 lbf, +/- 5000 lbf, 200-2000 lbf, 500-5000 lbf, -2000 to -200 lbf, -5000 to -500 lbf.
Also added a slow frequency "sweep" from 5 to 60 Hz at +/- 1000lbf.
Great additions! How did you do the 3x3 plot matrix?
That was actually a mistake. I uploaded several as a gallery instead of individually. I reuploaded them individually.
All data here was collected on a single test frame, a Landmark 370.10, 20-kip capacity.
Crosshead height was 29.5 inches.
Grips were MTS 647.10 wedge type, weight of each grip body is ~ 66 lbs.
Data was collected using MTS Basic Testware on an MTS Flextest 40.
Data channels are synchronized.
A single "HCF" type round dynamometer was used, steel, 1/2" diameter reduced section, 3.1 inches grip to grip, so fairly stiff.
Sample rate, test frequencies and applied force are as indicated.
"Noise" data recorded with dynamometer and force transducer unclamped without cycling, showing 5 seconds and also showing zoomed in to the first 0.03 seconds.
Shown is the difference "error" between force transducer and dynamometer for cycling as well as "noise".
Did not observe appreciable difference in noise between using very low dither (0.01 - 0.25 v 400-528 Hz) and no dither. Further data will be collected without dither.
My intention next is to "fill in the gaps" on what I have here, plus show fully reversed vs tension or compression only, at these and a few more frequencies and endlevels.
Great data Kirk! Thanks for posting. Are you available and willing to chat a bit about it in the Monday meeting?
Unfortunately, I might not be able to attend the meeting today.
Here's some pertinent info on the data for the group, though.
All data here was collected on a single test frame, a Landmark 370.10, 20-kip capacity.
Crosshead height was 29.5 inches.
Grips were MTS 647.10 wedge type, weight of each grip body is ~ 66 lbs.
Data was collected using MTS Basic Testware on an MTS Flextest 40.
Data channels are synchronized.
A single "HCF" type round dynamometer was used, steel, 1/2" diameter reduced section, 3.1 inches grip to grip, so fairly stiff.
Sample rate, test frequencies and applied force are as indicated.
"Noise" data recorded with dynamometer and force transducer unclamped without cycling, showing 5 seconds and also showing zoomed in to the first 0.03 seconds.
Shown is the difference "error" between force transducer and dynamometer for cycling as well as "noise".
Did not observe appreciable difference in noise between using very low dither (0.01 - 0.25 v 400-528 Hz) and no dither. Further data will be collected without dither.
My intention next is to "fill in the gaps" on what I have here, plus show fully reversed vs tension or compression only, at these and a few more frequencies and endlevels.
Kirk, this is wonderful!
I really like what you have done and where you are headed with this. This information more than covers what I was looking for. Pete and I can take it from here for today. Thanks!