New NTBW50AA
Moved antenna to fence line on a 10' emt. Nothing much seems to change no matter what I do, so I guess this is what I'm locked into unless I start getting carried away. I still get generally about a 400 ppt swing with the OSC, about 20 ns swing with the PPS and the other things are just what they are. Different filters and surveys don't seem to change anything that I've seen. I like the antenna better at the fence. but I had to add another 30' of RG-6. I have about 100' of FSJ4-50B that I have no use for right now, but I doubt that will do anything for me. http://i251.photobucket.com/albums/gg287/DogTi/time/20139-181346_zps31fb89be.jpg
Is this acceptable? From what I deduce, I can expect to get accuracy of +-0.2 Hz at 1GHz? I was hoping for another decimal. This should be plenty good to get mHz accuracy at 100 MHz if I can even measure that using SpecLab. Seems to me Arthur was getting much smaller swings on his ppt with his older Nortel.
I did figure out the temp thing and found a link where it's explained, but I have a new chip coming and expect to change it, but I doubt if anything will change except the reported temp will most likely track to less than 1° C unlike it's doing now.
Would it be worth looking into the MV89A OCXO as a replacement to get better ppt? I haven't done any looking into my OCXO yet to see what it's spec sheet may be.
Still waiting on a chip inductor that I messed up in the TS-2000 so I can get the XRef working.
Dave
N3DT
I still get generally about a 400 ppt swing with
the OSC, about 20 ns swing with the PPS and the
other things are just what they are. Different
filters and surveys don't seem to change anything that I've seen.
* * *
Is this acceptable?
You seem to have a reasonably decent view of the
sky, judging by the numbers of satellites you are
able to track, but the signal quality is lower
than I expect (note: my only Trimble GPSDO
experience is with TBolts). I'd expect the
signal reports to be mostly from 40 dBc upwards,
while yours run from low 30s to barely 40 (IIRC,
this hasn't changed materially as you've moved
the antenna multiple times). Your results look
consistent with those signal readings. I don't
recall -- have you tried a different
antenna? It's also possible that the front end
in that particular GPS is noisier than the TBolts I'm familiar with.
Also, I see that the oscillator loop gain is set
for 1.4 Hz/v. I have no idea what is normal for
the OCXOs used in that unit. Have you verified
this yourself? (Turn disciplining off, then
manually run the DAC voltage up and down while
watching the frequency on a counter.) Lowering
the GPSDO loop gain may improve matters. Or
not. (IIRC, this may mean setting "gain" to a
higher number -- the "gain" does not SET the loop
gain at that value, it tells the digital loop
"expect the oscillator to have this EFC characteristic.")
I did figure out the temp thing and found a link
where it's explained, but I have a new chip
coming and expect to change it, but I doubt if
anything will change except the reported temp
will most likely track to less than 1° C unlike it's doing now.
That would be my expectation, too.
Would it be worth looking into the MV89A OCXO as
a replacement to get better ppt? I haven't done
any looking into my OCXO yet to see what it's spec sheet may be.
That depends on whether or not the remaining
errors you are seeing are attributable to the
oscillator. The remaining errors can be classed
broadly as (i) noisy/wandering OCXO, and (ii)
adjustments to the DAC that the GPS thinks it
should make. In your case, they look more like
class (ii) than class (i) to me. Before you
switch out the OXCO, let the crystal that is in
the unit now settle down for a few months of
uninterrupted running. The class (i) errors
should go down by themselves. If the errors you
see do not go down, it is more evidence that your
limitation is noise in the GPS, not the OCXO.
Best regards,
Charles
Thanks Charles,
I have tried the &a command and that is where LH put the numbers, TC from 100 to 500, Damp from 1.2 (that's how it came) to 1.0, and it changed the gain from 1.2 to 1.403. I could certainly try playing with the DAC voltage. I haven't written anything to the prom yet, I think, when I turn the unit off and back on it seems to revert to where it was when I got it, but that's only been a few times and for a few minutes, it's been on nearly since I got the antenna installed on Sep 9, or 9 days. I haven't done a precision survey at this location either, but it didn't seem to make any difference at the other locations.
As far as dBc, I've seen numbers in the 50's at this location and I don't think I've seen that before, but nothing scientific about any of the locations or numbers. This location sees to the east and west probably better than the others, south should be about the same. Actually right now I'm seeing 2 in the 60's and one at 71.5.
I wasn't going to change the OCXO willy-nilly, I'll have to explore this one first. I'll see if I can find any info on this OCXO.
The antenna is a Lucent PCTEL "GPS-TMG-HR-26N, High Rejection 26dB With Enhanced Narrow Band Filtering", apparently for high RF environments, only one I've got or tried, it came with the Nortel from RDR. About 130' of RG-6 and with the 26dB gain and a 4 dB noise figure, it's probably swamping any loss in the cable or NF of the RX.
You didn't say, but I get the feeling maybe I should be getting smaller ppt numbers? When you say noisy, is high ppt generally a result of excess noise (front end NF)?
When I get a chance I'll play with the DAC voltage.
Dave
Hi
- Have you done a 3 day long survey at any of the locations? You need to do that to get the errors down.
- Fiddling with the gain and filter settings on a NTBW50 isn't useful. The unit does not work the same way as a TBolt. Just leave them alone.
- Because of the way the uint works there is no way to translate the LH display to frequency stability.
- The OCXO in that unit is at least as good as a typical MV89. Since you have no way to "tune" the unit, replacing the OCXO is not a good idea. Also you may have to buy (and run long term ADEV tests on) a dozen MV89's before you get a "good" one.
If you want to check the stability of the GSPDO, get a cheap Rb. That's the only way to know what's going on. There is no display / measurement / technique that will answer the question looking at the GPSDO alone.
Bob
On Sep 18, 2013, at 10:11 AM, quartz55 quartz55@hughes.net wrote:
Moved antenna to fence line on a 10' emt. Nothing much seems to change no matter what I do, so I guess this is what I'm locked into unless I start getting carried away. I still get generally about a 400 ppt swing with the OSC, about 20 ns swing with the PPS and the other things are just what they are. Different filters and surveys don't seem to change anything that I've seen. I like the antenna better at the fence. but I had to add another 30' of RG-6. I have about 100' of FSJ4-50B that I have no use for right now, but I doubt that will do anything for me. http://i251.photobucket.com/albums/gg287/DogTi/time/20139-181346_zps31fb89be.jpg
Is this acceptable? From what I deduce, I can expect to get accuracy of +-0.2 Hz at 1GHz? I was hoping for another decimal. This should be plenty good to get mHz accuracy at 100 MHz if I can even measure that using SpecLab. Seems to me Arthur was getting much smaller swings on his ppt with his older Nortel.
I did figure out the temp thing and found a link where it's explained, but I have a new chip coming and expect to change it, but I doubt if anything will change except the reported temp will most likely track to less than 1° C unlike it's doing now.
Would it be worth looking into the MV89A OCXO as a replacement to get better ppt? I haven't done any looking into my OCXO yet to see what it's spec sheet may be.
Still waiting on a chip inductor that I messed up in the TS-2000 so I can get the XRef working.
Dave
N3DT
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Dave wrote:
You didn't say, but I get the feeling maybe I should be getting
smaller ppt numbers? When you say noisy, is high ppt generally a
result of excess noise (front end NF)?
Hard to say about the ppt numbers -- my TBolts give somewhat lower
numbers than you are getting, but they use different OCXOs with
different phaselocking circuitry, so I can't say what would be
considered normal for the model you have. Also, as several have said
already, the GPSDO is measuring itself with itself (no independent
reference), so its proclamations need to be taken with more than some
caution.
Excess noise on the GPS signals (from whatever cause) causes timing
errors in the GPS, which makes the GPS think the oscillator needs
correction, which leads to DAC adjustments that aren't
necessary. But if you are seeing at least some numbers in the 50 dBc
and higer range, there is nothing further you can do in this
regard. (Well, if the model you have supports it, you can set a
"signal level" mask that will exclude satellites with signals below a
dBc threshold of your choice. As with the elevation mask, if you get
too aggressive you may force the unit into holdover when there are a
few OK but not great signals.)
As we all have said repeatedly, you need a good, accurate
survey. Also, it is the nature of crystal oscillators that when they
are disturbed (turned off and back on, frequency adjusted,
handling/moving, fast temperature changes, etc., etc.) that they take
time to settle back down -- often months. Now that you have a
permanent antenna location, do a precision survey, SAVE IT, then let
the unit sit undisturbed for several months and see where you are.
You are certainly in the ballpark -- there are clearly no gross problems.
Bob presumably knows more than I do about non-TBolt Trimbles (since I
know nothing about them). If he says you can't adjust the TC and
"gain" I'd be inclined to believe him. (And if you can't adjust the
"gain," there is no point in measuring the EFC characteristics of the
OCXO, except for curiosity.)
The antenna is a Lucent PCTEL "GPS-TMG-HR-26N, High Rejection 26dB
With Enhanced Narrow Band Filtering", apparently for high RF
environments, only one I've got or tried
You may have read on another thread about narrow bandpass filters
tending to have rising group delay at the passband edges. An antenna
with "Enhanced Narrow Band Filtering" very likely has more group
delay error than one with more gentle filtering, so it is possible
that satellites with high doppler shifts are causing the GPS timing
solution to shift around more than it would with a different
antenna. The one you have is presumably good enough for cell timing,
since that is what it was designed for. The question is whether the
lower group delay of an antenna without "Enhanced Narrow Band
Filtering" would be enough better that you could tell. Without
measuring the filters, we can't really predict what to expect. You
might ask the seller if he would exchange your antenna for one that
doesn't have the extra filtering. You almost certainly don't need it
in your rural location, and an antenna without it would at least not
be any worse, all else equal.
Best regards,
Charles
HI
What can be said with reasonable confidence:
- The unit isn't broken and it does not need to go back to RDR under their 30 day return policy.
- The OCXO isn't broken and it does not need to be replaced.
- The unit is functioning ok, but probably not as good as it could.
- Without comparison to an independent standard, there's no way to know how well it is or isn't doing.
Bob
On Sep 18, 2013, at 10:44 PM, Charles Steinmetz csteinmetz@yandex.com wrote:
Dave wrote:
You didn't say, but I get the feeling maybe I should be getting smaller ppt numbers? When you say noisy, is high ppt generally a result of excess noise (front end NF)?
Hard to say about the ppt numbers -- my TBolts give somewhat lower numbers than you are getting, but they use different OCXOs with different phaselocking circuitry, so I can't say what would be considered normal for the model you have. Also, as several have said already, the GPSDO is measuring itself with itself (no independent reference), so its proclamations need to be taken with more than some caution.
Excess noise on the GPS signals (from whatever cause) causes timing errors in the GPS, which makes the GPS think the oscillator needs correction, which leads to DAC adjustments that aren't necessary. But if you are seeing at least some numbers in the 50 dBc and higer range, there is nothing further you can do in this regard. (Well, if the model you have supports it, you can set a "signal level" mask that will exclude satellites with signals below a dBc threshold of your choice. As with the elevation mask, if you get too aggressive you may force the unit into holdover when there are a few OK but not great signals.)
As we all have said repeatedly, you need a good, accurate survey. Also, it is the nature of crystal oscillators that when they are disturbed (turned off and back on, frequency adjusted, handling/moving, fast temperature changes, etc., etc.) that they take time to settle back down -- often months. Now that you have a permanent antenna location, do a precision survey, SAVE IT, then let the unit sit undisturbed for several months and see where you are.
You are certainly in the ballpark -- there are clearly no gross problems.
Bob presumably knows more than I do about non-TBolt Trimbles (since I know nothing about them). If he says you can't adjust the TC and "gain" I'd be inclined to believe him. (And if you can't adjust the "gain," there is no point in measuring the EFC characteristics of the OCXO, except for curiosity.)
The antenna is a Lucent PCTEL "GPS-TMG-HR-26N, High Rejection 26dB With Enhanced Narrow Band Filtering", apparently for high RF environments, only one I've got or tried
You may have read on another thread about narrow bandpass filters tending to have rising group delay at the passband edges. An antenna with "Enhanced Narrow Band Filtering" very likely has more group delay error than one with more gentle filtering, so it is possible that satellites with high doppler shifts are causing the GPS timing solution to shift around more than it would with a different antenna. The one you have is presumably good enough for cell timing, since that is what it was designed for. The question is whether the lower group delay of an antenna without "Enhanced Narrow Band Filtering" would be enough better that you could tell. Without measuring the filters, we can't really predict what to expect. You might ask the seller if he would exchange your antenna for one that doesn't have the extra filtering. You almost certainly don't need it in your rural location, and an antenna without it would at least not be any worse, all else equal.
Best regards,
Charles
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Thanks Bob, Charles,
Well, here's my plan. Get a single piece of RG-11 direct burial. Get a more stable post for the antenna and perhaps a bit taller. Keep my eye out for a different antenna, right now they seem to be going for $50 and up for the 26 dB versions, I could contact RDR, but I didn't see anything on his site about other antennas. Yes, this is no high RF area except when maybe I turn on the 170 W 2M brick and that's nothing, I saw that thread about filters in the antenna. Finish this 48 hour survey and get these other things done and then do a 72 hour survey. Let it run for months and see where we are. Get a smaller switching type power supply.
I figured nothing was broken or not working right.
Bob, are you saying playing with the masks TC, gain and other stuff in LH does nothing to the NTBW50AA? It just affects how LH looks on my computer screen and I'm really not changing anything of how the Nortel works?
Thanks,
Dave
----- Original Message -----
From: Bob Camp
To: Discussion of precise time and frequency measurement
Sent: Thursday, September 19, 2013 7:23 AM
Subject: Re: [time-nuts] New NTBW50AA
HI
What can be said with reasonable confidence:
- The unit isn't broken and it does not need to go back to RDR under their 30 day return policy.
- The OCXO isn't broken and it does not need to be replaced.
- The unit is functioning ok, but probably not as good as it could.
- Without comparison to an independent standard, there's no way to know how well it is or isn't doing.
Bob
I was playing with SpecLab and my TS-2000 just to see how accurately I could measure frequencies in the HF region. I notice when I set the rx on cw and listen to the 750 Hz output of WWV at 15 or 20 MHz with SL, I get like 2 and sometimes more tracks about 2 Hz apart constantly shifting around. I assume this is Doppler going in and out? If that's so, how can anyone reasonably expect to measure any ionospheric HF signal in the mHz range?
Dave
N3DT
On 9/19/2013 6:48 AM, quartz55 wrote:
I was playing with SpecLab and my TS-2000 just to see how accurately I could measure frequencies in the HF region. I notice when I set the rx on cw and listen to the 750 Hz output of WWV at 15 or 20 MHz with SL, I get like 2 and sometimes more tracks about 2 Hz apart constantly shifting around. I assume this is Doppler going in and out? If that's so, how can anyone reasonably expect to measure any ionospheric HF signal in the mHz range?
Dave, it's not that easy. I did the ARRL FMTs a few years back when
they were transmitting from the other side of the States and thought I
was doing well to get within 0.1 Hz. The phase shifts on 7 and 14 MHz I
was seeing due to Doppler were up to 360 degrees, and quite rapid.
Mind you I was using non computer techniques [1], so averaging by
eyeball was possible.
Also if you are using WWV / WWVH, here at least, you often get both
signals at similar strength, so you might well be seeing two different
offsets.
Dan
ac6ao
[1] looking at the 100 kHz IF output of the receiver on a scope, both
scope trigger and Rx reference driven by homebrew GPS unit.
I knew it was not that easy. I didn't think about WWVB, and yes, I hear them quite often on 20. So it's just a matter of averaging what you can measure and assuming that the average will be close? I can imagine the shift can be all one way or the other for extended periods and how would anyone know which way? I'll have to check another signal that I know is not a double and see what I see. With SpecLab, it's easy to see mHz, but it's constantly changing so I guess one needs to log the data and use the spreadsheet to average it and hope it's close. The other thing is the 2000 wanders around especially when the fan goes on, so I've got a circuit to lock the MO to a GPS. XRef-VS, there are others for other radios too. We'll see what happens when I get the XRef installed and working.
When I saw the WWV/B? signals I figured it was just how lucky you were. If I'm seeing 2 Hz spread how can anything be measured with precision? You can get close to anything once, it's doing it all the time that counts.
Dave
N3DT
----- Original Message -----
From: Dan Rae
To: Discussion of precise time and frequency measurement
Sent: Thursday, September 19, 2013 10:07 AM
Subject: Re: [time-nuts] Doppler and FMT
On 9/19/2013 6:48 AM, quartz55 wrote:
I was playing with SpecLab and my TS-2000 just to see how accurately I could measure frequencies in the HF region. I notice when I set the rx on cw and listen to the 750 Hz output of WWV at 15 or 20 MHz with SL, I get like 2 and sometimes more tracks about 2 Hz apart constantly shifting around. I assume this is Doppler going in and out? If that's so, how can anyone reasonably expect to measure any ionospheric HF signal in the mHz range?
Dave, it's not that easy. I did the ARRL FMTs a few years back when
they were transmitting from the other side of the States and thought I
was doing well to get within 0.1 Hz. The phase shifts on 7 and 14 MHz I
was seeing due to Doppler were up to 360 degrees, and quite rapid.
Mind you I was using non computer techniques [1], so averaging by
eyeball was possible.
Also if you are using WWV / WWVH, here at least, you often get both
signals at similar strength, so you might well be seeing two different
offsets.
Dan
ac6ao
[1] looking at the 100 kHz IF output of the receiver on a scope, both
scope trigger and Rx reference driven by homebrew GPS unit.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
No virus found in this message.
Checked by AVG - www.avg.com
Version: 10.0.1432 / Virus Database: 3222/6180 - Release Date: 09/19/13