PEARS REPEATER MAINTENANCE LOG - 2013
A record of trips to/work at various repeater sites by PEARS members or helpers (reverse time order)
|
DATE |
SITE |
PERSON/S |
ACTION |
|
6 December |
- |
ZS2AE |
Wifi issues No improvement found at Slipper, so the lack of coverage persists. ICASA once again informed us that we were on the 5620MHz radar comms channel. Yet again we had to shuffle around the 5GHz band looking for a slot that would at least allow connectivity with some form of Circuit Quality. Then we experienced total interference at the best performing hight site (until then) at Lovemore. No slot could be found between 5625-5825MHz. in desperation, ICASA was contacted to verify the permitted channels, and we were told that it was 5470-5820. Since the default channels list for South Africa excludes certain channels seemingly allowed by ICASA, I set about adding these channels into the scan list manually, and configuring all users off Lovemore to include this channel, whereupon the Lovemore Access Point was moved to the new channel. Most users followed, one Ubiquiti user couldn't follow due to his settings, and another got left behind because he was oiffline when I did the scan list edits. In the end though, we have improved Quality figures drastically and are probably invsisble to all who use a degfault scan list!! Still need to repair Slipper though... :( |
|
5 December |
Slipper |
ZS2GV, ZS2AAW |
Mains checks, replace rusted wifi box, repair HF antenna We were met on site my the PE Munic Electricity Dept to do some tests on the power reticulation. An answer to the continuous mains fail alarms was discovered today. In single-phasing conditions, the whole mountain supply becomes unstable due to fluctuating loads on two of the three phases. This causes a cycling of power to all users, with a period of approx 10s. Hence the mains fail alarm clears and activates for the whole time that the single pahsing condition exists. (Not great for switch-mode power supplies!!). Steps have been taken to remove the identified contributors (not us!), which is a three phase customer that has individual charger modules on the three phases to Neutral, two of which cycle between themselves when the HT phase is lost. Glen did the climbing today, and firstly removed the rusted wifi box for rebuilding down at ground level. He then installed a new aluminium spreader pipe over the broken halves, and clamped the centre conductor to this spreader, to re-instate the wideband HF antenna top end. He also crimped a new lug onto the stainless wire for re-attachment to the balun. The wind was rather strong, as seen by the toolbag being swept off course.
ZS2GV planning his attack. Broken spreader pipe to the top RHS
Toolbag wind meter....
The old rusted wifi box
RB500 rebuilt into the stainless steel box Thanks Glen! |
|
14 November |
George |
ZS2AE, ZS2AAW |
Maitenance checks before the Garden Route Rally This site was visited in thick mist and rain, and during a power outage. The UHF link frequency error on TX was corrected by means of the adjustment cap, but the cxrystal appears to be a little erratic. TX power, RX sensitivity, audio levels all appear normal. The link was re-connected to the original UHF link antennas (one feed, split into two yagi's east and west) and signals from Riversdale were returned when testing DTMF. VSWR tests into this antenna were excellent. It was noted that the 7/8 heliax supplied for the Sentech mast installation was stored inside the building, ready for action. VHF repeater was tweaked for max performance through the multiple-pass cavity filters on site. All in order for the short term. |
|
14 November |
Knysna (Brenton) |
ZS2AE, ZS2AAW |
Maitenance checks before the Garden Route Rally The VHF repeater had been isolated for some time, with open-squelch noise persisting until time-out. The quick site check revealed a borderline squelch setting and a low recovered audio level, further investigation confirmed the failure of the DC decoupling electrolytic in the output of the receiver IF board being low value, which reduces the audio level and noise level into the squelch circuit, thereby opening the squelch. Fortunately a replacement 4.7uF 63V cap could be scrounged frm the toolkit, and this was fitted on site. RX audio levels returned to normal, as did the suelch control. Receive and transmit filters were re-aligned, all other specs being within tolerance on the repeater. Both UHF links were checked, and netted where necessary. Signals were tested to George, and also to a handheld from the Plett side, since Plett was already decommissioned. This ensured that the Knysna repeater could be linked to George and used for comms if necessary. |
|
13 November |
Slipper |
ZS2AAW |
BBS/Email PC offline Remote power-off power-on commands could not revive the PC, so a trip to site was called for. When on site, a simple power-on button press brought the PC to life. Since it could not be verified that there was no problem, it was taken back to the QTH for checking, where it can continue to perform the email server function when coupled to the wifi network there The Eastern end of the HF multiband antenna had broken loose - on the nylon rope originally part of this antenna (not the ski-rope used for the anchor point). Reshuffling of knots got the antenna back up and supporting the sectional pole mast (the antenna acts as the East-West stay wires too). The western end was checked and all appeared fine. Its high time that the stainless steel rope is fitted, to replace the nylon ropes originally used... |
|
1 October |
Mount Rd |
ZS2AAW |
Room being refurbished When visiting the site to reset the UHF Link channel after a power failure incident, I found the contractors hard at work clearing the repeater room of the broken toilet, cubicle and hand basin. Thankfully the sewer vent odour will now be a thing of the past :)
The wall demolished, toilet and basin removed
The Repeater still working in the dust
|
|
17 September |
Workshop |
ZS2AAW |
Wifi Frequency Audit by ICASA ICASA telephoned
me to verify all frequencies in use on the PEARS Wifi
network. They are shutting down users that are not within the following
limits: As it turns out, two of our frequencies are outside these limits due to moving away from interference in the past. Mount Rd AP12 was on 5880MHz, and Mount Rd - Slipper link was on 5900MHz. I undertook to move these back to within the legal channels. Mount Rd
AP12 is now on 5740 MHz Signal to noise ratios are a little worse, but CCQuality still seems acceptable. |
|
16 September |
Slipper |
ZS2AAW |
Install remote power control relay
for PC on mains
With the inverter being too small to run the BBS and email PC, there is no longer a remote reset function on the PC. The recent mains fail on Slipper on the weekend caused the PC not to reboot, so the site had to be visited to power off-on the PC. Therefore a 12V relay was installed in pace of the inverter, and the mains connection to the PC was routed through the NO contacts so that the PC can once again be controlled by DTMF 70014 * or #. All tested OK Noted that the HF antenna is now ripped off the balun feed, so urgent repairs are required. Also noted that the mains fail input floats at 20V! Seems a little weird. More investigation to be done next time, in an attempt to deduce why the mains alarm gets stuck into a repetitive loop each time the mains fails. |
|
14 May |
Workshop |
ZS2AAW |
MFJ TNC Repaired The faulty MFJ1274 TNC from Town Repeater site was inspected for the permanently-on DCD LED, which in turn caused it never to transmit. The fault was found to be the negative 5V rail regulator transistor Q2 which was open-cct from base to emitter. This line feeds the XR2206 and 2211 decoder and encoder chips that form the 1200bd modem cct. Rather a weird fault, considering the circuit is only designed for a few mA. Replacing it with a BC237 brought the -5V rail back to life, and the PLL tone decoder chip resumed operation. Calibration of the unit still needs to be done, which means changing the EPROM back to the standard TNC2 version which allows calibration (the UI-digi firmware does not support "normal command mode"). |
|
MFJ TNC Power Supply section schematic
|
|||
|
13 May |
Mount Road |
ZS2AAW |
UHF Link no RX fault Went to site to investigate the UHF link not receiving signals from 438.650, but still transmitting out, so Town repeater users are heard on Uitenhage and Longmore, but cannot hear these repeater's traffic. An incorrectly set "REPEATER/LOCAL" switch was to blame. |
|
4 & 11-12 May |
Mount Road |
ZS2RA |
Repeater and Wifi
hub down Seems to be a power failure at site. Battery is shot, so will not keep the site up for any length of time. Richard kindly went thru a number of times to reset a tripping supply breaker on the 8th floor, that feeds the sub-distribution board on the roof level. |
|
5 May |
Longmore |
ZS2RL & XYL, ZS2TED, ZS2AE,
ZS2AAW & Harmonic
|
VHF Repeater repaired The Kyodo repeater was showing similar symptoms to other Kyodo's at Slipper and Town, which was the intermittent loss of transmitted power. This was confirmed when we arrived, by the front panel TX PA bar graph meter not showing any indication, and a test set also showing no RF power out. A workbench was set up, and the repeater disassembled to re-solder the suspected power amplifier control transistor which regulates the RF power output. Once re-soldered, a quick check showed 15W TX again. This site is a solar-powered site, so transmitter power is kept to an economical level to conserve battery power.
The Longmore site
The repair team, ZS2AE, Robyn, ZS2AAW The remote controlled channel change function was also tested to ensure the Longmore repeater linking can be changed from PE Town and Uitenhage repeaters, to the Slipper repeater if required, for Rally purposes. All worked well. Thanks to Ted for the very tasty pastries he brought along! |
|
2 May
|
Lady's Slipper |
ZS2AAW
|
UHF Link fault, PC removed again, and earthing
done The UHF link 438.700 fault was investigated. The Tait T800 unit showed a failed power supply, and upon investigation it was discovered the fuseholder had a bad connection. The fuse was cleaned up but the power supply was also exchanged for one from the UHF packet 1200bd backbone (no longer in use)
Note the missing DC output LED on the Tait Power supply
The failed Email and BBS PC was removed from site again, after it crashed repeatedly. Site earthing was completed, with all equipment being bonded to the incoming supply and feeder entry earths.
16mm2 Earth conductor dropper connected to each unit
And 2.5mm2 earths to the HF rigs
|
|
29 April
|
Lovemore
Heights |
ZS2AAW
|
Gateway PC installed The DStar gateway PC was fetched from Henning. ZS2HK who had prepared it for service at this site. It was installed and connected to the PEARS wifi network, and the DStar controller on site. Henning will perform remote desktop configuration and testing.
DStar gateway PC on the
top shelf, Power Supply & LAN, 48-12V converter,
Rear view of the rack Also recently changed was the orientation of the two VHF antennas for the repeater - they are now on the South side of the mast, favouring Schoenies, for Ironman comms...
The tower looking up from the West side
Upper (RX) folded dipole
Lower (TX) folded dipole |
|
25 April |
Governorskop |
ZS2AAW |
Link to Cradock out of action UHF link from Cradock was one-way towards the North (again). When arriving on site, the RX was permanently open-squelch, and had timed out. Power off-on cured the problem, but further investigation was necessary. Tapping the receiver brought up the fault again, seems to be on the audio board, where the squelch circuit resides. Further manipulation of components and tapping could later not bring it up again, so a dry joint is assumed the cause, and will be monitored and replaced when it fails again. Site was left connected to Cradock and tests performed to Aliwal North were successful. |
|
18 April |
Governorskop |
ZR2DR |
Link to Cradock and Node PC
replaced David kindly offered to re-install the repaired link (a new RX crystal was installed). Tests done immediately thereafter showed everything working perfectly towards Cradock. The node PC also re-installed and hopefully working OK - no packet equipment on Slipper to test with yet, though. |
|
22 March |
Lady's Slipper |
ZS2AAW |
Email & BBS PC seems to have
crashed Went back to site today because the PC was offline, could not VNC into it, or receive emails. Black screen on boot-up, with a fast flash of the screen power LED, so it seems to not like the resolution. Removed and taken home to test on a newer monitor. Mouse also U/S. Seems to boot up direct into 1024x768 OK, so the screen up on Slipper must be old, or faulty. Lockup interference was on again all morning until I got to site, then it stopped! When I left, it started again, so I turned around and went back, it was reluctant to resurface, but when it did, I notice half-scale RX signal strength when TX'ing on the VHF repeater, and ALSO half scale open squelch receiving the SAME noise on the VHF packet Kyodo. Switching off the packet rig does not help, nor does switching off all the Tait mains supplies. Immediately the WIFi is killed, the noise stops. DIdn't come back when the wifi was turned on again either. Also noted is broadband spurious energy on 145.690MHz containing an FM station on 107.9MHz. Will be reported to ICASA. The Icom HF rig audio cable was wired into the conference network, and the RF fly lead fitted. |
|
21 March |
Lady's Slipper |
ZS2NT, ZS2RA, ZS2AE, ZS2EA,
ZR2CON, ZS2AAW
|
Email & BBS PC re-installed, Wifi antenna changed, HF Icom
re-installed
The Wifi Access Point 5GHz omni antenna was removed, extended, fitted with a heliax feeder and refitted atop the H-Frame by Nick, ZS2NT. Hoping this will improve AP signals from Slipper, which are poor everywhere. Also refitted was the BBS/email PC after running faultlessly at ZS2AAW's QTH for the last month. It was found that the PS2 mouse is u/s though, which makes navigation difficult... Audio line IN & OUT cables from the two sound cards were cabled to the conference rack, and for the first time, the 145.700 repeater audio is now available on Skype as ZS0NTP_PEARS. The Icom 718 Remote-controlled HF set was re-installed, a new feeder run up the tower for a 40m dipole, terminated on the feeder entry panel outside, but no connector fitted on the top end yet. The 40m antenna made by Ted, ZS2TED could not be installed due to a lack of centre mounting hardware. A fly lead from the feeder entry panel to the Icom also needs to be made up with N-type connector on the other end, it was cut to length on site in preparation. The Icom was powered up and connected to the control interface and to the PC. Audio cabling from the control box was now too short and must be remade. |
|
15 March |
Mount Kempt
Governorskop |
ZR2JH, ZS2AAW
ZR2JH, ZS2AAW |
New locks fitted Saltmine visit to site gave the chance to change all locks to the EC repeater standard, as well as fit one in the gate chain. No RF checks were done.
Mt Kempt Repeater cabinet...
...with new locks fitted
Cradock Link removed, Node PC removed The off-frequency UHF link to Cradock was removed on passing. Also retrieved was the Node PC that runs the digipeater, as the CPU fan was siezed and the PC shuts down after a while as a result. Site is in a poor condition, stay wires are broken, trees are tall, batteries need a top-up, and the door & frame needs painting. Tower needs a clean-up and the VHF co-linear needs replacing
Trees are taking over at Governorskop
|
|
19 February |
ZS2AAW QTH |
ZS2AAW |
Email & BBS PC revived Virus scans, Xoftspy scans revealed no serious threats. But the temp folder was filled with 65 500+ tmp files and when this happened, windows response was slowed thanks to 100% CPU usage. Clearing all tmp files cured the CPU load. But every 5 mins, 77 new files were generated. That means about 3 days before overflow again. Killing the email server program stopped the tmp files buildup. Reformatting the HDD is a LAST resort, thanks to the difficulty in loading WinFBB many years ago - let sleeping dogs lie! The PC temp folder has been manually cleared for the last few weeks, so it runs normally. Then ZS2GV complained that he was no longer receiving peham emails. Send and Receive crashed with an error. Checking the server showed three pages of emails waiting. The received emails on the server were sorted in date order and the oldest one for ZS2GV was deleted - it was very suspect, with no subject and no "from" address. Voila - the rest of ZS2GV's emails now downloaded fine. His outlook program was set to retrieve messages every 5 mins - and the server was generating those files during the retrieve process, which crashed due to the suspect file, thus not clearing the tmp files when done. The PC seems happy now, with only a handful of tmp files after a day's work. Ready to return to Slipper so that the Packet ports can go back on air... |
|
11
January |
Lady's
Slipper |
ZS2AE,
ZS2AAW |
BBS PC dead The PC running ZS0NTP and the PEARS wifi email server ".....@peham.co.za" went SK sometime on Wednesday evening. Remote power off-on resets did not bring it back to life. Boot failure with a black screen was found on arrival at site. The PC was removed for repair. Checks at home initially pointed towards the SATA HDD not responding. After booting from another HDD, the suspect HDD could be read. When returning to a single HDD, all worked normally, except for the NetMos PCI Serial card for COM3, COM4 that could not be revived. Reseating this card cured that problem. PC is now running on test at home. Standard padlocks were fitted to site, replacing the badly rusted shackles, and allowing the standard key access. |
