I agree I have good taste and you make good videos! Smith charts have confused me since university. I need to watch that R&S video just to conquer them finally!
QUCS (or qucs Studio which adds some nice features) should be much more suited for this job I haven't used it much myself, but for one, it comes with smith charts built-in...
I found out about the more advanced features (or maybe standard features) in qucs after I did all the work for this video. I actually used qucs for creating some of the other structures on my rf demo board (like the 20dB coupler). I did not run into the microstrip simulation part yet, but I saw a video on it a few days ago and I really like that. I want to make some stuff with this in the future.
Thanks for this video, great stuff! I posted it over at Reddit, on the RFElectronics subreddit. Hope I can drive some more views your way - you've certainly earned them with your amazing content!
Thanks, that was really informative. You mentioned ε is a bit uncontrolled in commercial FR4; is it possible to work back from your results to find the actual average value for an improved iteration? Presumably your looped traces give a better average value than a straight trace, due to the weave of the PCB.
I tried, but it is really tricky. You have to know the exact location of the reference plane of your network analyzer measurement. I did not include enough options in this board to be able to do that accurately. I'd also have to calibrate out the connectors. I thought it would be relatively easy, but turns out it isn't.
No problem, if you get me a N9918B FieldFox Handheld RF Analyzer (26GHz) then I'll make those videos :-D. This is way out of my budget I'm afraid :-) Would love to play with that though!
Oh wait, I see that you'd like to cover 40GHz as well...... it should be a N9953B FieldFox Handheld RF Analyzer then, no idea what it costs, but you can probably get a small house for it :-) I think it's just 61k USD :-D
I hit you up right when you started on LinkedIn saying you were a great instructor..... so I said I'm receiving a signal of one pulse per second or have gone ahead and put it through temperature controlled or ocxo module and have a 10 megahertz signal and I want to divide that up between 4:00 SMA outs! I I'm already using a really good pulse per second and a microcontroller to create my own network time so I would like to sync all of my RF devices with the same 10 MHz clock.
So you want to have a 10MHz splitter / distribution amplifier? Is this a digital signal, or a sinewave? What are the demands on phase noise? I guess a 4-way splitter with 4 amplifiers behind it would work well for a sinewave (you can use those standard mini-circuits transformers). The amplifiers should deliver enough signal strength and be relatively low noise. If it is a digital signal, you can use gates, but you have to make sure the supply is really clean, otherwise supply noise will translate to phasenoise......
Great info. This video gives me some cool ideas for the kinds of tools that are helpful in electronics. I really appreciate the exceptional knowledge that you are sharing. As far as attention span, 20 mins is a good soft "max" for a video. Of course, it does depend on the content. Do you have tools that allow you to show "viewing time" on a individual viewer basis or perhaps "average viewer" basis?
I can see the average view time, not individual. Many people click away after a few seconds because they find out the video is not for them. That really reduces the average view time. At the moment this time is 6m20s which is pretty high based on what I've learned 🙂
@@HansRosenberg74 That's also what I see on my pedagogical videos: A lot of people click away after a few seconds. A small group stays and watches the entire video.
attenuators are such a hack to use instead of 50 Ohm SMA loads... tho I've never entirely understood why they act as such. In my stupid brain, since they have a "characteristic impedance" Z0, they can be considered as a T-line. So they are an open terminated T-line in your setup.Why is there no reflection at the input? You are saying the signal goes thru on way, then backwards from the open terminatiation. Why does it goe thru on way in the first place??? Shouldn't it reflect back immediately?
Ehm, yeah ;-) If I'm not making any errors in calculation here, it would be around 3.75 meters for 1/4 wave at 10MHz assuming 1.5e8m/s signal speed :-) Board material is not that expensive these days but I still think this is going to hurt :-)
I was sure there was a python package, however, I wanted to fully control the appearance of the chart and lines on it so it would be clear in a video presentation like this. It was also nice to learn how to create these lines. I learned a lot :-)
I'm a little confused why you went through a long procedure at 9:30 for calculating the impedance of the quarter wavelength stub rather than just taking the geometric mean of 50 and 100 ohms. Anyway, very nice video! Thanks.
the x-axis is non linear in a smith chart. So that is what all the math is for. If you take the geometric average you would get 75 I assume (if that is what you meant) and we need 70.7. But I'm not a math genius so I may be missing something, an easier solution.....
To get the midpoint you have to calculate the mean in the Gamma domain, and then reconvert to impedance. That is what was done in the video, because X is Gamma as shown below. I think the geometric mean thing is a coincidence, in fact the correct impedance is 70 ohm when you have a Gamma of 1/6 (0.166...) and not 70.7. X = -1 + 2Z/(Z+Z0) X = (-(Z+Z0)+2Z) / (Z+Z0) X = (Z-Z0)/(Z+Z0) = Gamma
I checked that out. I'm showing a transformer there just to show that you can buy this kind of thing off the shelf, but I'm not actually using it in this video. Sorry to confuse you.
sir I will probably enroll to your course If you make it more affordable. 1euro = 40x our currency , so it becomes huge cost for me. Thank you for the great video
I know, there is a huge difference in income around the world. My country is quite expensive. Up till now I've invested around 6000 euro's in this channel and 5 months of costs without income (I'm doing this full time). I assume the total costs to be 15000 to 18000 euro's. So I have to make that back, and more so I can build up a pension and pay for unexpected costs like my car breaking down. You know how that goes ;-) Reality of life :-) . When I can cover my costs I will look for cheaper options in the future, maybe a book? Or mini-courses. This program I'm making now is really big, but that also makes it really valuable. Best regards, Hans
Time flies when you're having fun. That was a complicated video that really challenged me. Thanks!
You're welcome!
I have no idea what you’re talking about. But am interested in circuits. Excellent stuff.
Well, if my visuals are enough, maybe I should become an entertainment youtuber, lol
When there is RF stuff I have all the attention that you want, really nice done!
haha, thanks!
I agree I have good taste and you make good videos! Smith charts have confused me since university. I need to watch that R&S video just to conquer them finally!
good luck! It's not that hard if you watch a few good videos on it.
Really, really excellent presentation. Thank you!
You're welcome!
Great video, thanks for explaining RF topics in the way that is easy to grasp!
Thanks! That is a great compliment!
QUCS (or qucs Studio which adds some nice features) should be much more suited for this job
I haven't used it much myself, but for one, it comes with smith charts built-in...
I found out about the more advanced features (or maybe standard features) in qucs after I did all the work for this video. I actually used qucs for creating some of the other structures on my rf demo board (like the 20dB coupler). I did not run into the microstrip simulation part yet, but I saw a video on it a few days ago and I really like that. I want to make some stuff with this in the future.
@@HansRosenberg74 the filter generator tool in qucs can also generate microstrip versions
Excellent video, many thanks!
You are welcome!
Greetings from your former collegue!! Following you now and then in your video's, enjoy them very much.
Gerrie
Hee Gerrie, wat ontzettend leuk om van je te horen. Ik hoop dat alles goed gaat, ik heb even een berichtje naar Linkedin gestuurd! Groetjes, Hans
Thanks for this video, great stuff! I posted it over at Reddit, on the RFElectronics subreddit. Hope I can drive some more views your way - you've certainly earned them with your amazing content!
Thanks a lot!
Excellent video! Thanks.
You're welcome!
Thanks for your work!
You're welcome :-)
Great video! IMHO, Qucs-S is best suited for this. It also comes with a power combining design tool.
I'm found out too late that qucs has a lot more options. I did not hear about the combiner designer yet, thanks!
Thanks, that was really informative. You mentioned ε is a bit uncontrolled in commercial FR4; is it possible to work back from your results to find the actual average value for an improved iteration? Presumably your looped traces give a better average value than a straight trace, due to the weave of the PCB.
I tried, but it is really tricky. You have to know the exact location of the reference plane of your network analyzer measurement. I did not include enough options in this board to be able to do that accurately. I'd also have to calibrate out the connectors. I thought it would be relatively easy, but turns out it isn't.
Let's get to the more fun stuff. 20 to 40 ghz. The 2 ghz stuff is like working with audio! What do you do when the smith charts don't work any more!
No problem, if you get me a N9918B FieldFox Handheld RF Analyzer (26GHz) then I'll make those videos :-D. This is way out of my budget I'm afraid :-) Would love to play with that though!
Oh wait, I see that you'd like to cover 40GHz as well...... it should be a N9953B FieldFox Handheld RF Analyzer then, no idea what it costs, but you can probably get a small house for it :-) I think it's just 61k USD :-D
I see that you make your own Valves?! Really cool!
I hit you up right when you started on LinkedIn saying you were a great instructor..... so I said I'm receiving a signal of one pulse per second or have gone ahead and put it through temperature controlled or ocxo module and have a 10 megahertz signal and I want to divide that up between 4:00 SMA outs! I I'm already using a really good pulse per second and a microcontroller to create my own network time so I would like to sync all of my RF devices with the same 10 MHz clock.
So you want to have a 10MHz splitter / distribution amplifier? Is this a digital signal, or a sinewave? What are the demands on phase noise? I guess a 4-way splitter with 4 amplifiers behind it would work well for a sinewave (you can use those standard mini-circuits transformers). The amplifiers should deliver enough signal strength and be relatively low noise. If it is a digital signal, you can use gates, but you have to make sure the supply is really clean, otherwise supply noise will translate to phasenoise......
Good Stuff, and yes I have a Looooooooooooong attention span.
hahahaha, thanks! :-D
Great info. This video gives me some cool ideas for the kinds of tools that are helpful in electronics. I really appreciate the exceptional knowledge that you are sharing.
As far as attention span, 20 mins is a good soft "max" for a video. Of course, it does depend on the content.
Do you have tools that allow you to show "viewing time" on a individual viewer basis or perhaps "average viewer" basis?
I can see the average view time, not individual. Many people click away after a few seconds because they find out the video is not for them. That really reduces the average view time. At the moment this time is 6m20s which is pretty high based on what I've learned 🙂
@@HansRosenberg74 That's also what I see on my pedagogical videos: A lot of people click away after a few seconds. A small group stays and watches the entire video.
attenuators are such a hack to use instead of 50 Ohm SMA loads... tho I've never entirely understood why they act as such. In my stupid brain, since they have a "characteristic impedance" Z0, they can be considered as a T-line. So they are an open terminated T-line in your setup.Why is there no reflection at the input? You are saying the signal goes thru on way, then backwards from the open terminatiation. Why does it goe thru on way in the first place??? Shouldn't it reflect back immediately?
I'm assuming that a 10 MHz quarter wavelength it's going to be a pretty long transmission line
Ehm, yeah ;-) If I'm not making any errors in calculation here, it would be around 3.75 meters for 1/4 wave at 10MHz assuming 1.5e8m/s signal speed :-) Board material is not that expensive these days but I still think this is going to hurt :-)
Have you ever tried Sonnet Lite? It's a free RF electromagnetic planar structure simulator.
OK! thanks for the tip. I'll have a look at that!
Python can do schmitt charts quite niceliy with the scikit-rf package
I was sure there was a python package, however, I wanted to fully control the appearance of the chart and lines on it so it would be clear in a video presentation like this. It was also nice to learn how to create these lines. I learned a lot :-)
I'm a little confused why you went through a long procedure at 9:30 for calculating the impedance of the quarter wavelength stub rather than just taking the geometric mean of 50 and 100 ohms. Anyway, very nice video! Thanks.
the x-axis is non linear in a smith chart. So that is what all the math is for. If you take the geometric average you would get 75 I assume (if that is what you meant) and we need 70.7. But I'm not a math genius so I may be missing something, an easier solution.....
To get the midpoint you have to calculate the mean in the Gamma domain, and then reconvert to impedance. That is what was done in the video, because X is Gamma as shown below. I think the geometric mean thing is a coincidence, in fact the correct impedance is 70 ohm when you have a Gamma of 1/6 (0.166...) and not 70.7.
X = -1 + 2Z/(Z+Z0)
X = (-(Z+Z0)+2Z) / (Z+Z0)
X = (Z-Z0)/(Z+Z0) = Gamma
I get lost since 2:14
I checked that out. I'm showing a transformer there just to show that you can buy this kind of thing off the shelf, but I'm not actually using it in this video. Sorry to confuse you.
sir I will probably enroll to your course If you make it more affordable. 1euro = 40x our currency , so it becomes huge cost for me. Thank you for the great video
I know, there is a huge difference in income around the world. My country is quite expensive. Up till now I've invested around 6000 euro's in this channel and 5 months of costs without income (I'm doing this full time). I assume the total costs to be 15000 to 18000 euro's. So I have to make that back, and more so I can build up a pension and pay for unexpected costs like my car breaking down. You know how that goes ;-) Reality of life :-) . When I can cover my costs I will look for cheaper options in the future, maybe a book? Or mini-courses. This program I'm making now is really big, but that also makes it really valuable. Best regards, Hans
wow your english fluent not sound like a dutch
haha, really, thanks :-)
first ❤😂
Hahaha, congratulations 🎉🎉🎉