bg7tbl gpsdo master reference

[sarel.wagner]

Been running my GPSDO (non bg7tbl) Star4+ Oscilloquartz unit for a few days now. Very happy indeed. External 26dB ebay timing antenna on 8m of LMR400 cable. This all via RS232 to USB with PPS signal to Lady Heather. See attached.

I have as yet no means to measure Allen Variance. As this is a hobby only, what would be the easiest and/or cheapest way to do so? TICC? I do need to get a Counter still, only counter is the Siglent one in my SigGen and my (new to me) OScope Tek TDS2014 with FFT Best I can do under my budget, no SA for me. So for a counter I am considering a HP 53131A together with possibly a TICC.

[K1FPV]

Hmmmm,

I've noticed that with Heather 3.10, the program reads out to the exact U.T.C time right to the second. In the meantime, Heather 5.0 is off by 17 seconds. Anyone else notice that?

Bill,
K1FPV

[texaspyro]

I've noticed that with Heather 3.10, the program reads out to the exact U.T.C time right to the second. In the meantime, Heather 5.0 is off by 17 seconds. Anyone else notice that?

Which GPS receiver does it use?  I think BGLTBL has more versions than birthdays...

The 17 second offset is due to the receiver being in GPS time and Heather in UTC time (or the other way around).  Try the TG and TU keyboard commands.

[K1FPV]

OK....that did it! Now everything is on the same time!

Thanks,
Bill
K1FPV

[ted572]

Has anyone here used this GPSDO with a built in LCD Screen? It sells for $209 on Ebay.  Listed as - GPSDO-Symmetricom Inside, 10 MHz GPSDO, with Power Supply, Antenna, and Display.
Example: http://www.ebay.com/itm/GPSDO-Symmetricom-Inside-GPS-10MHz-1PPS-GPS-Disciplined-Clock-GPS-Ant-Display-/172166437092?hash=item2815eb58e4:g:Ze8AAOSw1DtXFbR8

It looks interesting with the Display.

[J Dennis]

When using Lady Heather's  '/rxu' switch with the BG7TBL  '3-29-2017,' the 'run' LED stops flashing (either turns off or steady-on, depending on state when command issued), and fewer satellites are displayed as 'green.'  More information is displayed, like the firmware revision and additional observations, but the reduced number of active satellites is weird.   A subsequent "/rxn" returns the 'RUN' LED to flashing, and the dropped satellites turn 'green.'  Is this normal?  This unit uses the NEO-7M receiver.  Anyone else see this?

[H.O]

Hi,
Please excuse the newbie type questions but I could use some helpful pointers here.
I've had the 2015-09-17 version of the BG7TBL unit (with an NEO-6M unit inside) for quite some time but only just now thought I'd have a play with the RS232 port on it to see what it can do. Problem is I can't seem to have it output anything what so ever.

I've scoped the TX-pin and it just sits idle high all the time, nothing's comming out. I've connected it to a PC and lanched Lady Heather which scans for modules but comes up empty. I've tried terminal program at various baudrate - nothing. Tried sending your typical characters like cr, lf, etc but it seems to not do anything what so ever. It is happily outputting the 10MHz and 1pps signals.

Questions:
1) What should come out the serial port of the unit? (I thought it was NMEA strings at 4800 baud)
2) Do I need to do some circuit surgury in order to get it talking to Lady Heather?
3) Did a new firmware version with a fix for the small frequency error ever surface?

Thanks and, once again, I appologize for the newbie question, I've browsed this thread for quite some time without arriving at any definitive answers.

[ted572]

Hi,
Please excuse the newbie type questions but I could use some helpful pointers here.
I've had the 2015-09-17 version of the BG7TBL unit (with an NEO-6M unit inside) for quite some time but only just now thought I'd have a play with the RS232 port on it to see what it can do. Problem is I can't seem to have it output anything what so ever.

The BG7TBL GPSDO can't be controlled via software that I know of, but you can read and use its received data in the following software.
Compatible Software: Freeware: VisualGPSView, http://www.visualgps.net  This software is great and should make you happy enough.
Note: Some USB/RS-232 Adapters, etc, require opening the RTS {7} and CTS {8} lines & connecting them together on the PC side. A DIY 9 Pin F to M adapter can be fabricated for this function.

EDIT: I don't think 'Lady Heather' will work with the BG7TBL GPSDO, but I know that 'VisualGPSView' I suggested above will.  You just may very well need to modify your RS-232 connections as I described.

[cdev]

You can use a GPSDO's 1PPS as a super accurate, low jitter source of timing data by using one of he flow control lines in the RS232 connection to connect it to your computer via the serial port.

The NMEA data is no more accurate than a net time server, but a 1PPS from a good GPSDO if it is locked to UTC time is likely to measurably better than the 1PPS from a GPS alone in terms of jitter.

Don't use USB for this, if you can avoid it, as it impacts accuracy a lot..


An RS-232 level PPS signal should be on the DCD line

Make sure you get the polarity right..  If you are using NTP and one of NTPs refclocks, read the documentation closely, there are a lot of settings which can tell your computer which edge of the signal and which signal to use as the beginning of the second.

Getting that "fudge factor" right is likely to heavily impact the accuracy of your setup.

A GPSDO is one of the most accurate, lowest jitter sources of timing data available. But all that accuracy is wasted as far as NTP goes by attempting to use it with a USB serial port.

The more bare metal the interrupt is, the better.


This is from David Taylor's web site on setting up NTP- This page is one of a great many on his site on NTP setup and accuracy. Its about a specific GPS but it contains a lot of useful info on serial port setups.
 
http://www.satsignal.eu/ntp/Sure-GPS.htm#pps

Krysztof Zaraska notes:

    USB-TTL converters (I'm using FT232) use "assert low" logic, meaning that low DCD is seen as asserted.  Your site mentions this, but you don't say that if the PPS signal is directly hooked to the DCD pin of FT232, this results in NTP detecting the PPS source, but showing that the PPS source is off by 100ms (as it is looking at the wrong edge of the PPS pulse).  Seeing this behaviour can be... confusing.

    You recommend using a CMOS inverter to flip the pulse polarity.  I have found that at least with the GPS receiver I am using (Neo-6M from ublox) there is a simpler way: there is manufacturer's application (u-center) which allows you to set PPS pulse polarity.

[nctnico]

You can use a GPSDO's 1PPS as a super accurate, low jitter source of timing data by using one of he flow control lines in the RS232 connection to connect it to your computer via the serial port.

The NMEA data is no more accurate than a net time server, but a 1PPS from a good GPSDO if it is locked to UTC time is likely to measurably better than the 1PPS from a GPS alone in terms of jitter.

That depends on what timescale you are looking. The 1PPS output of the BG7TBL drifts around 50ns compared to the 10MHz output so I wouldn't call it low jitter.

[DaJMasta]

You can use a GPSDO's 1PPS as a super accurate, low jitter source of timing data by using one of he flow control lines in the RS232 connection to connect it to your computer via the serial port.

The NMEA data is no more accurate than a net time server, but a 1PPS from a good GPSDO if it is locked to UTC time is likely to measurably better than the 1PPS from a GPS alone in terms of jitter.

That depends on what timescale you are looking. The 1PPS output of the BG7TBL drifts around 50ns compared to the 10MHz output so I wouldn't call it low jitter.

This.  I've actually been measuring a number of PPS outputs from GPS units and they are not particularly stable - they have a relatively coarse adjustment step and basically switch over the center in intervals proportional to how far away they currently are.  If you took a hardware counter that counted to 10 million and used the 10MHz reference out as the clock for it, you would get a nice, stable PPS output, but the BG7TBL unit (at least the one I have) just outputs the PPS signal directly from the GPS, so you don't get the smoothing of the disciplined oscillator.

As for talking to the unit, I would check the internal connections/solder joints, as other posters have run into comm problems related to that.  The RS232 on the unit in question should just be the TX/RX from the ublox module directly, so it would be a NMEA output signal, and you should be able to breakout the pins directly from the GPS module if you want to verify that there is data coming from the module.

[cdev]

That's not good.

Are you sure/

 the bg7tbl devices simply pass their internal GPS's pps to their 1pps output?

(without improving it?)

So, people are buying it simply for the 10 MHz and not caring about the time being no more accurate than it is on a GPS? (arguably less accurate because of the apparent inability to set the cable delay, etc.)

Admittedly, the difference is such a small amount its challenging to measure it but still.

If you pay >$100 for a GPSDO but only get half the benefit you might get if it was designed properly, that's not good.

[texaspyro]

I don't think 'Lady Heather' will work with the BG7TBL GPSDO, but I know that 'VisualGPSView' I suggested above will.  You just may very well need to modify your RS-232 connections as I described.

Heather will work with most GPS receivers including NMEA and Ublox.  You might not be able to control the GPSDO features, but if it brings out the receiver serial data Heather should display it.  A few GPS receivers power up "mute" and don't send any data until commanded.  Heather cannot auto-detect these and you need to specify the receiver type (with the /rx? command) and the GPS serial input line needs to be available.

[DaJMasta]

Well I don't know 100% if it does not pass through another chip, but the PPS output on my unit (with a Neo 7M) matches the behavior and performance of a Neo M8T and a Neo 7N.... an odd find given the price and the timing specification differences, but it seems like the PPS signal adjustment step is decent sized on all of them, so while the modules are constantly correcting to make the average PPS output be in the right place, the second to second variation is still somewhat noisy.  Maybe as part of my tests I'll hook up a couple of 12 bit counters to the output and see how smoothed it actually is, it's slow going because everything needs time to be warmed up and stabilized and then the sample size is around 14h, but I'll have results eventually.

[cdev]

I'm probably misunderstanding you.

Well I don't know 100% if it does not pass through another chip, but the PPS output on my unit (with a Neo 7M) matches the behavior and performance of a Neo M8T and a Neo 7N.... an odd find given the price and the timing specification differences, but it seems like the PPS signal adjustment step is decent sized on all of them, so while the modules are constantly correcting to make the average PPS output be in the right place, the second to second variation is still somewhat noisy.  Maybe as part of my tests I'll hook up a couple of 12 bit counters to the output and see how smoothed it actually is, it's slow going because everything needs time to be warmed up and stabilized and then the sample size is around 14h, but I'll have results eventually.

.... Similarly to how errors show up on the tracks from multiple anything, lets say planes in the same area at the same time, similar timing errors will likely occur on multiple GPSs.. and cancel out.

[DaJMasta]

Right, my point is that while they converge to exactly one second intervals as they should, the minimum adjustment step for timing a single second is fairly coarse, and it will bounce around between adjustment bits so that averaged over a long period, it's stable, but in the short term, it is not very finely smoothed.  If the adjustment step was smaller, the variation from the intended center point could be much smaller.  I'm measuring around +-25ns from an exact one second signal total variation, but the minimum adjustment step is around 20ns, so that individual short term measurements based off just the PPS signal could vary considerably, this has been the case across the ublox modules I've measured so far.

If that signal was disciplining an OCXO with a long integration time, then you used the OCXO's output to create another PPS signal, that one could be considerably more stable, probably an order of magnitude or two more than the 2x10^-8 worth of minimum adjustment variation I can see on the module's PPS output.  I think this is probably around as good as you can get from a commercial GPS signal without additional smoothing, but the hardware needed to smooth it further is probably already in the box.

[H.O]

The BG7TBL GPSDO can't be controlled via software that I know of, but you can read and use its received data in the following software.
Compatible Software: Freeware: VisualGPSView, http://www.visualgps.net  This software is great and should make you happy enough.
Note: Some USB/RS-232 Adapters, etc, require opening the RTS {7} and CTS {8} lines & connecting them together on the PC side. A DIY 9 Pin F to M adapter can be fabricated for this function.

EDIT: I don't think 'Lady Heather' will work with the BG7TBL GPSDO, but I know that 'VisualGPSView' I suggested above will.  You just may very well need to modify your RS-232 connections as I described.

Heather will work with most GPS receivers including NMEA and Ublox.  You might not be able to control the GPSDO features, but if it brings out the receiver serial data Heather should display it.  A few GPS receivers power up "mute" and don't send any data until commanded.  Heather cannot auto-detect these and you need to specify the receiver type (with the /rx? command) and the GPS serial input line needs to be available.

Thank you both. I probed around a bit more today and found that I do have NMEA data at 9600 baud going into the MAX3232 chip in the BG7TBL module but nothing's coming out, I'm reading stange voltages on the V+ and V- pins so something's up with the RS232 tranceiver chip for sure. Tapping off the 3.3V level signal into an FTDI chip gets me data into the PC and Lady Heather displays it without doing anything in particular to it.

I'll see if I have a replacement tranceiver chip.

Thanks again!

[nctnico]

That's not good.

Are you sure/

 the bg7tbl devices simply pass their internal GPS's pps to their 1pps output?

(without improving it?)

So, people are buying it simply for the 10 MHz and not caring about the time being no more accurate than it is on a GPS? (arguably less accurate because of the apparent inability to set the cable delay, etc.)

Admittedly, the difference is such a small amount its challenging to measure it but still.

The difference isn't challenging to measure. Most modern scopes easely show 1ns/div. This is a screenshot of the difference between a 1PPS which uses the 10MHz output as a reference compared to the 1PPS coming from the BG7TBL:

[cdev]

Why doesn't he build in a limiter (sine to square wave converter) divide by 10,000,000, and allow adjustment of the rising edge to the UTC second?

Even a square wave (50% duty cycle/500 ms positive) would be adequate, as long as the rising edge came at the transition to the second.

Perhaps that one is the harder of the two to get right?

[nctnico]

Why doesn't he build in a limiter (sine to square wave converter) divide by 10,000,000, and allow adjustment of the rising edge to the UTC second?

Even a square wave (50% duty cycle/500 ms positive) would be adequate, as long as the rising edge came at the transition to the second.

Perhaps that one is the harder of the two to get right?

The biggest question is: is that 1PPS exactly right? IMHO the conclusion is that GPS isn't somehow good enough to transfer time with better than (say) 100ns accuracy. For a lot of purposes that is good enough.

[cdev]

The ionosphere shifts kind of like waves in the ocean (it looks a bit like that and the time constant is not that dissimilar either) from one second to the next. Depending on the GPS, "the time" likely shifts with it.

You can see this motion in a scatter plot of your position in the three dimensions that leaves out time, or by plotting the three dimensions one on top of another in the vertical axis and time in the horizontal axis.

It helps greatly to tell the GPS that you are stationary so it can ignore the changes that clearly are inconsistent with a stationary or near stationary location. That's its job.

But you need a good signal, free of multipath, and you have to be able to tell the GPS to do that!

Ublox have multiple means of communicating to the GPS on their midrange models.. what amounts to the equivalent of multiple UARTs so it would be possible.

My guess is that BG7TBL is such a small customer that ublox "won't give him the time of day". 

(Quite an appropriate colloquial English expression, as they go, in this case)

[cdev]

BG7TBL should switch to Skytraq.

Skytraq would likely be the best match for his endeavors. 

They perform better for me. As the best I can do as far as measurement is NTP performance compared to other NTP servers, I have no way of doing precise measurement with the GPIOs that I know of.

Somebody should write that software.

However, this GPS is connected to my Raspberry pi GPIO by a wire approximately 2 cm long.  It was the shortest I could make it without making it structurally unsound.

The (NTP-via-ntpq) jitter measurement is always the same, some very low amount. ".004" I think.

The stats are always the same. Almost no variation.

[Tom45]

BG7TBL should switch to Skytraq.
  ...
The jitter measurement is always the same, some very low amount. ".004" I think.

The stats are always the same. Almost no variation.

Thanks for that. I hadn't looked into Skytraq before. The Venus precision timing receiver looks very nice:

  http://www.skytraq.com.tw/products/Venus838LPx-T_PB_v2.pdf

A search shows this board: https://www.tindie.com/products/nsayer/skytraq-venus838lpx-t-timing-gps-module-breakout/



Tom

[BradC]

BG7TBL should switch to Skytraq.

I thought the whole thing about those boxes was they used a complete surplus GPSDO. As far as I've seen there's no smarts on the bit BG7TBL puts together. It is entirely reliant on a complete and self-contained GPSDO module sitting on a carrier board that just supplies power and some level conversion.

[DaJMasta]

They've got a CPLD and a micro doing stuff in the current versions, while the older versions were basically carrier boards/adapters for commercial modules, the new ones definitely are controlling the disciplining of the OCXO with on their own.