|
In preparing to upgrade the GPS RF distribution system at my AI6VN/KH6 Maui site next month, I have been staging the new system here in California and thought it might be of interest to others. I purchased a 'bullet' GPS antenna for $50 from ebay which claims 40 LNA gain and adapt the N connector on it to a F connector which is compatible with the 100' run of RG-6  On my desk the RG6 goes into a Bias-T which gets 5VDC from the USB-A jack on one of the Kiwis. The Kiwi supplies 3 VDC Bias-T which is dropped to 2.4VDC by the steering diodes in the 8 way GPS distribution splitter. That 2.4V is not enough to power that bullet antenna, which seems very happy with the 5V from the Kiwi USB port  Even with the 11+ dB of RF signal loss in the splitter, the Kiwi is tracking 10-12 satellites at most times  Of course this configuration introduces a ground loop between the USB ground and the GPS RF SMA input of the Kiwi. I won't know how much RFI is introduced by that loop until I get to Maui when I'll add a report of the results in that installation.
|
|
|
I have had a similar system in place for a few years. Those 40 dB
bullets are getting harder to find on the cheap. I also found that
some GPS splitters nurture CMI and other noise
-Jim
On 8/20/22 10:58, Rob Robinett wrote:
toggle quoted message
Show quoted text
In preparing to upgrade the GPS RF distribution system at my
AI6VN/KH6 Maui site next month, I have been staging the new system
here in California and thought it might be of interest to others.
I purchased a 'bullet' GPS antenna for $50 from ebay which claims
40 LNA gain and adapt the N connector on it to a F connector which
is compatible with the 100' run of RG-6
On my desk the RG6 goes into a Bias-T which gets 5VDC from the
USB-A jack on one of the Kiwis. The Kiwi supplies 3 VDC Bias-T
which is dropped to 2.4VDC by the steering diodes in the 8 way GPS
distribution splitter. That 2.4V is not enough to power that
bullet antenna, which seems very happy with the 5V from the Kiwi
USB port
Even with the 11+ dB of RF signal loss in the splitter, the Kiwi
is tracking 10-12 satellites at most times
Of course this configuration introduces a ground loop between the
USB ground and the GPS RF SMA input of the Kiwi.
I won't know how much RFI is introduced by that loop until I get
to Maui when I'll add a report of the results in that
installation.
|
|
|
If anyone wants to build their own up to 8-way GPS distribution board having high isolation at HF, one minimizing CM currents among GPS devices, there is design information at https://www.dropbox.com/sh/qn5041sj70xnnl6/AADFTN-I4arbYaIe5KhQAHPXa?dl=0The board design ActiveSplit9brd.brd, can be dropped on oshpark.com website and the board ordered. Parts required are in the associated csv list. It is all SMD and some of the parts are pretty small so you need to be comfortable with reflow soldering on a hot plate or similar technique. Some of these are in use at various wsprdaemon sites. Glenn
On 8/20/22 10:58, Jim Lill wrote: I have had a similar system in place for a few years. Those 40 dB bullets are getting harder to find on the cheap. I also found that some GPS splitters nurture CMI and other noise
|
|
|
I can attest to the usefulness and quality of Glenn's GPS distribution board, one of which he contributed to the KPH site. But building a surface mount PCB is currently not part of my skillset. So the configuration I describe above is very much a second best solution which has the advantage of not requiring any soldering skills. Jim and Glenn are of course correct that my configuration is vulnerable to ground loop induced RFI in much the same way an LNA ahead of the Kiwi's RF input can also introduce RFI. So check the 0-30 Mhz spectrum of your installation before and after attaching the GPS RF input to your Kiwi to see if there is significant RFI introduced when the GPS RF is connected.
On Sat, Aug 20, 2022 at 10:07 AM Glenn Elmore < n6gn@...> wrote: If anyone wants to build their own up to 8-way GPS distribution board
having high isolation at HF, one minimizing CM currents among GPS
devices, there is design information at
https://www.dropbox.com/sh/qn5041sj70xnnl6/AADFTN-I4arbYaIe5KhQAHPXa?dl=0
The board design ActiveSplit9brd.brd, can be dropped on oshpark.com
website and the board ordered. Parts required are in the associated
csv list.
It is all SMD and some of the parts are pretty small so you need to be
comfortable with reflow soldering on a hot plate or similar technique.
Some of these are in use at various wsprdaemon sites.
Glenn
On 8/20/22 10:58, Jim Lill wrote:
> I have had a similar system in place for a few years. Those 40 dB
> bullets are getting harder to find on the cheap. I also found that
> some GPS splitters nurture CMI and other noise
-- Rob Robinett AI6VN mobile: +1 650 218 8896
|
|
|
Bruce KX4AZ
When I check my Kiwis are typically using 4-6 satellites for a fix. I am curious if there is a measurable improvement in the frequency stability or accuracy on the KiwiSDR when increasing the number of "good" satellites being used for the GPS? Or is it more of a insurance policy to assure there will always be some minimum number (3?) of "good" satellites available?
|
|
|
From my experience so far, I do not see a significant improvement
from 'good enough' to, say,a 10 satellite fix. Neither provides
enough short term stability to reduce spreading enough to remove
the Kiwi from the total 'spreading budget' on 20m and mid-long
FST4W modes. True GPS phaselock that provides external clocking
can do this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
toggle quoted message
Show quoted text
When I check my Kiwis are typically using 4-6 satellites for a
fix. I am curious if there is a measurable improvement in the
frequency stability or accuracy on the KiwiSDR when increasing the
number of "good" satellites being used for the GPS? Or is it more
of a insurance policy to assure there will always be some minimum
number (3?) of "good" satellites available?
|
|
|
As Glenn says, even locked on 12 satellites the Kiwi's internal frequency accuracy in nowhere go enough for FST4W-300/900/1800 reception on the HF bands. However at KPH when I went from an average of 4 satellites to 10-12 the spot reports went from +- 10 Hz to +- 1 Hz on 20M So I am not investing much time or effort in the GPS RF distribution system as I move on to modify my Kiwis to accept an external clock from a Bodar.
On Sat, Aug 20, 2022 at 11:51 AM Glenn Elmore < n6gn@...> wrote:
From my experience so far, I do not see a significant improvement
from 'good enough' to, say,a 10 satellite fix. Neither provides
enough short term stability to reduce spreading enough to remove
the Kiwi from the total 'spreading budget' on 20m and mid-long
FST4W modes. True GPS phaselock that provides external clocking
can do this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
When I check my Kiwis are typically using 4-6 satellites for a
fix. I am curious if there is a measurable improvement in the
frequency stability or accuracy on the KiwiSDR when increasing the
number of "good" satellites being used for the GPS? Or is it more
of a insurance policy to assure there will always be some minimum
number (3?) of "good" satellites available?
-- Rob Robinett AI6VN mobile: +1 650 218 8896
|
|
|
KD2OM
Won’t there be the same issue of CM when using a bodner as there would be with a gps splitter?
On Aug 20, 2022, at 15:49, Rob Robinett <rob@...> wrote:
As Glenn says, even locked on 12 satellites the Kiwi's internal frequency accuracy in nowhere go enough for FST4W-300/900/1800 reception on the HF bands. However at KPH when I went from an average of 4 satellites to 10-12 the spot reports went from +- 10 Hz to +- 1 Hz on 20M So I am not investing much time or effort in the GPS RF distribution system as I move on to modify my Kiwis to accept an external clock from a Bodar. On Sat, Aug 20, 2022 at 11:51 AM Glenn Elmore < n6gn@...> wrote:
From my experience so far, I do not see a significant improvement
from 'good enough' to, say,a 10 satellite fix. Neither provides
enough short term stability to reduce spreading enough to remove
the Kiwi from the total 'spreading budget' on 20m and mid-long
FST4W modes. True GPS phaselock that provides external clocking
can do this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
When I check my Kiwis are typically using 4-6 satellites for a
fix. I am curious if there is a measurable improvement in the
frequency stability or accuracy on the KiwiSDR when increasing the
number of "good" satellites being used for the GPS? Or is it more
of a insurance policy to assure there will always be some minimum
number (3?) of "good" satellites available?
-- Rob Robinett AI6VN mobile: +1 650 218 8896
|
|
|
I currently see no evidence of CM introduced by the bodnar on my gpsdo kiwi.
I’m also surprised in no not fixable added noise because I am powering it over usb to the pc in the shack.
toggle quoted message
Show quoted text
On Aug 20, 2022, at 4:20 PM, KD2OM <steve@...> wrote:
Won’t there be the same issue of CM when using a bodner as there would be with a gps splitter?
Steve
On Aug 20, 2022, at 15:49, Rob Robinett <rob@...> wrote:
As Glenn says, even locked on 12 satellites the Kiwi's internal frequency accuracy in nowhere go enough for FST4W-300/900/1800 reception on the HF bands.
However at KPH when I went from an average of 4 satellites to 10-12 the spot reports went from +- 10 Hz to +- 1 Hz on 20M
So I am not investing much time or effort in the GPS RF distribution system as I move on to modify my Kiwis to accept an external clock from a Bodar.
On Sat, Aug 20, 2022 at 11:51 AM Glenn Elmore < n6gn@...> wrote:
From my experience so far, I do not see a significant improvement from 'good enough' to, say,a 10 satellite fix. Neither provides enough short term stability to reduce spreading enough to remove the Kiwi from the total 'spreading budget' on 20m and mid-long
FST4W modes. True GPS phaselock that provides external clocking can do this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
When I check my Kiwis are typically using 4-6 satellites for a fix. I am curious if there is a measurable improvement in the frequency stability or accuracy on the KiwiSDR when increasing the number of "good" satellites being used for
the GPS? Or is it more of a insurance policy to assure there will always be some minimum number (3?) of "good" satellites available?
--
Rob Robinett
AI6VN
mobile: +1 650 218 8896
|
|
|
No noticeable added noise*
toggle quoted message
Show quoted text
On Aug 20, 2022, at 4:29 PM, WA2TP - Tom <myis300@...> wrote:
I currently see no evidence of CM introduced by the bodnar on my gpsdo kiwi.
I’m also surprised in no not fixable added noise because I am powering it over usb to the pc in the shack.
On Aug 20, 2022, at 4:20 PM, KD2OM <steve@...> wrote:
Won’t there be the same issue of CM when using a bodner as there would be with a gps splitter?
Steve
On Aug 20, 2022, at 15:49, Rob Robinett <rob@...> wrote:
As Glenn says, even locked on 12 satellites the Kiwi's internal frequency accuracy in nowhere go enough for FST4W-300/900/1800 reception on the HF bands.
However at KPH when I went from an average of 4 satellites to 10-12 the spot reports went from +- 10 Hz to +- 1 Hz on 20M
So I am not investing much time or effort in the GPS RF distribution system as I move on to modify my Kiwis to accept an external clock from a Bodar.
On Sat, Aug 20, 2022 at 11:51 AM Glenn Elmore < n6gn@...> wrote:
From my experience so far, I do not see a significant improvement from 'good enough' to, say,a 10 satellite fix. Neither provides enough short term stability to reduce spreading enough to remove the Kiwi from the total 'spreading budget' on 20m and mid-long
FST4W modes. True GPS phaselock that provides external clocking can do this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
When I check my Kiwis are typically using 4-6 satellites for a fix. I am curious if there is a measurable improvement in the frequency stability or accuracy on the KiwiSDR when increasing the number of "good" satellites being used for
the GPS? Or is it more of a insurance policy to assure there will always be some minimum number (3?) of "good" satellites available?
--
Rob Robinett
AI6VN
mobile: +1 650 218 8896
|
|
|
KD2OM
Are you feeding multiple kiwis with one bodner. I believe that would be the same as one gps feeding several receivers.
On Aug 20, 2022, at 16:31, WA2TP - Tom <myis300@...> wrote:
No noticeable added noise*
On Aug 20, 2022, at 4:29 PM, WA2TP - Tom <myis300@...> wrote:
I currently see no evidence of CM introduced by the bodnar on my gpsdo kiwi.
I’m also surprised in no not fixable added noise because I am powering it over usb to the pc in the shack.
On Aug 20, 2022, at 4:20 PM, KD2OM <steve@...> wrote:
Won’t there be the same issue of CM when using a bodner as there would be with a gps splitter?
Steve
On Aug 20, 2022, at 15:49, Rob Robinett <rob@...> wrote:
As Glenn says, even locked on 12 satellites the Kiwi's internal frequency accuracy in nowhere go enough for FST4W-300/900/1800 reception on the HF bands.
However at KPH when I went from an average of 4 satellites to 10-12 the spot reports went from +- 10 Hz to +- 1 Hz on 20M
So I am not investing much time or effort in the GPS RF distribution system as I move on to modify my Kiwis to accept an external clock from a Bodar.
On Sat, Aug 20, 2022 at 11:51 AM Glenn Elmore < n6gn@...> wrote:
From my experience so far, I do not see a significant improvement from 'good enough' to, say,a 10 satellite fix. Neither provides enough short term stability to reduce spreading enough to remove the Kiwi from the total 'spreading budget' on 20m and mid-long
FST4W modes. True GPS phaselock that provides external clocking can do this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
When I check my Kiwis are typically using 4-6 satellites for a fix. I am curious if there is a measurable improvement in the frequency stability or accuracy on the KiwiSDR when increasing the number of "good" satellites being used for
the GPS? Or is it more of a insurance policy to assure there will always be some minimum number (3?) of "good" satellites available?
--
Rob Robinett
AI6VN
mobile: +1 650 218 8896
|
|
|
Only 1 at this time. I have the two channel Bodnar. I am wondering if each Chanel has isolation.
toggle quoted message
Show quoted text
On Aug 20, 2022, at 4:57 PM, KD2OM <steve@...> wrote:
Are you feeding multiple kiwis with one bodner. I believe that would be the same as one gps feeding several receivers.
On Aug 20, 2022, at 16:31, WA2TP - Tom <myis300@...> wrote:
No noticeable added noise*
On Aug 20, 2022, at 4:29 PM, WA2TP - Tom <myis300@...> wrote:
I currently see no evidence of CM introduced by the bodnar on my gpsdo kiwi.
I’m also surprised in no not fixable added noise because I am powering it over usb to the pc in the shack.
On Aug 20, 2022, at 4:20 PM, KD2OM <steve@...> wrote:
Won’t there be the same issue of CM when using a bodner as there would be with a gps splitter?
Steve
On Aug 20, 2022, at 15:49, Rob Robinett <rob@...> wrote:
As Glenn says, even locked on 12 satellites the Kiwi's internal frequency accuracy in nowhere go enough for FST4W-300/900/1800 reception on the HF bands.
However at KPH when I went from an average of 4 satellites to 10-12 the spot reports went from +- 10 Hz to +- 1 Hz on 20M
So I am not investing much time or effort in the GPS RF distribution system as I move on to modify my Kiwis to accept an external clock from a Bodar.
On Sat, Aug 20, 2022 at 11:51 AM Glenn Elmore < n6gn@...> wrote:
From my experience so far, I do not see a significant improvement from 'good enough' to, say,a 10 satellite fix. Neither provides enough short term stability to reduce spreading enough to remove the Kiwi from the total 'spreading budget' on 20m and mid-long
FST4W modes. True GPS phaselock that provides external clocking can do this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
When I check my Kiwis are typically using 4-6 satellites for a fix. I am curious if there is a measurable improvement in the frequency stability or accuracy on the KiwiSDR when increasing the number of "good" satellites being used for
the GPS? Or is it more of a insurance policy to assure there will always be some minimum number (3?) of "good" satellites available?
--
Rob Robinett
AI6VN
mobile: +1 650 218 8896
|
|
|
Not quite but perhaps a risk.
CM noise current traversing the length of the KiwiSDR PCB goes
from the LAN/PS end of things, through the Z of the ground plane
and out the GPS/Antenna end. It's only about 4 milliohms real but
much more that due to the length/inductance reactance. I measure
about -80 dB coupling for conversion on that path to unwanted
differential preamp input. The Antenna and GPS grounds aren't
identical but their paths are similar. BTW, the wire input
terminal is significantly different/worse than the path to the SMA
ground and I avoid using it entirely. Generally I unsolder the
connector and throw it away.
When I modify a Kiwi for external clocking I run the cable from a
connector mounted on the LAN-end panel to the connection roughly
midway on the PCB. This path doesn't develop nearly as much IZ
drop as one going all the way to the SMA/GPS end through the
region of the preamp input and the sensitive area for
CM/differential conversion. I haven't carefully measured
conversion from CM to differential (Kiwi response) on this path
but it hasn't been noticeable.
On a related note, the SMA/GPS end of the PCB is "hot" not only
this way but as a source at GPS-frequencies of QRM. I think JKS
mentioned this somewhere or other in his design notes. Replacing
the enclosure Al end plate with a transmission line open quarter
wave helps this. It's important to keep the GPS antenna well away
from this region - 10-20 feet even - to avoid self-QRM.
Connecting the Antenna and GPS SMAs directly to ground is a bad
idea since it only improves the longitudinal path through the
board such that CM LAN and PS currents do more harm. It's for this
reason that I generally insert a high CM impedance, e.g. a T1-1 as
a flux coupled transformer with only 1 pF of inter-winding
capacitance. This produces much higher CM impedance than the best
of broadband ferrite chokes.
On 8/20/22 14:20, KD2OM wrote:
toggle quoted message
Show quoted text
Won’t there be the same issue of CM when using a bodner as there
would be with a gps splitter?
Steve
On Aug 20, 2022, at 15:49, Rob
Robinett <rob@...> wrote:
As Glenn
says, even locked on 12 satellites the Kiwi's internal
frequency accuracy in nowhere go enough for
FST4W-300/900/1800 reception on the HF bands.
However
at KPH when I went from an average of 4 satellites to
10-12 the spot reports went from +- 10 Hz to +- 1 Hz on
20M
So I am
not investing much time or effort in the GPS RF
distribution system as I move on to modify my Kiwis to
accept an external clock from a Bodar.
On Sat, Aug 20, 2022 at
11:51 AM Glenn Elmore < n6gn@...>
wrote:
From my experience so far, I do not see a
significant improvement from 'good enough' to, say,a
10 satellite fix. Neither provides enough short
term stability to reduce spreading enough to remove
the Kiwi from the total 'spreading budget' on 20m
and mid-long FST4W modes. True GPS phaselock that
provides external clocking can do this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
When I check my Kiwis are
typically using 4-6 satellites for a fix. I am
curious if there is a measurable improvement in the
frequency stability or accuracy on the KiwiSDR when
increasing the number of "good" satellites being
used for the GPS? Or is it more of a insurance
policy to assure there will always be some minimum
number (3?) of "good" satellites available?
--
Rob Robinett
AI6VN
mobile: +1 650 218 8896
|
|
|
My cluster of 4X Kiwi-BBAI are in a custom arrangement where the
SMA connectors are not grounded to the chassis. That seems to
mitigate the CMI. The type of GPS distro system can do terrible
things however. I found that the Time Machine splitter are awful
for example.
Currently all my RX, 2 ea on 2 antenna, suffer from external mix
products from the strong MWBC TX. Finding the mix mechanism,
diodes if you will, is a challenge. Fixing them may be impossible.
WRT to strong MWBC, both K2ZN and myself have both found that the
Mini-Circuits T1:1 saturate with the MW and other signals we have.
I switched to HB 1:1 that can handle 20W
-Jim
WA2ZKD
On 8/20/22 17:11, Glenn Elmore wrote:
toggle quoted message
Show quoted text
Not quite but perhaps a risk.
CM noise current traversing the length of the KiwiSDR PCB goes
from the LAN/PS end of things, through the Z of the ground plane
and out the GPS/Antenna end. It's only about 4 milliohms real
but much more that due to the length/inductance reactance. I
measure about -80 dB coupling for conversion on that path to
unwanted differential preamp input. The Antenna and GPS grounds
aren't identical but their paths are similar. BTW, the wire
input terminal is significantly different/worse than the path to
the SMA ground and I avoid using it entirely. Generally I
unsolder the connector and throw it away.
When I modify a Kiwi for external clocking I run the cable from
a connector mounted on the LAN-end panel to the connection
roughly midway on the PCB. This path doesn't develop nearly as
much IZ drop as one going all the way to the SMA/GPS end
through the region of the preamp input and the sensitive area
for CM/differential conversion. I haven't carefully measured
conversion from CM to differential (Kiwi response) on this path
but it hasn't been noticeable.
On a related note, the SMA/GPS end of the PCB is "hot" not only
this way but as a source at GPS-frequencies of QRM. I think JKS
mentioned this somewhere or other in his design notes. Replacing
the enclosure Al end plate with a transmission line open quarter
wave helps this. It's important to keep the GPS antenna well
away from this region - 10-20 feet even - to avoid self-QRM.
Connecting the Antenna and GPS SMAs directly to ground is a bad
idea since it only improves the longitudinal path through the
board such that CM LAN and PS currents do more harm. It's for
this reason that I generally insert a high CM impedance, e.g. a
T1-1 as a flux coupled transformer with only 1 pF of
inter-winding capacitance. This produces much higher CM
impedance than the best of broadband ferrite chokes.
On 8/20/22 14:20, KD2OM wrote:
Won’t there be the same issue of CM when using a bodner as there
would be with a gps splitter?
Steve
On Aug 20, 2022, at 15:49, Rob
Robinett <rob@...>
wrote:
As
Glenn says, even locked on 12 satellites the Kiwi's
internal frequency accuracy in nowhere go enough for
FST4W-300/900/1800 reception on the HF bands.
However
at KPH when I went from an average of 4 satellites to
10-12 the spot reports went from +- 10 Hz to +- 1 Hz
on 20M
So I am
not investing much time or effort in the GPS RF
distribution system as I move on to modify my Kiwis
to accept an external clock from a Bodar.
On Sat, Aug 20, 2022
at 11:51 AM Glenn Elmore < n6gn@...>
wrote:
From my experience so far, I do not see a
significant improvement from 'good enough' to,
say,a 10 satellite fix. Neither provides enough
short term stability to reduce spreading enough to
remove the Kiwi from the total 'spreading budget'
on 20m and mid-long FST4W modes. True GPS
phaselock that provides external clocking can do
this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
When I check my Kiwis are
typically using 4-6 satellites for a fix. I am
curious if there is a measurable improvement in
the frequency stability or accuracy on the KiwiSDR
when increasing the number of "good" satellites
being used for the GPS? Or is it more of a
insurance policy to assure there will always be
some minimum number (3?) of "good" satellites
available?
--
Rob Robinett
AI6VN
mobile: +1 650 218 8896
|
|
|
The additional paths created by connecting multiple GPS grounds
together from multiple devices and device types definitely
complicates a multi-Kiwi station. It's the reason I built the
distribution board.
That's a good caution about transformer distortion. It likely
takes a big signal but a preamp ahead of it compounds the problem.
I have not seen T1-1s do that here but it is certainly something
to look out for. If one has two large signals that are suspect for
generating IMD, looking for spurious products at (2Fa +- Fb) is
probably a good idea. Particularly with large antennas, a lot of
signal may be present.
I have a related problem with strong FM BCB signals putting
distortion products into HF from the remote on the hilltop. FM
signals at the active antenna run in the -10 dBm region some times
and can seriously degrade WSPR spots on the lower HF bands. These
can't be filtered because of the extremely high input impedance of
the active antenna.
Maybe not an issue for most of us unless using one for a notch
but crystals and crystal filters can also generate distortion.
Of course the LTC6432 preamp in the Kiwi isn't perfect either. It
has pretty good IMD specs but that is another area for concern
when signals get really big.
The proof of the pudding for WSPR in all this is to know the
transfer between the radiation resistance of your antenna and the
Kiwi to determine the system noise floor. If it's below ITU
numbers then, at least for WSPR, there isn't a problem.
On 8/21/22 05:17, Jim Lill wrote:
toggle quoted message
Show quoted text
My cluster of 4X Kiwi-BBAI are in a custom arrangement where
the SMA connectors are not grounded to the chassis. That seems
to mitigate the CMI. The type of GPS distro system can do
terrible things however. I found that the Time Machine splitter
are awful for example.
Currently all my RX, 2 ea on 2 antenna, suffer from external
mix products from the strong MWBC TX. Finding the mix mechanism,
diodes if you will, is a challenge. Fixing them may be
impossible.
WRT to strong MWBC, both K2ZN and myself have both found that
the Mini-Circuits T1:1 saturate with the MW and other signals we
have. I switched to HB 1:1 that can handle 20W
-Jim
WA2ZKD
On 8/20/22 17:11, Glenn Elmore wrote:
Not quite but perhaps a risk.
CM noise current traversing the length of the KiwiSDR PCB
goes from the LAN/PS end of things, through the Z of the
ground plane and out the GPS/Antenna end. It's only about 4
milliohms real but much more that due to the length/inductance
reactance. I measure about -80 dB coupling for conversion on
that path to unwanted differential preamp input. The Antenna
and GPS grounds aren't identical but their paths are
similar. BTW, the wire input terminal is significantly
different/worse than the path to the SMA ground and I avoid
using it entirely. Generally I unsolder the connector and
throw it away.
When I modify a Kiwi for external clocking I run the cable
from a connector mounted on the LAN-end panel to the
connection roughly midway on the PCB. This path doesn't
develop nearly as much IZ drop as one going all the way to the
SMA/GPS end through the region of the preamp input and the
sensitive area for CM/differential conversion. I haven't
carefully measured conversion from CM to differential (Kiwi
response) on this path but it hasn't been noticeable.
On a related note, the SMA/GPS end of the PCB is "hot" not
only this way but as a source at GPS-frequencies of QRM. I
think JKS mentioned this somewhere or other in his design
notes. Replacing the enclosure Al end plate with a
transmission line open quarter wave helps this. It's
important to keep the GPS antenna well away from this region -
10-20 feet even - to avoid self-QRM.
Connecting the Antenna and GPS SMAs directly to ground is a
bad idea since it only improves the longitudinal path through
the board such that CM LAN and PS currents do more harm. It's
for this reason that I generally insert a high CM impedance,
e.g. a T1-1 as a flux coupled transformer with only 1 pF of
inter-winding capacitance. This produces much higher CM
impedance than the best of broadband ferrite chokes.
On 8/20/22 14:20, KD2OM wrote:
Won’t there be the same issue of CM when using a bodner as
there would be with a gps splitter?
Steve
On Aug 20, 2022, at 15:49, Rob
Robinett <rob@...>
wrote:
As
Glenn says, even locked on 12 satellites the Kiwi's
internal frequency accuracy in nowhere go enough for
FST4W-300/900/1800 reception on the HF bands.
However
at KPH when I went from an average of 4 satellites
to 10-12 the spot reports went from +- 10 Hz to +- 1
Hz on 20M
So I
am not investing much time or effort in the GPS RF
distribution system as I move on to modify my Kiwis
to accept an external clock from a Bodar.
On Sat, Aug 20, 2022
at 11:51 AM Glenn Elmore < n6gn@...>
wrote:
From my experience so far, I do not see a
significant improvement from 'good enough' to,
say,a 10 satellite fix. Neither provides
enough short term stability to reduce spreading
enough to remove the Kiwi from the total
'spreading budget' on 20m and mid-long FST4W
modes. True GPS phaselock that provides external
clocking can do this.
On 8/20/22 12:37, Bruce KX4AZ wrote:
When I check my Kiwis are
typically using 4-6 satellites for a fix. I am
curious if there is a measurable improvement in
the frequency stability or accuracy on the
KiwiSDR when increasing the number of "good"
satellites being used for the GPS? Or is it
more of a insurance policy to assure there will
always be some minimum number (3?) of "good"
satellites available?
--
Rob Robinett
AI6VN
mobile: +1 650 218 8896
|
|
|
