bg7tbl gpsdo master reference

[kj7e]

You just need to set the NMEA Protocoll / Version to 4.1, than you can use all 72 SATs and GPS, Gallileo and GLONASS together.

I could be wrong, but I think the setting NMEA protocol to v4.1 enables the position fix flags for Galileo and BeiDou to be sent out of the receiver, but does not affect the Navigation input to the receiver.  Regardless, I found that little tidbit of info in Sec 31.1.3 of the Protocol manual, now u-center shows me all the tracked SV's.  In order for the BG7TBL MCU to run and update the LCD, I had to set the Main Talker ID to GPS, the MCU does not like the extra flags.

Edit: I had to set the main talker ID to GPS for my older 2017-12 version, the newer 2018-06 does not care and can be left to automatic, this is nice as u-center and LH will properly show the Galileo and BeiDou.

I am using a narrow band L1 Timing antenna now. Going to swap it out for a multiband GNSS antenna, although the L1 narrow band antenna does seem to be receiving Galileo well.  Right now I have the GNSS types set for GPS, SBAS, and Galileo.

Edit: Now using a PCTEL GNSS multi band antenna and see BeiDou, GLONAS and Galileo;
https://www.pctel.com/antenna-product/global-gnss-timing-reference-antenna-gnss1-tmg-26n/

[Ohm_My]

I am using a narrow band L1 Timing antenna now. Going to swap it out for a multiband GNSS antenna, although the L1 narrow band antenna does seem to be receiving Galileo well.  Right now I have the GNSS types set for GPS, SBAS, and Galileo.

If you want to perform precision timing be sure and disable the SBAS subsystem as per page 66 (19.2 Recommendations - "u-blox M8 Receiver Description Including Protocol Specification R16.")

[kj7e]

If you want to perform precision timing be sure and disable the SBAS subsystem as per page 66 (19.2 Recommendations - "u-blox M8 Receiver Description Including Protocol Specification R16.")

Yeah, I just read over that section again, I thought I saw that once before.  I'm not sure if the SBAS subsystem introduces jitter or just delay.

[kj7e]

Now I have two BG7TBL LCD GPSDO's.  One with a blue background LCD model 2017-12-16 and a new one with a green lcd background model 2018-06-05.  I wanted to update the display on both to show ppb, found a few differences in the display programing command between them, also some information that's not in the very nicely but rouge translated v1.1 manual found here;
https://www.eevblog.com/forum/testgear/bg7tbl-gpsdo-master-reference/msg1912184/#msg1912184

This is whats in the book;
2. Display UTC Time, with Frequency Error in PPB (recommended)
$GPGTC,2 UTC +0000T*$GPGTC,2 GMT +0000T*

However, I found the 2017 unit accepted the code but gave the time an offset of tens of minutes, this was due to a missing 0 in the time parameter.  The correct code was;
$GPGTC,2 UTC +00000T*$GPGTC,2 UTC +00000T*

Also corrected the repeated "UTC".

The 2018 unit also would now say ERROR if the format was not perfect and needed the additional 0 in the string, the 2017 would accept the string but would not display the correct time.  Last, the code above with the proceeding + in the time parameter would cause the LCD display to literally display a "+" before the time.  To correct this, I replaced the + with an additional leading white-space as such;
$GPGTC,2 UTC  00000T*$GPGTC,2 UTC  00000T*





Since I had a few extra NOS UCT 108663 ovens laying around, I pulled the as shipped Trimble 63090 and replaced it, had to move a dc block cap C34 to position C44 which was unpopulated to route the 10MHz sine wave correctly.  Also adjusted the EFC gain to 2x as the 108663 needed 5.6v, by default the EFC gain is 3.3v x 1.5 for a max of 5v.  Not sure yet if I will update the u-blox M7 for the M8T on this one.  Going to do some comparison between the two.



Also made a closed cell foam insulation hat for the OCXO, I believe the a bit of insulation to reduce external thermal influences is a good thing here.


Last note, while working on the new 2018 unit, I found one row of pins on the LCD connector header on the main board were not soldered, easy fix but this caused the display to blank if you moved the cable at all.

[Ohm_My]

Also made a closed cell foam insulation hat for the OCXO, I believe the a bit of insulation to reduce external thermal influences is a good thing here.

I'm curious to hear the temps that Lady Heather is reporting with your foam hat. I have a similar oscillator (Oscilloquartz 8663-XS DOCXO) in my unit that runs at about 49.1 - 50.3 while the unit is in it's aluminum enclosure.

[texaspyro]

Also made a closed cell foam insulation hat for the OCXO, I believe the a bit of insulation to reduce external thermal influences is a good thing here.

No, it's probably a bad idea.   The oscillator ovens were designed to run without any external insulation (and work quite well as is).   Adding insulation can upset the balance of the thermal gains of the ovens and interfere with the operation of the internal oven controllers. 

[kj7e]

I'm curious to hear the temps that Lady Heather is reporting with your foam hat. I have a similar oscillator (Oscilloquartz 8663-XS DOCXO) in my unit that runs at about 49.1 - 50.3 while the unit is in it's aluminum enclosure.

You must have a OSA Star 4 board, there is no way to pull OCXO temp from the BG7TBL units.

No, it's probably a bad idea.   The oscillator ovens were designed to run without any external insulation (and work quite well as is).   Adding insulation can upset the balance of the thermal gains of the ovens and interfere with the operation of the internal oven controllers.

Those are some good points, wish I had some way to verify or measure the effects.  The same argument has been made with the LTZ1000, the chip works best when it can dissipate some heat and be allowed to regulate itself.  My thought was to keep the environment around the OXCO steady, but perhaps this will do more harm than good, simple enough to remove.

[German_EE]

I have one of the original BG7TBL units without a display and this afternoon it failed on me. The red ALM light stays on and the GPS Lock light is off. The Status LED is however flashing and the OCXO is warming up. I see 3.29V at the antenna socket.

Does anyone have some fault finding information?

[kj7e]

I have one of the original BG7TBL units without a display and this afternoon it failed on me. The red ALM light stays on and the GPS Lock light is off. The Status LED is however flashing and the OCXO is warming up. I see 3.29V at the antenna socket.

Does anyone have some fault finding information?

The first thing I would check is if your OCXO has aged out and can no longer be netted back to 10.000MHz.  Check the EFC voltage, be default, the EFC voltage range is 0-4.9v.  Happened to me not long ago.

This video may help you;


I made a post on this here;
https://www.eevblog.com/forum/testgear/bg7tbl-gpsdo-master-reference/msg2255781/#msg2255781

Post a photo of which OCXO you have.

[German_EE]

I have one of the CTS OCXO devices fitted but the problem has been located, my antenna has died after being knocked from the window frame one too many times by the cats. At least, I think it's the antenna as they are not supposed to rattle.

Off to EBay for a replacement, I may get one of the marine units and mount it outside.

[usagi]

my antenna has died after being knocked from the window frame one too many times by the cats.

double sticky tape.

[texaspyro]

my antenna has died after being knocked from the window frame one too many times by the cats.

double sticky tape.

For the antenna or the cats?   

[jonese]

I received my 2019 model blue LCD a day ago (PCB says 2019, but the case back says 2018).  The 10 MHz functions and the LCD works, but it's very easy to scramble or blank the LCD.  All I have to do is touch the top of the unit.  I believe it's a ESD issue.  If I leave the unit untouched, it's fine.  10 MHz output works fine during the LCD issue however.

I emailed the seller and they mentioned unplugging the antenna and back in again.  And, yes, after 5 seconds or so the display sorts it self out.  Not sure I'm finding that as an acceptable situation.  The fact the seller came back quickly and with a decisive answer leads me to think this is not the first time they have seen this.

Anyone else?


Update:
Figured out a fix.  Shielding.  The end plates weren't electrically connected reliably.  The whole top and front were floating.  Once I sanded the ends of the aluminum and put some conductive tape to allow the end plates to make contact with them, the problem went away.

The design could use some improvement in ensuring all the surfaces make contact with each other.

[usagi]

my antenna has died after being knocked from the window frame one too many times by the cats.

double sticky tape.

For the antenna or the cats?   

yes.

[kj7e]

I received my 2019 model blue LCD a day ago (PCB says 2019, but the case back says 2018).  The 10 MHz functions and the LCD works, but it's very easy to scramble or blank the LCD.  All I have to do is touch the top of the unit.  I believe it's a ESD issue.  If I leave the unit untouched, it's fine.  10 MHz output works fine during the LCD issue however.

I emailed the seller and they mentioned unplugging the antenna and back in again.  And, yes, after 5 seconds or so the display sorts it self out.  Not sure I'm finding that as an acceptable situation.  The fact the seller came back quickly and with a decisive answer leads me to think this is not the first time they have seen this.

Anyone else?


Update:
Figured out a fix.  Shielding.  The end plates weren't electrically connected reliably.  The whole top and front were floating.  Once I sanded the ends of the aluminum and put some conductive tape to allow the end plates to make contact with them, the problem went away.

The design could use some improvement in ensuring all the surfaces make contact with each other.

The ground for the LCD should be through the ribbon cable and not via chassis ground.   I would inspect the ribbon cable connectors and insure they are properly soldered to the boards.  I had one where an entire row of pins was not soldered.  I have run both of mine on the bench completely apart many times and chassis ground between the front and rear covers is not needed for proper LCD operations.

[jonese]

The ground for the LCD should be through the ribbon cable and not via chassis ground.   I would inspect the ribbon cable connectors and insure they are properly soldered to the boards.  I had one where an entire row of pins was not soldered.  I have run both of mine on the bench completely apart many times and chassis ground between the front and rear covers is not needed for proper LCD operations.

Yes, thank you for comments.  There is indeed ground connectivity to the LCD PCB.  I had at the time checked both connector solder joints and they were pretty good quality.  Mine also works without any casing around the electronics.

Humidity at the time was a bit low (< 40%) so touching most anything in the house would have caused some static discharge while moving around.  This is what I'm fairly sure was causing the LCD to scramble (I have a feeling it was the actual the HD44780 module that was having issues).  If I grounded the top case and front panels, no more scrambled LCD's with subsequent touching.

[nealix]

Hi Guys:

I notice from the postings that there is now a 2019 version of the BG7TBL GPSDO.
For those here who have purchased it, do you know if the frequency is accurate, and
if the previous software bug that caused a small offset has been fixed?
What improvements are in the 2019 version?
Are there any specifications or user manual?

All of the current ebay vendors that have "2019 10MHZ OUTPUT SINE WAVE GPS DISCiPLINED CLOCK GPSDO"
in their product listing, are actually showing a photo of the 2017 unit :-) :-)

Thanks for any initial impressions, opinions, etc.

Neal

[Amateurtelecom]

Hello!

For anyone who's interested in the EFC behaviour of the (2016-05-31) BG7TBL GPSDO are here some test measurment results:



I replaced the 5 VDC Bliley OCVCXO by a 12 VDC Trimble 34310-T OCVCXO (as shown below) and I wondered what the EFC voltage behaviour was...



The replacement was quite easy. The EFC amplification isn't changed and only jumper R5 is moved to position R4 to get 12 VDC to the OCVCXO instead of 5 VDC.



I won't bother you with the details here, so if you're interested in all the details it can be found here: http://www.amateurtele.com/index.php?artikel=203&id=#1380 (Due to the international group of users, I translated most if it already from Dutch to English, but it's still work in progress.) 

[timber23]

I notice from the postings that there is now a 2019 version of the BG7TBL GPSDO.

All of the current ebay vendors that have "2019 10MHZ OUTPUT SINE WAVE GPS DISCiPLINED CLOCK GPSDO"
in their product listing, are actually showing a photo of the 2017 unit :-) :-)

Thanks for any initial impressions, opinions, etc.

I received my GPSDO within a week from China to Germany. Super fast delivery. On the front the datecode 2019/03/25 is printed.
The parcel included the BG7TBL GPSDO, a DC wall plug and an GPS antenna. I payed $148 on ebay. Link to eBay offer: https://ebay.us/9W2CF7 Seller: thanksbuyer-hobby

There is no manual in the parcel. I do not have any equipment which is capable of veryfining the 10MHz output. On the back, there is a sticker with a link to a taobao site: https://bg7tbl.taobao.com/

Best regards
Timber

[panman]

Sorry in advance for the "ugly" formatting. Not used to posting via this format...

I had a single BG7TBL Symmetricom Inside GPSDO as a instrument time base. I wanted the ability to test it for accuracy and recently purchased a BG7TBL 2019-03-25 GPSDO for comparison. Initial tests showed the 2019-03-25 unit was about 900uHz low. To confirm accuracy (looks like the song “Does Anyone Really Know What Time It Is” is true), I borrowed a Rhode & Schwarz ED170MP GPSDO and a friend’s (Doug) BG7TBL Symmetricom Inside GPSDO. This post summarizes the results and I have come to the conclusion that the 2019-03-25 unit is not as accurate as the Symmetricom Inside units with a difference of 1E-2 to 1E-3. I don’t profess to be an expert in this area so please let me know your thoughts.

Test Setup
All GPSDOs were powered for at least one week before conducting any tests. All tests run between 24 and 48 hours. The Rohde & Schwarz has its own down converting antenna. The three BG7TBL GPSDOs share a common antenna (Symmetricom 58532A) via a 4 port GPS distribution amplifier (Symmetricom 58536A). Since all 3 share the same antenna, the survey data should be extremely close with respect to position.

Fluke PM6690 Setup

• Time base reference: Rohde & Schwarz ED170MP w/20dB attenuator
• Measurement Time: 10 seconds
• Device under test: 1 of 3 BG7TBLs GPSDO

        -BG7TBL GPSDO Date code 2019-03-25 (internal board date code of 2019-01-07 just like the pictures Involute posted)

Photos of BG7TBL 2019-03-25 (1/2)

        -BG7TBL GPSDO “Symmetricom Inside” with integral display (Ron’s)
        -BG7TBL GPSDO “Symmetricom Inside” (Ron’s)

Results Summary:
Before drilling down into the results of the it appears as though the newest BG7TBL GPSDO (2019-03-25) is only good to 1E-11. In contrast, the earlier “Symmetricom Inside” units which are good to 1E-13 to 1E-14. The presumption here is the Rhode & Schwarz is an accurate reference.

• BG7TBL GPSDO Date code 2019-03-25:

10MHz output: Consistently about 900uHz low, 2 samples provided
1PPS output: Terrible short term (1 minute) stability per the attached scope capture (persistence set to infinite). 2.5 divisions at 20ns per division for a 50ns jitter over a 1-minute interval.
Survey data: Since unit outputs NEMA sentences via the DB9 connector, an analysis of the “$GPGGA” records over about 1000 seconds results in a wide range of positions. This summary and all GPGGA records are in the attached spreadsheet.

   Latitude   Longitude   Altitude
Average   4308.83108   7610.25686   125.8
Min   4308.83043   7610.25545   122.6
Max   4308.83189   7610.25804   131.0
StdDev   0.000430624   0.000811474   1.885299917

• BG7TBL GPSDO “Symmetricom Inside” with integral display (Ron’s)

10MHz output: 14uHz high in the sample provided and typical of other tests
1PPS output: Minimal short term (1 minute) stability with about 1/4 division or 5ns jitter over a 1 minute interval.
Survey data: LAT(N 43:08:49.884) LON(W 76:10:15.426) H(124.50 m MSL)

• BG7TBL GPSDO “Symmetricom Inside” (Doug’s’s)

10MHz output: 1uHz high in the sample provided.
1PPS output: Minimal short term (1 minute) stability with about 1/4 division or 5ns jitter over a 1 minute interval.
Survey data: LAT(N 43:08:49.848) LON(W 76:10:15.384) H(126.00 m MSL)

Ron
WB2WGH

[r o b e r t]

Another BG7TBL,  this dated 2019-02-28 on the enclosure, 2019-07-05 on the board. The antenna voltage is 5v not 3.3v as indicated in the sales info.

[panman]

I sent BG7TBL an email regarding the accuracy of the 2019-03-25 unit referencing my eevblog post. Here is his response:

"hello this normal.there is deviation with this type gpsdo.you can see other friend test result.this version is peak to peak is small value. lcd-gpsdo not this problem. thanks bg7tbl"

When queried as to which models do not have the problem the response was:

"this version" with the image included below:

[Note image appears not to be posting. The image is of a unit with a 2017-12-16 date code with the one line display on the side opposing the connectors.]

[Hans_18T]

"should" be posted at https://www.eevblog.com/forum/testgear/new-gpsdo-bg7tbl/ ?
Thanks for the post, I am curious for the difference between the old and the new version (technically), although for my purpose (external reference frequency counter) I am very happy wilt the "old" version.

[n3mmr]

I sent BG7TBL an email regarding the accuracy of the 2019-03-25 unit referencing my eevblog post. Here is his response:

"hello this normal.there is deviation with this type gpsdo.you can see other friend test result.this version is peak to peak is small value. lcd-gpsdo not this problem. thanks bg7tbl"

When queried as to which models do not have the problem the response was:

"this version" with the image included below:

[Note image appears not to be posting. The image is of a unit with a 2017-12-16 date code with the one line display on the side opposing the connectors.]

I find BG7TBL's reply difficult to understand, me not being a native english speaker and not familiar with chinese either.

Can someone clarify?

Also, a single unit of a specific date code is not statistically significant.
Have you tested any other identical units?

[kj7e]

I sent BG7TBL an email regarding the accuracy of the 2019-03-25 unit referencing my eevblog post. Here is his response:

"hello this normal.there is deviation with this type gpsdo.you can see other friend test result.this version is peak to peak is small value. lcd-gpsdo not this problem. thanks bg7tbl"

When queried as to which models do not have the problem the response was:

"this version" with the image included below:

[Note image appears not to be posting. The image is of a unit with a 2017-12-16 date code with the one line display on the side opposing the connectors.]

I find BG7TBL's reply difficult to understand, me not being a native english speaker and not familiar with chinese either.

Can someone clarify?

Also, a single unit of a specific date code is not statistically significant.
Have you tested any other identical units?

I believe he means the frequency offset is typical on the version in your thread.  The 2017-12 LCD versions do not seem to have this offset.