Are you sure the TVS' aren't in parallel? They typically shunt the excess
current. They are really one or two diodes back to back. Not necessarily the
same characteristics for both diodes in a bidirectional TVS when
manufactured.
N2TK, Tony
-----Original Message-----
From: TowerTalk <towertalk-bounces@contesting.com> On Behalf Of
wc1m73@gmail.com
Sent: Thursday, September 5, 2019 2:13 PM
To: john@kk9a.com; towertalk@contesting.com
Subject: Re: [TowerTalk] Grounding
Not thyristors. They use Transient Voltage Suppressor (TVS) diodes in series
with the motor control and pot lines. Now that you mention it, I may have
had to replace the TVS diodes, too, but I can't remember. I know I have some
on hand, as well as some replacement FETs in case it happens again. As I
recall, the TVS diodes and FET are mounted on the terminal strip in the
back, so easy to replace without removing any circuit boards.
I do remember telling Jeff about the strike and the symptoms, and he pointed
me right at those components. Cheap and easy to fix.
Evidently, the transient was too fast and/or below the TVS threshold (82V),
and got to the FET. As with MOVs, there's a tradeoff between the activation
threshold of the protection device, the operating voltage of the
semiconductor and how much back voltage it can withstand.
It's not clear to me why the GH RS-232 interface and CPU weren't damaged. As
I recall, they were the only devices connected to the computer where that
was the case. The schematic doesn't show enough detail to tell.
73, Dick WC1M
-----Original Message-----
From: john@kk9a.com <john@kk9a.com>
Sent: Wednesday, September 4, 2019 8:53 PM
To: towertalk@contesting.com
Subject: Re: [TowerTalk] Grounding
I thought that Green Heron has built in thyristors to protect the
controller's components.
John KK9a
Quoting wc1m73@gmail.com:
> I agree that simply disconnecting the coax isn't enough, and that a
> quality whole house surge protector is a good idea. It certainly can't
> hurt. But I highly recommend designing the station so that
> *all* wires between the tower(s)/antenna(s) and the building can be
> easily disconnected, with the possible exception of the ground cables
> (see discussion below.) Belt and suspenders are worth every penny.
>
> 12 years ago I had a lightning surge that caused over $10K in damage
> to $30K worth of equipment (had I not done all of the troubleshooting
> and repair myself, the bill would have been much
> higher.) Luckily, my home insurance covered it. I've posted
> extensively on TT about that event, so I won't repeat all of the gory
> details here, but wanted to point out that the strike did not hit the
> tower. It hit a tree several hundred feet from the tower, and the
> damage was caused by an induced surge. Damage on the tower was pretty
> minimal -- just the destruction of four relay diodes in the stack
> match and complete vaporization of a wide ground trace on its PCB (the
> only place where I found any scorch marks or physical evidence of the
> surge.) Other than that, none of the hardline, coax, cables or
> antenna switch relays were damaged. FWIW, the shields of all the
> hardline and coax were bonded to the tower at the top and bottom, and
> the relays were configured to be open rather than shorted when not
> selected. None of the 24 SteppIR motors and none of the four rotors on
> the tower were damaged. The rest of the damage was indoors, as
> described below.
>
> The tower that took the surge has a Ufer ground with three 50'
> radials of 1/0 buried stranded cable, each Cadwelded to four 8'
> ground rods spaced 16' apart (12 rods total). As it's 225' from the
> shack, there's no ground cable running between the tower and the SPG
> at the shack entrance (Polyphaser says that past 75' the grounds don't
> "see" each other.) That said, there are two runs of 1-5/8"
> heliax between the tower ground and shack SPG, and the huge corrugated
> copper shields probably have much lower inductance than a
> 1/0 ground wire. So if the grounds can see each other past 75', I
> think they're pretty-well bonded together. I have bulkhead Polyphaser
> coaxial surge suppressors for every RF cable mounted through the
> bottoms of metal utility cabinets at both ends (tower and SPG), as
> well as panel-mounted surge suppressors in the utility boxes with two
> MOVs and a fuse on both ends of each control wire.
> All shields and ground wires are grounded at both ends. The tower and
> SPG ground systems are bonded to the metal cabinets via 1/0 stranded
> wire. The SPG is also bonded to all the other ground systems that
> enter the house -- telco, cable, satellite, electric.
> Unfortunately, they external drops for those grounds are on the other
> side of the house, so I had to run the bonding cable from the utility
> cabinets directly a cross the basement ceiling to the utility entry
> panel. It wasn't clear to me that installing a couple hundred feet of
> ground wire around the house with 15-20 ground rods Cadwelded to it
> was a better way to bond the grounds. Open to discussion on that.
>
> My theory on how the destruction occurred is that the surge was
> handled by the coax surge protectors, and most of the MOVs, but the
> MOVs connected to my three SteppIR controllers were of too high a
> value to protect the delicate ICs. They were more suitable for
> protecting old-style rotor controllers, which didn't have any delicate
> ICs. As it was, the motor control FETs in my three Green Heron
> controllers were destroyed, but I believe the GH circuit design is
> such that the FETs acted as fuses to protect the CPUs and other
> electronics. Not so the SteppIR controllers. Their delicate SMD driver
> chips were fried, and the surge made it past them to the
> RS-232 transceiver interface. From there it got to the computer,
> destroyed the motherboard, CPU and multiport RS-232 card, and then
> propagated to every device connected to that card, including a Ten-Tec
> Orion (totaled), an Alpha 87A (RS232 IC fried and CPU partially
> damaged) an Acom 2000A (RS232 IC and CPU fried), a MicroHam StackMaX
> controller (totaled), a W5XD keyer (RS232 IC and CPU). The StackMax
> could have been damaged by the voltage threshold of the MOVs in the
> suppressors being rated too high. I also lost the
> RS-232 level converter between an FT-1000D and the computer, but there
> was no damage to the radio. Evidently, the level converter saved the
> radio.
>
> I also had damage to a very expensive high-end CD player, a C/Ku-band
> satellite dish LNB, a C/Ku-band 4DTV satellite receiver, a C/Ku-band
> satellite TV DVB PCI card in the computer, a Dish satellite receiver,
> a garage door opener, a phone line, a digital scale (totaled via RS232
> from the computer), and a ROKU media processor. I believe the damage
> to those devices came via the C/Ku-band satellite TV DVB PCI card in
> the aforementioned computer, which was on the same PCI bus as the
> RS-232 card. It's possible the surge was picked up by one or both
> satellite dishes, damaging the entertainment system devices and
> propagating to the computer via the DVB card, and to all the station
> devices, but given the damage and the destruction of the SteppIR and
> GH motor drivers, I think it was the other way around (or maybe both.)
> No power supplies were damaged, and no devices outside the shack and
> entertainment center were damaged, except the garage door opener and
> telco line. I've lost several garage door openers after lightning
> storms, and have read that they're quite susceptible to lightning (I
> suspect the long wires to the manual switches acts like an antennas.)
> The damage to the telco line was outside the house, so evidently it
> picked up the surge there and the indoor suppressors did their job.
>
> After repairing all the damage, I designed and installed a large patch
> panel so I could quickly disconnect the many cables between the tower
> and shack. It wasn't feasible to locate the panel outside, which would
> have been preferable, so I installed it in a basement closet. All the
> cables passing through the utility cabinets that are connected to the
> SPG enter the house just above the sill, easily routed down to the
> panel. The station side cables run from the patch panel up through the
> basement ceiling into the shack. There are about 20 or so cables on
> the patch panel, most of them multi-wire.
> It takes about five minutes to disconnect all of them. I also
> installed electrical disconnect panels so I can quickly cut off the
> shack circuits from the house AC. My two amps and a couple of power
> supplies, routers, etc. are also located in the closet, where I can
> quickly unplug them from the AC as well.
>
> I built a smaller panel on the other side of the wall behind our
> entertainment center so I could disconnect it from the rooftop TV/FM
> and satellite TV antennas, and there's also an AC disconnect box in
> there.
>
> Normally, I leave all the cables disconnected when the station isn't
> in use. I only disconnect AC and TV cables when we go on vacation or
> there's a really nasty storm in the area.
>
> Note that I replaced the MOVs in the surge suppressors connected to
> the SteppIR and (I think) GH controllers with MOVs that have as low a
> value as I could get without going below the operating voltage of the
> driver outputs.
>
> This setup has worked well for 12 years, but I hasten to point out
> that although we've had our share of lightning storms since then, our
> area isn't prone to them and we haven't seen anything near as nasty as
> the one that hit my station in 2007. So I'm not resting on my laurels
> and am always on the lookout for anything I can do to reduce the risk
> of lightning damage.
>
> [BTW, if TT was on groups.IO, as I've pleaded for, I would be able to
> attach photos of my patch panel, surge suppressors, tower, etc.
> And likely the text of my post would have been shorter, one picture
> being worth the proverbial 1,000 words!]
>
> Hope this is helpful to someone.
>
> 73, Dick WC1M
>
> -----Original Message-----
> From: john@kk9a.com <john@kk9a.com>
> Sent: Wednesday, September 4, 2019 11:45 AM
> To: towertalk@contesting.com
> Subject: Re: [TowerTalk] Grounding
>
> GL rebuilding Keith.
>
> A tower lightning strike will cause an induced voltage spike in a
> home's electrical system, even if nothing is connected to the tower.
> I have had this occur while building a new tower and the control
> cables and coax were not yet connected or even touching the tower.
> Simply disconnecting the coax which many hams do is ineffective
> lightning protection. There is a lot of valuable protection
> information on this list. Assuming that you have a bunch of ground
> rods and a SPG, I would suggest a good whole house surge protector.
>
> John KK9A
>
> from [Keith Dutson] NM5G
>
> My experience is that lightning can be managed, to a point. A direct
> strike to your tower can be catastrophic. I suffered one on May 9.
> My daughter lives about 600 feet away and witnessed the strike. She
> said sparks flew in all directions. My losses are huge. I am working
> with the insurance company, and expect the total damage to be well in
> excess of $50K. There is evidence that high voltage appeared on the
> tower, power lines, phone lines
> (fiber) and even on my wi-fi connection from the house to the shack.
> Every gigabit switch was toasted, and every PC connected was blown.
> Another big loss was two large screen TVs, both connected to internet
> via gigabit switches. Repairs were successful by replacing the power
> supplies.
>
> All this, plus my main two stations, both with Yaesu FTDX9000D and
> Alpha 87A, are gone. Coax was not even connected because I had just
> reconfigured the station and had not completed the coax switching
> section. However, units were plugged into power, and the Yaesu
> transceivers were connected to the PCs.
>
> 73, Keith NM5G
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|