Interesting point about the 'speedup' circuit. I am using a vacuum relay in series with a
reed relay in a circuit similar to the Measures circuit. The voltage across the relays is
normal when measured in a steady state condition, but the open circuit voltage supplied
through a dropping resistor is higher. The idea is that the filter capacitor in the power
supply discharging through the dropping resistor and relay coils provides the 'goose', as
you call it. Now I'm thinking that this may not be such a good idea. I have used similar
circuits over the past 15 years or so and only once replaced a relay, due to the receive
contact becoming unreliable.
I admit to not worrying about the sequencing, depending on my K3, which provides a
reasonable delay of about 8 ms before and after producing RF.
For contests, I operate the K3's PTT from a Winkey device. This outputs CW directly from
the computer and generates a PTT signal that holds during CW generation but drops quite
quickly after the transmission finishes. That way I don't have to use the semi-QSK in the
transceiver, which will stay in transmit longer than needed. I save the full QSK for
pileups.
On 11/10/2013 6:58 AM, Jim Garland wrote:
Interesting discussion, re vacuum relay failure modes. I've zapped two RJ-1a
relays over the years, but both times it was my fault. The first time was in
the 1980s, when I used one as an HV interrupter in a power supply, not
realizing they're not designed for DC load switching. The second time was in
the 1990s, when I used one to switch an extra capacitor in parallel with
the tune cap on an 8877 amplifier, exceeding the relay's voltage rating.
The failure mode in each case was to weld the NO contacts together.
I've used RJ1a relalys in QSK circuits for many years, including a half
dozen contests each year, but nothing like the contest involvement some
others have reported. During the past year, I've gotten interested in QSK
circuits for a homebrew project and made many tests on a variety of vacuum
relays. Most types, I've found are not well-suited for QSK use, either
because they're too noisy, too slow or, more commonly, have excessive
contact bounce. For example, the Gigavac G41, used in the Alpha 9500, can
take 7-10 mS to settle down and can have twenty or more contact bounces when
the NO contacts make. The problem, I believe, is that the coil produces too
much magnetic force on the movable contact and really slams it into its
mating contact. If the bouncing happens while hot-switching RF, the relay
will have a very short life. So-called "speedup" circuits, that goose the
relay coil with a voltage pulse to accelarate the switching time, can really
exacerbate the contact bounce problem. As I mentioned before, the Jennings
RJ1a appears to be the best relay for QSK use at amateur power levels
(though I've not checked its Kilovac or Gigavac clones). The RJ1a is not
only very fast (1-2 mS closing, and 500uS opening times), but has absolutely
no contract bounce.
In researching the subject during the past year, I've looked at the relay
circuits of a number of commercial amplifiers. Most seem to treat T/R
switching pretty cavaliearly, without adequate protection for the relays. In
my own design (the circuit mounts on a 3"x5" PCB), I've tried to prevent any
situation that can damage the contacts. This involves several precautions:
(1) sequencing the amp so the output relay, bias switch, and input relay
operate sequentially in that order, after allowing for the relay closing
times and contact bounce. (2) Having an RF "lockout" sensor that prevents
the amplifier switching sequence from beginning if RF is present on the
input line. (3) Having an RF "lockon" sensor that keeps the relays closed so
long as RF is present. My new amplifier completes its T/R switching in 4 mS,
which is suitable for QSK speeds up to about 60 wpm, far beyond it's
builder's capabilities.
73,
Jim W8ZR
-----Original Message-----
From: Amps [mailto:amps-bounces@contesting.com] On Behalf Of Larry Benko
Sent: Saturday, November 09, 2013 11:38 PM
To: amps@contesting.com
Subject: Re: [Amps] Best source for Gigavac GH-1 ?
Jim,
What is failing on the relays? Is it the coil opening up, the contacts
sticking, the receive side contact appearing open, or the pivot getting
sloppy and the contacts bouncing too much? I have tested many relays
for mechanical life (low current in the contacts) in previous employment
and many exceed 20 million cycles. All the relays I have tested were
made by reputable manufacturers and were generally tested at about a
15Hz rate. In a well designed amp guaranteed to not hot switch driven
with the proper coil voltage the life should be very close to the
mechanical life.
There is a problem using a relay as a T-R device where the receive side
of the contact sees only very low current (gold plated or gold flashed
contacts are best) and the transmit side sees high current (definitely
not gold plated). The dilemma is what type of contact material to use.
Generally the high current side rules and the low current side may need
a bias current to be reliable. Often using 2 relays in parallel or
bifurcated contacts will give the desired reliability.
73,
Larry, W0QE
--
Vic, K2VCO
Fresno CA
http://www.qsl.net/k2vco/
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|