[Amps] Input relay for QSK amp

Jim Garland 4cx250b at miamioh.edu
Sun Dec 9 10:22:45 EST 2018 

Jim, You're right in principle, of course. The problem is that a 
well-designed amplifier relay circuit shouldn't depend on what 
transceiver is used. The amplifier itself should be immune to 
hot-switching in all situations.

Here's an extreme example. The Flex 6000 series transceivers have a 
menu-set delay time adjustable from 0 -1000 mSec, which is the time RF 
appears after the amp relay keying line is closed. The default setting 
for this delay is zero mSec, which means RF arrives at the amplifier at 
the exact same time the amp relay line is closed. A well-designed amp 
QSK circuit should ignore that RF and not key up the amplifier at all, 
until the Flex delay is adjusted to be long enough for the amp relays to 
accommodate it.

Similarly, if a key is closed, for example to allow tuning of the 
amplifier, and then the relay keying cable is accidentally pulled loose, 
then a proper QSK circuit will keep the amplifier on-line until the RF 
disappears, even though the relay cable is disconnected. Keeping the 
amplifier on-line so long as RF is present is a key requirement of a 
good QSK circuit

Furthermore, for high speed QSK CW operation, the amp has to be fast 
enough to transition from receive to transmit and back in a few mSec, 
all the time sequencing the relays properly to avoid hot switching. 
Doing that involves some circuit complexity that can't be avoided. So, 
yes, in principle all that's needed is to keep RF out of the amplifier 
until the amp is on-line and ready to receive it, and to hold the amp 
on-line until after the RF disappears. It's implementing that 
requirement for all possible situations that's the challenge. What on 
the surface seems simple, in practice actually involves a bit of circuit 
complexity in the QSK design.

73,

Jim W8ZR

On 12/9/2018 7:08 AM, Bob Maser wrote:
> I agree with everything you said.  The problem with some radios is 
> they start sending'
> RF down the line at the same time they are giving the relays the 
> energize signal.  I have built
> some big amplifiers in my day (don't forget the saying that in 
> California too much power is just
> enough) and I always made sure that the output relay was quicker than 
> the input/bias relay
> in fact I usually used open frame P&B for them and RJ2B on the 
> output.  RJ1's just don't
> instill confidence  HI.
> Bob
>
> On Sat, Dec 8, 2018 at 11:49 PM Jim <jimw7ry at gmail.com 
> <mailto:jimw7ry at gmail.com>> wrote:
>
>     Its actually very simple Jim.
>
>     Don't send RF until the amplifier is ready. Not doing so, makes the
>     relays hot switch. It matters not which one is energized before RF is
>     sent. As long as they both are before RF is sent from the exciter.
>     Same
>     on dekey. RF MUST stop flowing from the exciter BEFORE the relays
>     (either one of them) DE-energize.
>
>
>     Thanks
>     73
>     Jim W7RY
>
>     On 12/8/2018 11:37 AM, MU 4CX250B wrote:
>     > Bob,
>     > Years ago, I bought a lifetime supply of RJ1As for my own projects,
>     > and have tested roughly two dozen on my homebrew relay tester. The
>     > RJ1As typically close in about 2msec, with the fastest closing I
>     > measured at 1.2msec. Their opening time is typically about 500uSecs.
>     > None of them show any contact bounce. Other vacuum relays, e.g.,
>     RF1D,
>     > RJ1. GC1, etc., are more like you remember.
>     >
>     > Here's a link to a 2015 QSK article I wrote for QST,
>     > that goes into detail on amp relay timing issues, including the
>     > crucial need to avoid hot-switching. Doing so is more
>     complicated than
>     > merely choosing a slow input relay, or slowing down the input relay
>     > transistor driver with an RC circuit. Many purported "QSK" designs
>     > ignore these other considerations.
>     > 73,
>     > Jim w8zr
>     >
>     >   from my iPhone
>     >
>     >> On Dec 7, 2018, at 8:16 PM, Bob Maser <bob.maser at gmail.com
>     <mailto:bob.maser at gmail.com>> wrote:
>     >>
>     >> I must be old school.  The input relays need to be slower than
>     the output
>     >> relay.
>     >> Seems to me that a RJ-1 takes 8ms to close and stop bouncing
>     around. I have
>     >> always
>     >> used a P&B miniature open frame DPDT for that purpose.  The
>     output of the
>     >> amp needs
>     >> to be closed before the input turns on.
>     >> Bob
>     >>
>     >>> On Fri, Dec 7, 2018 at 9:39 PM Jim <jimw7ry at gmail.com
>     <mailto:jimw7ry at gmail.com>> wrote:
>     >>>
>     >>> Those relays that Rich (and I ) used are mostly unavailable. I
>     used to
>     >>> buy them 100s at a time. No more of them can be had.
>     >>>
>     >>> Most of what you find on eBay are seconds and most don't work.
>     Or they
>     >>> are re-labeled counterfeits.
>     >>>
>     >>>
>     >>> Thanks
>     >>> 73
>     >>> Jim W7RY
>     >>>
>     >>>> On 12/4/2018 4:03 PM, Alek Petkovic wrote:
>     >>>> G'day Vic.
>     >>>>
>     >>>> The Matsushita/Panasonic relay that Rich Measures used in his QSK
>     >>>> circuit, published at somis.org <http://somis.org>, is still
>     available on Ebay. I bought
>     >>>> 10 of them a couple of years ago
>     >>>>
>     >>>> Otherwise, Jim, W7RY, uses a small relay in his QSK kit. He
>     laso deals
>     >>>> on Ebay.
>     >>>>
>     >>>> 73, Alek.
>     >>>> VK6APK
>     >>>>
>     >>>>> On 4/12/2018 6:21 PM, Victor Rosenthal 4X6GP wrote:
>     >>>>> I'm trying to choose a relay to use as the input relay in a
>     QSK amp.
>     >>>>> I have been very satisfied with the Jennings RJ1-A as an output
>     >>>>> relay. The input relay in this circuit needs to be DPDT, or
>     I'd just
>     >>>>> use two similar vacuum relays.
>     >>>>>
>     >>>>> Here are my needs:
>     >>>>>
>     >>>>> 1. DPDT.
>     >>>>> 2. 12 or 24v coil.
>     >>>>> 3. Operate/release time, including bounce <= 4 ms. Faster is
>     better.
>     >>>>> 4. Should carry 100w RF in 50 ohm system. Not expected to
>     make/break
>     >>>>> live circuit.
>     >>>>> 5. Must have low resistance for low-level signals (receive
>     path).
>     >>>>> 6. Must have long life >= (10 x 10**6 operations)
>     >>>>> 7. Must not be insanely expensive, and available from Mouser
>     or RS.
>     >>>>>
>     >>>>> I found this relay:
>     >>>>>
>     <https://www.mouser.co.il/datasheet/2/418/NG_SS_108-98005_L-1019204.pdf
>     >>>>> But maybe there's a better choice? I'm overwhelmed by the
>     number and
>     >>>>> varieties of relays in the catalog.
>     >>> _______________________________________________
>     >>> Amps mailing list
>     >>> Amps at contesting.com <mailto:Amps at contesting.com>
>     >>> http://lists.contesting.com/mailman/listinfo/amps
>     >>>
>     >> _______________________________________________
>     >> Amps mailing list
>     >> Amps at contesting.com <mailto:Amps at contesting.com>
>     >> http://lists.contesting.com/mailman/listinfo/amps
>

More information about the Amps mailing list