10MHz GPSDO by "BH3SAP"

[Jdw1953]

I’m continuing to review comments about this specific GPSDO on eBay reviews as well as here. Some are saying that the OFFSET AND PWMSET values that you save via menu are only used as initial values whenever doing your next GPS calibration. A completed calibration is active only as long as GPSDO remains powered up (and locked by saving it in menu at that time) the next power-up however you first need to do another GPS calibration before use (but happens faster this time because you have the previous values for the calibration procedure to start with).

Some people (like myself) have tested this GPSDO with Oscilloscope/Frequency counter and not seen any change in output by just saving the offset (say like 80 Hz for example) but no way it influences measured output but the GPS calibration procedure takes longer to complete when starting with a large saved offset.
These crystal oscillator circuits change some with age, temperature, etc. so makes sense that whatever is learned during calibration is trusted by whatever application that is needing a near perfect 10 MHz at that point in time. Just do not power it down until you are finished using it.

These are just my observations and thinking so far.

[flash2b]

I have tested the accuracy (see some posts above) against a Miller GPSDO and it was within 0.001Hz.

I have also tested the OFFSET function and that changed the output accordingly (see some posts above). I used -0.3Hz. There is however a bug for negative offsets stored, and I do not really use it since I want 10MHz output exactly. I believe the OFFSET should be saved -after- PWMSET is saved. There is no mention of OFFSET in the manual whatsoever.

The PWMSET value after correctly saving, so dial to PWMSET, push the knob, dial to EXIT, push the knob, works and the value is stored. This is however for use after power up without the antenna, so as a "normal" XO. Then this value is used a reference. Once you reconnect the antenna, the GPSDO starts to align the PWMSET again. This is described in the manual (which I posted above).

Above all applicable to the stock firmware.

[Gertjan]

I've now also added access to the GPS module via two pins on the bluepill. The module (ATGM336H version) can be configured via GnssToolkit and it seems like you can configure it for stationary usage.....

Do you think it is possible to get NMEA data out of this GPSDO?
It would be a big improvement to be able to monitor reception (number of satellites, signal strength) etcetera. That would make one more confident in the quality and reliability of it's output signals...

That has been done, but you need to solder wires onto two of the pins of the MCU board (PA2 (TX) and PA3 (RX), 9600 baud). Unfortunately there is no easy way to use the USB port on the board for that. My firmware also shows the number of satellites on the display.

Excellent news that serial communication is available on pins of the MCU board!
Indeed a pity that the standard USB port can not be used. But this is easily overcome with a small Serial to USB board wired to those RX/TX pins.

Off coarse NMEA is not the only way to monitor a GPSDO. Setup and monitoring software of the GPS module manufacturer would probably work too. But still you need access tot the data stream out of the GPS module...

Do you think the signal on the RX/TX pins on the MCU board you mentioned are a direct copy of the serial data out of the GPS module? (Or are these pins simply the MCU pins where the serial port of the GPS module is connected to?  )

regards, Gertjan.

[dankar]

Excellent news that serial communication is available on pins of the MCU board!
Indeed a pity that the standard USB port can not be used. But this is easily overcome with a small Serial to USB board wired to those RX/TX pins.

Off coarse NMEA is not the only way to monitor a GPSDO. Setup and monitoring software of the GPS module manufacturer would probably work too. But still you need access tot the data stream out of the GPS module...

Do you think the signal on the RX/TX pins on the MCU board you mentioned are a direct copy of the serial data out of the GPS module? (Or are these pins simply the MCU pins where the serial port of the GPS module is connected to?  )

regards, Gertjan.

The RX/TX pins is specific to my alternative firmware, and it's a direct copy of the communication with the GPS module. So anything sent will be sent directly to the module and anything out of the module will be sent directly back. You can probably attach directly to the GPS module as well, or at the pins where it's connected to the MCU (should work with the original firmware), but I felt it was safer to "relay" the communication via the MCU.

Using this it's possible to connect with GnssToolkit3 (for the atgm336h module) or u-center (for any u-blox) and test/configure the module.

[tvmanmarks]

hi
  i am a very old hobbyist who needs an accurate 10 mhz standard to set up my frequency counter because i no longer receive wwv at my location.  now i'm having doubts about this device. will it do that simple task ?
                                            barry

[flash2b]

Yes it will do. My GPSDO had a difference of only  0.001Hz compared to known-to-accurate counterpart as well as a rubidium XO.

[tvmanmarks]

i saw a difference in price of almost double if purchased through an Ebay vendor VS direct from china. i guess with Ebay you are guaranteed satisfaction.
              anyone know why WWV at any freq is almost not available here in south fla ?

[flash2b]

I am in Europe, we have DCF77 instead so I cannot help you with that.

Ebay = AliExpress + Markup for new chinese goods. For discount buy them on single's day 11-11, sure there will be some vendors or coupons on Ali.

[Solder_Junkie]

hi
  i am a very old hobbyist who needs an accurate 10 mhz standard to set up my frequency counter because i no longer receive wwv at my location.  now i'm having doubts about this device. will it do that simple task ?
                                            barry

I built a GPSDO based on the Lars Walenius circuit elsewhere on this forum. I am puzzled about the "offset" adjustment on the BH3SAP GPSDO, surely the whole idea of a GPSDO is that it is locked to GPS and is inherently self calibrated.

Like many low cost GPSDO units, this Chinese one uses a basic navigation GPS module, you can obtain a noticeable improvement in short term stability by using a timing module. I see many constructors refer to stability over many tens of seconds, or even over a day or two... Not of much relevance when you want a reference for a frequency counter or signal generator where you want short term stability too.

The GPSDO I built is this one, which adds time/date display to the original Lars design:
https://www.qsl.net/zl1bpu/PROJ/NGPSDO/New%20GPSDO.htm

This is the main thread on the eevblog forum:
https://www.eevblog.com/forum/projects/lars-diy-gpsdo-with-arduino-and-1ns-resolution-tic/


SJ

[tvmanmarks]

i received mine and because i have read all posts, i had a basic understanding of how to operate it. one matter not discussed here or mentioned on the instructions was proper antenna placement to receive satellite signal indoors etc. just for fun i stuck my antenna to the side on my microwave oven and the search began. i stopped when it settled down to 0.0 .  imagine that
                                                                 barry

[tvmanmarks]

update is not good.   
 I purchased the gpsdo because i already knew my frequency counter was a bit off.  Before today counter at 10mhz after warm up read 10000.140 . so leaving it connected to gpsdo i expected frequency counter display  to change as a warmed up gpsdo went through its paces. when its display read PPB 0.0  i went to menu found pwmset pressed button to save value. pressed button again, rotated and turned to exit.
 exact same reading on freq counter as was before, during and after the adjustment of sat. receiver. tomorrow with an extension cord i'll take it outside and place sat receiving ant on roof  of my car and start over. after i'm done i'll leave  12 volt connected and with sat rcvr still on  reconnect my frequency counter.  even though at start counter display was wrong it should have changed +/- of the current display regardless.
                                                                      barry 

[bingo600]

update is not good.   
 I purchased the gpsdo because i already knew my frequency counter was a bit off.  Before today counter at 10mhz after warm up read 10000.140 . so leaving it connected to gpsdo i expected frequency counter display  to change as a warmed up gpsdo went through its paces. when its display read PPB 0.0  i went to menu found pwmset pressed button to save value. pressed button again, rotated and turned to exit.
 exact same reading on freq counter as was before, during and after the adjustment of sat. receiver. tomorrow with an extension cord i'll take it outside and place sat receiving ant on roof  of my car and start over. after i'm done i'll leave  12 volt connected and with sat rcvr still on  reconnect my frequency counter.  even though at start counter display was wrong it should have changed +/- of the current display regardless.
                                                                      barry

Or you just have an accurate gpsdo , that has the error correction better than -1e9 , and your counter can't display the change.

[tvmanmarks]

update is not good.   
 I purchased the gpsdo because i already knew my frequency counter was a bit off.  Before today counter at 10mhz after warm up read 10000.140 . so leaving it connected to gpsdo i expected frequency counter display  to change as a warmed up gpsdo went through its paces. when its display read PPB 0.0  i went to menu found pwmset pressed button to save value. pressed button again, rotated and turned to exit.
 exact same reading on freq counter as was before, during and after the adjustment of sat. receiver. tomorrow with an extension cord i'll take it outside and place sat receiving ant on roof  of my car and start over. after i'm done i'll leave  12 volt connected and with sat rcvr still on  reconnect my frequency counter.  even though at start counter display was wrong it should have changed +/- of the current display regardless.
                                                                      barry

Or you just have an accurate gpsdo , that has the error correction better than -1e9 , and your counter can't display the change.

which makes you an optimist and confirms me to be a pessimist :+)
   Saturday i  will take everything outdoors and place the antenna on the roof of my car (just like my sxm ant) and with my gpsdo andport freq counter preheated i'll have it another go at it.    hope you are right
                                                                      barry

[flash2b]

My Racal Dana 1998 using its internal OCXO is also a little bit off. However I use the BH3SAP after warming up as Ext Std so that measurements made with the 1998 are referenced that GPSDO instead. The 1998 sees if an external reference is connected, but on some freq counters you need to push a button.

In the photo I showed above, with the 3 counters, the BH3SAP is compared to a Miller GPSDO on the Ext Std. The difference is very very little.

[tvmanmarks]

went outdoors Saturday, warmed up rcvr and then placed antenna on car roof .  i allowed the setup screen time to do its thing.
right away i saw the difference as it was running its numbers to 0.1 etc.
 after finishing and 12 volts on it i checked the frequency on my un calibrated frequency counter. the reading was even more than  it was the day before but i was able to calibrate to 10mhz w/a bunch of zero's to follow.
 so your assumption was correct. disciplined or not the receiver seems to have a very accurate crystal osc. to begin with.
                          thanks again, Barry

[Pilatus]

I received this BH3SAP gpsdo last week, and so far it looks quite good on my Racal Dana 1999C frequency counter (stable to the 1 mHz over a 20s period).

I do get a gpsdo "PPb display" excursion of +/- 0.5 or so. This number seems to be updated every second on that display. Typically it is +/- 0.2 (alternating between positive and negative every couple of seconds), but I have seen it as high as 0.7 PPb.

The unit has been running for several days. My lab is located on the 2nd story of a 2 story house, sloped roof. I was thinking this may be due to poor antenna reception so I  tried moving the antenna closer to the window (on a steel plate, tried other locations within the house) with no noticeable improvement. It is snowing here so going outside is not an option for now.

I see other people's displays for this device showing a solid 0.0 PPb. Is this (a solid "0.0 PPb" display) what I should be expecting?

edit: corrected 20s counter average from 1 Hz to 1 mHz.

[tvmanmarks]

what made it work for me was first placing antenna on my micrwave but better yet my  suv's roof. with an extension cord for the receivers power supply i allowed it to warm up before adjusting settings.
                                   barry

[Pilatus]

Thanks for the reply Barry. I will try your suggestion when the weather improves.

Still, an impressive little device for the money (paid $52).

[HappySundays]

Hi all. I just received my GPSDO and was overjoyed to see that new firmware was available. I have the latest binary, but I'm not at all familiar with the bluepill. I own all manner of programmers including SWD-based ones for ST development. Haven't touched them for years though and am rusty!  A lot of sites mention the ST-Link V2 but that seems to be just a USB to SWD interface. Is the SWD interface the only way to program it? Does that USB port have no programming capability? No bootloader in there? Do I need to move any of those jumpers from the BOOT position? So sorry for the noob questions.

[HappySundays]

Update: I put on my big boy pants, did some research and used a J-Link and the ST cube programmer to get it done with no issues. New firmware works… but I might volunteer some tweaks over the break. On my unit, the default contrast is way off and there doesn’t appear to be a contrast adjustment. Not a show stopper. Initial settling time is still ongoing after an hour. I understand that once the PWM start value is saved, it will be quicker in the future. Certainly not as wobbly as the original. Happy days, and one more firmware project to work on.

[G4ELJ]

Just programmed up the alternative firmware and it is working well.

Same issue though with lack of display contrast.

Tracing the pin outs.

Pin 3 of the LCD display does to R8/C17 as an RC filter. The other end of R8 goes to Pin A10 on the Blue pill board.
Presumably in the original BH3SAP firmware a PWM voltage output from A10 is smoothed by the RC filter and gives a decent adjustment of the contrast.
Currently there is only ~100mV on pin 3 of the LCD.

Unfortunately I am not a programmer :-(

Of course one could remove R8 and just use a 10K preset with Pins 1, 2 and 3 on the LCD display to manually set the optimum contrast voltage on Pin 3.
 
Might be nice to also activate the PC13 LED to flash the PPS. One would then have another visual way of knowing that the code is running even if one can't read the display very well.

Duncan

[daddygo]

Hi @dankar

This is some frenetic work with FW, thanks.
I will now take both of our units off the shelf during the break and convert to 0.0.12 FW and they will get a buffer for their 1PPS outputs..
1PPS via MCU is also a great idea. (if you have the time)

With these modifications I look at these units with a much better eye. 


+Edit: (4ch driver on the way).... - Schmitt trigger + Quad buffer /w 3-state out


+++Edit2: under reconstruction (yes the pin headers are a bit high, 90 degrees should be installed next time)

[daddygo]

Dear  @flash2b

This osciloscope (FNIRSI) is a "undemanding - punchy unit" unit that we use in the field to avoid having to carry millions of dollars worth of lab equipment into dirty transmitter towers 
Otherwise we have one of these for more serious things:

Teledyne LeCroy Digital Oscilloscope WAVEPRO 960.

Yes I know that the other is a toy  , but suddenly measuring something 40 metres up in a tower is mostly appropriate..

[Ghostrider66]

I got mine last week and changed the firmware immediately. Nice work, many thanks for that.
For all of you who want to replace the BluePill Board before flashing the firmware from dankar:
- beware of the wire below the board, which connects the output of the OCXO with the OSC-IN Pin of the STM32 Controller (Pin 5) - or should connect
- Check the soldered wire. On my board it was soldered to the OSC-Out Pin (Pin 6). This only worked because the 8 MHz crystal was still in place. On my replacement board I removed the crystal and the two caps to gnd (not needed). Took me a whole weekend to figure out, that the controller did not get the necessary clock signal.

[fredo_]

Got mine today and flashed @dankar's firmware.
Thanks for the great job @dankar !
I put the original firmware binary file here in case someone needs it.

Same issue here with contrast being set to maximum, making it impossible to read in direct sight :-/
@dankar, I didn't see a "setContrast()" method in the LCD driver you used.
Any idea of the LCD type that is used ?