Subject: RE: TAKENAKA
Date: Wed, 19 Jan 2000 07:34:55 -0600
From: "Rood, Omar" <Omar.Rood@mts.com>
To: "Imafuku, Nobuhiro" <Nobuhiro.Imafuku@mts.com>, "Larson, Rod"
     <Rod.Larson@mts.com>
CC: "Katabami, Yoshihiro" <Yoshihiro.Katabami@mts.com>

Hello Fuku,

1.  I do not owe Takenaka any help on repair of their table.

This job was agreed upon that acceptance would be before they connected
to the table.  Then Bami said Takenaka would connect to the table at
their risk as the easiest way to verify performance.

It is absurd that they have bearings with less stroke than the actuator!
MTS goes to the expense to make hydraulic cushions on the actuator for a
good reason.  Takenaka knew the stroke of the actuator.

2.      Anticipation should be 1/2 of  impulse response.  I think your
training time should have been good enough.  You should check the IRF
plotter for a good training before running.

3.      While training, did you turn down convergence before stopping
the program?

4.      Did you preview before you ran?

5.      Were you in tracking? Or frozen mode.  You should go to frozen
mode for an earthquake record.

6.  Velocity, the 800 gpm size valve was needed to meet the velocity
spec at the acceleration corner with a maximum specimen.
Omar

> -----Original Message-----
> From: Imafuku, Nobuhiro
> Sent: Wednesday, January 19, 2000 6:03 AM
> To:   Rood, Omar; Larson, Rod
> Cc:   Katabami, Yoshihiro
> Subject:      TAKENAKA
>
> Hello, OMAR & ROD san.
>
> 1. We had a trouble this Monday. After tuning system with table, we
> did maximum performance test and 100 gal THD test (0.5, 1, 3, 5, 8 &
> 10 Hz) for our reference. Both test results were quite good (I want to
> send you test results but I'm sorry I can NOT send you today because I
> forget floppy disk at TAKENAKA site). Then, after finishing AIC
> tuning, we run KOBE earthquake. We use 2 minutes 33 second length
> earthquake.
> 1)    When we tuned system with "Impulse Response : 1 sec" &
> "Anticipation : 0.33 sec", I think these parameters worked OK
> (normally we use these parameter for existing table tuning).
> 2)    When we tuned system with "Impulse Response : 4 sec" &
> "Anticipation : 1.33 sec", I'm not sure these parameters worked OK or
> not. We tuned system about 15 minutes at AIC training. After passing
> about 50 second, actuator suddenly extended about + 43 cm position
> then retracted about - 51 cm position (I assume actuator moves maximum
> velocity). Then, we pushed E-Stop button.
> 3)    We checked system, but unfortunately table slide bearings were
> broken. The reason why actuator stopped about - 51 cm position, we had
> steel stopper at this position.
> 4)    We disassembled table yesterday and sent all slide bearings to
> vender factory.
> 2. We are still NOT sure why actuator suddenly goes unstable. But, I
> assume one reason is caused by AIC control loop. I still do NOT
> understand AIC control concept completely, but I think we have some
> limitation to use AIC such as noise level because this trouble happens
> at very low level (Table noise level is less than 5 gal). I understand
> we have no responsibility for table design. But, I think we have to
> help TAKENAKA to fix this trouble. Please reply your comment.
> 3. I have some questions.
> 1)    What kind of controller did you use to calibrate load cell ?
> Because this system controller is TestStar IIs but we have calibration
> disk and shunt cartridge for TestStar II. Is this OK or not ?
> 2)    Both actuators performance is quite good and it is so easy to
> perform 140 - 150 kine at 0.3 - 1 Hz range. Are these actuator so high
> performance ? I think this is so dangerous for this system.
>
> REGARDS     FUKU
>
