dr_phil

Hi there. I hacked two of these HDMI switches and they both had the same architecture. Each input socket has a large surface-mount diode attached nearby. The cathode of each diode was routed onto a common track on the PCB. In the case of the rectangular switcher, this is really conspicuous as it runs round the top and right hand side of the PCB. The idea of this is when you plug in any HDMI source, this track becomes energised with +5 V and allows the switcher to power up. The three diodes are quite large so you should be able to find them quite quickly. Good luck and let me know how you get on...

Hi. You need to build the hacked HDMI switcher as described on one of my other instructables. This actually performs the switching operation. Then you need to make sure the control line for this is wired to Pin 8 of the nano. The sketch segment which does the switching is:allow = (parentalOverride || timerOverride || (userLoggedIn > 1)); // set HDMI switch port to NOT(allow) if (allow == 1) { digitalWrite(8, LOW); } else { digitalWrite(8, HIGH); }Decide which version of the HDMI switch you are going to build (+ on board or off board transistor). Then follow the instructions for that particular version.Hope this helps...

Thanks for the compliment. Aesthetics was a big thing for me during the design. Generally I don't have the free time to devote to beautifying things but this was always going to end up on the mantle-piece so I made the effort!Sorry to hear you had to get rid of your tools. 3D printers are pretty cheap nowadays (but have a long learning curve) and you don't need a lathe! Perhaps give it a go anyway?

Sorry - you have to make it... but I guarantee you will value it more once you have :-).

Good to hear from you again. I totally understand your concerns. For me it's just a bit of a hobby but Instructables (or Github in your case) lets me establish my 'priority' in the novel aspects of the work. Really nice work on that key. I'm just getting to grips with CW (very slow!) and a straight key, so it will be some time before I need one like yours! Perhaps a new instructable is called for?Keep up the good work...

Hi BI3QWQ. Nice to hear from you again. I often point people in the direction of your work. I didn't realise you'd put the source code on GitHub. Thanks very much for sharing it with everyone as it looks much better than my version (which I keep telling myself I would update but now perhaps I don't need to bother!). I recently 'blew the dust off' my analyser (with original sketch) to tune up a new magnetic loop (80-40m) and it worked really well, so it's a useful bit of kit to have around whatever the firmware...All the bestdr_phil

Thanks very much. It's proved to be a popular design. Anyway, you should have no problems at all with a standard 3D printer. I use a Wanhao Duplicator I3 with a build plate of 210 mm x 200 mm. The zodiac dial is about 155 mm diameter as downloaded. Good luck and please post an 'I made it' picture when you've finished.

Hi ximulf - glad you found it interesting. Given that your requirement is slightly different, your approach should probably be different too. If you look at my other instructable (kids screen-timer with arduino) I describe a different HDMI switch. This actually came with an IR remote for changing channels I think, so you could make something work, although the last time I looked the LIRC package for the RPi was broken!The hard-wired remote method you suggest seems rather ingenious and is probably simpler to implement (once you've found the right signal lines) as you would just command the channel you wanted and probably don't even need to check you've done it. It would be interesting to see if you can get it to work - getting the signal levels right at the receive end might be a challenge…

Hi ximulf - glad you found it interesting. Given that your requirement is slightly different, your approach should probably be different too. If you look at my other instructable (kids screen-timer with arduino) I describe a different HDMI switch. This actually came with an IR remote for changing channels I think, so you could make something work, although the last time I looked the LIRC package for the RPi was broken!The hard-wired remote method you suggest seems rather ingenious and is probably simpler to implement (once you've found the right signal lines) as you would just command the channel you wanted and probably don't even need to check you've done it. It would be interesting to see if you can get it to work - getting the signal levels right at the receive end might be a challenge.I hope this helps...

Excellent work! Well worth using headphones to listen to this though - laptop speakers completely murder the sound. I'm no expert but this little circuit sounds wicked to me, and nice presentation too! It would be interesting to see some zoomed-in waveforms for some of the settings to get an appreciation of how the sound is generated. I look forward to seeing the future builds.

Hi and thanks. Please check out the video made by 123Karl123 (at : https://youtu.be/A301G1TldFo). He modified the design to run from a motor and has a static earth/moon and rotating baseplate, but it will give you the idea. I made a video for the hand-cranked original but couldn't upload it to this site and didn't want to subscribe to Youtube. Hope this meets your needs.

Hi again - it's at the bottom of the instructable, before the recommendations and comments. There's a nice thumbnail of your work. What is the logo / sticker on top of the motor housing, or is that the back of the motor?

Not at all - no offence taken, but I couldn't think of a useful response as it was a matter of personal taste for me. When I bought the marbles, I thought the earth was really well done and the sun had a lot of artistic licence but was generally pleasing to the eye. For me the moon was spoiled by the big black blob (which I couldn't easily correlate with known features) so I felt I had to hide it with the shaft mount. For the casual observer it more closely resembles the real thing now, IMHO. Great work, again :-). Please post your model on the 'I made it' part of the page if you can.

Really nice work! Well done and a good video too. It's nice to see one working. The motor gear looks really good, whizzing round so fast. One observation - you used the same moon marble as me. It has a big black spot on one side. I couldn't see how this resembled any moon feature so I made sure I put the glue on the black spot and pressed that onto the shaft, so it couldn't be seen. That's down to personal taste...Once again, excellent job. Thanks for posting the video.dr_phil

Wow - this looks really nice. Good solution on the zodiac too!. From your pictures and description, am I right in thinking that in your model, the orbit arm, earth/moon axis and motor are effectively static and the sun / zodiac plate rotates as the motor turns (along with the earth axial tilt and moon). That's a good approach to motorisation that I haven't seen before. Does it all run smoothly?Well done for a great build!

Hi and thanks for your kind comments. I'm glad you are able to find positives from the current virus thing. I am picking up a few old abandoned projects myself. Good spot on the synodic month. I can't believe I missed that at the time. I will update the text. Maybe I was getting blase after the third significant figure. The gear-trains are never a perfect match so things probably even out. The machine was never intended to be super-accurate, just accurate enough to demonstrate the concepts. This was how I was able to justify ignoring the solar/sidereal day problem with the crank. During testing I made sure I had the right number of lunar months in a year but there was no way I was going to count the number of earth rotations :-).So you are suggesting to keep the crank as one turn per side…

Hi and thanks for your kind comments. I'm glad you are able to find positives from the current virus thing. I am picking up a few old abandoned projects myself. Good spot on the synodic month. I can't believe I missed that at the time. I will update the text. Maybe I was getting blase after the third significant figure. The gear-trains are never a perfect match so things probably even out. The machine was never intended to be super-accurate, just accurate enough to demonstrate the concepts. This was how I was able to justify ignoring the solar/sidereal day problem with the crank. During testing I made sure I had the right number of lunar months in a year but there was no way I was going to count the number of earth rotations :-).So you are suggesting to keep the crank as one turn per sidereal day and modify the existing lunar train to correct for it? That's an elegant solution. I think if you were using the tellurion as an educational tool you'd still have to pretend the crank represented a solar day (how many people have heard of sidereal days, after all?). The alternative would be to do your gear mod and replace the crank with a motor (more gears, shafts etc.) to avoid confusion.Good luck with your project and the price of brass!

Hi - i don't know. As far as I can tell, Instructables sometimes loses comments or discussions - they just disappear. I get the email notifications but unless I copy and paste the embedded link I can't reply. I'm afraid there is no easy fix for the zodiac dial - I'm sure no-one will notice if you put it on upside down. Worked for me!

Hi again. If you are driving this with a motor under the sun axis, you will need to rethink the whole thing carefully to maintain proper operation. The only way this works properly is for the sun to be static and the crank at the end of the orbit arm to be rotated to work the machine. Either put your motor here (it will not look nice and wiring it will be fun) or make some complex radial linkage from the sun axis to the end of the orbit arm. These are non-trivial changes.Well spotted on the zodiac plate. I think this was designed for an astro-clock, which is an earth-centric model. When you have a sun-centric model, the sun effectively moves in the opposite direction so you need to reverse the dial so that the sun passes through the zodiac in the right order. You're the only person to not…

Hi again. If you are driving this with a motor under the sun axis, you will need to rethink the whole thing carefully to maintain proper operation. The only way this works properly is for the sun to be static and the crank at the end of the orbit arm to be rotated to work the machine. Either put your motor here (it will not look nice and wiring it will be fun) or make some complex radial linkage from the sun axis to the end of the orbit arm. These are non-trivial changes.Well spotted on the zodiac plate. I think this was designed for an astro-clock, which is an earth-centric model. When you have a sun-centric model, the sun effectively moves in the opposite direction so you need to reverse the dial so that the sun passes through the zodiac in the right order. You're the only person to notice this in two years...

Hi again. The sun, sun axis, 136m1 wheel and 39m1 wheel do not rotate relative to the baseplate. I think the 'sleeve' you refer to is actually the axis of rotation of the orbit arm and is a clearance fit around the sun axis. The sun pointer locates in a slot at the base of this and is also a clearance fit to the sun axis. Thus as the orbit arm rotates round the static sun axis, the sun pointer moves round the zodiac at the same rate. Make sense?

Hi again. I tried posting a 15 second video before but you can only do it via Vimeo or YouTube - you can't just attach a .mov file. Sorry about that...

Hiya. The separation between the top of the 136m1 wheel and the base of the sun holder happens to be about 51 mm. Note that the exact length of this shaft won't affect the function of the machine; it's solely an aesthetic issue. 140 mm should be fine (worked for me and apparently for many others) but perhaps don't cut it until you know how long you want it to be. My baseboard is 15 mm thick and the shaft extends a further ~2 mm to engage in a 'dog' in the 'baseclamp' part to stop it rotating. I don't think you would notice a mismatch of one or two mm between the heights of the sun and earth.In summary, the length of all the vertical shafts which support the orbs can be changed according to taste (but not the inclined Earth axis). The nominal lengths given in the table will probably work j…

Hiya. The separation between the top of the 136m1 wheel and the base of the sun holder happens to be about 51 mm. Note that the exact length of this shaft won't affect the function of the machine; it's solely an aesthetic issue. 140 mm should be fine (worked for me and apparently for many others) but perhaps don't cut it until you know how long you want it to be. My baseboard is 15 mm thick and the shaft extends a further ~2 mm to engage in a 'dog' in the 'baseclamp' part to stop it rotating. I don't think you would notice a mismatch of one or two mm between the heights of the sun and earth.In summary, the length of all the vertical shafts which support the orbs can be changed according to taste (but not the inclined Earth axis). The nominal lengths given in the table will probably work just fine.

Hi again,The table on step 4 shows which bearings or shafts (both made from brass tube) need to be pressed into which gears. Even if a gear is 'free to move' it still has to have a bearing pressed into it. Does this make sense? If you cross-reference the table with the diagram in step 2, you should see how things go together. Sometimes a gear needs to be pressed onto a shaft once a number of other gears have been slid on loosely, meaning that the assembly won't easily come apart again. This is particularly true of the earth/moon assembly.Revisit the instructions and let me know if you are still having problems. I'll take you through it if I can remember...Cheers

Hi there. Erm, not sure what you mean. I've put loads of photos in the description, such that many people have been able to build this already. I'd have to dismantle it to take new ones. Which part of the build is giving you problems? Perhaps I can help verbally?

Hi there - sorry I missed this - there are several ways to get comments through instructables which I'm not familiar with.Your build looks amazing - you should consider publishing the PCB design as I know loads of people who will love to get hold of it. I hope it proves useful for you.Cheersdr_phil

Sorry CO7KD I only just saw this comment. What do you mean exactly? Do you want to extract the measurement so you can plot your own graph on a PC?

Hi - you'd need to find a programmable signal generator or fast DDS module that works in the relevant frequency band. I've just had a quick look on Ebay and I don't think such things are readily available. Sounds like there's a new instructable there, waiting to happen...

Yeah, but a bit of rope didn't meet certain elements of my specification so I had to do something a bit more involved. Such is life...

Hi there. Sorry this took so long to respond - I don't see questions in my email! I think the key thing is to maintain all the gear train ratios. You can either make the teeth bigger when you increase the diameter, or increase the number of teeth. The first method is easy but might be a bit clunky. The second method will involve you recalculating new gear ratios (using the recommended website!) to match the 'accepted' celestial ratios to an acceptable level of accuracy (which only you can decide). At this point you should possibly consider designing a new tellurion from scratch, rather than modifying my design. Good luck whatever you choose and drop me a line to let me know how it turns out.

Whoops - sorry for missing this. Thanks for the comment. I haven't posted a video (although I had prepared one) because you have to link to YouTube rather than just embedding a file within the Instructable. I didn't fancy that as I prefer to limit my internet footprint. Sorry about that.

Sounds like you are doing a proper job! I made all the gears (cut the hard way with a fly-cutter and dividing head) in styrene for an astronomical clock a few years back. The thought of getting even one of these wrong would be enough to make me abandon it. I discovered that the 3d printed gears meshed easily with the real ones, and the rest is history.The reason I mention this is that I made three cycloidal fly-cutters for the wheels of different modules, but they were pretty much identical to first order and so on reflection it probably wasn't worth the effort.

Hi there. Well done on the build and I hope it is useful (and sufficiently accurate) for you.I'm no expert but I believe that the VSWR is independent of applied power (except at extremely high power where dielectrics might be degraded and change their impedance).

Hiya and thanks for the compliment. I hesitate to answer your question as I haven't been 'in the zone' on this job for ages. Skimming through my write-up, it looks like the earth is turning at the sidereal (relative to the stars) rate with each turn of the crank rather than the solar rate, where the day is 4 minutes longer. So the earth is rotating too slowly with each turn of the crank. Thus you would need a gear ratio of about 1.00275 between the crank and the earth axis (but not the rest of the model). Use the 'gear train calculator' to work out the gears; for a two stage train it gives 67:69, 38:37 (to 5 d.p.). Fitting it into the existing model would be interesting.Do think through these suggested numbers very carefully won't you before undertaking a build, as they could well be wron…

Hiya and thanks for the compliment. I hesitate to answer your question as I haven't been 'in the zone' on this job for ages. Skimming through my write-up, it looks like the earth is turning at the sidereal (relative to the stars) rate with each turn of the crank rather than the solar rate, where the day is 4 minutes longer. So the earth is rotating too slowly with each turn of the crank. Thus you would need a gear ratio of about 1.00275 between the crank and the earth axis (but not the rest of the model). Use the 'gear train calculator' to work out the gears; for a two stage train it gives 67:69, 38:37 (to 5 d.p.). Fitting it into the existing model would be interesting.Do think through these suggested numbers very carefully won't you before undertaking a build, as they could well be wrong. It's so easy to make mistakes as I have found out several times, but the only way I discovered them was by working the model for a long time, whilst observing the motions carefully and discovering that something was not moving at the right rate. I decided that an error of a fraction of a percent was more than acceptable for a 3D-printed model.One kind commenter was just happy that the earth rotated at all, as it was quite a rare feature. Good luck and let me know how it turns out...

Sounds really good! Nice idea to use two switches in order to get each one to deliver a specific function. Hopefully they are cheap enough!I found that the HDMI switches from Ebay occasionally corrupted the HDMI signal slightly, so that PC video signals ended up being displayed incorrectly on my screen after passing through it. I suspect it comes down to the sort of display one uses and the quality of the HDMI switch. Something to watch out for anyway...

Hi - I think math.h is a standard library which is included with the IDE. You shouldn't need to go hunting for it. I think you can even omit the #INCLUDE statement!

HiSorry for tardy response. I think the Uno and nano have broadly equivalent pin-outs in terms of the pin names (e.g D3, A5). Just print out a crib-sheet from the internet (or just look at the PCB) to make sure you are using the right pins on the UNO. Then just connect the relevant pins from the PCB to the OLED and other peripherals as set out in the instructable.Good luck with the project...Cheers

Hi thereBoth parts were available on ebay when I looked this evening. I wouldn't like to try recommending alternatives without comparing spec sheets. The diodes are germanium, I think, so switch on at a lower voltage which is important for this application.Cheers

HiThese error logs are always a bit opaque. I'll take a punt and ask whether you have correctly loaded the Oli Kraus library into your IDE? I ask because 'cppsrc' appears in your path, even though it just one subdirectory of the U8glib package. Your path should probably read '...\Arduino\libraries\U8glib'Universal 8bit Graphics Library, https://github.com/olikraus/u8glib/ Copyright (c) 2012, olikraus@gmail.comCheers...

Hi. Can you be more specific? I have sometimes had problems uploading to a Nano and need to specify the ATMega328P (old bootloader) option. Also there are two flavours of Nano; for this code you'll need the 328P because the other nano (168) has less RAM. Just some suggestions...

Hi again! That weavefurniture website is really interesting and particularly the HDMI switch. Nothing new under the sun eh? I notice they use four leads rather than three as in my scheme, which perhaps suggests a higher level of control. Your investigations so far have been very thorough and I guess the switch you finally bought was probably much more expensive given the tech support you've managed to obtain. The line of enquiry is really interesting though and it would be valuable to see whether it's applicable to the cheap switches (if someone has the time to look!) Good work and please publish your findings...

Hi. I don't think your suggestion would work. The diodes take power from the input leads and use it to drive the switching circuitry. That's why you have to remove them so you can provide an external on/off control.If you wanted to set to a given channel (and you are using a microcontroller to do the work), you could just pulse the button signal (it probably pulls some control line LOW) and monitor the indicator LED lines until the one you want is HIGH.Cheers

Two detents per pulse is required (though on reflection I can't see why). The KY-040 modules don't work, that's all. So I think the one you suggest sounds okay. I used :ALPS - EC11B152442D - ENCODER, ROTARY, 11MM, SWITCH, 15PPRfrom good old Ebay. I'll update the text and thanks for pointing that out. I've since used cheaper encoders on other stuff and really ought to recode this, but it works and time is precious...

Hi again. I think the main problem is the earth axis shaft. This is long, thin and supports several other shafts and gears. Unfortunately it's only supported at its base by a 5 mm brass tube. So the whole earth-moon assembly can rock from side to side when you move the orbit arm. This is enough to disengage occasionally from the 11m1 tall gear and the 127m1 gear. Because the 11m1 gear is supported by a much larger gear at its base, I don't find this rocks too much.In an ideal world I would redesign the orbit arm by changing the 2 off 10m1 gears so that the 12m1 bevel gear is moved away from the earth-moon assembly axis. Then the earth-moon assembly bearing could be doubled or tripled in length which would help significantly. I can't promise that this will happen in the near future though,…

Hi again. I think the main problem is the earth axis shaft. This is long, thin and supports several other shafts and gears. Unfortunately it's only supported at its base by a 5 mm brass tube. So the whole earth-moon assembly can rock from side to side when you move the orbit arm. This is enough to disengage occasionally from the 11m1 tall gear and the 127m1 gear. Because the 11m1 gear is supported by a much larger gear at its base, I don't find this rocks too much.In an ideal world I would redesign the orbit arm by changing the 2 off 10m1 gears so that the 12m1 bevel gear is moved away from the earth-moon assembly axis. Then the earth-moon assembly bearing could be doubled or tripled in length which would help significantly. I can't promise that this will happen in the near future though, as other projects are vying for my attention :-(.cheers

Thanks. Your message tailed off... Do you mean it is too loose or too tight? Have you drilled the bearing hole straight? One future improvement area in my model is that the bearings of some of these key gears are a bit short and therefore the shafts can precess slightly. This leads to a fair bit of slop in the design, which is visible if you move the orbit arm from side to side. This can sometimes cause the moon wheel to slip round the large year wheel. Generally speaking though, all the parts were printed as published and no special fettling was required during assembly. Please send more information if you need guidance.Cheersdr_phil

Hi ZoliHe released an early version of his sketch which is attached to one of the comment chains below, but was reluctant to share the full version openly on the forum. He hasn't posted for many months now, unfortunately. Having said that, I didn't really find the manual calibration method too onerous, and once it's done you don't need to repeat it.Cheersdr_phil

Hi - yes I suppose you could use a 16x2 LCD, but you will have to completely rewrite those parts of the code that address the screen. I used the OLED screen because it has high resolution, good visibility and viewing angle and uses I2C. The LCD screen (if I recall correctly) uses many wires to the Arduino, which I probably didn't have spare. Oh, and forget the graphing. Other than that, I can't see why it wouldn't work on a Mega. Some of the pins might need to be changed for SPI / I2C / interrupts etc.Good luck and please post when you've finished...

Thanks for that - I'll give it a try as soon as I get chance!

Dear bi3qwq - this looks really good and thanks for sharing. Have you had problems with people misusing your code? This is a real shame if true but I'm sure amateurs visiting this page really appreciate your generosity. I haven't had time to try loading your script onto my machine yet. Do I need to reduce the size of the font library, as you described previously, or does it all fit onto the nano without adjustment now?

Hi again and thank you. Do you mean the HDMI splitter or the whole thing? The splitters appear to be readily available on e-bay as of five minutes ago. Even if you can't find an identical splitter, there is a good chance that the internals will work in a similar way. My screen-timer worked with two different HDMI splitters, which could both be modified in the same way even though they had different circuit boards.If you are trying to buy the whole timer system, you might be disappointed. Similar devices which only work on composite video were available in the US up to a years ago, I think, but most stuff is HDMI nowadays so this wouldn't really suit. Many smart TVs have parental controls nowadays (but I'm not sure this extends to time intervals for use). Remember also that if the kids hav…

Hi again and thank you. Do you mean the HDMI splitter or the whole thing? The splitters appear to be readily available on e-bay as of five minutes ago. Even if you can't find an identical splitter, there is a good chance that the internals will work in a similar way. My screen-timer worked with two different HDMI splitters, which could both be modified in the same way even though they had different circuit boards.If you are trying to buy the whole timer system, you might be disappointed. Similar devices which only work on composite video were available in the US up to a years ago, I think, but most stuff is HDMI nowadays so this wouldn't really suit. Many smart TVs have parental controls nowadays (but I'm not sure this extends to time intervals for use). Remember also that if the kids have hand-held consoles they can fall back on, this device may not be the complete answer. I still find it useful for screen management, at the moment. If you want one, I am afraid you will have to make one yourself. Sorry!

Hi - yes! A five-second search on ebay for "hdmi switch 3" reveals several hits on the first page. Whilst I can't guarantee that the internals are identical, these units are so cheap that it's worth a punt IMHO. Even if the circuit board is different, the approach outline in this instructable may well still be valid - you might just need to do a bit of tracing on the PCB.Cheers

Hi again bi3qwq. You have done a really nice job on the improvements to your user interface and I notice that 66 ohms gives an SWR of 1.33, so the algorithm seems to be working fine. Are you willing to share your code with the community please? I for one would love to try it out.I liked the antenna pictures, but I guess this is a temporary installation? There must be a lot of neighbours to worry about!Cheers

Hi again. The core of this system is based on a resistive bridge circuit and firmware developed by Beric Dunn. The calculation is taken directly from his code. If you look at Wikipedia, under VSWR, you'll see that:VSWR = |Vmax| / |Vmin| = (|Vf|+|Vr|) / (|Vf|-|Vr|)The Arduino sketch uses the second part of this formula. I think Beric's resistive bridge uses diodes and low pass filters to obtain the modulus (magnitude) of the forward and reverse voltages (FWD and REV respectively), then calculates:VSWR = (FWD+REV) / (FWD-REV).I'd need to check this on LTSPICE to satisfy myself 100%, but I think this is what is happening in the circuit. The Dunn circuit and code is very widely used by amateurs.Wikipedia also reports that for a purely resistive load,VSWR = (Rload/Zline) ^ (+/- 1),with the sig…

Hi again. The core of this system is based on a resistive bridge circuit and firmware developed by Beric Dunn. The calculation is taken directly from his code. If you look at Wikipedia, under VSWR, you'll see that:VSWR = |Vmax| / |Vmin| = (|Vf|+|Vr|) / (|Vf|-|Vr|)The Arduino sketch uses the second part of this formula. I think Beric's resistive bridge uses diodes and low pass filters to obtain the modulus (magnitude) of the forward and reverse voltages (FWD and REV respectively), then calculates:VSWR = (FWD+REV) / (FWD-REV).I'd need to check this on LTSPICE to satisfy myself 100%, but I think this is what is happening in the circuit. The Dunn circuit and code is very widely used by amateurs.Wikipedia also reports that for a purely resistive load,VSWR = (Rload/Zline) ^ (+/- 1),with the sign of the exponent selected such that VSWR>1. So a 200 ohm resistor presented to the analyser should show a reading of (200/50) ^ (+1) = 4. A 12.5 ohm resistor should give the same answer (using an exponent of -1).If you have access to a commercial antenna analyser it would be very interesting to carry out a comparison with a real antenna and report back. I don't, so I have to rely on fixed resistors. I know that my rig's ATU won't work above about 3:1 SWR, and this is consistent with the results shown on my analyser.I hope this helps :-)

Thanks - it seems still to be generating a lot of interest. I think you could use the clock generator you suggest, but you may wish to add some sort of filter as the output will be closer to a square wave and thus will be rich in harmonics. The AD9850 just does the job 'out of the box' and doesn't cost that much. It will take me a while to check your schematic, particularly as I don't know much about the Si5351. Thanks for sharing it though.The AD9850 module is controlled via four pins in the current design: FQ_UD, SCLK, SDAT and RESET. The module uses SPI comms. The subroutines SetDDSFreq and send_byte send the frequency information to the module. I can't help you with the specifics of your clock generator I'm afraid - I know nothing about it. Unless money is very tight, I would recommen…

Thanks - it seems still to be generating a lot of interest. I think you could use the clock generator you suggest, but you may wish to add some sort of filter as the output will be closer to a square wave and thus will be rich in harmonics. The AD9850 just does the job 'out of the box' and doesn't cost that much. It will take me a while to check your schematic, particularly as I don't know much about the Si5351. Thanks for sharing it though.The AD9850 module is controlled via four pins in the current design: FQ_UD, SCLK, SDAT and RESET. The module uses SPI comms. The subroutines SetDDSFreq and send_byte send the frequency information to the module. I can't help you with the specifics of your clock generator I'm afraid - I know nothing about it. Unless money is very tight, I would recommend using the AD9850 which does exactly what you need - what's your time worth?Good luck...

Ah - that makes sense. I'm happy to help with debugging the SWR calc, but I'm sure you've got it covered. I still don't understand the graph data storage though. Because your code will just sweep the frequency until you tell it to stop, the code doesn't know in advance how often to sample the SWR reading to generate the graph array. Or am I misunderstanding?Cheers

Hi bi3qwq. I had some problems getting your sketch to fit on my Nano (max 30720 bytes). I had to go through it trimming out lots of text etc. before I could get it to fit. Now it's installed, I can have a proper play.* The display looks great, although the bar graph and figures overflow when the antenna port is open circuit.* The encoder button and DDS module use different pins on my hardware, but this was easily changed in your sketch.* I think there is definitely a problem with the calibration routine, as your screen (for example) reads an SWR of about 7 for a measured resistance of about 150 ohms. The actual value should be around 3.I think if you can revisit the SWR calculation and calibration aspect, and perhaps investigate why it doesn't fit on some Nanos (this could be a problem wi…

Hi bi3qwq. I had some problems getting your sketch to fit on my Nano (max 30720 bytes). I had to go through it trimming out lots of text etc. before I could get it to fit. Now it's installed, I can have a proper play.* The display looks great, although the bar graph and figures overflow when the antenna port is open circuit.* The encoder button and DDS module use different pins on my hardware, but this was easily changed in your sketch.* I think there is definitely a problem with the calibration routine, as your screen (for example) reads an SWR of about 7 for a measured resistance of about 150 ohms. The actual value should be around 3.I think if you can revisit the SWR calculation and calibration aspect, and perhaps investigate why it doesn't fit on some Nanos (this could be a problem with my set-up, admittedly), then it all looks really good. The bar display and the graph zooming feature are cool, and it's much easier to pick an operating frequency than on mine. Nice work...Cheersdr_phil

Hi bi3qwq. I had some problems getting your sketch to fit on my Nano (max 30720 bytes). I had to go through it trimming out lots of text etc. before I could get it to fit. Now it's installed, I can have a proper play.* The display looks great, although the bar graph and figures overflow when the antenna port is open circuit.* The encoder button and DDS module use different pins on my hardware, but this was easily changed in your sketch.* I think there is definitely a problem with the calibration routine, as your screen (for example) reads an SWR of about 7 for a measured resistance of about 150 ohms. The actual value should be around 3.I think if you can revisit the SWR calculation and calibration aspect, and perhaps investigate why it doesn't fit on some Nanos (this could be a problem wi…

Hi bi3qwq. I had some problems getting your sketch to fit on my Nano (max 30720 bytes). I had to go through it trimming out lots of text etc. before I could get it to fit. Now it's installed, I can have a proper play.* The display looks great, although the bar graph and figures overflow when the antenna port is open circuit.* The encoder button and DDS module use different pins on my hardware, but this was easily changed in your sketch.* I think there is definitely a problem with the calibration routine, as your screen (for example) reads an SWR of about 7 for a measured resistance of about 150 ohms. The actual value should be around 3.I think if you can revisit the SWR calculation and calibration aspect, and perhaps investigate why it doesn't fit on some Nanos (this could be a problem with my set-up, admittedly), then it all looks really good. The bar display and the graph zooming feature are cool, and it's much easier to pick an operating frequency than on mine. Nice work...Cheersdr_phil

Hi again. I'm really impressed with your improved user interface; I particularly loved the graph zooming! I would be delighted if you would share your work with the community. Perhaps add the sketch as an attachment to a comment on this page? It's a good idea to put the calibration resistors on the board. Presumably you can switch these into the antenna port at will? Anyway, to get the calibration values you need to do a logarithmic curve fit on the measured versus true VSWR curve. I think it would be difficult to do this on the Arduino, whereas most spreadsheets have the necessary algorithms built in. I have looked at the instructions I originally posted for calibration and I think they are complete - I will try to help with specific questions however.

Hi BI3QWQ. I think you have done a beautiful job. There must have been at least $192 of labour in there, but the end result looks great. I hope it meets your needs as an analyser. I am sure a few of the people who have previously commented will be queuing up to get your schematic and PCB! I like using modules because much of the hard stuff is already done for me and they are easy to replace if I blow them up.You mentioned that you were going to rewrite the code. Which areas did you think needed changing? Please post when you are done so we can all try it out.Cheersdr_phil

Hi - really nice job! I like the transparent sun orb. It's good to see how people vary the design based on their tastes and component/PLA availability. Thanks for the positive report on the knob and for sharing the file, which I've now referred to in the main text. I'm still determined to get the crank working but when I've fixed the burned-out wires in my printer (a rite of passage, apparently) I'll probably make and fit the knob. This will make the tellurion easier to use for not-so-delicate fingers.

Hi and thank you.1. The first bevel gear STL I uploaded was incorrect, as pointed out by Styxman53. I corrected this error the day it was pointed out, so just download the bevel file again (it's got a different name, slightly).2. I also included a revised orbit arm with increased 48m1 clearance on the final page, along with some improved drive-train gears with dogs to prevent slipping on machines that have been dismantled several times.3. Have you used bigger marbles than me? I think the spindles are okay - any larger and they will detract from the appearance IMO. Hot melt glue worked really well for me and there's plenty of contact area for a reasonable bond.4. I agree but a crank is aesthetically preferable. I think the crank will work really well if some weight (not much) could be adde…

Hi and thank you.1. The first bevel gear STL I uploaded was incorrect, as pointed out by Styxman53. I corrected this error the day it was pointed out, so just download the bevel file again (it's got a different name, slightly).2. I also included a revised orbit arm with increased 48m1 clearance on the final page, along with some improved drive-train gears with dogs to prevent slipping on machines that have been dismantled several times.3. Have you used bigger marbles than me? I think the spindles are okay - any larger and they will detract from the appearance IMO. Hot melt glue worked really well for me and there's plenty of contact area for a reasonable bond.4. I agree but a crank is aesthetically preferable. I think the crank will work really well if some weight (not much) could be added to the end of the orbit arm. I'm wondering if some washers could be embedded in the crank housing or something. Try gently pressing down on the crank housing as you work the crank and you will see what I mean.I'm delighted you are building one of these. Please post an 'I made it' photo when you are done.

Hi Brian and thanks. I'm delighted to see you are making your own. Sorry to hear you can't find a baseplate; try looking at bread/chopping boards on ebay. Your alternative sounds good though. I considered some sort of ornate 3D printed base but my imagination failed me. Ultimately I just wanted to finish the job quickly and didn't want the base to distract from the mechanism. You will need to ensure that the base is fairly substantial, otherwise working the crank will just make the whole tellurion shift around which would be very annoying.I think that 'rounded section' of brass you refer to is the bearing assembly for the orbit arm. It actually shows (on my original) a piece of 7 mm tube x ~10 mm long surrounding the sun axis of 6 mm OD. In practice there should be no gap between the orbi…

Hi Brian and thanks. I'm delighted to see you are making your own. Sorry to hear you can't find a baseplate; try looking at bread/chopping boards on ebay. Your alternative sounds good though. I considered some sort of ornate 3D printed base but my imagination failed me. Ultimately I just wanted to finish the job quickly and didn't want the base to distract from the mechanism. You will need to ensure that the base is fairly substantial, otherwise working the crank will just make the whole tellurion shift around which would be very annoying.I think that 'rounded section' of brass you refer to is the bearing assembly for the orbit arm. It actually shows (on my original) a piece of 7 mm tube x ~10 mm long surrounding the sun axis of 6 mm OD. In practice there should be no gap between the orbit arm axis and the spacer above it, but then this detail would be lost from the schematic. Good luck with the build and please post an 'I made it!' when you are done.

Arie - that place looks great. I will add it to my list! I love this sort of stuff...

No problem - it was the least I could do after you went to the trouble of making it all!I have added dogs to some of the gears and shafts in the drive chain which has eliminated the occasional slippage. I didn't want to use adhesive as it's effectively irreversible, whereas these gears can still be removed in principle (with a screwdriver blade). I will add an extra page to this instructable with the revised parts and also a new orbit arm with extra clearance for the 48m1 wheel.

Wow! What an amazing piece of work and faithfully reproduced so fast, too. Thanks for your kind comments. I have checked out the files that I uploaded and I put the wrong version of the bevel gear on the page, so I'm going to correct that now. It should be 12bevel.stl. Sorry about that and well spotted. If you want to check operation just wind the 54/11 gear clockwise with your fingers. You will get plenty of torque this way. I am experimenting with a modified drive arrangement that will get rid of the occasionally slippage of the bevel gears on their shafts and will post when I have it.Beautiful job, again....

Hi and thank you. I don't have any videos at present but will look at making a short one. I think donating a tellurion to your local school is really laudable. This version is quite a delicate machine though, so you'd need to consider what the likely usage might be (or be prepared to make plenty of spares!).

Thank you. I am a bit limited on free time so my typical working pattern was to design a new or modified part every Friday late afternoon and set it running on the printer during the evening. Then I'd have the rest of the weekend to incorporate it into the tellurion. Even so, the most complex part probably only takes a hour and a half (at the speed I print). Take it slow and enjoy it. You can build a simple version (just showing how the months work) with a subset of the parts, to keep you motivated...

Thanks! I think these devices are more for casual demonstration of phenomena than for making predictions, so I didn't really consider a motor to be necessary. I also didn't want to mess around with power supplies, switches, wiring etc. which would detract from the overall appearance of the device. I'd also need to shift the drive from the circumference to the sun axis. This job was effectively a feasibility study for an astronomical clock build that stagnated a few years ago. My plan for that device is to have an earth-centric depiction of the heavens that will be motorised and will reflect the current state of the sky. Watch this space...

Thanks. I think it's a pain to have to rotate the earth because a day is 1/365th of a year and the gears involved can get onerous. Nevertheless, when you see it going with the earth whizzing round it's definitely worth the effort...

Thanks for your kind comments. I should probably have varnished the baseplate before I published but I was worn out by that stage!

Thanks Yoruk.

Thank you very much. The module 1 gears have quite small teeth (~3 mm pitch around the circumference) so you'd need a very dense wood to do it. You could always scale everything up I suppose. Then there's the press-fit bearings, which might also be more challenging in different materials. Perhaps start with PLA to get the orrery bug out of your system! You might decide that it's sufficient to tick it off your 'bucket list'. Best of luck and please post an 'I made it!' when you are done, whatever medium you go for.

Now that would be an interesting challenge, but not for me! :-) Thanks for the kind words, anyway!

Thanks. Give it a year or two for me to design and build the astro-clock. As this will be earth-centric, not only will all the sums be easier but it won't matter whether the earth is flat or round. Then everyone's world view is accorded equal respect and they can still predict astronomical phenomena with some accuracy!

Thank you and I wish you well with your own version. Please get in touch if you need any help.

Thanks WannaDuino - have a go at making one yourself!

Thank you. Yes, I thought about it. The sun tube is hollow so it would be an easy matter to route some wires through to an LED or something similar in a hollow yellow orb. I wanted to keep it all low-tech and hand-powered though, not wanting to mess around with wires and batteries, which would need a thicker baseplate, on/off switch.... and I'd already paid for the Sun marble!

Thanks for your kind comments. I'm glad it's off the bench now so I can look at something else...

Good work - you are clearly more patient than me! I said I would do something similar probably a year ago, but never got round to it. So sincere thanks for getting me off the hook! :-)

Hi PaoloI only just noticed your questions. You do have to replace the 3rd and 4th values as they appear in the relevant line of the array declaration. These are referred to as columns 2 and 3 in the text because the first column on Arduino IDE is column 0. Sorry for the confusion - I might clarify this. Anyway, it sounds like you sorted it all out in the end. Good work and I hope you find it useful...

HiYep I'm sure these will be fine. I think I used 47Rs instead of 50Rs. This is not a high precision instrument, after all. Cheers...

Hi Robert - interesting to hear that you've successfully modded a DDS module (have you published?). It's really hard for me to help remotely on this one, particularly without screenshots etc. The best I can suggest is that you attempt an all-band sweep and monitor the serial output. At least you can determine whether it's the display routine or the maths which is not working properly. Let me know how you get on...

Oh yeah - sorry. You need the adaptor to go from the nano to the rear panel, so it's a mini plug. I'll correct the text.Thanksdr_phil

Hi hl2daa.Looks fantastic - it's amazing to see someone on another continent using my instructable and modifying it for their own ends, particularly as those broad-band graphs with multiple resonances look really good!I probably haven't quite caught the meaning of everything in your message but please ask again if I have missed anything.Cheers

Hi PaoloV22I couldn't get this one (it's on a little PCB, correct?) to work reliably with the existing code, which annoyed me because it was cheap. Currently the code needs a two detent per step encoder, because I borrowed someone else's code to use it. Since then I've worked out how to get a standard encoder to work but haven't bothered changing the code. You should be able to hack it without too much trouble I think.Cheers,

Thanks - mine haven't collapsed yet! Not even close, as far as I can tell!

Hi Edward - you could probably substitute the high-side PNP transistor switch with a relay, if you wanted a conceptually simpler circuit (albeit with an extra wire). The transistor method really is easy though (even I could manage it...!). I don't know how the current draw would compare between a transistor and a relay coil, however, but suspect that the relay will be more power-hungry.Good luck.

Thanks Paulo. Sounds like you made it? If so I hope you find it useful (and worth the effort of building!).I have just updated the firmware to include 160 m and 60 m (and all the bands still just about fit onto the OLED display). When I've done the calibration I'll upload V7. Watch this space...

Hi Chris - thanks for letting me know. I hope it proves useful to you. I've bought a phase-sensitive detector module from ebay recently which I'm planning to incorporate into the design instead of the bridge. Watch this space (eventually!).As far as the box goes, I just picked sometime that looked okay off ebay uk. It's broadly classed as an instrument enclosure (abs body with aluminium ends). I don't think it was branded in any way. Just looking at the layout of your board, you might be better off using a simple acrylic (plexiglass) sandwich with holes for the encoder and the antenna connector. Use PCB standoffs to support the plates. You get screen protection for free if you do this!Thanks again...

Good work and thanks for letting me know! Did you reverse engineer my stripboard layout or do your own from the K6BEZ layout? I'm guessing that the arduino is hidden underneath the copper groundplane?

Hi Steve. You make some excellent points. I like the idea of an in-built cal routine (but I have to remind myself all the time that better is the enemy of good enough!). Anyway I've just found out how to extract the hex file and will email you one if you want (if it's not a breach of protocol and I can figure out how to attach stuff!).I tend to put my projects together with cheap, compact Nanos and other import modules. Once the firmware is finalised I box it up and do nothing except change the battery occasionally. Your point about the OLED is a good one - in the driver file there is a huge range of options and you almost have to guess which display you have. It's worth it though, as these are crystal-clear, bright displays and even the graphing works well - I'm not sure how a classic 'l…

Hi Steve. You make some excellent points. I like the idea of an in-built cal routine (but I have to remind myself all the time that better is the enemy of good enough!). Anyway I've just found out how to extract the hex file and will email you one if you want (if it's not a breach of protocol and I can figure out how to attach stuff!).I tend to put my projects together with cheap, compact Nanos and other import modules. Once the firmware is finalised I box it up and do nothing except change the battery occasionally. Your point about the OLED is a good one - in the driver file there is a huge range of options and you almost have to guess which display you have. It's worth it though, as these are crystal-clear, bright displays and even the graphing works well - I'm not sure how a classic 'lined' LCD would work. Much of my stuff ends up using every pin so I'm a bit of an I2C fan. This project has quite a few pins left over at the moment.I'm keen to update this device to measure the magnitude of reactance, so I can plot minimum SWR and reactance null simultaneously. If it's easy it could happen quite quickly. If not, I'll move on to the next project in my brimming in-tray! Watch this space.Best regards, dr_phil