[TenTec] Pegasus/Jupiter keying - In search of Perfect CW :-)
Robert & Linda McGraw K4TAX
RMcGraw@Blomand.Net
Sun, 03 Feb 2002 12:52:05 -0600
Lee:
Very good presentation! I'd like to see more high level, concise
presentations like this. Even I could understand what was measured and what
was observed.
73
Bob K4TAX
----- Original Message -----
From: "WA3FIY" <wa3fiy@radioadv.com>
To: <tentec@contesting.com>
Sent: Saturday, February 02, 2002 8:39 PM
Subject: [TenTec] Pegasus/Jupiter keying - In search of Perfect CW :-)
> There are [or have been] several interesting threads running which
> seem to be slowly converging into a good discussion of keying
> characteristics of Pegasus/Jupiter [PJ's for short], Omni VI+, Icom
> 756Pro and also some general comments regarding key clicks, etc.
>
> The thing I find interesting about the latest round of discussions is
> that we are beginning to take the subjectivity out of the discussion
> and are looking at some numbers that describe what we are hearing.
>
> So in that spirit, I decided to run some tests on my PJ's much like
> George, W5YR has been running on his Pro. In short, the PJ's look
> pretty good. Perfect? No, but pretty good when held up in
> comparison to the rest of the pack.
>
> The test set-up consisted of my Jupiter running in both native
> Jupiter mode and in Pegasus emulation mode. The Jupiter was
> connected to a dummy load with a -20 db coupler to my triggered
> oscilloscope. I operated the Jupiter with the internal keyer enabled
> and had a Radio Adventures Co. CodeBoy keyer connected to the PTT
> line on the accessory connector. The 'scope was being triggered by
> the keyer output when I used the external keyer and by the RF
> envelope when operating the internal keyer.
>
> I am not done with the tests but here is what I have found so far.
>
> Sidetone - Sidetone output timing from the PJ's EXACTLY tracks the
> RF envelope, even to the point of having a smooth rise and fall time at
> make and break. So, you don't really need to connect into the rf
> output to know what your keying waveshape looks like. Just connect
> into the line out and monitor the sidetone to see what the keying
> timing is like. AND, you do not need any off air monitor to hear what
> your signal is doing. Just listen to the sidetone.
>
> Rise and Fall times - The rise and fall times are 3mS on all but the
> first element in a transmission. The first element has a 2 mS rise
> time on the first element and a 3 ms fall time. This finding exactly
> duplicates the photos in the QST reviews as best I can tell from the
> pictures. I was a little confused by this. This is QSK.......why would
> the first element be different from any other? And, what is
> considered a "first element"? I think I have the answer which I will
> get into a little later.
>
> Keying spike - Although difficult to see in the QST photos, the PJ's
> do have a VERY SMALL keying spike on the first element. The first
> element has about a 2% voltage overshoot which is a 4% power
> overshoot! The percentage of overshoot remains constant over the
> full power output range. So at 10 watts out, the overshoot reaches
> 10.4 watts and at 100 watts output, the overshoot hits 104 watts. I
> cannot imagine this small amount causing any amplifier to trip or do
> anything else nasty. Subsequent elements have a barely detectable
> overshoot which is so small I can't measure it with any accuracy. I
> believe the overshoot and the faster rise time of the first element
> are related and I will get to that matter later as mentioned above.
>
> Weighting - Here is where it gets a little complicated because of the
> flexibility of the Ten Tec and N4PY control software user
> programming and my ignorance. I am not exactly clear how to set
> the user selectable weighting to "nominal" values, meaning values that
> would yield the 1:1 and 3:1 dit and dah to space ratios. In fact, I got
> bogged down on these tests and moved on to testing with an external
> keyer. Basically though, the Jupiter was shortening the dits a
> noticeable amount at very high speeds, speeds in the 60+ WPM range.
> It also shortened the dits at slower speeds but, although
> measurable, I could not detect it by casual listening at speeds below
> 30 WPM. I need to learn more about the setup though before I can
> draw any meaningful conclusion other than, at the speed you usually
> operate, set the weighting to a value that sounds good in the sidetone
> and your signal will sound the same on the air.
>
> External keyer - I confirmed something which may already be
> broadly known here, that is, you can key the PJ's nicely by keying
> the PTT line. I connected my CodeBoy both the official Ten Tec way
> thru the key jack and the unofficial way thru the PTT line on the
> accessory connector, and all measurements were IDENTICAL! That
> is a nice undocumented feature. The CodeBoy has exactly 1:1 and
> 3:1 dit and dah to space ratios in it's default mode and that is what I
> started testing with. When keying the PJ's, I noted a couple things.
>
> First, the delay between key closure and RF [and sidetone] output is
> 20mS at all speeds. That can be seen in the QST photos. [I just
> realized that I did not test for delay when operating split.....I'll do
> that and report later]. BTW, this has nothing to do with keyer
> weighting. This 20mS is the time required to shift internal circuitry,
> most especially the PLL, from receive to transmit condition and is not
> user adjustable.
>
> Secondly, the dit to space ratio drops below 1:1 by a greater degree
> as the speed increases, indicating that there are some uncompensated
> timing issues in the firmware of the PJ's. However, the CodeBoy has
> a "space" trimmer pot which when properly adjusted yielded a 1:1
> ratio over the full speed range of the CodeBoy which is 10 to 40
> WPM. In uncompensated form, the dah's became a little light at high
> speeds but after the space adjustment, the dah's were again at 3:1.
> At this point, using the CodeBoy, I have the recommended 1:1 and 3:1
> ratios over the full speed range of the CodeBoy [which BTW exceeds
> my speed range :-) ] and I have sidetone reflecting the transmitted
> code elements exactly and I have Ten Tec QSK about which I need
> say no more because it is hard to improve upon and I used too many
> ands in this sentence. Sorry! :-)
>
> Now.............about the rise time and "spike" such as it is, on the
first
> element. First of all, what is a "first" element? Since this is a full
> QSK rig are not all elements a "first" element? NO, they are not!
> The first element is the element sent after the rig has been at rest
> for some period of time, maybe in the neighborhood of a second or
> two. Ten Tec would know better than I what that time is. It is the
> time in which the ALC circuit timing resets after sending elements. I
> believe this explains the different rise time and the spike.
>
> IMHO, one of the outstanding features in several generations of Ten
> Tec rigs is the ALC stabilized output at ALL power levels. This
> feature is present in the PJ's and I hope TT includes it in the new
> rigs coming. BUT, there is one VERY MINOR drawback.......and that is
> the effect on the first element. Let me repeat, the effect on the
> first element is measurable but NEGLIGIBLE! It is a fly speck!
>
> Here is what I think is happening. After resting, the ALC circuit is
> reset. When the first key closure comes along, the transmitter is
> running "open loop" with respect to ALC stabilized power output. TT
> has very nicely built in a controlled rise time in the open loop mode so
> key clicks are minimized. As the power is rising, it eventually hits
> ALC threshold at which point the loop closes and stabilizes output
> power at the set level. That transition from detection of ALC
> threshold to stabilized output takes a finite amount of time and it is
> during that time that the overshoot, the very small overshoot, occurs.
> Once the ALC loop is closed, the transmitter system gain is reduced
> by some apparently small amount and subsequent rise times are a
> little longer as a result of the reduced system gain. This attention to
> small detail is not found on may rigs I have owned and that have been
> discussed here. A tip of the hat to Ten Tec as a friend of mine
> would say. Can this be improved on? Sure.......at a cost. Does it
> need to be improved on? In my opinion, not at all!
>
> The rise and fall times of 3mS seem to be a good compromise too.
> Short enough to create distinct elements to cut through QRN and
> QRM but long enough to minimize key clicks, hence being a good
> neighbor.
>
> The PJ's are nice CW radios!
>
> ********REQUIRED DISCLOSURE************
>
> I do not have a financial interest in Ten Tec except by virtue of
> being a customer of both ham radio gear and OEM enclosures for the
> CodeBoy and DigitalDial which my company Radio Adventures Co.
> manufactures. Obviously I do have a financial interest in the
> CodeBoy, a nice little keyer. :-)
>
> **********PERSONAL NOTE*************
>
> I commented early in this post about taking subjectivity out of our
> discussions of some of these matters. I would like to qualify that
> statement somewhat. Subjectivity is certainly important because it
> pretty much tells us what we like and don't like about our gear. It
> really doesn't matter much what the numbers say if I don't like this
> thing! If I don't like it......I don't like it. Or.........If it like
it, so
> what about the numbers. It is just that sometimes I think we go
> rambling on talking about our subjective likes and dislikes almost as
> though they were statistical, mathematical facts, when in fact, they
> are simply likes and dislikes. So please understand that I know the
> importance of subjective opinion but also I know the importance of
> statistical fact. Both have their place in our discussions and I hope
> more of us will take the time to dig through the rhetoric and find the
> facts.
>
> This is a great group. Keep up the lively discussions on all the
> important matters like the dots, etc. :-) Oh, my Pegasus is black
> and my Jupiter is red and my .................................... oh
well, not now.
> maybe another time.
>
> -73-
>
> -Lee-
> _______________________________________________
> TenTec mailing list
> TenTec@contesting.com
> http://lists.contesting.com/mailman/listinfo/tentec