3 Checking the installation

Contents of this section

Installation verification procedures can be divided into two main categories:

1. Initial quick checkout which can be performed with either Mark~III or Mark~IV formatter connected. The procedures presented below assume a Mark~III formatter. Equivalent ``form4='' command strings must be substitute if a Mark~IV formatter is being used instead. (A listing of Mark~IV allowed commands can be found at ftp://dopey.haystack.edu/pub/mark4/formater/mk4vocab.txt .)
2. A more detailed verification procedure which exercises all signal paths of HAM, RDM and WRMs. Mark~IV formatter is connected at this stage.

3.1 Initial quick checkout

When you change the recorder type in ``equip.ctl'', you also need to change the definition of R/W electronics channels in ``head.ctl'' (Mark~III definitions shown here):

---

* Write Heads  Read Heads  Reproduce: Electronics  Detector
      all        odd                      odd        odd

---

---

* Write Heads  Read Heads  Reproduce: Electronics  Detector
      all        all                      all        odd

---

After editing the files, please restart the FS (``terminate'', ``fs'').

The purpose of the following tests is to verify that the modules have been manufactured according to the design documents and that there are no defective components, solder joints, or other manufacturing errors and deviations.

PWM

Functionality can be tested by connecting the power module to the Metrum/Honeywell 96 recorder power supplies and verifying the output voltages available at front panel jacks. Please test the functionality of FS9 ``enable='' command as follows. Please note that you have to issue ``st=for,0,on'' after each enable command for the new setting to take effect. ``enable='' updates only FS internal variables and only ``st='' command sends the appropriate bits to the PWM via IOM.

enable=

All four WRM write voltage LEDs should be off.

enable=stack1

WRM 1 and WRM 3 write voltage LEDs should be on.

enable=stack2

WRM 2 and WRM 4 write voltage LEDs should be on.

enable=stack1,stack2

All four WRM write voltage LEDs should be on.

Using low-level MAT commands to test IOM.

You can use the FS ``mat='' command to directly access individual bits on the IOB according to the protocol described in Haystack drawing ``4712D001'', ``Mark~IV Recorder Communication Protocol'' which is provided in the IOM section of the documentation binder. For example:

---

mat=#94=00000034.$

---

You can interrogate the current bit values with:

---

mat=#94.>,?

---

Other useful MAT commands for testing the IOM:

---

mat=#94=80000101%$     all WRM voltage LEDs of PWM on
mat=#94=80000001%$     LEDs 1 and 3 on
mat=#94=80000100%$     LEDs 2 and 4 on
mat=#94=00000x0x%$     all LEDs off

---

3.2 Detailed verification procedures

If you have been using the old Mark~III formatter for the previous tests, now it is time to complete signal cabling as instructed in section sig-cables . Please do not forget to update ``/usr2/control/equip.ctl'' according to instructions in section edit-equip and restart the FS.

HAM

Dismounted HAMs can be tested by connecting a spare Metrum head stack to the HAM and verifying the (relatively simple) write circuitry of HIBs (see the HIB schematics in documentation binder slot HAM) with an ohmmeter, inspecting HIBs for shorts and bad joints. The write circuit actually encompasses a back-to-back diode and a single 1~kohm resistor in series with the head coil: the transistors do not actively participate in writing, so this simple conductivity test should be sufficient.

At Metsähovi, the read preamplifiers on HIB and HPB boards were verified with an IFR AN 920 spectrum analyzer with a tracking generator. Both read and write cables of the HAM under test were connected to the corresponding CPM, to verify the cables associated with them. A 4cm open ``antenna'' wire was connected to the tracking output generator coaxial cable. The wire was placed in the vicinity (about 1 cm) of the head stack to invoke aerial pickup sufficient to create a small signal for the preamplifiers. The bandpass curve which is characteristic to this setup shows that all amplifier channels are performing in the same way. The resulting graphs can be found in station-specific section of the documentation binder.

WRMs

The modules were connected to the Metsähovi VLBA formatter and HAM #006 which is permanently installed in our recorder. (Our HAM uses the same design and is identical to the HAMs supplied.) The formatter was set up to record in Mark III mode A mode at sampling rate of 8~MHz, resulting in track bit rate of 9~Mbit/s. The VLBA formatter emulation of Mark III embeds VLBA track information in the auxiliary field of all output frames, thus making it possible to detect at each output from which input each signal originates. The exact setup is shown in the SNAP procedure file ``mdextr.prc'' provided in the station-specific section of the documentation binder. The procedure file is also available at ftp://kurp-ftp.hut.fi/pub/mk4/mdextr.prc .

The ``repro=byp,...'' command exercises the bypass reproduce selectors of WRMs to route the signal present at WRM write outputs back to the recorder decoding logic. Any faulty components, shorts, bad joints etc. should manifest themselves as wrong track numbers when decoded with the procedure ``md2trk''. This procedure gives output similar to the following:

9608509090343/mcb/0001    <--- aux field captured, 0000==not captured
9608509090346/mcb/2304    <--- '04' in hex: VLBA track 4 in channel A
9608509090348/mcb/0100    <--- rest of 32-bit aux field
9608509090458/mcb/0001      <--- capture channel A again
9608509090461/mcb/2304      <---    
9608509090464/mcb/0100      <---
9608509090580/mcb/0001    <--- capture channel B
9608509090583/mcb/0205    <--- notice '05' in hex: VLBA track number 5
9608509090585/mcb/8100    <---
9608509090696/mcb/0001      <--- capture channel B again
9608509090700/mcb/0205      <---
9608509090702/mcb/8100      <---
9608509090822/dqa/1, 1us,04,0.,0, 0.0, 148, 2us,05,0.,0, 0.2, -86
                         ^  ^  ^                ^  ^  ^
                         !  !  !                !  !  !--- B sync errors
                         !  !  !                !  !--- B parity errors
                         !  !  !                !--- B track number
                         !  !  !--- A sync errors
                         !  !--- A parity errors (per 1 megabyte)
                         !--- A track number

The log listing generated by ``md36trk'' procedure shows that write signals of all VLBA/Mark IV data tracks 0..35 pass through the write signal path correctly and that both A and B bypass reproduce channels are capable of sampling and reproducing back to the recorder any write output channel. Please note that VLBA system tracks 0..1 and 34..35 are routed to VLBA formatter output tracks of 2..3 and 32..33, respectively, and thus the decoded aux field track numbers are the latter. This rerouting and duplication is performed in the VLBA recorder and thus the formatter cannot embed separate identification to system tracks.

An experimental version of the test procedure file for Mark~IV recorder has been provided as ftp://kurp-ftp.hut.fi/pub/mk4/m4extr.prc . To use this (or the VLBA version ``mdextr.prc''), copy the file into ``/usr2/proc'' directory and make it the current procedure file:

---

proc=m4extr

---

---

log=wrm012       or your WRM serial number

---

You can invoke the procedure ``tr36init'' to set up the formatter and the recorder ready for manual ``repro=byp,...'' commands and signal inspection with an oscilloscope at WRM monitor BNC connectors. To test all track combinations of WRM #1 (or the only WRM in a VLBA drive), invoke ``testwrm'' (no ``tr36init'' is required before this). ``testwrm=1'' should add 100 to all ``repro=byp,...'' track numbers and thus it should test WRM #2 if you are using ``m4extr.prc'' with Mark~IV equipment. This test procedure uses the Mark~III decoder with the ``decode=a,aux'' command to dump the auxiliary field in hexadecimal into the log file. Decoding the hexadecimal representation which includes parity bits has been discussed in section pre-checks .

After the decoded track numbers show that all Mark~IV tracks 02--33 of both WRM modules are functional you should write some thick tape at 4MHz sampling (4.5Mbits/s track output) at 135ips, and at 8MHz sampling (9Mbits/s track output) at 270ips. Please make sure that the rack/VC IF inputs are not at zero level: this causes recorded data to contain lots of zeroes which is the worst case for NRZ-M encoded data. If your write head has been contoured for thin tape according to Hans Hinteregger's recommendations, you can set the Mark~IV formatter to a canned mode which will result in 18Mbits/s track output rate and record at 320ips. An excerpt from Hans' message on 13-Oct-1996 about the importance of head contour:

        The same recorded 3M5345 tape that registered 26 dB in good contact 
on one machine, was down to 13 dB on a new head that was not fully run-in 
after 50 hours, after a single 'morning test' run with a thick relatively 
abrasive FujiH621 tape at 15".  In this case error rates went to about 
10**-2.  SNR was restored to 20 dB with the contour tape, still far from 
optimum but useable.

You can verify the recordings either manually (see section first-tests ) or with the automatic procedure described in the next section test-RDM .

A new ``4hdcal.snp'' automatic narrow-track calibration schedule file does not exist yet. It should be possible to use the ``track 15/18'' adapter cable described in section withmk3 in WRM #1 even input connector J5 to achieve a ``forced'' single track output as in Mark~III mode D with only head 15 (VLBA/Mark~III head 18) enabled. (Remove all other write inputs at WRM front panels during the calibration.) It may also be possible to command the Mark~IV formatter to enable only track 18.

With the adapter cable in place, you may want to attempt to edit the standard Mark~III ``hdcal.snp'' into a ``4hdcal.snp'' by:

1. Checking that all ``form4='' commands result in track 18 being output at 4.5Mbits/s.
2. Incrementing all track numbers in ``repro='' commands by 3.
3. Adding ``,1,4'' to all repro commands to select the 135ips equalizer and 4MHz clock recovery rate.
4. Checking that all ``stack='' commands reference write (#1) and read/write (#2) stacks correctly.

3.3 Performance tests

3.4 Writing a Mark III test tape

Next Chapter, Previous Chapter

Table of contents of this chapter, General table of contents

Top of the document, Beginning of this Chapter