@@IMSAIGuy Chinese New Year Wed, Jan 29, 2025 Year of the snake is coming. The phrase is Cantonese and basically means "Wishing you prosperity and wealth."
@@atakaravelioglu301 Yes. If those itty bitty SMD inductors are replaced with hand-wound air core inductors for instance, the bandpass shape will be better and the loss will be a lot less.
I'd like to look at S11 too. I'm worrying if trace width was calculated with the mask on top or just a bare board. I woukdn't guess that much change at VHF but the RF gods may be angry.
Sometimes its good to add a through with just your trace. SMA -> "50" trace -> SMA to measure the actual trace impedance and loss. You see this on dev boards quite often. Also, you have some stubs on the L and C, you could put the pads directly on the trace. This might help.
That's what I figured, pretty low frequency. Maybe try just putting 0 ohm resistors on just a bare board and measuring the trace loss. Would give you a good idea if it's your trace or something else.
That's kinda cool, I suspect it would be a tricky algorithm but I'm wondering if it could be coded so you feed it with a clean signal from a generator and it auto tunes for max output
I was wondering that too so I went back to the pinout in the first video and it looks like VDD and/or SCLK. No idea what the capacitance would be without power or programming. It seems like if you put a dab of solder paste on that trace and heated the board from the bottom that it would pull it under since solder flows toward the heat.
Don't be afraid of 0.5mm pitch, even with 0.2mm (probably 0.15) spacing QFNs with modern fabricators PCB you got enough space for solder mask between pads, so flux will do its job just fine. And with a stencil should be a walk in a park..
Each stage has a slightly different centre frequency. When you tune all four you should be able to get them to agree.
The parasitics may play a role.
Thanks for sharing information for many years. The red card looks very nice. Btw, I also have added my dogs on many of my cards for many years.
Congrats on the PCBway. Chinese New Year is coming up soon. Gung hay fat choy!
what day? and what animal
@@IMSAIGuy Chinese New Year Wed, Jan 29, 2025
Year of the snake is coming.
The phrase is Cantonese and basically means "Wishing you prosperity and wealth."
Great job with the PCB design and build. Excellent. I'm also curious what you will share with us about the filter insertion loss.
Did you short the two SMA cables, just as a sanity check? Maybe you forgot a 10 dB pad or reference settings somewhere...
yes, I did a thru cal
Congratulations. I was just thinking how much you've stepped up your pcb design projects.
10 db is a LOT of loss! Especially for that poor of a filter.
Because of low Q SMD inductors.
@@atakaravelioglu301 Yes. If those itty bitty SMD inductors are replaced with hand-wound air core inductors for instance, the bandpass shape will be better and the loss will be a lot less.
One possibility is that the PCB input / output is not properly designed to match a 50 ohms load.
I'd like to look at S11 too. I'm worrying if trace width was calculated with the mask on top or just a bare board. I woukdn't guess that much change at VHF but the RF gods may be angry.
Did you use the small hot plate you made a video on? I’m thinking of ordering one
No. Used my oven
@@IMSAIGuy I gather the wife was out shopping that day.
@@Enigma758 I have an IR reflow oven, not the wife's
Am I too much of a dainty flower if I complain about how I don't like boards with sharp corners? :) :) :)
are you still using those rounded scissors? 😎
Nice. But I was hoping to see the hotplate solder porn!
Sometimes its good to add a through with just your trace. SMA -> "50" trace -> SMA to measure the actual trace impedance and loss. You see this on dev boards quite often. Also, you have some stubs on the L and C, you could put the pads directly on the trace. This might help.
at 150MHz those are not active stubs
That's what I figured, pretty low frequency. Maybe try just putting 0 ohm resistors on just a bare board and measuring the trace loss. Would give you a good idea if it's your trace or something else.
@@redrighthand9122 trust me, the trace is not the problem. I've done lots of microstrip designs to many GHz
That's kinda cool, I suspect it would be a tricky algorithm but I'm wondering if it could be coded so you feed it with a clean signal from a generator and it auto tunes for max output
I'd want to know more about the missing pin/pad.
What’s missing?
I was wondering that too so I went back to the pinout in the first video and it looks like VDD and/or SCLK. No idea what the capacitance would be without power or programming. It seems like if you put a dab of solder paste on that trace and heated the board from the bottom that it would pull it under since solder flows toward the heat.
Don't be afraid of 0.5mm pitch, even with 0.2mm (probably 0.15) spacing QFNs with modern fabricators PCB you got enough space for solder mask between pads, so flux will do its job just fine.
And with a stencil should be a walk in a park..