Cable length for KPA/KAT100 remote; importance of KPA100 upgrade

[Chris Kimball]

Just bought a fine used remote KPA100/KAT100 in an EC2 cabinet.  It's neat, but I'd like to put it under the operating desk, rather than beside the K2.  Can the K2 serial bus and the other 3 signals be extended for a total of 6 ft?  Any problem remoting the RF by the same 6 ft?

Alas, the KPA100 is of very early production, 2002.  How urgent is it to apply the level Revision A, 2006 upgrade?  That is, what's the probability of significant smoke production in reasonable (CW) use without first applying the upgrade?

Thanks,

Chris
WB4WZR

[Don Wilhelm-4]

Chris,

Why not try locating the KPA100/KAT100 at 6 cable feet from the base K2
and see what happens.  There is no real information available on that
subject that I am aware of.
There are cautions available - those cautions say that there *may* be
timing problems on the Auxbus signalling - if an option does not respond
in time, the K2 will conclude that option either is not present or is
not properly responding (Error 080).

The upgrade to the KPA100 is highly recommended.  However (and I am
going far out on a limb here), the upgrade helps to solve coupling
problems between the base K2 and the KPA100, and those conditions may
not be present with a remotely mounted KPA100.  If you detect any sign
of instability for the KPA100, please install the KPA100UPKT.

73,
Don W3FPR

On 2/22/2012 8:33 PM, Chris Kimball wrote:

> Just bought a fine used remote KPA100/KAT100 in an EC2 cabinet.  It's neat,
> but I'd like to put it under the operating desk, rather than beside the K2.
> Can the K2 serial bus and the other 3 signals be extended for a total of 6
> ft?  Any problem remoting the RF by the same 6 ft?
>
> Alas, the KPA100 is of very early production, 2002.  How urgent is it to
> apply the level Revision A, 2006 upgrade?  That is, what's the probability
> of significant smoke production in reasonable (CW) use without first
> applying the upgrade?
>
> Thanks,
>
> Chris
> WB4WZR
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Chris Kimball]

Just made up a new cable of length approximately 15 ft. (That's the length that was available.)  It's foil-shielded, 4 conductor cable probably equal or equivalent to that provided by Elecraft.

The KPA-100 / KAT-100 combination appear to be working normally.  Is this a new world record for internal bus length?  Or does the smoke appear later?

Chris
NQ8Z

[Don Wilhelm-4]

Chris,

I don't think you will get any smoke, but if you get "strange
happenings" - like the KPA100 is not recognized, or doesn't give a
response back to the K2 in sufficient time (command flow over the
AUXBUS) and things get "mixed up" or you get ERR 080, the cure will be
to shorten the cable.

73,
Don W3FPR

On 3/23/2012 2:49 PM, Chris Kimball wrote:
> Just made up a new cable of length approximately 15 ft. (That's the length
> that was available.)  It's foil-shielded, 4 conductor cable probably equal
> or equivalent to that provided by Elecraft.
>
> The KPA-100 / KAT-100 combination appear to be working normally.  Is this a
> new world record for internal bus length?  Or does the smoke appear later?
>
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jim Brown-10]

On 3/23/2012 12:21 PM, Don Wilhelm wrote:
> I don't think you will get any smoke, but if you get "strange
> happenings" - like the KPA100 is not recognized, or doesn't give a
> response back to the K2 in sufficient time (command flow over the
> AUXBUS) and things get "mixed up" or you get ERR 080, the cure will be
> to shorten the cable.

I would not expect timing problems, but parallel wire cables are recipe
for noise and RFI problems, and making them longer makes that worse. I
don't understand why, but Elecraft seems still not to have gotten the
message that TWISTING is FAR more important than shielding in providing
immunity from RFI and noise.

There are simple instructions on my website for using CAT5 cable for
this sort of application.  
http://audiosystemsgroup.com/HamInterfacing.pdf   The specific wiring
instructions are given for RS232 connections, and your cable is NOT, but
if you will be in good shape if you follow the conceptual instructions
of using one twisted pair for each signal circuit with the "colored"
conductor used as the numbered pin and all four "white/color striped"
conductors twisted together and tied to the connector shell.  If you
have shielded CAT5, tie those four striped whites to the designated
return pin and tie the shield to the connector shells.

73, Jim Brown K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Don Wilhelm-4]

Jim,

While you may be correct about the noise pickup, there is also concern
about timing issues - if only because the performance has not been
tested at longer lengths by observing the bit timings on the AUUXBUS.

Some time back I had discussions with Jack Brindle about about extending
the cable length to the KRC2.  He did indicate that there could be
timing issues on the AUXBUS.  Being a single wire interface, when a
command is sent, the sending device waits for the expected response, and
if it is not received in time, an error condition will be declared.
Those are the facts as I know them, and I  recall that 6 feet was
considered an OK length because it had been tested - longer lengths had
not been tested.
Of course, this applies to the K2.  The K3 may or may not be different.

73,
Don W3FPR

On 3/23/2012 6:04 PM, Jim Brown wrote:

> On 3/23/2012 12:21 PM, Don Wilhelm wrote:
>> I don't think you will get any smoke, but if you get "strange
>> happenings" - like the KPA100 is not recognized, or doesn't give a
>> response back to the K2 in sufficient time (command flow over the
>> AUXBUS) and things get "mixed up" or you get ERR 080, the cure will be
>> to shorten the cable.
> I would not expect timing problems, but parallel wire cables are recipe
> for noise and RFI problems, and making them longer makes that worse. I
> don't understand why, but Elecraft seems still not to have gotten the
> message that TWISTING is FAR more important than shielding in providing
> immunity from RFI and noise.
>
> There are simple instructions on my website for using CAT5 cable for
> this sort of application.
> http://audiosystemsgroup.com/HamInterfacing.pdf   The specific wiring
> instructions are given for RS232 connections, and your cable is NOT, but
> if you will be in good shape if you follow the conceptual instructions
> of using one twisted pair for each signal circuit with the "colored"
> conductor used as the numbered pin and all four "white/color striped"
> conductors twisted together and tied to the connector shell.  If you
> have shielded CAT5, tie those four striped whites to the designated
> return pin and tie the shield to the connector shells.
>
> 73, Jim Brown K9YC
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jim Brown-10]

On 3/23/2012 3:24 PM, Don Wilhelm wrote:
> I  recall that 6 feet was
> considered an OK length because it had been tested - longer lengths had
> not been tested.
> Of course, this applies to the K2.  The K3 may or may not be different.

Thanks Don.  At the very least, if this is a critical issue it ought to
be tested and the results noted.  There's also the issue of cable
capacitance, which can be a big deal for high speed lines that are not
impedance-matched (like RS232).  CAT5 has a big advantage there too --
it's pretty low capacitance as cables go. Perhaps Jack can tell us if
that circuit is impedance matched (that is, treated as a transmission
line).

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Don Wilhelm-4]

Jim.

Jack Brindle and I just had an off-list email exchange in which he did
state that the conductor capacitance and resulting waveform distortion
may be a factor.

I really don't know the relative conductor to shield capacitance of the
4 conductor cable may be, but common sense tells me that it is variable
depending on which conductor you select - at least your implementation
of CAT5 twisted pair provides consistency.

As far as your "transmission line" suggestions go, I am going to go out
on a limb and say that YES, the AUXBUS signalling is a transmission line
when external to the K2 (even though it is uncontrolled).  The fact is
that it "floats around" inside the K2 to all the option processors and
the KIO2, it is not treated as a transmission line, and level sensing at
the receivers is sufficient.  When one goes external with that signal,
the conditions change - to get the signal from "here to there", it must
appear on some sort of transmission line (the cables in the Elecraft
KRC2, XVxxx transceivers, KPA100 for some examples) do constitute a
transmission line whether that is intended or not - but bottom line, the
receiver is an input to a CMOS firmware chip, and that input is level
sensitive.
While the drivers and receivers may not have been designed for a
transmission line ( I believe they were not), they are acting as drivers
and receivers on whatever transmission line is presented between them.  
Be careful - amateurs often think of transmission lines as coax and
parallel lines to antennas, but the world of transmission lines is much
greater then that.

Given the characteristics of transmission lines regarding distortion of
digital square wave signals, the receiver "interpretation"
characteristics become important.  I do not see exotic line drivers and
line receivers in any Elecraft devices, and so the obvious conclusion is
that the "receivers" are level driven.  If the data pulse is distorted,
the timing may be off, and if the levels are influenced by line
capacitance, switching may not occur at all.

In other words, Jack Brindle has tested the K2 AUXBUS up to 6 feet with
the standard 4 conductor cable, and all seems to be in order.  For those
who would want to go beyond that length, a lot of careful observation
for potential problem situations may be required - in other words, go
ahead and try it - it may work in your setup. but I cannot provide any
assurance that it will work at that length for all situations.  I am am
a retired IBM Assurance engineer whose business was testing that a
product met specifications under all possible conditions.  So I can say
that it might work for you, but that is no proof that it will work in
another controlled setting unless you provide detailed data (well beyond
"it works") to me and the general Elecraft community.

I have not personally tested the command/response timing, but I do
reflect Jack's comments which I do consider as regsarding "good
engineering practice".

73,
Don W3FPR


On 3/23/2012 8:01 PM, Jim Brown wrote:

> On 3/23/2012 3:24 PM, Don Wilhelm wrote:
>> I  recall that 6 feet was
>> considered an OK length because it had been tested - longer lengths had
>> not been tested.
>> Of course, this applies to the K2.  The K3 may or may not be different.
> Thanks Don.  At the very least, if this is a critical issue it ought to
> be tested and the results noted.  There's also the issue of cable
> capacitance, which can be a big deal for high speed lines that are not
> impedance-matched (like RS232).  CAT5 has a big advantage there too --
> it's pretty low capacitance as cables go. Perhaps Jack can tell us if
> that circuit is impedance matched (that is, treated as a transmission
> line).
>
> 73, Jim K9YC
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Jim Brown-10]

On 3/23/2012 6:17 PM, Don Wilhelm wrote:
> I really don't know the relative conductor to shield apacitance of the 4 conductor cable may be,

That's something I had to learn about from a practical, hands-on point
of view for my work in pro audio, and also as a member of the AES
Standards Committee.

> but common sense tells me that it is variable depending on which conductor you select - at least your implementation of CAT5 twisted pair provides consistency.

In the world of pro audio, for a 10 year period beginning about 15 years
ago we used RS232 a lot to communicate with DSP used in sound systems,
and a common connection was for a computer in the audience connected to
a DSP unit in an equipment rack buried in the bowels of the building. We
learned by thoughtful experimentation that we could reliably do that
over 200 ft or more of cable if it was low capacitance cable like CAT5,
but were limited to far less distance with ordinary cables. The
capacitance of CAT5 is on the order of 12 pF/ft, while the capacitance
of ordinary twisted pair is on the order of 35 pF/ft, and it's even
greater for commonly used cables.

Cable capacitance can be quite important, even with baseband audio,
where harmonic components coming out of a mic picking up musical
instruments can extend to 50 kHz and above, and an 38,400 Hz RS232
circuit will have components well above that..Some output stages can
ring and even go unstable when driving highly capacitive loads.

BTW -- if you dig up a recent copy of the standard that defines RS232,
it talks about cable length limitations in terms of cable capacitance,
not length. In simple terms, it boils down to waveform distortion based
on capacitive loading of the output stage and high frequency attenuation.

> As far as your "transmission line" suggestions go

What I'm getting at here is that to treat it as a transmission line it
simply needs to be approximately matched at each end, and the output
stage needs to be capable of driving the impedance it sees. CAT5 (and
digital audio cables) are about 100 ohms; most practical jacketed
cables, twisted or not, are on the order of 70-80 ohms.

With un-matched circuits like RS232, there are three key issues beyond
timing -- capacitance, loss, and noise. Another important benefit of
CAT5/6 cables is that their HF attenuation is a lot less than ordinary
cables.  The method I described for wiring CAT5 at each end helps
prevent the Pin One Problem from coupling noise, and also minimizes IR
drop in the return conductor that can be present if the units being
connected are powered or grounded at different points.

My comment that Elecraft ought to be testing longer cable lengths is
based on the expectation that, especially with the KX3, some users will
want to separate their KX3 from their power amp and antenna tuner by
sticking the KX3 on the dash and the amp/tuner in the trunk. I could see
cable lengths on the order of 15 ft in a typical car, longer in a few.

I'm with you on specifications too, and simply saying that the bus will
work for 6 ft is hardly a specification, or even an adequate
description. What is needed is a more thorough technical description of
the bus from a SYSTEM point of view, so that a technically educated user
can look at it, understand its limitations, and understand how to
implement an optimum interconnection when he needs to approach or
stretch those limits.

73, Jim K9YC
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html