10MHz GPSDO by "BH3SAP"

[flash2b]

I noticed on my unit that the green led on the BP, so PC13, is never lit or blinking (PPS) using v0.1.9. For the rest everything is working properly.

Have others the same experience or is maybe my led is defective ?

Edit: Confirmed that the led on my BP is physically damaged (not by me )

[daddygo]

Thank you all for your help and understanding, I still need a way to trigger an led tho when the gpsdo has  got lock.

HI @ME,

something like this might be good for your Lock LED project, this is sized above the rottary switch, with a 3mm LED

[daddygo]

I have successfully installed firmware v0.1.9 from fredzo/fredo_ into my BH3SAP GPSDO 

First I did make a backup of the original firmware and compared this to a file shared within this topic. They were identical so that was a success. I used a WeActStudio STM32 Mini Debugger on the original BP. I was unable to read stable when the BP was still in the socket (sometimes lost connection to the target), so I pulled it out the socket carefully while I did the reprogramming. I noticed that when I tried the programming while the BP was still in the socket, the GPS module was also powered since the LED was lit.

Flashing the BP was easy and took 2 seconds to complete.

Here are some pictures:
(Attachment Link)
(Attachment Link)

The most satellites I got were 24 using the antenna stuck to the window of my loft.

Thank you Dankar, 2A03 and fredzo/fredo_ for all you effort and daddygo for the tips.

Hi @flash2b

the reason you couldn't program in the slot might be that the WeActStudio STM32 Mini Debugger couldn't drive 3.3V well,
if drop the voltage, normal to a certain extent, even at 2.6V the programming can be faulty...

ST-Link V2 which I use, so I don't have to remove the BP from the slot

[daddygo]

I noticed on my unit that the green led on the BP, so PC13, is never lit or blinking (PPS) using v0.1.9. For the rest everything is working properly.

Have others the same experience or is maybe my led is defective ?

just tested with the latest FW (v0.1.9) the PC13 LED is blinking, anyway, it's driven by the MCU PPS signal since @fredo_'s FW mod
this signal exists even if there is no GPS PPS signal, as it is generated by MCU and synchronized with the GPS PPS if there is incoming PPS from GPS

[flash2b]

Thanx for you photo. Seems my led is dead 

I am however not so much worried by that since everything else is working great, a super improvement compared to the original firmware.

Your comment on my WeActStudio STM32 Mini Debugger could be true, popping the BP out of the socket I did carefully since I knew about the wire at the bottom.

Edit:
Just confirmed with a magifying glass that the led is damaged no transparent case and does not have any bond wires so therefore no light. The red power led is working ok.

[HappySundays]

Hi @fredo_

I've loaded up 0.1.9 and love the time offset mode but... when I set 11 hours to get the time correct for my time zone (GMT +11), it has yesterday's date. We live in the future here in Australia. It's already the 17th but the date is shown as the 16th.
Should it not be +/-12 instead of 0 to 23? Both should take care of dates rolling over/under.

David

[ME]

Do these use an active antenna?, My time screen just has a sort of flashing wave/search icon followed by  TIME:   but no time displayed, does that mean no sat lock?.

[ME]

Can the original gps modual be replaced with one of these? https://www.ebay.co.uk/itm/356533964466?_skw=Neo-M8+module&itmmeta=01JPHHGVMMTYBCXE9HM8K3F0Z1&hash=item5303148eb2:g:rIwAAOSwKBRnogaF&itmprp=enc%3AAQAKAAAA0FkggFvd1GGDu0w3yXCmi1dI7B5KL24X5kdZraYGLZRwd1N7Mahc7kDXVNA0wHi1WQBFoOvCcRNY2iQ9UgBzuK%2FDOOWsLFzpd%2B21iDB%2BGicESk6gvG6IQZ2auGCNglymN5ZDA97q3YV39wxpWJnXSSZze7pDYMkfZ0l5AwAto0WtVk8ZPp8eMXKhwzNd299B%2Ba0M7NsHzKUboNieE%2BK%2BHGOpUtVwaP2sletzyD7ztPeumFmx0kNIE8PLoianhUHYlIMaHeBep4nt19KXF502kfw%3D%7Ctkp%3ABk9SR8C6w7G0ZQ
 Sorry but its the only way i see of showing the link.

[daddygo]

Do these use an active antenna?, My time screen just has a sort of flashing wave/search icon followed by  TIME:   but no time displayed, does that mean no sat lock?.

Yes these are active antennas, fed by the VCC_RF pin of the u-blox or ATGM (CH) module, depends on the type, but usually up to 50mA
pls. flip the magnetic antenna from the factory GPSDO package over and you will see that it can supply up to 3-5V DC from NEO or ATGM modules

after the icon you mentioned, you can see the number of SAT(s) you have reached, this will tell you if you have a connection to GPS modul / SAT
usually if you have no time display you have no GPS lock

[daddygo]

Can the original gps modual be replaced with one of these? https://www.ebay.co.uk/itm/356533964466?_skw=Neo-M8+module&itmmeta=01JPHHGVMMTYBCXE9HM8K3F0Z1&hash=item5303148eb2:g:rIwAAOSwKBRnogaF&itmprp=enc%3AAQAKAAAA0FkggFvd1GGDu0w3yXCmi1dI7B5KL24X5kdZraYGLZRwd1N7Mahc7kDXVNA0wHi1WQBFoOvCcRNY2iQ9UgBzuK%2FDOOWsLFzpd%2B21iDB%2BGicESk6gvG6IQZ2auGCNglymN5ZDA97q3YV39wxpWJnXSSZze7pDYMkfZ0l5AwAto0WtVk8ZPp8eMXKhwzNd299B%2Ba0M7NsHzKUboNieE%2BK%2BHGOpUtVwaP2sletzyD7ztPeumFmx0kNIE8PLoianhUHYlIMaHeBep4nt19KXF502kfw%3D%7Ctkp%3ABk9SR8C6w7G0ZQ
 Sorry but its the only way i see of showing the link.

you can't use this module because there is no PPS output to PIN header

two more things:

1.this is a "fake" relabeled NEO-M6 module PCB anyway, I've been dealing with a lot of u-blox stuff these days and immediately know that the NEO 8 series has built in EEPROM in the modules, this fake NEO 8 PCB has an eeprom which is unnecessary, the label underneath it is definitely a NEO6
(in addition, the NEO-M8L module in the picture is more for automotive use, DEAD RECKONING module + the 3D sensor in GPSDO is pointless)

2. it's pointless to choose a newer NEO series for this GPSDO, because they are not really better or more stable for this purpose
there are u-blox moduls / chips that are specifically designed for timing that can be a step forward, but the price starts at $250
https://www.u-blox.com/en/time

I made a PCB with NEO-M9N module (in production now) for this GPSDO unit, because we use GPSDO in the transmitter environment and there was a lot of interference with mobile operators' systems
the NEO-M9 series has a built-in LTE notch filter and this is an advantage for us, and we use special "mushroom" GPS antennas with 25-40m coax cable so I also designed an external LNA drive with 5V PSU

Oh and yes I almost forgot another big advantage of the NEO-M9 is that it runs on TCXO and is more stable.

SUM: for simple purposes, it is not worth installing a GPS module more serious than the original NEO-M6 or ATGM336H

[fredo_]

I just released v0.1.10: https://github.com/fredzo/gpsdo-fw/releases/tag/v0.1.10
Changes :
- [@HappySundays] Changed time offset management to fix date display (offset now ranges from -14 to +14 hours and date is calculated accordingly)
- [@ME] Added two additional GPIO outputs :
  - PA0 can be used to drive a led to display GPS lock status (active when low)
  - PA1 can be used to drive a led to display GPSDO lock status (based on the PPM threshold setting in PPM menu, active when low)
- GPSDO lock status is also shown in the main screen with a padlock icon

If someone has the time to check the date calculation logic for possible bugs: https://github.com/fredzo/gpsdo-fw/blob/d2f90bad53e456235a9e02e673a8f621ccd348b7/src/gps.c#L360

[flash2b]

I flashed it. The time was set to -13 at first (was +1 with 0.1.9) but I could adjust it to my timezone again.

GPSDO is running again... thank you fredo_ !

[daddygo]

I just released v0.1.10: https://github.com/fredzo/gpsdo-fw/releases/tag/v0.1.10
Changes :
- [@HappySundays] Changed time offset management to fix date display (offset now ranges from -14 to +14 hours and date is calculated accordingly)
- [@ME] Added two additional GPIO outputs :
  - PA0 can be used to drive a led to display GPS lock status (active when low)
  - PA1 can be used to drive a led to display GPSDO lock status (based on the PPM threshold setting in PPM menu, active when low)
- GPSDO lock status is also shown in the main screen with a padlock icon

If someone has the time to check the date calculation logic for possible bugs: https://github.com/fredzo/gpsdo-fw/blob/d2f90bad53e456235a9e02e673a8f621ccd348b7/src/gps.c#L360

Hi fredo_

thanks again 

now that you've "pushed out" these two GPIOs, you're encouraging me again not to stop and even though one half of me is resisting, so I decided to put the two LEDs on the front panel 
I'll check the two LEDs and the new FW version in the evening

hmmm, it never ends, just when you think you have nothing more to add it turns out you do

PS:
you write that the GPIOs are active at "low" level, so need an inverted buffer

[flash2b]

I like the little padlock overlayed icon. It was shown before PPB was 0.00 and the save message was shown, is that OK ?

The PPB was 0.22 at that time. I also noticed that PPB goes from 0.78 -> 0.66 -> 0.55 -> 0.44 ->0.33 ->0.22 -> 0.11 -> 0.00 so always those numbers, is that related to the precision of the calculation ?

[fredo_]

now that you've "pushed out" these two GPIOs, you're encouraging me again not to stop and even though one half of me is resisting, so I decided to put the two LEDs on the front panel 

Well send us a picture if you do so, I'll add it to the documentation!

you write that the GPIOs are active at "low" level, so need an inverted buffer

I figured we would drive the led directly from the GPIO like it's done with LED1 on the bluepill board.

[fredo_]

I like the little padlock overlayed icon. It was shown before PPB was 0.00 and the save message was shown, is that OK ?

Yes, the lock threshold is configurable in the PPB sub-menu ("PPB Lk" entry), default is 0.50.

The PPB was 0.22 at that time. I also noticed that PPB goes from 0.78 -> 0.66 -> 0.55 -> 0.44 ->0.33 ->0.22 -> 0.11 -> 0.00 so always those numbers, is that related to the precision o the calculation ?

Yes that's because this value is a 128 second running average.
The bluepill board is running at 70Mhz and we count the number of clock cycles over a 1 second period, so the smallest PPB value we can measure is 1000/70 = 14.28 (you can check the instant PPB value in the PPB sub-menu).
We get more precision by averaging this instant values over time.

[daddygo]

If someone has the time to check the date calculation logic for possible bugs: https://github.com/fredzo/gpsdo-fw/blob/d2f90bad53e456235a9e02e673a8f621ccd348b7/src/gps.c#L360

@fredo_

I don't believe in Ai yet, it's too early for this planet, but I pushed your "date calculation logic" request to ChatGPT, it came out as I'm not a SW guy at this level, maybe full of crap...

the GPT screen responses are in a .docx and the resulting code is in *.txt(s) in the whole in ZIP

ChatGPT:
Would you like me to add alarm-based auto-increment (e.g., auto-increment at midnight)? 🚀

here I have gone no further

might be good for a bit of exploration, but I say there might not be a point to it  - it's a machine, even may contain code ideas

[ME]

Does the led on pb13 illuminate when gps has lock or unlock?.

[fredo_]

I hate AIs because they give you a bunch of wrong answers around correct ones, and you spend more time checking which answer is correct than figuring it out by yourself...
ChatGPT didn't disappoint me on this one 
5 remarks, 4 wrong or irrelevant in the context BUT one was correct! I mistakenly added September in the list of 31 day months...
I'll fix it in the next release, thx @daddygo!

[daddygo]

now that you've "pushed out" these two GPIOs, you're encouraging me again not to stop and even though one half of me is resisting, so I decided to put the two LEDs on the front panel 

Well send us a picture if you do so, I'll add it to the documentation!

you write that the GPIOs are active at "low" level, so need an inverted buffer

I figured we would drive the led directly from the GPIO like it's done with LED1 on the bluepill board.

Yeah, - this can also be a good solution (like LED1 on BP) and the simplest, but limits the type of LEDs that can be used
I haven't even looked at how much the peripherals so far load the BP with something like 20mA / recom. I see 8mA as PIN

for example, most LEDs, and my favourite Lite-On blue LEDs have a voltage drop of 3 - 3.5V, which schema. is limited by the 3.3V PSU
so as I would put the LEDs on a mini PCB anyway for correct mounting, it doesn't take long to add a small circuit and power the LEDs from a 5V PSU base board + SN74LVC2G06

to avoid the need to fix the mini PCB, two 3mm LEDs would hold it in place

[daddygo]

I hate AIs because they give you a bunch of wrong answers around correct ones, and you spend more time checking which answer is correct than figuring it out by yourself...
ChatGPT didn't disappoint me on this one 
5 remarks, 4 wrong or irrelevant in the context BUT one was correct! I mistakenly added September in the list of 31 day months...
I'll fix it in the next release, thx @daddygo!

I absolutely agree with you, it's too early for humanity,  but, you know, there's money to be made from everything in this current mess,
kids can't even learn to read / count but they have the AI controled gadgets in their hands 

at least we caught the end of September, hihihih

[fredo_]

Does the led on pb13 illuminate when gps has lock or unlock?.

The led for GPS lock is on PA0 (see here: https://github.com/fredzo/gpsdo-fw?tab=readme-ov-file#gps-lock-and-gpsdo-lock-outputs).
PA0 is pulled low when the GPS is locked so that you can drive a LED with it like this:

[2A03]

Hi, fredo_

Thank you for merging my source code.
The date display is also good. It would be even more perfect if it had the ISO 8601 format (YYYY/MM/DD).

It seems that the code to save the ATGM336H baud rate setting to the EEPROM of the ATGM336H was deleted when it was merged.
It would be helpful if you could modify it so that the command to save to the EEPROM is sent after the baud rate change command for the ATGM336H model.

I checked the time it takes for the settings to be lost with another ATGM336H. The baud rate setting was only retained for about 30 minutes.
Therefore, if you shut down for a certain period of time, there should be a discrepancy between the baud rate of the BluePill and the ATGM336H.

I also think it would be better to save to the BluePill when the GPS model is detected. Currently, it is only saved when some setting change is written to the BluePill.

[ME]

Did this topic get moved ? , can only get this page via a saved link!.

[fredo_]

Hello there,
v0.1.11 has been released: https://github.com/fredzo/gpsdo-fw/releases/tag/v0.1.11
Changes are:
- Added ISO date format (yy/mm/dd) [@2A03]
- Added config save command for ATGM336H GPS module after baud rate change [@2A03]
- Added GPS module model save after auto-detection [@2A03]
- Fixed day calculation (for month = September)