Elecraft

Requirements for keying a KXPA100

Classic

List

Threaded

8 messages Options

John Marvin

Requirements for keying a KXPA100

So, I got my KXPA100 kit and have assembled it. But I'm having trouble
keying it with my transceiver (HPSDR Hermes with alterations to support
driving the KXPA100). My circuit didn't have any trouble keying my
previous amp (a HFPacker V4). If I manually short the KXPA100 key input
things appear to work in the limited testing I've done, so I don't think
there is anything wrong with the KXPA100.

I've measured 5V as expected at the key input, and I've measured the
current while connecting the KXPA100 key input to my transceiver. I'm
measuring 0.58 mA when the transceiver attempts to key the amp. The
KXPA100 manual says that the key current will be 1 mA max (if I am
reading that correctly) so this seems to be about right. Is there a
minimum current requirement? There is a 560 ohm resistor in series
internally, along with diodes (one in line with the key output to the
KXPA 100 and one inline with the key output to the internal tx preamp)
to prevent interaction between the two.

Any ideas? My next experiment will probably involve changing or
bypassing the 560 ohm resistor, but I have doubts that will fix the problem.

John
AC0ZG
______________________________________________________________
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

Lyle Johnson

Re: Requirements for keying a KXPA100

John,

The diode drop is likely the primary culprit. If you must use a series
diode, use a Shottky diode (like a 1N5711) and not a silicon diode (like
a 1N4148).

The key input goes to a microprocessor digital I/O pin inside the
KXPA100, and a diode drop - especially with a significant series
resistance -- puts the detected voltage in the indeterminate region.

73,

Lyle KK7P

> So, I got my KXPA100 kit and have assembled it. But I'm having trouble
> keying it with my transceiver ...There is a 560 ohm resistor in series
> internally, along with diodes...

______________________________________________________________
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

John Marvin

Re: Requirements for keying a KXPA100

Lyle,

Thanks. I would have thought that a transistor driver would have been
used to drive the I/O pin. Anyway, I measured the voltage drop across
the key input of the KXPA100 when the transceiver was attempting to key
the amp. The drop was 2.1 volts. Since the voltage for the circuit is
internal to the KXPA100, this is the voltage drop within my transceiver
(i.e. the switching diode, the 560 ohm current limiting resistor and the
keying circuit on the Hermes board), leaving 2.9 volts for the KXPA100.
I don't know what else is in the circuit inside the KXPA100, so I don't
know if the microprocessor is seeing that entire 2.9v. Do you know
whether or not that is the case?

I could get another .3 volts or so by switching to a schottky diode, and
possibly another .15 volts if I cut the resistor value in half (I'd like
to guarantee a maximum current of about 50ma (when keying a 12-13.8V
circuit), since the Hermes keying circuit is rated for up to 100ma).
That would give about 3.35 volts, which still is under the 3.5v normally
required to trigger a high state for cmos. That might work, but I'm not
sure I want to live with something that might be flaky.

I want to preserve the capability of triggering an amp with a relay coil
in the keying circuit, so I don't want to drop the diode. I fried the
tiny IC switch on the Hermes board once, and although I'm reasonably
competent when it comes to surface mount soldering, I don't relish doing
that repair again. I haven't had another failure (not entirely sure of
the cause of the first failure) since adding the additional protection
in the circuit (the diode, resistor and a bypass capacitor). I may have
to consider a different design where the Hermes keys a transistor switch
which then provides the path to ground for the tx preamp and the
external amp. That way the transistor can be my "fuse" and I can then
reduce the circuit elements in the keying path.

John
AC0ZG

On 1/14/2014 5:28 AM, Lyle Johnson wrote:

> John,
>
> The diode drop is likely the primary culprit. If you must use a
> series diode, use a Shottky diode (like a 1N5711) and not a silicon
> diode (like a 1N4148).
>
> The key input goes to a microprocessor digital I/O pin inside the
> KXPA100, and a diode drop - especially with a significant series
> resistance -- puts the detected voltage in the indeterminate region.
>
> 73,
>
> Lyle KK7P
>
>> So, I got my KXPA100 kit and have assembled it. But I'm having
>> trouble keying it with my transceiver ...There is a 560 ohm resistor
>> in series internally, along with diodes...
>

______________________________________________________________
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

John Marvin

Re: Requirements for keying a KXPA100

On 1/14/2014 2:08 PM, John Marvin wrote:

> Lyle,
>
> Thanks. I would have thought that a transistor driver would have been
> used to drive the I/O pin. Anyway, I measured the voltage drop across
> the key input of the KXPA100 when the transceiver was attempting to
> key the amp. The drop was 2.1 volts. Since the voltage for the circuit
> is internal to the KXPA100, this is the voltage drop within my
> transceiver (i.e. the switching diode, the 560 ohm current limiting
> resistor and the keying circuit on the Hermes board), leaving 2.9
> volts for the KXPA100. I don't know what else is in the circuit inside
> the KXPA100, so I don't know if the microprocessor is seeing that
> entire 2.9v. Do you know whether or not that is the case?
>
> I could get another .3 volts or so by switching to a schottky diode,
> and possibly another .15 volts if I cut the resistor value in half
> (I'd like to guarantee a maximum current of about 50ma (when keying a
> 12-13.8V circuit), since the Hermes keying circuit is rated for up to
> 100ma). That would give about 3.35 volts, which still is under the
> 3.5v normally required to trigger a high state for cmos. That might
> work, but I'm not sure I want to live with something that might be flaky.
>
> I want to preserve the capability of triggering an amp with a relay
> coil in the keying circuit, so I don't want to drop the diode. I fried
> the tiny IC switch on the Hermes board once, and although I'm
> reasonably competent when it comes to surface mount soldering, I don't
> relish doing that repair again. I haven't had another failure (not
> entirely sure of the cause of the first failure) since adding the
> additional protection in the circuit (the diode, resistor and a bypass
> capacitor). I may have to consider a different design where the Hermes
> keys a transistor switch which then provides the path to ground for
> the tx preamp and the external amp. That way the transistor can be my
> "fuse" and I can then reduce the circuit elements in the keying path.
>
> John
> AC0ZG
>
> On 1/14/2014 5:28 AM, Lyle Johnson wrote:
>> John,
>>
>> The diode drop is likely the primary culprit. If you must use a
>> series diode, use a Shottky diode (like a 1N5711) and not a silicon
>> diode (like a 1N4148).
>>
>> The key input goes to a microprocessor digital I/O pin inside the
>> KXPA100, and a diode drop - especially with a significant series
>> resistance -- puts the detected voltage in the indeterminate region.
>>
>> 73,
>>
>> Lyle KK7P
>>
>>> So, I got my KXPA100 kit and have assembled it. But I'm having
>>> trouble keying it with my transceiver ...There is a 560 ohm resistor
>>> in series internally, along with diodes...
>>
>

John Marvin

Re: Requirements for keying a KXPA100

In reply to this post by John Marvin

Oops, sorry for the prior mistaken posting.

I just realized I can try an experiment directly shorting the key input
of the KXPA100 through a variable resistor, in order to find what is
required to trigger it. I can then do the math and add another .1v
safety margin in order to determine whether I can go with a simple
change or something a little more involved.

John
AC0ZG

On 1/14/2014 2:08 PM, John Marvin wrote:

Don Wilhelm-4

Re: Requirements for keying a KXPA100

In reply to this post by John Marvin

John,

For reliable operation, the input to a CMOS device must see a low level
of 1.5 volts or less. Your 2.9 volts is in the indeterminate region.
As a side note, a high level should be at least 3.5 volts (up to 5
volts), but in this case, the high level is determined by the pullup
resistor in the amp.

You may want to redesign your keying interface to eliminate the diodes.
If I can suggest using a switching FET such as the 2N7000. If you need
isolation between 2 sinks, use 2 2N7000s to accomplish that isolation.
If you need more current or voltage handling than the 2N7000 will do,
look at a HEXFET.

73,
Don W3FPR

On 1/14/2014 4:08 PM, John Marvin wrote:

Don Wilhelm-4

Re: Requirements for keying a KXPA100

In reply to this post by John Marvin

John,

Rather than rely on the particular switching point of a sample of 1
KXPA100, why not measure the voltage across the variable resistor - when
it is 1.3 volts or lower (1.0 volt if you are adding the voltage drop of
a schottky diode), measure the value of the resistor - that is the
maximum resistance that can be in the circuit.

73,
Don W3FPR

On 1/14/2014 4:21 PM, John Marvin wrote:

> Oops, sorry for the prior mistaken posting.
>
> I just realized I can try an experiment directly shorting the key
> input of the KXPA100 through a variable resistor, in order to find
> what is required to trigger it. I can then do the math and add another
> .1v safety margin in order to determine whether I can go with a simple
> change or something a little more involved.
>
> John
> AC0ZG
>
> On 1/14/2014 2:08 PM, John Marvin wrote:
>> Lyle,
>>
>> Thanks. I would have thought that a transistor driver would have been
>> used to drive the I/O pin. Anyway, I measured the voltage drop across
>> the key input of the KXPA100 when the transceiver was attempting to
>> key the amp. The drop was 2.1 volts. Since the voltage for the
>> circuit is internal to the KXPA100, this is the voltage drop within
>> my transceiver (i.e. the switching diode, the 560 ohm current
>> limiting resistor and the keying circuit on the Hermes board),
>> leaving 2.9 volts for the KXPA100. I don't know what else is in the
>> circuit inside the KXPA100, so I don't know if the microprocessor is
>> seeing that entire 2.9v. Do you know whether or not that is the case?
>>
>> I could get another .3 volts or so by switching to a schottky diode,
>> and possibly another .15 volts if I cut the resistor value in half
>> (I'd like to guarantee a maximum current of about 50ma (when keying a
>> 12-13.8V circuit), since the Hermes keying circuit is rated for up to
>> 100ma). That would give about 3.35 volts, which still is under the
>> 3.5v normally required to trigger a high state for cmos. That might
>> work, but I'm not sure I want to live with something that might be
>> flaky.
>>
>> I want to preserve the capability of triggering an amp with a relay
>> coil in the keying circuit, so I don't want to drop the diode. I
>> fried the tiny IC switch on the Hermes board once, and although I'm
>> reasonably competent when it comes to surface mount soldering, I
>> don't relish doing that repair again. I haven't had another failure
>> (not entirely sure of the cause of the first failure) since adding
>> the additional protection in the circuit (the diode, resistor and a
>> bypass capacitor). I may have to consider a different design where
>> the Hermes keys a transistor switch which then provides the path to
>> ground for the tx preamp and the external amp. That way the
>> transistor can be my "fuse" and I can then reduce the circuit
>> elements in the keying path.
>>
>> John
>> AC0ZG
>
> ______________________________________________________________
> 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
>

John Marvin

Re: Requirements for keying a KXPA100

In reply to this post by John Marvin

I got a variety of suggestions, and decided to go the easiest route
(since it worked) of removing the current limiting resistor and
switching the diode to a schottky diode. The extra diode is required in
my case because the PTT also keys an internal TX preamp whose keying
circuit is run on 13.8V. Without the diode this circuit (unless I chose
another suggestion of creating two separate PTT drivers on the
transceiver) would see the floating 5V in the KXPA100 as a "relative"
ground, so just connecting the KXPA100 key line would have caused the
internal TX preamp to be keyed.

It was clear that a current limiting resistor was not required for
either the internal TX keying circuit or the KXPA100 (they key current
is about 1 mA for each of them). I decided if a current limiting
resistor is required for keying another amp I can build an external
adapter cable that includes the resistor for that use case.

With the schottky diode in place the Hermes transceiver pulls the
KXPA100 key line down to about 0.39 volts, which should be enough to
trigger it with hopefully a comfortable margin. Everything appears to be
working properly now, so now I get to play!

Anyway, thanks to all for the helpful suggestions.

73,

John
AC0ZG
______________________________________________________________
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