Q: Increase transmit power/range

[denniskarlsson]

If I want to be able to send the signal to my entire house, how do I build such a thing?

[Mercoory]

I recommend you to check out Andreas Spiess's video based on @hzeller 's work : https://youtu.be/6SHGAEhnsYk

[denniskarlsson]

Yes, that's more like it, but it seems like the range still is very limited. According to the video it's just a few centimeters.

[hzeller]

Few centimeters is what the simple wire already provides, so the ferrite will possibly better.

I strongly suggest to ask some local ham radio person to help you with that. They know what can be done technically while still staying within legal limits.

[denniskarlsson]

ok, thank you. I will see what I can up with.
I live far from anyone else so it shouldn't be a problem with a little more power on the transmitter.

[Mercoory]

In the video Andreas shows a circuit. And says the voltage for the amp circuit is set for his country's law ( Switzerland ). So check your country's regulation, and calculate with a formula the allowed voltage on that simple circuit.

[electrogoodie]

You circle the whole periphery of your house with four loops of solid wire like AWG14 (just look at the VLF antenna in airports) or follow Andreas Spiess's video on youtube on how to tune your humungous antenna for proper transmission line impedance matching or resonance

[pierow2k]

As @Mercoory mentioned earlier, Andreas Spiess makes a reference to Vcc at about 9:52 in the Remote Controller for Clocks video:

Please do not increase the supply voltage of the transistor to more than 3.3 volts as this will increase the transmitted power and therefore the reach of the signal.

[smeisner]

If anyone is interested, I created a PCB for the circuit above and will be receiving them in ~ a week (although with COVID-19, it may be held up at US Customs). The minimum order was 5, so I may have a few left over if anyone wants one. I made the board the same size as a Raspberry Pi Zero.

Top of PCB being made:
https://www.meisners.net/images/wwvb-pcb.jpg

My plan is to solder a ferrite core antenna to the L1 pads.

Steve

[LIBHUT]

I'd be really interested in one of your PCB's -in need of another project for my lockdown!

[privatesam]

Hi @smeisner the comment from LIBHUT was me on my work account by accident. I would be very interested in one of the PCB's if we can arrange?

[smeisner]

Hi! Thanks for the second email. I meant to reply to your first post, but completely forgot. Sorry about that. What is your shipping address? I have 1 left, so it's yours! Steve

…

On Mon, May 4, 2020 at 11:39 AM privatesam ***@***.***> wrote: Hi @smeisner <https://github.com/smeisner> the comment from LIBHUT was me on my work account by accident. I would be very interested in one of the PCB's if we can arrange? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABVHM3B2SNO7KYOPLJCFNSLRP3OTBANCNFSM4JGIRRMA> .

[privatesam]

Hi Steve - thanks for this. Let me know your paypal so I can drop you some money for this. Just realised you're in the US and I'm in the UK so if its too much hassle no problem but my address is:

Sam Manley
5 Perceval Avenue
London
NW3 4PY

Thanks.

[smeisner]

@privatesam I put the envelope with a board in the mail today. I missed the postal pick up, so it will start moving tomorrow.

If you want to email directly, feel free to send me an email by going to my web site and using the email form:

https://www.meisners.net (use the "Email me" link at the bottom)

Steve

[privatesam]

@smeisner Thank so much! That is so amazing of you to go to the trouble. Will drop you an email now.

[GOTO-GOSUB]

I have also made some PCB's for this application. Please check my Github for more details.

A big thank you to Henner for his work on this excellent solution to a problem.

[harlock974]

Beware of typos in the above schematic. Component values should be :

R3 : 4.7 kΩ
R2 : 560 Ω
R4 : 200 Ω
C1 : 1 nF
L1 : 4.5 mH

L1 is Andreas' ferrite antenna inductance. C1 have to be calculated from L1 and from the requested frequency (40, 60 or 77.5 kHz) with the following formula :

C = 1 / (4×π²×L×f²)

[smeisner]

Beware of typos in the above schematic. Component values should be :

R3 : 4.7 kΩ R2 : 560 Ω R4 : 200 Ω C1 : 1 nF L1 : 4.5 mH

L1 is Andreas' ferrite antenna inductance. C1 have to be calculated from L1 and from the requested frequency (40, 60 or 77.5 kHz) with the following formula :

C = 1 / (2×π²×L×f²)

Thanks so much for the corrected details! I will give this a try at some point soon.

[smeisner]

Using the values in the schematic (as an example) and assuming freq of 60Hz and the units for L are in H, C is farad), I get:

C = 1 / (2 * π² * L * 60²)
C = 1 / (19.739208802 * 3600 * L)
C = 1 / (71061.151687843 * L)
if L = 1 nH (0.000000001 H),
C = 1 / 71061.151687843 * 0.000000001
C = 14072 F
That would be a huge value! Obviously I am doing this wrong.

[harlock974]

We both made mistakes. The formula is C = 1 / (4×π²×L×f²)
(and not 1/(2×π²×L×f²) , I edited my last post).
And I suppose your frequency is 60 kHz (60000 Hz).
So you should use a capacitor of 7036 µF.
Still quite large...

[smeisner]

We both made mistakes. The formula is C = 1 / (4×π²×L×f²) (and not 1/(2×π²×L×f²) , I edited my last post). And I suppose your frequency is 60 kHz (60000 Hz). So you should use a capacitor of 7036 µF. Still quite large...

Yes, sorry. 60kHz here in the 'states'.

That seems wrong....could it really call for such a large cap? I don't know enough about generating radio signals to really judge that.

[harlock974]

That seems wrong....could it really call for such a large cap? I don't know enough about generating radio signals to really judge that.

May be the inductance of your antenna is too low. As a comparison I can give you the values for the antennas I successfully used and the theoretical values of C for 60kHz :

• AM Loop antenna : 27 µH / 261 nF
• Ferrite antenna from an old AM receiver : 504 µH / 14 nF
• Inductor coil : 0.1H / 70 pF

The ferrite antenna give the best results with the longest range.

[jrotaetxe]

@smeisner

If anyone is interested, I created a PCB for the circuit above and will be receiving them in ~ a week (although with COVID-19, it may be held up at US Customs). The minimum order was 5, so I may have a few left over if anyone wants one. I made the board the same size as a Raspberry Pi Zero.

Top of PCB being made: https://www.meisners.net/images/wwvb-pcb.jpg

My plan is to solder a ferrite core antenna to the L1 pads.

Steve

Hi, Steve.
Sorry for disturb you after so long, but I wonder if you still have docs to let me order somewhere a PCB to make the antenna circuit.
Best regards
Jon Rotaetxe

[smeisner]

I do have a Kicad 6 project. But so far, I can only find a newer version I made. Do you want this?

Feel free to email me directly at: steve at meisners dot net

Having direct access will be easier to share files.

Edit: @jrotaetxe ... sorry I didn't tag you in my reply

[nickapos]

For anyone who wants to calculate the components they can use this https://www.omnicalculator.com/physics/resonant-frequency-lc

[smeisner]

Looking at the second circuit described in #34 it looks like achive many meters of range is possible.

I'd be willing to make a handful of these (as SMD designs) populating everything (the N- and P-MOSFETs, the 10ohm and 100mF (that is 100 millifarad, right @harlock974 ?). The antenna and corresponding cap will be left off as it would potentially be different for everyone.

I could even make the cap & antenna pads thru holes to make it easy to populate. ...or I could make the very large SMD. And again, same form factor to fit the RPi Zero.

If I get enough interest, I will make a batch. Let me know by replying to me at: txtempus@meisners.net

BTW, thank you @nickapos for that link. Very helpful!!!

[harlock974]

I'd be willing to make a handful of these (as SMD designs) populating everything (the N- and P-MOSFETs, the 10ohm and 100mF (that is 100 millifarad, right @harlock974 ?).

No it is 100 nanoFarad. Sorry my handwriting is very poor (I edited #34 it is clearer now).

[smeisner]

ok, I have created the initial schematic & pcb in Kicad and have the files up for review:
https://www.meisners.net/files/txtempus-sch.png
https://www.meisners.net/files/txtempus-pcb.pdf
https://www.meisners.net/files/txtempus-3d.jpg

C1 & C2 are user supplied and can be thru hole or SMD. The idea being the end user adds one of these caps. L1 is the antenna. Everything else will be populated.

The two MOSFETs were changed to AO3400A (N-channel) & AO3401A (P-channel) as the PCBS house has these on hand.

@harlock974 Can you verify these parts are OK as substitutions? Also, the orientation of Q1? It doesn't quite match your drawing.
Here's the datasheets:
https://datasheet.lcsc.com/lcsc/1811081213_Alpha---Omega-Semicon-AO3400A_C20917.pdf
https://datasheet.lcsc.com/lcsc/1810171817_Alpha---Omega-Semicon-AO3401A_C15127.pdf

If anyone has feedback for the design, please let me know. I included 2 holes to be able to tie wrap a ferrite core antenna down.

So far, I have interest from @jrotaetxe and @nickapos ... anyone else?

FYI: I will charge actual cost for manufacturing and shipping to your location...that's all. The PCBs cost ~$2 USD for 5, so if I get 10 with assembly, it will be cheap per board. Shipping to US is ~$25.

[smeisner]

Well, that was quick. I already discovered a problem; Kicad has the wrong footprint for the 40 pin header. The orientation is off 180 degrees. I will fix this and repost later today.

[nickapos]

Hi Steve is the 2$ for 5 boards with the assembly including the various components?Can you also please let me know what is the shipping cost to the UK?Thanks and regards --------------------------------------------- Nick Apostolakis, MM0NKKOn 30 Oct 2023, at 18:10, Steve Meisner ***@***.***> wrote: ok, I have created the initial schematic & pcb in Kicad and have the files up for review: https://www.meisners.net/files/txtempus-sch.png https://www.meisners.net/files/txtempus-pcb.pdf https://www.meisners.net/files/txtempus-3d.jpg C1 & C2 are user supplied and can be thru hole or SMD. The idea being the end user adds one of these caps. L1 is the antenna. Everything else will be populated. The two MOSFETs were changed to AO3400A (N-channel) & AO3401A (P-channel) as the PCBS house has these on hand. @harlock974 Can you verify these parts are OK as substitutions? Also, the orientation of Q1? It doesn't quite match your drawing. Here's the datasheets: https://datasheet.lcsc.com/lcsc/1811081213_Alpha---Omega-Semicon-AO3400A_C20917.pdf https://datasheet.lcsc.com/lcsc/1810171817_Alpha---Omega-Semicon-AO3401A_C15127.pdf If anyone has feedback for the design, please let me know. I included 2 holes to be able to tie wrap a ferrite core antenna down. So far, I have interest from @jrotaetxe and @nickapos ... anyone else? FYI: I will charge actual cost for manufacturing and shipping to your location...that's all. The PCBs cost ~$2 USD for 5, so if I get 10 with assembly, it will be cheap per board. Shipping to US is ~$25. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: ***@***.***>

[harlock974]

Hi @senza-lattosio

I realized a very gross circuit usign the original Andreas Spiess' work and a ferrite antenna from AM radio. It works quite well but Im looking for a more elegant implementation and possibly wider coverage. Right now Im able to reach 5 to 6 meters.

You should reach a better range with a ferrite antenna and 5V from Raspberry GPIO. Check the schematic with mosfet here. If you have an oscilloscope, check signal quality at the antenna and adjust capacitor accordingly. For DCF77, you will have a longer range with Raspberry Pi 3 than with a Pi 4 (because of available internal clocks).

[senza-lattosio]

Hi @senza-lattosio

I realized a very gross circuit usign the original Andreas Spiess' work and a ferrite antenna from AM radio. It works quite well but Im looking for a more elegant implementation and possibly wider coverage. Right now Im able to reach 5 to 6 meters.

You should reach a better range with a ferrite antenna and 5V from Raspberry GPIO. Check the schematic with mosfet here. If you have an oscilloscope, check signal quality at the antenna and adjust capacitor accordingly. For DCF77, you will have a longer range with Raspberry Pi 3 than with a Pi 4 (because of available internal clocks).

@harlock974
Thank you. Right now I'm actually using the 5V from GPIO, but I'm using a rpi zero WH; are you suggesting that using an rpi 3 will improve coverage? I didn't use rpi4 because when I frst launched txtempus there I read a message like "txtempus is known for not working properly on rpi4, go back to a previous model", so I settled for a rpi zero bc I guess is the same form factor as the @smeisner pcb.

btw, Im just realizing now that your setup is a bit different from smeisner's: is it because of the changes you made in the code?

[senza-lattosio]

@senza-lattosio Sure! What would be easiest; gerber files to merely submit an order or would you like the KiCad files I created? The latter will allow you to customize the PCB and create your own production files...assuming you're using KiCad (v6)

I also built a transmitter based on Andrea's video (with the ferrite core antenna), but did not get satisfactory coverage. This new board does work quite well, but I have not done an in-depth analysis.

Thank you so much. I'm planning something like this: use Kicad to add the required caps for the DCF77 signal and get a gerber file to submit to pcbway to get a handful of pcb's.
The only problem with my plan is that I never used kicad before, but Im very much willing to learn.
Somewhere here I also read the proper values for the caps to be added to your configuration, so I will look for it.
Thank you

[harlock974]

@harlock974 Thank you. Right now I'm actually using the 5V from GPIO, but I'm using a rpi zero WH; are you suggesting that using an rpi 3 will improve coverage?

Pi Zero has the same clocks than the Pi 3 so you should get the same range.

btw, Im just realizing now that your setup is a bit different from smeisner's: is it because of the changes you made in the code?

No, smeisner removed the 10 Ohms resistor because its Mosfet are more powerfull than mine. Except for this the circuits are the same. Do you run my code or txtempus ?

[senza-lattosio]

@harlock974 Thank you. Right now I'm actually using the 5V from GPIO, but I'm using a rpi zero WH; are you suggesting that using an rpi 3 will improve coverage?

Pi Zero has the same clocks than the Pi 3 so you should get the same range.

btw, Im just realizing now that your setup is a bit different from smeisner's: is it because of the changes you made in the code?

No, smeisner removed the 10 Ohms resistor because its Mosfet are more powerfull than mine. Except for this the circuits are the same. Do you run my code or txtempus ?

I'm running txtempus bc I discovered yours just now. Do you think that by using your code I will get improvements ? Of course by also changing the setup since Im using the attenuation pin right now)

Here's my ugly setup (sorry)

[harlock974]

I'm running txtempus bc I discovered yours just now. Do you think that by using your code I will get improvements ? Of course by also changing the setup since Im using the attenuation pin right now)

Your setup is fine, that's experimenting !

I don't know if you will have better result but you could try. If it is a transistor in your circuit, just put a 4.7K resistor between its base and GPIO4 and remove the attenuation pin connection.

[smeisner]

@senza-lattosio Here's my txtempus PCB files (including Gerber files in the production folder). Please share any improvements with the group.
TxtempusTransmitter.zip

The BOM has a column called LCSC Part number. This is for PCB assembly by jlcpcb.com. So you will need to modify these entries in the Kicad files before regenerating the BOM.

I am using KiCad V6.0, but you can import this project into KiCad 7. If you do, it will then be incompatible for anyone using 6 though.

[senza-lattosio]

@senza-lattosio Here's my txtempus PCB files (including Gerber files in the production folder). Please share any improvements with the group. TxtempusTransmitter.zip

The BOM has a column called LCSC Part number. This is for PCB assembly by jlcpcb.com. So you will need to modify these entries in the Kicad files before regenerating the BOM.

I am using KiCad V6.0, but you can import this project into KiCad 7. If you do, it will then be incompatible for anyone using 6 though.

Very much thankful for the share. Since i'm planning to place an order quite soon, will be glad to send spare PCBs to anyone who will ask so.

[harlock974]

Hi guys
What range do you get with Smeisner board ?
I tried with DCF77 and pi zero, two 4.7nF, and get 20 cm range.

[senza-lattosio]

Hi guys What range do you get with Smeisner board ? I tried with DCF77 and pi zero, two 4.7nF, and get 20 cm range.

Hi, sorry for the late reply, I had to wait for my pcb to be shipped.
Same here, about 20 cm range.
The "old" setup still working with ~6 meters range.

Here is the short range setup: I just modified the gerber file and added 2 4.7 nF caps.

In case someone needs the modified file are here, including bom and cpl
HPT_pcb.zip

[harlock974]

Thanks, I will cut the trace of the onboard antenna and connect a ferrite coil there. I should get the meters range.

[senza-lattosio]

Thanks, I will cut the trace of the onboard antenna and connect a ferrite coil there. I should get the meters range.

Right now with the "original" Spiess setup (transistor), txtempus executable and a ferrite coil I can get 6 meters range.
I will try to replace the trace and the caps and see what happens (with your time-signal executable and only GPIO4 connected).
Maybe I will get ~ 10 meters.

[jrotaetxe]

Remember...
Not only the antenna, but the encapsulated ferrite....
So from C2 to the pin

[davoau]

@senza-lattosio Did you have any luck with getting 10M distance ? I have tried to download the HPT_pcb.zip from the link above but it doesn't work - is it correct or can you upload somewhere else
Thanks
David

UPDATE: I was able to download HPT_pcb.zip just now

[LouisLee985]

@smeisner

The PCB may be wrong?





https://www.digikey.com/en/products/detail/alpha-omega-semiconductor-inc/AO3401A/1855773


Perhaps:

[smeisner]

If what you're saying is correct, then the footprints in KiCad are wrong. Not saying you aren't correct...it'll just be surprising.

I can verify with the datasheet later today.

[smeisner]

@LouisLee985 You are 100% correct. The datasheet shows the KiCad footprint was indeed wrong. I have created a custom footprint and regenerated the PCB. I have already submitted an order for 10 PCBs and I will be shipping updated units to everyone that ordered the first round...after I verify they work properly.

I feel as though this was my fault for not verifying the footprint. I never used this component before and didn't realize the footprint was wrong. So for everyone that ordered the first round, I will take care of shipping costs as well.

For anyone else, after I get these 10 boards and verify them, I will submit another order for anyone that wants to purchase one (at cost for PCB and actual shipping). JFYI: I am not making any money on this.

[smeisner]

Well, bad news. I sent the PCB as @LouisLee985 described above to a fab house and received them. After installing it on a RPI Zero, the Pi wouldn't power on, but it immediately overheated. Just powering up the WWVB board (no RPi) using my bench power supply, the new circuit shows a short between 5V and gnd.

Did I screw up the PCB:

[LouisLee985]

Well, bad news. I sent the PCB as @LouisLee985 described above to a fab house and received them. After installing it on a RPI Zero, the Pi wouldn't power on, but it immediately overheated. Just powering up the WWVB board (no RPi) using my bench power supply, the new circuit shows a short between 5V and gnd.

Did I screw up the PCB:

There may be other bugs on your PCB.

[ads1230]

@smeisner any luck getting the new board to run? Interested in trying it out!

[smeisner]

@smeisner any luck getting the new board to run? Interested in trying it out!

Not at all. This board is not my design. I do not know RF communications. And given the last reply from LouisLee985, it appears as though he is not putting any further effort into this either. His last change ended up frying a RPi of mine.

I have dumped a good deal of money & time into this and cannot keep doing this sort of trial and error process. Sorry.

[nickapos]

For what is worth I would like to thank @smeisner for the time and effort he has put into this. Even though I do not get the range I was after, his board has a similar range to my own custom solution but in a lot more compact form. I use it all the time with my various watches.
I would advise whoever is interested in evolving this design to build their own based on the published schematics and then share their results. Then if people are interested in making that more compact I am sure they will say so.

[smeisner]

To be honest, if anyone comes up with a design (and is willing to work with me), I am fine creating additional PCBs to further this effort. I just don't want a (as we say in the states) "Drive By" attempt.

[ads1230]

I'm not a PCB / RF board expert but the schematics for the MOSFETs show

• Drain / Out (top pin)
• Gate / Control (bottom left)
• Source / In (bottom right)

On the latest board design Q1's OUT has 5V power and Q1 IN is connected to Q2 OUT (fed by ground) so I assume thats the short you're finding.

If we're using 2 MOSFETs

• Q1 linked to 5V
• Q2 linked to ground

Surely they should connect to C3 on alternate sides?
Please someone check my logic. Have I missed the point of the two MOSFETs?

If we de solder components from a previous board they could be hooked up to a bread board to test rather than getting new boards printed.

Some other thoughts (but baby steps) if this design works:

• The board can be compacted as the RPi pinouts are now clustered together
• A Dip Switch could be added to switch frequencies - essentially one board for all regions (2x 15nf, 1x 10nf)
• Need a solder point for an external antenna / possible cutoff for the internal antenna

[smeisner]

If we de solder components from a previous board they could be hooked up to a bread board to test rather than getting new boards printed.

I can take the last rev of boards that caused a short between 5V & GND (which I did not throw away) and cut the traces and use wrap wire (thin gauge) to try out your idea. If this works better than the first version we did, I have no problem submitting a new order for PCBs.

It will be close to a week before I can complete this testing due to other priorities.

Some other thoughts (but baby steps) if this design works:

• The board can be compacted as the RPi pinouts are now clustered together

Agreed. This current PCB design was done merely as a prototype and could easily be made smaller.

• A Dip Switch could be added to switch frequencies - essentially one board for all regions (2x 15nf, 1x 10nf)

Yes! Another great idea if the changed design works.

• Need a solder point for an external antenna / possible cutoff for the internal antenna

Again, good idea that's easy to add.

Thanks for adding to the effort with great ideas!!

[ads1230]

If we de solder components from a previous board they could be hooked up to a bread board to test rather than getting new boards printed.

I can take the last rev of boards that caused a short between 5V & GND (which I did not throw away) and cut the traces and use wrap wire (thin gauge) to try out your idea. If this works better than the first version we did, I have no problem submitting a new order for PCBs.

It will be close to a week before I can complete this testing due to other priorities.

Some other thoughts (but baby steps) if this design works:

• The board can be compacted as the RPi pinouts are now clustered together

Agreed. This current PCB design was done merely as a prototype and could easily be made smaller.

• A Dip Switch could be added to switch frequencies - essentially one board for all regions (2x 15nf, 1x 10nf)

Yes! Another great idea if the changed design works.

• Need a solder point for an external antenna / possible cutoff for the internal antenna

Again, good idea that's easy to add.

Thanks for adding to the effort with great ideas!!

As a quick amendment
From the original circuit diagram, Q1 and Q2's outputs should both go to the same point on C3 and C3 be connected to ground.

BS250 / 401
Source 5V
GPIO
Out = BS270 OUT

BS270 / 400
Source Ground
GPIO
Out = BS250 OUT

[smeisner]

So what you are describing above is the first version we made, based on #34 (2nd schematic)

As an exercise, I used the schematic described in harlock's post above and started from scratch, using the Alpha & Omega datasheets as reference. I came up with the same design you last posted (the V1.0 PCB). But this design has a very short transmission range.

So it seems we have come full circle ... or am I missing something?

I did use one of my V1.1 boards and cut traces and soldered wirewrap wire to make the circuit you cited earlier (#8 (comment)) but it didn't transmit a proper WWVB signal.

[harlock974]

Hi Smeisner,
You cannot expect a longer range for these low frequencies with a PCB antenna. See my note about antennas in https://github.com/harlock974/time-signal

[smeisner]

OK, this makes sense. I used your note to refer to the schematic, but never read the section at the bottom! Thank you!!

I will see if I can produce a half-decent version using the dual MOSFETs and a ferrite core antenna like this one which is tuned for 60Hz or WWVB)

[ads1230]

So what you are describing above is the first version we made, based on #34 (2nd schematic)

As an exercise, I used the schematic described in harlock's post above and started from scratch, using the Alpha & Omega datasheets as reference. I came up with the same design you last posted (the V1.0 PCB). But this design has a very short transmission range.

So it seems we have come full circle ... or am I missing something?

I did use one of my V1.1 boards and cut traces and soldered wirewrap wire to make the circuit you cited earlier (#8 (comment)) but it didn't transmit a proper WWVB signal.

I'm sorry the range did not improve, however we haven't come full circle! We've cleared the MOFSET wiring issues of the original PCB design (and cleared the short on v1.1!). If we add antenna positions for a ferrite core antenna and a DIP switch I think you'll have created the perfect board.