Відео

Link is in the description. Best regards, Hans

Please really helps :-) Link is in the description. Best regards, Hans

Link is in the description. Best regards, Hans

Link is in the description. Best regards, Hans

Hi, powerplanes are only really needed if currents are way over 10A. I would use wide traces for supplies. That limits the resistance and inductance. A 2.5mm trace has no problem dealing with 10A. The most used stackup is s-g-g-s. Where s=signal, g=ground.

@@HansRosenberg74 Thanks a lot for your quick answer. I ask this because, a PCB with many ICs, there is the posibility to have many VDD pins and it is easier to have a VDD plane for a quick connection, but there is a better way to do this? Thanks again

Link is in the description. Best regards, Hans

Thanks a lot! Link is in the description. I'm also building a course at the moment that takes it to a whole other level. Best regards, Hans

I didn't study how to make this thing more wideband. I think that you could use the same trick that I'm using in the wilkinson combiner video, multistage 1/4 wave transformers. That would mess up the phase however, unless you go for 3-stage transformers, then you end up at -90 degrees instead of 90 degrees for each section which should work in theory. However, I'd test this first in a simulator. Using google might reveal how to make these things wideband although this kind of info is usually hard to find.

I don't think so. This structure will hardly radiate either so I don't think it would greatly improve. The difficulty of a multilayer would be how to get the signal from the inner layer to the top layer where you can connect it to an sma connector without problems. Or you could use a through-hole sma connector and directly connect to the center-pin, but you might have to control the ground on all layers to make sure that you don't add a large parasitic capacitance there. All in all, this approach is safer. And, if you seriously want to use one of these, it should be in a metal case. It's not needed, I did some very accurate measurements on a radar system with one of these without a case, just make sure no metal gets close to it.

@@HansRosenberg74 thanks for the very comprehensive answer :)

Thanks a lot!

You're welcome. Link is in the description. Best regards, Hans

Thanks a lot. Link is in the description. Best regards, Hans

I would just google it, I don't have a specific recommendation

Hi, thanks for the ideas. I don't have 2 signal generators so I cannot do IP3. For noise I was a bit worried about the accuracy. Matching would have been an easy one, I forgot about that one!

I think electromagnetically, the current might travel on the inside of the ring, since that would be the path of least inductance. But I'm not sure if the ring is maybe too big to cause that. It probably follows the inner diameter because of smallest resistance and inductance. the error is most likely caused by the FR4 not being very accurate with its Er (dieelectric constant)

Link is in the description. Best regards, Hans

you're welcome :-)

Haha, yes, it's in between a checklist and a book for sure :-) If you're sure you never want to buy one of my courses, you could unsubscribe. I don't spam a lot. Once a month, but it will be a bit more when I launch my course in April. Those courses are hopefully my way to pay for making these videos and earn a living. Hope the 'booklist' helps you with your projects :-)

Thanks!

Link is in the description. Best regards, Hans

This is a new feature of youtube, I'll look into it, thanks.

UA-cam is really dumb.

Thanks for the tip!

Link is in the description, best regards, Hans

That's a nasty one. I"ve learned to always triple check those kinds of nasty things, hope you find the right footprint!

You're welcome!

hahaha, yeah, I was waiting for someone to make that remark :-) Luckily I don't take myself too seriously :-) (I hope). These testboards are soo cheap and I had so many circuits on them that I did not spend my normal time double checking everything :-). Thanks for your sign-up. I've just finished the first 2 complete modules of the course, I think I'll get to roughly 15 modules, I have 12 prepared already. It is way way way more work than i thought, but I'll get there. Hopefully many people will buy it otherwise I'm quite screwed with 8 months of work into it :-)

I am curious about that mature product. Can you share what kind of product it is?

@@HansRosenberg74 Yeah, makes sense. I've done that as well with JLC boards. As long as its not a board for a client of course. Regarding the course - Have you considered splitting the modules in different courses? The stated price is quite steep and a lot of people may be put off by it. That way if someone cant afford it all at once, they can spread out the expense over several months. Obviously do your own business analysis on this one, but at first glance it seams like it will be more profitable that way.

@@HansRosenberg74 The product is a NFC smart card reader for payment applications. I cant go into too much details, but the problem was with a filter on the output. Two inductors from different filters were coupling to each other due to bad layout. After watching your video on the filter design I put two and two and two together and I got quite a nice performance improvement. The certification process for those products is quite strict and usually all of the products on the market are on the edge of not passing. This really helped with the last certification we had to go to and its the way we will do the layout for this part of the reader in all future products.

Hi Nicola, I'm definately thinking about that. I don't know who my customers will be. For a professional (boss pays) this is a bargain. This will save a lot more than the costs of the course. For people who buy this privately, it is more expensive relatively. Allthough it depends, if you can boost your knowledge 10 years with this course, it might be worth it from a salary point of view. I can definitely make modules, i've already set it up with different modules. I'll just have to see how it goes.

Haha, thanks a lot

Glad you liked it

Thanks!

Link is in the description. Best regards, Hans

Link is in the description. Best regards, Hans

You are very welcome

What you could do is calculate the current that you want this thing to carry. Then you can plug this effective current into a trace calculator. So an 0.74mm trace with 17um thickness (or 35 if you use 1oz copper) and see how much temperature increase you get. 30-40 Degrees should not be a problem at all. You can then increase the current and see how bad it gets. You also lose 0.9dB max, this is a percentage of the power, say 10% roughly. So if you plan to put 1000 watts through here, expect 100 watts in loss (that would pretty surely fry this thing :-) ).

Actually, up to 5 GHz, fr4 works fine. Ok, you have a bit more loss, and the Er isn't constant across the pcb, causing a bit of ringing of the signal on transmission lines (the impedance shows small waves / ringing). However, if you just want a pretty good coupler, combiner or hybrid, it works good enough for most applications.

You have to use 90 degrees. If you don't, the 1/4 wave transformers won't work well and the whole thing comes apart.

Well, it's good that there is a correlation between price and quality :-) I am using cheap connectors, but they are a pain to solder (the solder gets repelled by the connector surface, if you look closely the soldering looks messy) and I have to tighten the cables really well otherwise I get an unreliable connection. I tried rework flux, but that also does not help the solder to flow. The upside is: 10 connectors are 2$ :-)

You're welcome!

These are exceptions. When you're working with such a high bandwidth feedback circuit, even the smallest amount of parasitic capacitance can ruin the phase margin, bringing it closer to instability. Now these gaps are very very small, so the frequency where they would cause problems with the return current path would probably be above 50GHz so that is nothing to worry about. Great question!

@@HansRosenberg74 Huh, so that is an exception. Do you think that might be a good Flawless PCB Design video topic? Where would you cut the ground plane? In the middle of the baluns like in the AD8352 datasheet? Or perhaps after the 25 ohm resistors?

@@cdroidno8 I would follow the datasheet.

Haha, maybe you should buy some lottery tickets :-) Usually everything works when you do this, just with reduced performance. And often when you need to pass EMC testing the problems show up. However, if you don't qualify a design, you'll never see this.

You're welcome! Link is in the description. Best regards, Hans

Link is in the description. Best regards, Hans

Haha, thanks a lot :-) You can find the link to the checklist in the description. I'm also building a (not free) course based on this material that is much more detailed and will show how to apply all this to a practical design. www.hans-rosenberg.com/product_development_course . Best regards, Hans

Hi Andrew, thanks for your compliments :-) Those traces are not 50 Ohms, that would be a bit wide. Furthermore, at these low frequencies the traces are not yet transmissionlines. The filter is meant to operate up to 100-200Mhz. The 0.25Wavelength at 1GHz is around 40mm which is the size of the filter. So reflection problems would not start untill 1GHz. Link to the checklist is in the description. I'm also building a course at the moment. www.hans-rosenberg.com/product_development_course .Best regards, Hans

You're welcome 😊

You're welcome 😊

Link is in the description. Best regards, Hans

You're welcome. Link is in the description. Best regards, Hans

Please always helps ;-) Link is in the description. Best regards, Hans

Link is in the description. Best regards, Hans