EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: suspension on December 28, 2022, 12:05:07 pm
-
Hi all
I have a requirement to build a passive low pass filter with passband of 15MHz and a stop band attenuation of -70DB at-least at 30MHz. Can someone who has experience in this area please advise me if it is possible to build such a filter with cheap inductors/caps available from aliexpress which do not have any data sheet?
Thanks
Sus
-
Hi all
I have a requirement to build a passive low pass filter with passband of 15MHz and a stop band attenuation of -70DB at-least. Can someone who has experience in this area please advise me if it is possible to build such a filter with cheap inductors/caps available from aliexpress which do not have any data sheet?
Thanks
Sus
What is the frequency that you require >70dB attenuation at?
The likelihood of success largely depends on how rapidly you want the filter to roll off to its stop-band attenuation. A filter that cuts off very rapidly will require high quality, precision components and you may not have much success obtaining such components from Aliexpress. However, a filter that has a much more relaxed roll-off rate will be more tolerant of lesser quality components.
-
I updated the question with that info. -70DB at 30MHz is what I am looking for.
-
Try this website: https://rf-tools.com/lc-filter/
Chebyshev topology, 6th or 7th order, choose "Standard" component values using E12 series.
PS: -70dB at 30MHz with a -3dB point at 15 MHz is very, very steep. It will take a lot of experimentation to achieve that with random Aliexpress components. I hope you have access to a VNA.
-
70db per octave is a pretty tall order for a passive filter. If you don't mind my asking what is the use case?
-
honestly, -70DB was bit arbitrary. All I want is to satisfactorily filter out 2nd+ harmonics from a signal which can have a max fundamental frequency of 15MHz. Apart from harmonics, I want to remove a 50MHz clock signal and its harmonics (which are clearly visible in scope and SA) as well.
I used a filter designer software and created an elliptical filter of 7th order and built it with aliexpress components. However, the response was nowhere close to a low pass. A forum discussion in a different site made me to believe that self-resonance is kicking in with inductors and caps making the filter useless. My only option seems to be to use branded components with low tolerances. However, many Chinese equipments seemingly built with cheap unbranded components have this sort of filters with satisfactory performance . I thought of posting this question here to get some idea from people who may be experienced in this type design process.
-
What you experienced is filter response element sensitivity, the Elliptical types are very sensitive to component variations/features. Chebyshev and Inverse Chebyshev are also sensitive, Butterworths & Bessels are less sensitive and can usually tolerate reasonable component variations, however have shallower transition to stopband.
Generally the steeper the passband to stopband transition the more sensitive to component effects. If you can tolerate a higher insertion loss then a pair of lower order LPFs separated by a attenuator can realize a good passband to stop band transition with standard components. You can't just cascade passive filters as they need to see a lossey source and load impedance to operate properly and why the attenuator between the two filter sections is necessary.
Best,
-
wouldn't the sensitivity of component values result in a filter that has different stop/pass band frequencies but still has a shape of a low pass? In my case however, when plotted with a SA, I see a wired curve which looks arbitrary and has no resemblance to a low pass/hi/band pass filter. All I see is a response curve which reduces to around -20DB around 8-12MHz and increase to around -17dB above 30MHz and stays there untill 100MHz.
-
Depending on how far off the components are from "ideal" and what other artifacts they inhibit (SRF, ESR, ESL), and the type filter, you may see all kinds of strange behavior.
As mentioned the Elliptical types are very component sensitive, they utilize transmission zeros in the transfer function to create the stepper stop to passband transitions, which tend to be very component sensitive in frequency location in the stop band and thus have a strong effect on the pass to stopband transition, not to mention how "deep" the transmission zeros achieve.
You likley will achieve a better representation of your simulated ideal filter response with a Butterworth or Bessel type. We utilize the Butterworth filter types often, because it usually tolerates non-ideal components well and these often can be standard values without significantly altering the filter response.
Best,
-
Thanks, will try out a Butterworth or Bessel type and see.
Do you think self-resonance can come in to play at these frequencies/inductances/capacitances (2.2uH, 50-500pf) ? I guess if that's the case, even Butterworth or Bessel types wouldn't give desired results?
-
Backlink for those curious, https://electronics.stackexchange.com/questions/647769/elliptic-filter-implementation-deviates-from-the-design-and-simulation (https://electronics.stackexchange.com/questions/647769/elliptic-filter-implementation-deviates-from-the-design-and-simulation)
At least, it sounds like the same question.
A good idea is to find models for all components, as well as you can; and tweak the SPICE model until it matches your observations. The problem with inductors of that type is they probably have a low SRF so you're dealing with not just equivalent R and C but higher order (series resonant or other TL like effects) modes too. And power inductors, if they have models or characteristics at all, likely don't show those higher modes at all. (For example, Coilcraft parts typically end just before or just after the first (parallel mode) SRF.)
Like, I made this filter,
(https://www.seventransistorlabs.com/Images/FilterBreadboarded.jpg)
Left side is a VCO, the right 10k 1/4W axial resistor is R8 and so on to the output at the right. So it's filtering the VCO output.
Equivalent circuit,
(https://www.seventransistorlabs.com/Images/FilterCircuit.png)
with design response,
(https://www.seventransistorlabs.com/Images/FilterResponse.png)
or zoomed out,
(https://www.seventransistorlabs.com/Images/FilterResponse1G.png)
Notice that copper is removed around L2 and L3, and gimmicks have been added (bent wires, bits of copper clad). C4 is the varicap.
C6-C2-C5 is an unavoidable capacitor divider, necessary to get the lowpass edge but the stray capacitance (along with the inductors' capacitances) causes an impedance transformation. In this case that's helpful, because Q1 output impedance is quite high, and it's just signal level so I can get a bit more level out that way. (The impedance at VC1 is, err what was it, a kohm, or more like 10? I forget.) The impedance at Q2 base is fairly low (50 or 500 ohms? I forget, again); it's also adjustable by R13, if you don't mind the change in output gain as well).
I don't have a proper spectrum of it but this was the amplitude while sweeping the VCO over its range. So the axis is nonlinear, but you can see the peaks and relative flatness:
(https://www.seventransistorlabs.com/Images/FilterSpectrum.png)
Hmm, that must've been before I got my spec... circa 2017. Hmm no, I got that 2015, why didn't I plot it that way?... :shrug:
Anyway, as you might guess, this model is quite a bit different from what you'd have designed normally. This is at 100MHz so of course RF type inductors are used (SMT chip air core) and their behavior is pretty simple, but at somewhat lower frequencies you'll still want to use similar types -- air core is still suitable (give or take Q, you might be looking at needing quite high Q here) or you might use ferrite or powder types, usually still in single layer windings, chip style.
Tim
-
Thanks, will try out a Butterworth or Bessel type and see.
Do you think self-resonance can come in to play at these frequencies/inductances/capacitances (2.2uH, 50-500pf) ? I guess if that's the case, even Butterworth or Bessel types wouldn't give desired results?
You are going to need a fairly complex 12th order Butterworth filter to meet your stop-band specs but meeting those same specs with a Bessel response is going to be practically impossible due to the very slow roll-off rate of a Bessel filter.
Self resonances will still affect the response but if you can chose your inductors such that their SRF is well above your desired stop-band then that effect should be minimised in your frequency range of interest. However, continuing to increase the frequency may result in the filter attenuation actually reducing due to SRF effects where your inductors stop behaving like inductors and start behaving more like capacitors and start passing the higher frequencies. This may be a problem if you still have significant frequency components well above your stop-band that you still need to get rid of.
-
I also suggest you think of avoiding magnetic coupling between the inductors of the filter. I've seen discussions here in the forum about that aspect, keep the individual filter stages well separated. For example, don't put the inductors all in a straight line, instead angle them 90 degrees against each other or apply other tricks to isolate the stages.
-
Interestingly, that might help, since it does a similar thing to the capacitors. It's not going to be a calibrated amount, though(!).
Tim
-
Just to understand this bit more, I built a simple air core (not surface mount) inductor (See below picture) and measured its inductance using the method described here: https://www.youtube.com/watch?v=iQQe8uSZ8xc. (https://www.youtube.com/watch?v=iQQe8uSZ8xc.) This gave me roughly 0.000887808 mH. Used this inductor in the following simple low pass filter and simulated. Simulation shows a fairly ok low pass response. In this circuit, R1 represents the internal 50 Ohm resister in function generator, R2 represent the input impedance of my TinySA spectrum analyzer. C1 is the external cap.
Buit this circuit and plot the response using TinySA and this also does not show any low pass profile. (Please see the photo).
I am not sure if this is also due to low SRF of the inductor I built?
Would the inductor with digikey part number LQG15HZ2N2S02D work for my purpose?
-
When you experiment with circuits in the range of RF frequencies that you are playing with, you need to minimise stray capacitance & stray radiation.
The input to the circuit needs to be fed with 50 Ohm coax, not unscreened leads. Same with the output lead to the SA. Reduce those unscreened leads to an absolute minimum, or even better, use a piece of coax soldered to the board with an SMA plug on the other end. Get rid of that big piece of copper & solder all earth connections to 1 point.
-
Hello there,
The actual success of this project as stated seems very unlikely. Perhaps with relaxed specifications.
A 70db cut from a frequency F to a frequency 2*F (as you stated 15MHz to 30MHz) means a 70db per octave roll off which is probably impossible.
Maybe using two or more filters with buffers in between might work.
You have to realize that a first order filter gets you -6db per octave and Nth order filter gets you N*6db down, so considering 1st is -3db and 70db would require 67db that would take you to around an 11th order filter to get -70db down at one octave.
The clock frequency is a little easier to get rid of but not much. It's roughly 4 times the fundamental which is better than 2 times.
Using a 2nd order filter gets you down to about -24db and -20db is usually good enough.
You have to realize that -70db is a cut by a factor of about 1/3000, while -20db is a cut of about 1/100. Big difference.
Some of the DSP type sine wave generators use passive filtering on the output to get a cleaner sine wave from a PWM digitally generated signal. You could check into those to see how they manage the filter design which is higher order. They post the schematic online but it is interesting to see how they did the component layout also on the PC board. If you cant find one online i'll look up a part number for the board and post it here.
-
Note that your circuit looks more like this, with ballpark values added, and some coupling between the pairs of inductors (say k=0.1 for the two on the right, k=0.3 on the left??). Further, they will be good antennas, and pick up ambient fields, of which, you're probably seeing commercial FM or TV in the 100MHz range? That will mostly be picked up by the longer wires, then the shorter ground wire acts as an impedance divider to it, of course still bringing plenty of signal into the receiver.
The ground node is meaningless, just a reference node for SPICE purposes of course; the voltage on that PCB with respect to the surroundings may involve common mode signals from both the source, and induced from ambient fields.
Ground becomes more meaningful if you solder that coax connector right onto the board, all ground pins touching it; or better yet, on its side so a complete face is joined all around. That essentially eliminates the ground side inductor, and the signal can be routed close to the board, greatly reducing coupling from ambient fields, greatly increasing CMRR, and giving its length a more controlled impedance (round wire over ground plane, at a height equal to typical hookup wire insulation, gives something like 60 ohms characteristic impedance).
Tim
-
Thanks everyone for all the valuable input.
I still do not understand why none of these problems seem to arise for the circuit explained in the following video.
https://www.youtube.com/watch?v=04GbWVfJGZk (https://www.youtube.com/watch?v=04GbWVfJGZk)
He gets a nice looking low pass response with -3dB at 7MHz. He does not even know the value of his inductors and from which material the core is constructed.
-
Hello there,
The actual success of this project as stated seems very unlikely. Perhaps with relaxed specifications.
A 70db cut from a frequency F to a frequency 2*F (as you stated 15MHz to 30MHz) means a 70db per octave roll off which is probably impossible.
Maybe using two or more filters with buffers in between might work.
You have to realize that a first order filter gets you -6db per octave and Nth order filter gets you N*6db down, so considering 1st is -3db and 70db would require 67db that would take you to around an 11th order filter to get -70db down at one octave.
The clock frequency is a little easier to get rid of but not much. It's roughly 4 times the fundamental which is better than 2 times.
Using a 2nd order filter gets you down to about -24db and -20db is usually good enough.
You have to realize that -70db is a cut by a factor of about 1/3000, while -20db is a cut of about 1/100. Big difference.
Some of the DSP type sine wave generators use passive filtering on the output to get a cleaner sine wave from a PWM digitally generated signal. You could check into those to see how they manage the filter design which is higher order. They post the schematic online but it is interesting to see how they did the component layout also on the PC board. If you cant find one online i'll look up a part number for the board and post it here.
Thanks for pointing out this. Yes this may be too steep. But what I am trying out now is a pretty simple 2nd order low pass filter with one L and one C just to see a low pass response without much regard to the actual slope. Still what I get is garbage - mostly a random response.
My actual use case is actually to Filter out various components from a DDS generated sine wave. This includes quantization noise, phase truncation noise, clock signal and its harmonics, non-linearities in DAC. If you can post a schematic of an output stage of such a system, that would be great.
-
Note that your circuit looks more like this, with ballpark values added, and some coupling between the pairs of inductors (say k=0.1 for the two on the right, k=0.3 on the left??). Further, they will be good antennas, and pick up ambient fields, of which, you're probably seeing commercial FM or TV in the 100MHz range? That will mostly be picked up by the longer wires, then the shorter ground wire acts as an impedance divider to it, of course still bringing plenty of signal into the receiver.
The ground node is meaningless, just a reference node for SPICE purposes of course; the voltage on that PCB with respect to the surroundings may involve common mode signals from both the source, and induced from ambient fields.
Ground becomes more meaningful if you solder that coax connector right onto the board, all ground pins touching it; or better yet, on its side so a complete face is joined all around. That essentially eliminates the ground side inductor, and the signal can be routed close to the board, greatly reducing coupling from ambient fields, greatly increasing CMRR, and giving its length a more controlled impedance (round wire over ground plane, at a height equal to typical hookup wire insulation, gives something like 60 ohms characteristic impedance).
Tim
I've actually simulated this with LTspice and still get a nice-looking low pass response (please see attached). But in the actual physical circuit, the response is random pass. That is what I do not understand :(
-
...
But what I am trying out now is a pretty simple 2nd order low pass filter with one L and one C just to see a low pass response without much regard to the actual slope. Still what I get is garbage - mostly a random response.
...
Looking at your construction practices, I assume this RF stuff is new to you. You may want to have a look at something like the ARRL handbook where they show various construction practices.
Here is a simple LC filter you asked about. I just grabbed an inductor and cap out of the drawer and stuck them onto a breadboard. There was no attempt to calculate or measure the components and the VNA isn't calibrated. The only goal is to show you a simple LCish network acting as a lowpass.
I use the term "ish" because while it looks simple, there's a lot going on here. To give you some clue as to the effects of construction, in the attached video I am using a low cost VNA to attempt to look a simple RLC network. It's built on a this same breadboard. At 33:30, I show the network being measured. I then cut the leads and remeasure. Note the difference in results.
https://www.youtube.com/watch?v=Y8ouApeex78 (https://www.youtube.com/watch?v=Y8ouApeex78)
-
Thanks everyone for all the valuable input.
I still do not understand why none of these problems seem to arise for the circuit explained in the following video.
:-DD He literally turns off the lights and the noise floor drops 20dB! What a terrible setup, huh?! (With solidly grounded coax connectors, and maybe some shielding over the filter, that would completely go away.)
I don't know why he doesn't know the material; yellow-gray or yellow-clear is a standard Micrometals part (and most other manufacturers seem to follow the same color coding): https://www.micrometals.com/products/materials/pc/ (https://www.micrometals.com/products/materials/pc/) and https://www.micrometals.com/products/materials/rf/ (https://www.micrometals.com/products/materials/rf/) I would guess #35 would give way too much inductance and loss, so it's probably the #6. Which I think he said was in the kit? So it would be weird if it wasn't?
And the valley at, what, 13MHz, then it rebounds! That's clearly due to the capacitor lead length, and maybe something about the clip leads connecting to it. It would be even worse if expanded to show the response at higher frequencies -- not necessarily more transmission, but more lumpy (uncontrolled) response.
The prototype also seems to be pretty gentle, Butterworth or so, plus the losses of the inductors (notice the 1dB or so insertion loss in the transition band i.e. 7MHz +/- a few times), which makes it more tolerant of errors. The outermost capacitors are way out of tolerance, even! A return loss plot would be more illustrative, probably showing where some of that insertion loss is coming from (i.e. it's reflecting off the filter rather than being passed through); best tuning is had by measuring return loss.
Mind, I'm just skipping through here, I don't feel a need to watch 28 minutes of something I've done many times before. Let me know any highlights to look at.
Tim
-
Outside of maybe learning some of the basics, you don't normally want to mix RF with breadboards. Here's a low power 12MHz Chebyshev 5th order lowpass using three inductors and two capacitors. It never had a purpose other than for demonstration and I never took the time to try and improve the ripple. It has a nasty dip. The filter is made using surface mount parts and is roughly 15mm by 10mm.
-
Hi all
I have a requirement to build a passive low pass filter with passband of 15MHz and a stop band attenuation of -70DB at-least at 30MHz. Can someone who has experience in this area please advise me if it is possible to build such a filter with cheap inductors/caps available from aliexpress which do not have any data sheet?
Thanks
Sus
I think you have to define what you mean by cheap. If basic 0805 C0G SMD caps are allowed and Micrometals powdered iron toroids are allowed then 70dB rejection is going to be really easy to achieve at 30MHz whilst maintaining a low insertion loss. 90dB rejection at 30MHz would require an 11th order Cheby filter. You could expect to see about 0.7dB insertion loss at 15MHz according to a filter simulation based on 0805 C0G caps and toroid inductors. This is for an 11th order cheby filter.
The more challenging part would be maintaining a 70dB stopband performance out to (say) 1GHz. This would require a decent layout and probably a tiny roofing LPF if you were aiming/hoping for >70dB stopband performance out past 1GHz.
-
I would start by proving your setup. Maybe just measure a thru to start with, forgetting the filter and make sure the generator and SA provide a flat responseish.
Here's another breadboard filter. Chebyshev, 9th order, series, 0.1dB ripple. No attempt to trim the leads or improve the layout. I swept it to 100MHz, then to 2.4GHz. Construction practice would greatly improve it.
-
Here's another Chebyshev, 9th order, shunt, 1.0dB ripple (based on parts I had), built onto a cheap Chinese attenuator board. All 0805 parts, nothing exotic. Inductors are from Coil Craft. I swept it to 30MHz then to 300MHz. Fc is a bit off but fairly close to the target. Of course, you hook this to your 200W ham radio, bad things are going to happen. Lot's to consider.
-
...
But what I am trying out now is a pretty simple 2nd order low pass filter with one L and one C just to see a low pass response without much regard to the actual slope. Still what I get is garbage - mostly a random response.
...
Looking at your construction practices, I assume this RF stuff is new to you. You may want to have a look at something like the ARRL handbook where they show various construction practices.
Here is a simple LC filter you asked about. I just grabbed an inductor and cap out of the drawer and stuck them onto a breadboard. There was no attempt to calculate or measure the components and the VNA isn't calibrated. The only goal is to show you a simple LCish network acting as a lowpass.
I use the term "ish" because while it looks simple, there's a lot going on here. To give you some clue as to the effects of construction, in the attached video I am using a low cost VNA to attempt to look a simple RLC network. It's built on a this same breadboard. At 33:30, I show the network being measured. I then cut the leads and remeasure. Note the difference in results.
https://www.youtube.com/watch?v=Y8ouApeex78 (https://www.youtube.com/watch?v=Y8ouApeex78)
Yes, I am new to RF. My main field of work is software and digital systems and this stuff I do for fun. Analog always is more interesting than digital electronics to me.
Looking at your LC filter, the main differences I can spot are the type of capacitor, maybe the value of inductor (Mine was around .88 uH) and perhaps the probes that you are using.
For the cap also, I used a ceramic one bought from aliexpress. And I soldered parts closer to the cap, so length of the leads does not matter much. Can this ceramic cap be the culprit?
-
Thanks everyone for all the valuable input.
I still do not understand why none of these problems seem to arise for the circuit explained in the following video.
:-DD He literally turns off the lights and the noise floor drops 20dB! What a terrible setup, huh?! (With solidly grounded coax connectors, and maybe some shielding over the filter, that would completely go away.)
I don't know why he doesn't know the material; yellow-gray or yellow-clear is a standard Micrometals part (and most other manufacturers seem to follow the same color coding): https://www.micrometals.com/products/materials/pc/ (https://www.micrometals.com/products/materials/pc/) and https://www.micrometals.com/products/materials/rf/ (https://www.micrometals.com/products/materials/rf/) I would guess #35 would give way too much inductance and loss, so it's probably the #6. Which I think he said was in the kit? So it would be weird if it wasn't?
And the valley at, what, 13MHz, then it rebounds! That's clearly due to the capacitor lead length, and maybe something about the clip leads connecting to it. It would be even worse if expanded to show the response at higher frequencies -- not necessarily more transmission, but more lumpy (uncontrolled) response.
The prototype also seems to be pretty gentle, Butterworth or so, plus the losses of the inductors (notice the 1dB or so insertion loss in the transition band i.e. 7MHz +/- a few times), which makes it more tolerant of errors. The outermost capacitors are way out of tolerance, even! A return loss plot would be more illustrative, probably showing where some of that insertion loss is coming from (i.e. it's reflecting off the filter rather than being passed through); best tuning is had by measuring return loss.
Mind, I'm just skipping through here, I don't feel a need to watch 28 minutes of something I've done many times before. Let me know any highlights to look at.
Tim
I completely ignored about up to which frequency he scans the SA. So re-tested my simple LC filter in 100Hz to 2MHz range and yet it gives a nice low pass response down to around -40DB. As you suggested, in the above video the repose is rebounding and will be as bad as mine if he scanned to 100MHz.
-
Outside of maybe learning some of the basics, you don't normally want to mix RF with breadboards. Here's a low power 12MHz Chebyshev 5th order lowpass using three inductors and two capacitors. It never had a purpose other than for demonstration and I never took the time to try and improve the ripple. It has a nasty dip. The filter is made using surface mount parts and is roughly 15mm by 10mm.
This is really nice. Do you mind sharing some more information about the components used (so I can order similar types) and maybe some pics of the construction?
-
I would start by proving your setup. Maybe just measure a thru to start with, forgetting the filter and make sure the generator and SA provide a flat responseish.
Here's another breadboard filter. Chebyshev, 9th order, series, 0.1dB ripple. No attempt to trim the leads or improve the layout. I swept it to 100MHz, then to 2.4GHz. Construction practice would greatly improve it.
I guess this nice response is due to the selection of caps and inductor types with high SRF?
Could I please know which types of caps and inductors these are?
-
Here's another Chebyshev, 9th order, shunt, 1.0dB ripple (based on parts I had), built onto a cheap Chinese attenuator board. All 0805 parts, nothing exotic. Inductors are from Coil Craft. I swept it to 30MHz then to 300MHz. Fc is a bit off but fairly close to the target. Of course, you hook this to your 200W ham radio, bad things are going to happen. Lot's to consider.
Thanks again.
Out of these RF capable inductors available from coilcraft, which type do you suggest?
https://www.coilcraft.com/en-us/products/rf/ (https://www.coilcraft.com/en-us/products/rf/)
My requirement is to be used in an output stage of a DDS to filter out various spurs due to DAC non linearities, quantization noise, phase quantization error and 50MHz clock and its harmonics. Max signal frequency is 15MHz.
And for caps, do you think I can just use regular 1205 caps from aliexpress to get this sort of response?
-
Hello there,
The actual success of this project as stated seems very unlikely. Perhaps with relaxed specifications.
A 70db cut from a frequency F to a frequency 2*F (as you stated 15MHz to 30MHz) means a 70db per octave roll off which is probably impossible.
Maybe using two or more filters with buffers in between might work.
You have to realize that a first order filter gets you -6db per octave and Nth order filter gets you N*6db down, so considering 1st is -3db and 70db would require 67db that would take you to around an 11th order filter to get -70db down at one octave.
The clock frequency is a little easier to get rid of but not much. It's roughly 4 times the fundamental which is better than 2 times.
Using a 2nd order filter gets you down to about -24db and -20db is usually good enough.
You have to realize that -70db is a cut by a factor of about 1/3000, while -20db is a cut of about 1/100. Big difference.
Some of the DSP type sine wave generators use passive filtering on the output to get a cleaner sine wave from a PWM digitally generated signal. You could check into those to see how they manage the filter design which is higher order. They post the schematic online but it is interesting to see how they did the component layout also on the PC board. If you cant find one online i'll look up a part number for the board and post it here.
Thanks for pointing out this. Yes this may be too steep. But what I am trying out now is a pretty simple 2nd order low pass filter with one L and one C just to see a low pass response without much regard to the actual slope. Still what I get is garbage - mostly a random response.
My actual use case is actually to Filter out various components from a DDS generated sine wave. This includes quantization noise, phase truncation noise, clock signal and its harmonics, non-linearities in DAC. If you can post a schematic of an output stage of such a system, that would be great.
That's interesting, that's exactly what the AD9850 board does. See image.
You could look that board up on the web and maybe even buy one to examine the layout and actual parts. About $20 USD.
-
Out of these RF capable inductors available from coilcraft, which type do you suggest?
https://www.coilcraft.com/en-us/products/rf/ (https://www.coilcraft.com/en-us/products/rf/)
My requirement is to be used in an output stage of a DDS to filter out various spurs due to DAC non linearities, quantization noise, phase quantization error and 50MHz clock and its harmonics. Max signal frequency is 15MHz.
And for caps, do you think I can just use regular 1205 caps from aliexpress to get this sort of response?
132-20L seems the most promising single part, if its value were suitable, but it isn't a very large value and is quite large physically (and probably expensive, and maybe not often stocked, I didn't check). 1812CS is probably what you would use otherwise, but the Q is fairly mediocre for something like this.
Not that power inductors are straight out, either, but you will have to search around for high Q and high SRF parts. How high is "high"? More than 4x Fc would seem a good start... otherwise, higher the better.
The required Q factor is at least the highest Q pole in the transfer function; and component Q = pole Q means losing 3dB on that pole alone (hand waving). Component Q say 10 times higher, means losing about 1/10th the signal or a dB, so you can need quite high Q to get a sharp and low-loss response.
Note that the transfer function can be modified to account for losses. The design Q can be raised to compensate for the losses, effectively pushing the "droop" from the transition band into the stop band; but this requires even higher Q, so can only be done so far within limited component Q. Filter tables are available for fixed-loss inductors: see the classic Zverev; or Williams and Taylor, Electronic Filter Design Handbook, 4th Ed.; or probably many others. Or you can just putz around with the values yourself until the response looks right (preferably in sim, then back and forth IRL as well), but this is increasingly difficult with higher order filters.
Capacitors: C0G ceramic, or PP or PPS film. It should be hard for anyone to screw these up (but never underestimate cheap suppliers' talents to do so), a Q of a few hundred would be embarrassingly low. Inductors are by far the dominant loss element. These types are also quite stable (C with respect to voltage and temperature); inductors also dominate the drift / tolerance error of the design.
Type 2 dielectrics (X7R, Z5U, etc.) are poor, typically with a Q of 10-30; plenty for supply filtering, and can be used for signal filtering in a pinch, but clearly this is quite limited as far as sharpness. They also introduce distortion(!). If you need bigger values and your eyes start to water from C0G prices, use film instead, or an active filter (RC + opamps, can use smaller capacitors at higher impedances -- typically for signal purposes under some MHz).
Tim
-
Here's another Chebyshev, 9th order, shunt, 1.0dB ripple (based on parts I had), built onto a cheap Chinese attenuator board. All 0805 parts, nothing exotic. Inductors are from Coil Craft. I swept it to 30MHz then to 300MHz. Fc is a bit off but fairly close to the target. Of course, you hook this to your 200W ham radio, bad things are going to happen. Lot's to consider.
Thanks again.
Out of these RF capable inductors available from coilcraft, which type do you suggest?
https://www.coilcraft.com/en-us/products/rf/ (https://www.coilcraft.com/en-us/products/rf/)
My requirement is to be used in an output stage of a DDS to filter out various spurs due to DAC non linearities, quantization noise, phase quantization error and 50MHz clock and its harmonics. Max signal frequency is 15MHz.
And for caps, do you think I can just use regular 1205 caps from aliexpress to get this sort of response?
I strongly suggest you start with something much more simple until you have some idea on how to setup such a test and can get meaningful data. As others have suggested, wires hanging out in the air as you have shown at 30MHz isn't the best setup. From my breadboard filters, at these low frequencies you can use some fairly poor construction if you don't require very good performance. I have no idea what you are using for a source. How level is it? How are you compensating for the system errors in your test setup? My guess, your not. it will be difficult to compensate for dangling wires like you show. Again, start with a thru and forget the filter. In other words, just connect your dangling wires to the dangling wires. Make sure you can measure that before you start looking at individual components.
I was your post on the VNA. IMO, you are now heading down the right path. At these lower frequencies, I would stay with the original NanoVNA. They have better performance than the V2+4 and LiteVNA and are much lower cost. If you want to play above 300MHz, I recommend the LiteVNA. Attached showing the last filter being swept from 1M-1G.
Construction wise, I'm sure the breadboards I show are self explanatory. The first cap you asked about was a silver mica. Again, I was not attempting to replicate your setup, but just show a simple LC filter as you asked. The 12MHz 5-pole filter is constructed on a coplanar waveguide. That last one I showed on made out of cheap Chinese attenuator is a microstrip.
Seeing that you are working with small signals I suggest staying with surface mount. Sounds what I have shown with that last filter may be close to your requirements. Capacitors are nothing special. 0603 AVX parts. Sadly as companies are bought and sold, the top management guts them and we loose a lot of the secrete sauce. If you bought parts from ATC today vs 15 years ago, you may not find the same performance. For the Coil Craft inductors, see:
https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/0805-(2012)/0805hp/ (https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/0805-(2012)/0805hp/)
***
Want to be clear that I am not suggesting using these inductors for your project. You had asked what I used. If you look at the datasheet, the DCR for these is over 2 ohms.
-
If the aim is to achieve a DDS output with a low spurious level then I think it's worth mentioning which DDS you are using.
Contrary to most of the advice on here, it isn't difficult to make a nice LPF at 15MHz using cheap parts. I'm guessing there's a lot of inexperienced people here trying to give advice who have close to zero practical experience. I put together a basic LPF at 15MHz using cheap 0805 caps and some tiny toroids and I managed to get the same result as the simulation.
See below. This managed about 100dB rejection at 30MHz (the cyan trace below is scaled at 10dB/div). There is just under 0.7dB insertion loss at 15MHz on the yellow trace. This stuff isn't difficult, there is no voodoo magic.
Generally, it's best to keep the RF output to less than a fifth of the DDS clock frequency if you want low spurious output. What DDS are you using? If you end up using tiny SMD inductors then it might be a good idea to include a passive compensation network to offset the droop in the LPF and the sinx/x rolloff in the DDS response by 15MHz.
-
I suspect they are going to want to see your construction.
It should be clear that to measure 100dB may require you spend a bit more cash. The LiteVNA costs about $120. Attached showing the floor of the VNA compared with that last filter. Note the IF was set to 400kHz which with this VNA takes a fair amount of time to sweep. Still, if the goal is 70, good enough.
-
Contrary to most of the advice on here, it isn't difficult to make a nice LPF at 15MHz using cheap parts. I'm guessing there's a lot of inexperienced people here trying to give advice who have close to zero practical experience. I put together a basic LPF at 15MHz using cheap 0805 caps and some tiny toroids and I managed to get the same result as the simulation.
Actually rather curious now, to which advice you're referring -- there's only, what, ten people in this thread, OP included? And most of the respondents have adequate to expert level experience in this subject, based on what I know of them from here or in other threads; and those that don't, haven't provided much quantitative or detailed information, so can be judged proportionately, if you will. (And, while I can't speak for others -- if it was me, I for one would appreciate a critique of communication / tone / factuality. :) )
Tim
-
(And, while I can't speak for others -- if it was me, I for one would appreciate a critique of communication / tone / factuality. :) )
Critique away!
Time for another bread board bashing and why they should never be used, especially for beginners.... :-DD
-
Contrary to most of the advice on here, it isn't difficult to make a nice LPF at 15MHz using cheap parts. I'm guessing there's a lot of inexperienced people here trying to give advice who have close to zero practical experience. I put together a basic LPF at 15MHz using cheap 0805 caps and some tiny toroids and I managed to get the same result as the simulation.
Actually rather curious now, to which advice you're referring -- there's only, what, ten people in this thread, OP included? And most of the respondents have adequate to expert level experience in this subject, based on what I know of them from here or in other threads; and those that don't, haven't provided much quantitative or detailed information, so can be judged proportionately, if you will. (And, while I can't speak for others -- if it was me, I for one would appreciate a critique of communication / tone / factuality. :) )
Tim
Don't be too curious.... it's just my opinion. For the sake of the thread it wouldn't be wise to comment any more other than to say that some answers were OK :)
What was missing was the reassurance that it is possible to make something fairly decent using cheap parts. See my earlier plot.
Aliexpress do sell MM powdered iron toroids and they do sell 0805 SMD C0G caps. I made my filter using just two cap values to try and make it 'cheap' and lower the risk of having some duff cap values. I could have done it just using 120pF caps (in series and parallel) and Aliexpress advertise 100 x 120pF 0805 caps for £1. That's enough for several filters for just £1. Having multiple caps in parallel can be good for UHF stopband performance as long as the parts are laid out carefully. Having a single cap value also helps with tolerancing issues especially if they all come off the same SMD tape feed.
Of course, I don't know if Aliexpress parts are real or if they are made from dried custard. I used C0G caps and some old MM toroid cores and I managed to get 100dB rejection at 30MHz. I aimed for this in order to counter the claim that 70dB rejection was difficult or impossible to achieve at 30MHz. I also achieved a very low passband ripple and low insertion loss. I didn't have to tweak any cap values to get this performance. If I press the toroids together such that they touch each other in a line (either at 90degrees or all in line) , the worst stopband performance I could get at 30MHz was about 90dB.
I think it would be useful to know what DDS is being used as this can help define the LPF requirements. It could be that having 70dB rejection at 30MHz isn't strictly necessary.
-
Ahh, alright then. The main trouble I think would be with your definition of "cheap parts" -- if one isn't practiced with (nor has the time to spare for...) coil winding, off-the-shelf is obviously valuable; but one will have to go shopping to find parts of comparable quality. And that's about all of that.
And for sure, layout is critical, as the two counter-examples here have illustrated. And, mind, not that the video example was at all terrible with respect to the transition band -- you can use quite sloppy layout and still get okay response around there. It's the loose leads that make a minefield of passband(s) in the (supposed-to- ;D )stop band, and let in interference. And putting everything inside a shield (coax cable, connectors bonded widely to ground plane/enclosure, etc.) handily solves the latter.
This is indeed the frequency range where signals "don't like to stay in wires" -- which is to say, human-scale wiring lengths are sufficient impedance/mismatch/coupling to run afoul of these kinds of symptoms. So, make those errors shorter, and your stopband will be low to a proportionally higher cutoff. (Literally: if 50mm lead lengths give bumps around 30MHz, 5mm lead lengths will give bumps around 300MHz, etc.!)
Also, good to know random cheap C0Gs are good. Like I said, I'd guess it's hard for them to screw that up... they can really only be the wrong material entirely. Well, if you've not measured tempco, or breakdown voltage, I guess that'd be something, but yeh. They're definitely not type 2, that's for sure.
Tim
-
Until they can measure the filter, the component costs and availability seems to be the least of their concerns. Hopefully they can get that part sorted out and still have enough cash to buy some parts.
-
OK, now I'm afraid to comment, but as has been mentioned, generic SMT COG capacitors are pretty good -- at least for low-power levels. I'm using 0603, 0805, and 1206 COG multilayer ceramic caps from suppliers like Murata, Yaego, TDK, for HF and low VHF filters and matching networks in 1W and below designs. The caps are the cheap and easy part, pennies or tenths of a cent each.
Inductors are slightly more difficult. You can wind your own on iron-powder toroid cores (micrometals, etc) and get Qs of 100 or better with good self-resonant frequencies. If you can relax your design a bit you can use fixed surface-mount inductors. I'm using SMT solenoid inductors from Murata like this one: LQW2UAS1R0G0C (about 9 cents ea from LCSC), which is 1uH with a Q of 35 (min, @50 MHz) and a min SRF (Self-Resonant Frequency) of 290 MHz. Since these are unshielded solenoid inductors you need to be careful with layout, but you can build quite respectable filters with them.
You can also get multilayer ferrite SMT inductors such as the TDK MLF2012 parts, but the SRF tends to be lower.
Neither of these inductors will easily give you the filter performance you are asking for, but you might check to see if a less stringent requirement might be in order. You can get very useful performance using fixed-value 5% tolerance, medium Q parts. Low cost and zero tuning are nice things to have.
BTW, sometimes you can use SRF to your advantage, in that it can steepen the roll-off of a LPF or provide a notch, of course with the drawback of worse ultimate attenuation.
-
If the aim is to achieve a DDS output with a low spurious level then I think it's worth mentioning which DDS you are using.
Contrary to most of the advice on here, it isn't difficult to make a nice LPF at 15MHz using cheap parts. I'm guessing there's a lot of inexperienced people here trying to give advice who have close to zero practical experience. I put together a basic LPF at 15MHz using cheap 0805 caps and some tiny toroids and I managed to get the same result as the simulation.
See below. This managed about 100dB rejection at 30MHz (the cyan trace below is scaled at 10dB/div). There is just under 0.7dB insertion loss at 15MHz on the yellow trace. This stuff isn't difficult, there is no voodoo magic.
Generally, it's best to keep the RF output to less than a fifth of the DDS clock frequency if you want low spurious output. What DDS are you using? If you end up using tiny SMD inductors then it might be a good idea to include a passive compensation network to offset the droop in the LPF and the sinx/x rolloff in the DDS response by 15MHz.
Hi there,
Great, now maybe you could show some pics of the layout so we can see what it looks like in real life, and maybe some measurement pics of the performance.
Also, how many caps are you using and how many inductors, and any resistors?
And how about the values of the caps, inductors, and resistors if present?
Of course it goes without saying a schematic would be nice.
Sounds interesting.
-
OK, now I'm afraid to comment, ...
With it being in the beginners section, hopefully the OP doesn't feel the same.
I have no problem showing off Z5Us with long leads stuck in some breadboard and measured with a cheap VNA. IMO, better to sort out some of the other problems first.
-
Hi there,
Great, now maybe you could show some pics of the layout so we can see what it looks like in real life, and maybe some measurement pics of the performance.
Also, how many caps are you using and how many inductors, and any resistors?
And how about the values of the caps, inductors, and resistors if present?
Of course it goes without saying a schematic would be nice.
Sounds interesting.
Hi, the filter isn't anything special, it's just an 11th order cheby LPF. I tried to comply with the requirement that the parts should be available cheaply from Aliexpress and this meant I had to use T37-6 toroids. These are quite big as you can see in the image of the measurement pic below. I would have liked to have used T30 or T27. In this pic I've deliberately put all the toroids in line and pressed them up against each other so they are actually touching each other. The stopband performance isn't so good now but it still achieves about -90dBc at 30MHz as you can see in the image.
I did want to design it using just 120pF caps but this would have used up a lot of my 120pF 0805 caps. I used 270pF and 120pF caps in SMD 0805 packages and obviously I needed a lot of them. The caps in the centre sections of the filter are all 270pF and 120pF in parallel. The outer caps are 120pF in parallel with two 120pF caps in series.
This use of multiple caps may seem crazy, but it helps minimise the impact of capacitor tolerance and it also helps with the stopband performance up at UHF. The number of turns for each toroid was calculated using Micrometals' own software and this usually gives a very good estimate for the turns required. I had to adjust nothing apart from squishing the turns of each toroid slightly to get the flattest passband ripple. There's a VNA plot of the filter a few posts back.
I'm not suggesting that this filter design should be used, as it has more rejection than required. Maybe a 9th order filter could be used. However, I think it's worth checking out the DDS setup before designing the filter. Normally having a 15MHz output with a DDS clock of only 50MHz isn't ideal if the aim is to also have low 'in-band' spurious from the DDS. That's why I'm wondering if the 70dB filtering requirement at 30MHz is really needed. It would be be interesting to know which DDS is being used.
-
Contrary to most of the advice on here, it isn't difficult to make a nice LPF at 15MHz using cheap parts. I'm guessing there's a lot of inexperienced people here trying to give advice who have close to zero practical experience. I put together a basic LPF at 15MHz using cheap 0805 caps and some tiny toroids and I managed to get the same result as the simulation.
Actually rather curious now, to which advice you're referring -- there's only, what, ten people in this thread, OP included? And most of the respondents have adequate to expert level experience in this subject, based on what I know of them from here or in other threads; and those that don't, haven't provided much quantitative or detailed information, so can be judged proportionately, if you will. (And, while I can't speak for others -- if it was me, I for one would appreciate a critique of communication / tone / factuality. :) )
Tim
Don't be too curious.... it's just my opinion. For the sake of the thread it wouldn't be wise to comment any more other than to say that some answers were OK :)
What was missing was the reassurance that it is possible to make something fairly decent using cheap parts. See my earlier plot.
Aliexpress do sell MM powdered iron toroids and they do sell 0805 SMD C0G caps. I made my filter using just two cap values to try and make it 'cheap' and lower the risk of having some duff cap values. I could have done it just using 120pF caps (in series and parallel) and Aliexpress advertise 100 x 120pF 0805 caps for £1. That's enough for several filters for just £1. Having multiple caps in parallel can be good for UHF stopband performance as long as the parts are laid out carefully. Having a single cap value also helps with tolerancing issues especially if they all come off the same SMD tape feed.
Of course, I don't know if Aliexpress parts are real or if they are made from dried custard. I used C0G caps and some old MM toroid cores and I managed to get 100dB rejection at 30MHz. I aimed for this in order to counter the claim that 70dB rejection was difficult or impossible to achieve at 30MHz. I also achieved a very low passband ripple and low insertion loss. I didn't have to tweak any cap values to get this performance. If I press the toroids together such that they touch each other in a line (either at 90degrees or all in line) , the worst stopband performance I could get at 30MHz was about 90dB.
I think it would be useful to know what DDS is being used as this can help define the LPF requirements. It could be that having 70dB rejection at 30MHz isn't strictly necessary.
This sounds great. For my case, I would like to avoid toroid cores if that at all is possible. Do you think I can get similar performance by using something like https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/0805-(2012)/0805hp/ (https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/0805-(2012)/0805hp/) for the inductor? Or better yet, do you think it would be possible to get a similar performance using some of the surface mount inductors available from aliexpress?
The DDS is a custom built one using a Xilinx FPGA. And the DAC used is 16bit one with clock frequency of 50MHz. I would like to get 15MHz sine out with low noise, if possible, but can reduce it to 10MHz max if 15MHz turns out to be rather difficult. I am planning to do various compensations digitally by using Tayler series and inverse sinc filtering. But these are not yet incorporated in the digital design.
-
Here's another Chebyshev, 9th order, shunt, 1.0dB ripple (based on parts I had), built onto a cheap Chinese attenuator board. All 0805 parts, nothing exotic. Inductors are from Coil Craft. I swept it to 30MHz then to 300MHz. Fc is a bit off but fairly close to the target. Of course, you hook this to your 200W ham radio, bad things are going to happen. Lot's to consider.
Thanks again.
Out of these RF capable inductors available from coilcraft, which type do you suggest?
https://www.coilcraft.com/en-us/products/rf/ (https://www.coilcraft.com/en-us/products/rf/)
My requirement is to be used in an output stage of a DDS to filter out various spurs due to DAC non linearities, quantization noise, phase quantization error and 50MHz clock and its harmonics. Max signal frequency is 15MHz.
And for caps, do you think I can just use regular 1205 caps from aliexpress to get this sort of response?
I strongly suggest you start with something much more simple until you have some idea on how to setup such a test and can get meaningful data. As others have suggested, wires hanging out in the air as you have shown at 30MHz isn't the best setup. From my breadboard filters, at these low frequencies you can use some fairly poor construction if you don't require very good performance. I have no idea what you are using for a source. How level is it? How are you compensating for the system errors in your test setup? My guess, your not. it will be difficult to compensate for dangling wires like you show. Again, start with a thru and forget the filter. In other words, just connect your dangling wires to the dangling wires. Make sure you can measure that before you start looking at individual components.
I was your post on the VNA. IMO, you are now heading down the right path. At these lower frequencies, I would stay with the original NanoVNA. They have better performance than the V2+4 and LiteVNA and are much lower cost. If you want to play above 300MHz, I recommend the LiteVNA. Attached showing the last filter being swept from 1M-1G.
Construction wise, I'm sure the breadboards I show are self explanatory. The first cap you asked about was a silver mica. Again, I was not attempting to replicate your setup, but just show a simple LC filter as you asked. The 12MHz 5-pole filter is constructed on a coplanar waveguide. That last one I showed on made out of cheap Chinese attenuator is a microstrip.
Seeing that you are working with small signals I suggest staying with surface mount. Sounds what I have shown with that last filter may be close to your requirements. Capacitors are nothing special. 0603 AVX parts. Sadly as companies are bought and sold, the top management guts them and we loose a lot of the secrete sauce. If you bought parts from ATC today vs 15 years ago, you may not find the same performance. For the Coil Craft inductors, see:
https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/0805-(2012)/0805hp/ (https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/0805-(2012)/0805hp/)
***
Want to be clear that I am not suggesting using these inductors for your project. You had asked what I used. If you look at the datasheet, the DCR for these is over 2 ohms.
I actually measured the direct response (no filter) at the start of the setup. This was done by connecting the out of the function generator directly to the input of tinySA. Set the tiny SA to attenuate to 10db (so that its input impedance become 50Ohms based on its spec). Set the generator to scan from 1MHz to 100MHz linearly for 10s. Setup tinySA to show the same range and used two traces, one with no calculation and other with max hold calculation. After a while I get a nice flat response. So, I guess this measurement setup is ok for this range of frequencies?
Regarding the NanoVNA, out of the two sites https://nanorfe.com/nanovna-v2.html (https://nanorfe.com/nanovna-v2.html) and https://nanovna.com/, (https://nanovna.com/,) which one should I use? The one from https://nanorfe.com/nanovna-v2.html (https://nanorfe.com/nanovna-v2.html) is pretty expensive.
-
Hi there,
Great, now maybe you could show some pics of the layout so we can see what it looks like in real life, and maybe some measurement pics of the performance.
Also, how many caps are you using and how many inductors, and any resistors?
And how about the values of the caps, inductors, and resistors if present?
Of course it goes without saying a schematic would be nice.
Sounds interesting.
Hi, the filter isn't anything special, it's just an 11th order cheby LPF. I tried to comply with the requirement that the parts should be available cheaply from Aliexpress and this meant I had to use T37-6 toroids. These are quite big as you can see in the image of the measurement pic below. I would have liked to have used T30 or T27. In this pic I've deliberately put all the toroids in line and pressed them up against each other so they are actually touching each other. The stopband performance isn't so good now but it still achieves about -90dBc at 30MHz as you can see in the image.
I did want to design it using just 120pF caps but this would have used up a lot of my 120pF 0805 caps. I used 270pF and 120pF caps in SMD 0805 packages and obviously I needed a lot of them. The caps in the centre sections of the filter are all 270pF and 120pF in parallel. The outer caps are 120pF in parallel with two 120pF caps in series.
This use of multiple caps may seem crazy, but it helps minimise the impact of capacitor tolerance and it also helps with the stopband performance up at UHF. The number of turns for each toroid was calculated using Micrometals' own software and this usually gives a very good estimate for the turns required. I had to adjust nothing apart from squishing the turns of each toroid slightly to get the flattest passband ripple. There's a VNA plot of the filter a few posts back.
I'm not suggesting that this filter design should be used, as it has more rejection than required. Maybe a 9th order filter could be used. However, I think it's worth checking out the DDS setup before designing the filter. Normally having a 15MHz output with a DDS clock of only 50MHz isn't ideal if the aim is to also have low 'in-band' spurious from the DDS. That's why I'm wondering if the 70dB filtering requirement at 30MHz is really needed. It would be be interesting to know which DDS is being used.
Hello again,
That's very nice, but we cant see the board or the components as it is just a giant blur. It's even a bit hard to see the measurement scales and values.
Worse yet, no schematic, no specific component values (inductors especially), and how many inductors and capacitors are required, etc.
For someone to reproduce this work they would need to see all this specific info. Also, some impedance matching info might be nice.
Also, i think it would be very interesting to look at the design mathematically but for that we need a schematic and component values.
If you can provide that information this would become much more interesting. For myself the math analysis would be most interesting but i need more info for that.
Why do you say you dont think he would want to use this?
-
Hi there,
Great, now maybe you could show some pics of the layout so we can see what it looks like in real life, and maybe some measurement pics of the performance.
Also, how many caps are you using and how many inductors, and any resistors?
And how about the values of the caps, inductors, and resistors if present?
Of course it goes without saying a schematic would be nice.
Sounds interesting.
Hi, the filter isn't anything special, it's just an 11th order cheby LPF. I tried to comply with the requirement that the parts should be available cheaply from Aliexpress and this meant I had to use T37-6 toroids. These are quite big as you can see in the image of the measurement pic below. I would have liked to have used T30 or T27. In this pic I've deliberately put all the toroids in line and pressed them up against each other so they are actually touching each other. The stopband performance isn't so good now but it still achieves about -90dBc at 30MHz as you can see in the image.
I did want to design it using just 120pF caps but this would have used up a lot of my 120pF 0805 caps. I used 270pF and 120pF caps in SMD 0805 packages and obviously I needed a lot of them. The caps in the centre sections of the filter are all 270pF and 120pF in parallel. The outer caps are 120pF in parallel with two 120pF caps in series.
This use of multiple caps may seem crazy, but it helps minimise the impact of capacitor tolerance and it also helps with the stopband performance up at UHF. The number of turns for each toroid was calculated using Micrometals' own software and this usually gives a very good estimate for the turns required. I had to adjust nothing apart from squishing the turns of each toroid slightly to get the flattest passband ripple. There's a VNA plot of the filter a few posts back.
I'm not suggesting that this filter design should be used, as it has more rejection than required. Maybe a 9th order filter could be used. However, I think it's worth checking out the DDS setup before designing the filter. Normally having a 15MHz output with a DDS clock of only 50MHz isn't ideal if the aim is to also have low 'in-band' spurious from the DDS. That's why I'm wondering if the 70dB filtering requirement at 30MHz is really needed. It would be be interesting to know which DDS is being used.
Hello again,
That's very nice, but we cant see the board or the components as it is just a giant blur. It's even a bit hard to see the measurement scales and values.
Worse yet, no schematic, no specific component values (inductors especially), and how many inductors and capacitors are required, etc.
For someone to reproduce this work they would need to see all this specific info. Also, some impedance matching info might be nice.
Also, i think it would be very interesting to look at the design mathematically but for that we need a schematic and component values.
If you can provide that information this would become much more interesting. For myself the math analysis would be most interesting but i need more info for that.
Why do you say you dont think he would want to use this?
Yes that would be nice. Even though I may not be able to use toroids in the final version, my current interest is to build a low pass filter with whatever (cheap) components available to understand the feasibility.
-
I am thinking of buying these caps from digikey:
https://www.digikey.co.uk/short/zv2n5qnn (https://www.digikey.co.uk/short/zv2n5qnn)
All are C0G Ceremic ones. Similar ones are available at little lesser price at aliexpress, but first I want to try with these and switch to cheaper ones later.
Also, I ordered some free samples from following series of coil craft. I chose 1206 as it easier to handle.
https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/1206-(3216)/1206cs/ (https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/1206-(3216)/1206cs/)
Appreciate if someone can let me know if this section make sense. For caps, I covered 1pf to 1000pf. For inductors, its few nH to close to 1uH.
-
Hi there,
Great, now maybe you could show some pics of the layout so we can see what it looks like in real life, and maybe some measurement pics of the performance.
Also, how many caps are you using and how many inductors, and any resistors?
And how about the values of the caps, inductors, and resistors if present?
Of course it goes without saying a schematic would be nice.
Sounds interesting.
Hi, the filter isn't anything special, it's just an 11th order cheby LPF. I tried to comply with the requirement that the parts should be available cheaply from Aliexpress and this meant I had to use T37-6 toroids. These are quite big as you can see in the image of the measurement pic below. I would have liked to have used T30 or T27. In this pic I've deliberately put all the toroids in line and pressed them up against each other so they are actually touching each other. The stopband performance isn't so good now but it still achieves about -90dBc at 30MHz as you can see in the image.
I did want to design it using just 120pF caps but this would have used up a lot of my 120pF 0805 caps. I used 270pF and 120pF caps in SMD 0805 packages and obviously I needed a lot of them. The caps in the centre sections of the filter are all 270pF and 120pF in parallel. The outer caps are 120pF in parallel with two 120pF caps in series.
This use of multiple caps may seem crazy, but it helps minimise the impact of capacitor tolerance and it also helps with the stopband performance up at UHF. The number of turns for each toroid was calculated using Micrometals' own software and this usually gives a very good estimate for the turns required. I had to adjust nothing apart from squishing the turns of each toroid slightly to get the flattest passband ripple. There's a VNA plot of the filter a few posts back.
I'm not suggesting that this filter design should be used, as it has more rejection than required. Maybe a 9th order filter could be used. However, I think it's worth checking out the DDS setup before designing the filter. Normally having a 15MHz output with a DDS clock of only 50MHz isn't ideal if the aim is to also have low 'in-band' spurious from the DDS. That's why I'm wondering if the 70dB filtering requirement at 30MHz is really needed. It would be be interesting to know which DDS is being used.
Hello again,
That's very nice, but we cant see the board or the components as it is just a giant blur. It's even a bit hard to see the measurement scales and values.
Worse yet, no schematic, no specific component values (inductors especially), and how many inductors and capacitors are required, etc.
For someone to reproduce this work they would need to see all this specific info. Also, some impedance matching info might be nice.
Also, i think it would be very interesting to look at the design mathematically but for that we need a schematic and component values.
If you can provide that information this would become much more interesting. For myself the math analysis would be most interesting but i need more info for that.
Why do you say you dont think he would want to use this?
OK, but please bear in mind that I only tacked this filter together quickly to prove that getting 70dB rejection at 30MHz wasn't 'impossible'. I aimed for over 90dB to try and show this. This filter isn't meant to be the solution for the 70dB requirement. It offers about 90dB rejection and not 70dB rejection. I tacked a quick and dirty LPF together on a bit of spare PCB material. The construction method I used is only suitable for prototypes. I stuck microstrip pads onto a plain copper PCB to mimic a microstrip circuit. It looks horrible but is quicker than making a microstrip PCB.
I've attached the schematic and simulation below. I used 120pF and 270pF caps to keep the number of cap parts low. Each cap value costs £1 per 100 from Aliexpress so I designed it this way to keep the parts cost down. I'm not sure what you mean by mathematical analysis but here's my initial simulation below and also a copy of my earlier VNA plot. The agreement is quite good. If I press the toroids together so the windings touch within each other the stopband gets worse at 30MHz but it is still about 90dB.
The reason each cap value has 1nH next to it is because this is the package inductance of the SMD part. A real PCB construction would need lots of via holes very close to each cap if the stopband performance is to stay below about 70dB up through VHF and into UHF.
-
I used a Q of 150 as I used thinner wire than the example toroid in the Q curve below. If you search online you should be able to find the Micrometals catalogue and Q curve catalogue for their powdered iron toroids.
I'd definitely recommend getting a nanovna to assist with the design of filters like this. Even the original Hugen nanovnaH is pretty good at these frequencies. I have one here. It should also be possible to measure the inductance (and get an idea of the Q) of this little toroid inductor at 15MHz with a nanovnaH as long as the calibration is done carefully.
I've not used a TinySA but you may be stretching its limits with some of your tests. Even some regular lab spectrum analysers need to be used with attention to the settings if the aim is to measure harmonic distortion at -70dBc with low uncertainty.
-
I actually measured the direct response (no filter) at the start of the setup. This was done by connecting the out of the function generator directly to the input of tinySA. Set the tiny SA to attenuate to 10db (so that its input impedance become 50Ohms based on its spec). Set the generator to scan from 1MHz to 100MHz linearly for 10s. Setup tinySA to show the same range and used two traces, one with no calculation and other with max hold calculation. After a while I get a nice flat response. So, I guess this measurement setup is ok for this range of frequencies?
Regarding the NanoVNA, out of the two sites https://nanorfe.com/nanovna-v2.html (https://nanorfe.com/nanovna-v2.html) and https://nanovna.com/, (https://nanovna.com/,) which one should I use? The one from https://nanorfe.com/nanovna-v2.html (https://nanorfe.com/nanovna-v2.html) is pretty expensive.
Does "direct" include all the dangling wire you show? That wire is the circuit. You can try spreading it apart, moving it very close together. Get a feel how it effects the measurement. I thought about trimming the leads from that breadboard filter just to make the point on how much this effects the performance. Best if you experiment on your own.
The expensive VNAs were fairly inexpensive when they were first introduced. These, like the LiteVNA I have been using are not as good as the original NanoVNA when it comes to narrow band measurements or low frequency/noise measurements. I suggest the low cost unit. Be aware that all of the low cost VNAs I looked at use a square drive which can cause some problems. So Consider what connectors you plan to use with the filter and maybe pick up some cheap adapters while you are at it.
Attached showing the original NanoVNA. I no longer support these and suspect the open sourced software is much better. For $50, hard to beat.
-
OK, but please bear in mind that I only tacked this filter together quickly to prove that getting 70dB rejection at 30MHz wasn't 'impossible'. I aimed for over 90dB to try and show this. This filter isn't meant to be the solution for the 70dB requirement. It offers about 90dB rejection and not 70dB rejection. I tacked a quick and dirty LPF together on a bit of spare PCB material. The construction method I used is only suitable for prototypes. I stuck microstrip pads onto a plain copper PCB to mimic a microstrip circuit. It looks horrible but is quicker than making a microstrip PCB.
Looks like you have the wrong attachment. You included the original data from the VNA rather than a clear photo of the construct.
-
I am thinking of buying these caps from digikey:
https://www.digikey.co.uk/short/zv2n5qnn (https://www.digikey.co.uk/short/zv2n5qnn)
All are C0G Ceremic ones. Similar ones are available at little lesser price at aliexpress, but first I want to try with these and switch to cheaper ones later.
Also, I ordered some free samples from following series of coil craft. I chose 1206 as it easier to handle.
https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/1206-(3216)/1206cs/ (https://www.coilcraft.com/en-us/products/rf/ceramic-core-chip-inductors/1206-(3216)/1206cs/)
Appreciate if someone can let me know if this section make sense. For caps, I covered 1pf to 1000pf. For inductors, its few nH to close to 1uH.
The capacitors look good; 5% COG from known suppliers. Yes, for breadboarding 1206 parts are quite easy to handle. As you get more comfortable with SMT you will find that 0802 and 0603 packages aren't much harder to work with. I do need magnification when using the smaller packages. Make sure you have a way to keep track of what part is which, as the capacitor markings are essentially absent or invisible.
Those Coilcraft inductors are also decent SMT parts. The SRF and Q are as good or better than the similar Murata ones I've been using. None of them beat a small iron-powder toroid for performance though, and in high-performance filters those parameters matter. The toroid cores aren't particularly expensive, if you plan to play with RF you might want to get an assortment. We can advise you on the materials and sizes that are most useful, and this depends on the frequency range and power levels.
I've found that a capacitor inventory of the 10% or 20% series from 10pF to 10nF (cap tolerance 5%) is handy. I keep only a handful of different SMT inductor values in the 100nH - 2uH range, and use hand-wound toroids or capacitive dividers to match impedances (caps are cheap, inductors not so much). Some filter designs are more particular about topology, but my needs are generally pretty simple. If I have a design that I want to make more than one or two of I will buy the exact parts needed.
Build and measurement technique are very important. You have seen some good examples here showing how to carve up double-sided PCB for use at RF frequencies. There are discussions of various techniques here and elsewhere on the internet. Get some coax jumpers and some connectors that you can solder directly to the board. I like edge-mount SMA connectors -- you can get a bag of adequate ones from ebay or Amazon. These cheap SMAs aren't metrology-grade, but are 1000% better than using jumper wires.
Attached image shows what I mean by "capacitive divider"
-
I've attached the schematic and simulation below.
Could I know which software you used for this design? Or did you choose the topology and values by yourself?
-
I used Agilent/Keysight Genesys to design the filter although it could also have been designed using the free alternatives like Elsie or RFSIM99. RFSIM99 is really old and might not run on a modern PC though.
One thing to be mindful of with 1206 SMD caps is that this package size is one of the most brittle in terms of generating internal stress cracks. This is due to the long and skinny shape. If you use 1206 MLCC caps then be very careful of how you handle the PCB. Use a fairly thick PCB material and don't put the caps near any RF connectors where the twisting stresses of connector removal can crack the caps. Once they generate microscopic internal cracks the capacitance value becomes intermittent and this can cause all kinds of confusion and frustration. Sometimes it can take months or even years for the cracks to cause problems. However, the usual symptom is erratic filter performance when the PCB is handled or moved. Once this happens it usually means removal and replacement of all the caps. Adding a stiffener plate under the PCB can help a lot here.
-
To give you some idea how the low cost VNAs compare, I first bent some metal for the repurposed Chinese attenuator, 9-pole filter and then ran it on three different analyzers. I have heard people complain about the original NanoVNA's phase error. You can see while it has only 101 data points, the noise is a bit lower. As you move below 1MHz, the NanoVNA's noise is much better than the LiteVNA (can't tell from these plots). If I normalize the two low cost VNAs with the HP, you can get a better idea on the errors between these three setups.
-
Using the old HP, I added some BNC to banana adapters along with some jumpers to simulate your setup. I tried it with the two pairs separated but with the jumpers somewhat close to one another as shown. I then spread the wires apart something like what you show in your photo. Difficult to offer any advice until you move beyond this. Hope it helps.
-
To give you some idea how the low cost VNAs compare, I first bent some metal for the repurposed Chinese attenuator, 9-pole filter and then ran it on three different analyzers. I have heard people complain about the original NanoVNA's phase error. You can see while it has only 101 data points, the noise is a bit lower. As you move below 1MHz, the NanoVNA's noise is much better than the LiteVNA (can't tell from these plots). If I normalize the two low cost VNAs with the HP, you can get a better idea on the errors between these three setups.
The way I see, NanoVA is good enough for me as a start. Later, if I feel like I am into RF stuff more, perhaps I will buy a better one - may be liteVNA or V4 or a used HP.
From which site did you buy your NanoVA? From this (https://nanovna.com/) site (or any seller the site recommends) or from this (https://nanorfe.com/nanovna-v2.html) site?
-
Using the old HP, I added some BNC to banana adapters along with some jumpers to simulate your setup. I tried it with the two pairs separated but with the jumpers somewhat close to one another as shown. I then spread the wires apart something like what you show in your photo. Difficult to offer any advice until you move beyond this. Hope it helps.
Honestly, I am surprised. My 50 Ohm BNC cables for function generator are on the way still. In the meantime, I will re-test the setup with a coaxial cable that I have already, but I think its 75 Ohm. Would this also ruin the measurements totally?
Also could I know which connected you have used in this (https://www.eevblog.com/forum/beginners/high-frequency-passive-low-pass-filter-with-cheap-components/?action=dlattach;attach=1675468;image (https://www.eevblog.com/forum/beginners/high-frequency-passive-low-pass-filter-with-cheap-components/?action=dlattach;attach=1675468;image)) circuit to connect the cable to the breadboard?
Out of all 50 Ohm cables, which one works best for this type of work? RG-58/RG-178/Rg-400/etc...
-
Out of all 50 Ohm cables, which one works best for this type of work? RG-58/RG-178/Rg-400/etc...
RG58 is probably most commonly used with BNC connectors, but for short cables on the benchtop (and for low-power), I like RG174 because of the flexibility. RG174 is often used with SMA connectors.
For short runs at HF frequencies 75 Ohm coax won't make a huge difference but it will be measurable, especially with the type of filter you are considering. More critical is that 75 Ohm BNC connectors have a slightly different-sized center pin/socket, and mixing the two types may damage your connectors.
There is a big difference in cables and connectors when you get into precision RF measurements, especially at the higher frequencies. Test cables can cost hundreds of dollars. I don't need that level of performance, and use the cheap stuff.
-
The way I see, NanoVA is good enough for me as a start. Later, if I feel like I am into RF stuff more, perhaps I will buy a better one - may be liteVNA or V4 or a used HP.
From which site did you buy your NanoVA? From this (https://nanovna.com/ (https://nanovna.com/)) site (or any seller the site recommends) or from this (https://nanorfe.com/nanovna-v2.html (https://nanorfe.com/nanovna-v2.html)) site?
A friend of mine bought two of those NanoVNAs and sent me the one I have so I could show them how to use it. I'm not sure where they came from. I did buy the H4 and V2+ / V2+4 direct from the suppliers stores. Should cost you about $50 USD.
Honestly, I am surprised. My 50 Ohm BNC cables for function generator are on the way still. In the meantime, I will re-test the setup with a coaxial cable that I have already, but I think its 75 Ohm. Would this also ruin the measurements totally?
Also could I know which connected you have used in this circuit to connect the cable to the breadboard?
Out of all 50 Ohm cables, which one works best for this type of work? RG-58/RG-178/Rg-400/etc...
Well, I don't think the 75ohm cable is going to be any worse than what you have now but I suggest getting something better suited.
See the following post that talks about some of the cable I use for small signal measurements:
https://www.eevblog.com/forum/rf-microwave/nanovna-custom-software/msg4063489/#msg4063489 (https://www.eevblog.com/forum/rf-microwave/nanovna-custom-software/msg4063489/#msg4063489)
That interface came from a fun little contest we had some time ago where I had the need to measure some fairly fast signals on a breadboard. This thread shows how it evolved.
https://www.eevblog.com/forum/projects/challenge-thread-the-fastest-breadboard-oscillator-on-the-mudball/msg3113578/#msg3113578 (https://www.eevblog.com/forum/projects/challenge-thread-the-fastest-breadboard-oscillator-on-the-mudball/msg3113578/#msg3113578)
***
Be aware the Digi-Key used to offer a service where they would make cables. I am not sure if this is still offered but you may want to look into it. Getting the proper crimp tools can add up and if you only need a few sets, this may save you some cash.
-
I did buy the H4 and V2+ / V2+4 direct from the suppliers stores. Should cost you about $50 USD.
Not sure I can buy for that price. The one I can find is 269USD and is available here (https://www.tindie.com/products/hcxqsgroup/4-nanovna-v2-plus4/ (https://www.tindie.com/products/hcxqsgroup/4-nanovna-v2-plus4/)).
-
OK, but please bear in mind that I only tacked this filter together quickly to prove that getting 70dB rejection at 30MHz wasn't 'impossible'. I aimed for over 90dB to try and show this. This filter isn't meant to be the solution for the 70dB requirement. It offers about 90dB rejection and not 70dB rejection. I tacked a quick and dirty LPF together on a bit of spare PCB material. The construction method I used is only suitable for prototypes. I stuck microstrip pads onto a plain copper PCB to mimic a microstrip circuit. It looks horrible but is quicker than making a microstrip PCB.
I've attached the schematic and simulation below. I used 120pF and 270pF caps to keep the number of cap parts low. Each cap value costs £1 per 100 from Aliexpress so I designed it this way to keep the parts cost down. I'm not sure what you mean by mathematical analysis but here's my initial simulation below and also a copy of my earlier VNA plot. The agreement is quite good. If I press the toroids together so the windings touch within each other the stopband gets worse at 30MHz but it is still about 90dB.
The reason each cap value has 1nH next to it is because this is the package inductance of the SMD part. A real PCB construction would need lots of via holes very close to each cap if the stopband performance is to stay below about 70dB up through VHF and into UHF.
That looks very interesting. Why do you say he cant use it though if the cut is 90db and he only needs 70db i would think 90db would be even better. That should get the clock feedthrough factor even lower.
The math analysis would be something along the lines of a Nodal analysis for both AC and transient responses.
-
I did buy the H4 and V2+ / V2+4 direct from the suppliers stores. Should cost you about $50 USD.
Not sure I can buy for that price. The one I can find is 269USD and is available here (https://www.tindie.com/products/hcxqsgroup/4-nanovna-v2-plus4/ (https://www.tindie.com/products/hcxqsgroup/4-nanovna-v2-plus4/)).
I wouldn't expect you to find a V2Plus4 much cheaper. The first sentence only states that I bought the H4 and two V2+ VNAs direct. I think I paid about $170 for both V2s combined and a bit over $100 for the H4.
If you have changed your mind and now want the V2Plus4, I recommend you look at the LiteVNA64. The LiteVNA64 offers several advantages over the V2Plus4 regardless if you plan to use them standalone or attached to a PC. I paid $120 for my LiteVNA64. They also offer a smaller version for a lower cost.
My last sentence " Should cost you about $50 USD. " was based on your comments about getting a first gen NanoVNA. A quick look on Amazon and eBay, there are several still available for this price.
Another option you should consider is staying with the equipment you have and just learning how to use it. It may be more than capable of making these measurements. Address the dangling wires and see where you're at. You would need to do this even if you get a VNA. They can't just magically correct for the errors you are introducing with your poor setups. As I have stated a few times, I suggest you start by addressing some of these major problems.
-
Once you get rid of those dangling wires, the next step is address the construction. I showed that breadboard as an example of what I would consider a very poor construction practice.
For test1, I put the filter back together and ran a baseline with the old HP. Again, the goal isn't to show you a filter that would match your constraints but rather show the effects of the build quality.
For test2 I trimmed the inductor's leads which also allowed me to pack things closer together.
For test3 I trimmed all of the disc capacitor's leads.
Next I took a section of scrap PCB and cut it to roughly the length I needed. I then cut off a strip and made 6 short sections. These were soldered to the board. Next I cut all of the components leads as short as possible and soldered these to the PCB. To finish up, I soldered on a couple of sections of coax.
For test4, it's back to the network analyzer. Assuming the EEVBLOG website hasn't screwed up the photos again, S21_compare is the money shot showing all the data from the original breadboard to the PCB.
***
T3sl4co1l found an error in the first test where one of the caps wasn't attached. The filter was rebuilt using new, but the same parts and the test was repeated.
-
Oh yeah, a parallel thread that may be of interest: https://www.eevblog.com/forum/projects/design-review-needed-tps65131-buck-and-inverter-(single-supply-to-dual)/msg4613371/#msg4613371 (https://www.eevblog.com/forum/projects/design-review-needed-tps65131-buck-and-inverter-(single-supply-to-dual)/msg4613371/#msg4613371) (and related comments by myself and others)
Tim
-
Once you get rid of those dangling wires, the next step is address the construction. I showed that breadboard as an example of what I would consider a very poor construction practice.
For test1, I put the filter back together and ran a baseline with the old HP. Again, the goal isn't to show you a filter that would match your constraints but rather show the effects of the build quality.
For test2 I trimmed the inductor's leads which also allowed me to pack things closer together.
Hmmmmm, no way in heck those two should be so different. Actually the first I think should be a little better, since there's less capacitance between nodes (adjacent slots on solderless breadboard are usually about 4pF; 5-6 vs. 3-4 positions between nodes gives a bit less coupling between them; but that should only be relevant at much higher frequencies (this is just capacitance across the inductors, so, zeroes like of an elliptical filter, but placed way high). Maybe the length of the ground bus is more significant? Nah, but not by THAT big a difference...
Wait..! I can just make it out?
(https://www.eevblog.com/forum/beginners/high-frequency-passive-low-pass-filter-with-cheap-components/?action=dlattach;attach=1678342;image)
D'oh! Well, good illustration of what really bad wrong layout can do as well I guess, hah. That explains the significantly different passband too.
Yeah, and then the rising / HF asymptote for the breadboarded cases, should basically be down to the loop inductance along the ground bus I think? Which the ground plane handily eliminates, hence the stopband in the noise floor.
Tim
-
S21_compare is the money shot showing all the data from the original breadboard to the PCB.
Nice! Now do it using SMT parts! (just kidding, although it would be interesting to see the effect of changing out those leaded disc caps.)
What are the Q/SRF specs on those axial inductors? I haven't used those since I was a technician, about 100 years ago. Looking at a few almost-similar axial inductors the Q seems to be around 50 or less, and the SRF 150MHz or more, which is fairly close to the SMT parts mentioned previously.
-
Wait..! I can just make it out?
No idea what you are referring to. Maybe some circles and arrows.
Nice! Now do it using SMT parts! (just kidding, although it would be interesting to see the effect of changing out those leaded disc caps.)
What are the Q/SRF specs on those axial inductors? I haven't used those since I was a technician, about 100 years ago. Looking at a few almost-similar axial inductors the Q seems to be around 50 or less, and the SRF 150MHz or more, which is fairly close to the SMT parts mentioned previously.
The problem with all of these TH parts, I have no idea what they are. Just old scrap laying in the bins.
-
Oh, the cap lead! :-DD
I didn't save the data from the first time I put it together. While I had used different caps, S21 is much closer.
Advice to OP, The hurrier I go, the behinder I get.
double check your work.
-
OK, but please bear in mind that I only tacked this filter together quickly to prove that getting 70dB rejection at 30MHz wasn't 'impossible'. I aimed for over 90dB to try and show this. This filter isn't meant to be the solution for the 70dB requirement. It offers about 90dB rejection and not 70dB rejection. I tacked a quick and dirty LPF together on a bit of spare PCB material. The construction method I used is only suitable for prototypes. I stuck microstrip pads onto a plain copper PCB to mimic a microstrip circuit. It looks horrible but is quicker than making a microstrip PCB.
I've attached the schematic and simulation below. I used 120pF and 270pF caps to keep the number of cap parts low. Each cap value costs £1 per 100 from Aliexpress so I designed it this way to keep the parts cost down. I'm not sure what you mean by mathematical analysis but here's my initial simulation below and also a copy of my earlier VNA plot. The agreement is quite good. If I press the toroids together so the windings touch within each other the stopband gets worse at 30MHz but it is still about 90dB.
The reason each cap value has 1nH next to it is because this is the package inductance of the SMD part. A real PCB construction would need lots of via holes very close to each cap if the stopband performance is to stay below about 70dB up through VHF and into UHF.
That looks very interesting. Why do you say he cant use it though if the cut is 90db and he only needs 70db i would think 90db would be even better. That should get the clock feedthrough factor even lower.
The math analysis would be something along the lines of a Nodal analysis for both AC and transient responses.
OK but I didn't actually say it couldn't be used, I jut clarified that it was a demo filter to show what could be achieved with a filter built using just two different component types that are readily available from Aliexpress. 0.7dB insertion loss at 15MHz and over 90dB rejection at 30MHz.
The problem with using SMD inductors will be the the huge reduction in Q compared to the toroids. The Q would collapse from about 150 to something under 30 if Coilcraft 1206 series inductors were used. SMD inductors are generally fixed in value as well. So this is a big disadvantage compared to the toroids. The toroids are fiddly to wind and can look untidy on a PCB though.
I didn't think these 1206 Coilcraft inductors were available from Aliexpress but maybe I'm not looking and searching correctly. If SMD inductors were used with my filter then the insertion loss at 15MHz would probably increase to over 3dB. There's also going to be some sinx/x droop in the DDS output by 15MHz which will make the droop appear even worse. In this case though, it might be possible to correct the droop in the digital domain as the DDS design looks to use an external DAC.
Otherwise, if the 3dB droop is a problem, and SMD parts are mandatory then it might be worth re-visiting the elliptic filter design. This should give a lower loss compared to a 9th or 11th order cheby filter. The risk here is from the component tolerances. I think it would be OK to do this for a one-off filter, especially if some fine tuning of component values is needed.
I did a quick simulation to show what I mean. The filter is a 7th order design and it only just achieves 70dB rejection at 30MHz. The other issue with this type of filter is that it reduces to a ladder network of capacitors up at UHF and just the package inductance of the caps will cause stopband problems up at UHF. When the via hole inductance is added to this, things will get even worse. If the filter was built using 1206 SMD caps I think it would be even worse again. The stopband would probably reduce to 20-30dB by 500MHz.
The simulation below assumes the filter is built using three different 0603 cap values and one Coilcraft SMD inductor value. This will help with costs and also with tolerancing and maybe help with the UHF stopband issues if several 0603 caps are used in parallel for the shunt caps. However, if good stopband performance is needed above about 500MHz a roofing filter would be needed.
The filter response is as below. I don't think I'd get this through a design review at work as the component tolerances would cause the stopband to sometimes fail to meet -70dB at 30MHz. So no good for mass production. However, it would be OK for a one off filter design I think. You can see that the insertion loss is predicted to be just over 1dB at 15MHz. This is fairly reasonable I think.
-
Remember that the -70dB requirement was just the OPs WAG (Wild-Ass Guess), and perhaps not what is actually necessary for his actual needs.
But this remains a very interesting discussion and demonstration of how much component characteristics and construction / measurement techniques matter (they do, a lot).
-
Remember that the -70dB requirement was just the OPs WAG (Wild-Ass Guess), and perhaps not what is actually necessary for his actual needs.
But this remains a very interesting discussion and demonstration of how much component characteristics and construction / measurement techniques matter (they do, a lot).
Agreed, thanks. That's partly why I was interested in the DDS type. It's many years since I last designed a post DDS LPF at work, but I have designed quite a few. The first one would have been in late 1996 soon after the AD9850 DDS was released. A colleague also used an AD9850 DDS in a hybrid synth. I've also experience with the AD9851, AD9858, AD995x and the AD9914/5 chips.
Normally, the weak point for in band spurious from a DDS is from LO - nRF terms and these are typically at -55dBc to -65dBc. To avoid the worst of them the sample frequency needs to be at about 4.5 times higher than the highest RF output frequency. 50MHz is only 3.33 times greater than 15MHz. Maybe the DDS architecture that suspension is using is able to suppress these terms in some way. Otherwise, 70dB rejection for out of band terms seems a bit high to me if the in-band terms could be at -55dBc for example. These terms can be present even with DDS devices with a 14 bit DAC.
I've always included a passive compensation network with the post DDS LPF and this allows for the droop in the LPF and also the sinx/x rolloff in the DDS output. Going back further in time I've designed a lot (and I mean a LOT) of elliptic LPFs down at 1MHz or less. These were used in the final IF of various RF downconverters and were used as cascaded elliptic HPF and LPF to do the alias filtering ahead of a digital IF. I think the biggest LPF I did was a 13th order elliptic although a designed a lot of 11th order elliptic lowpass filters. The trick back then was to design for a 500 ohm impedance as this allowed the use of inductors with really high Q. An inductor Q of 500 wasn't uncommon. The stopband requirements in those days was -72dBc so I usually designed the filter for -77dBc. The passband droop had to be less than 1dB. A typical filter might have an Fc of 1MHz and a -72dBc stopband of about 1.15MHz or even tighter. I think I needed the 13th order filter when the spec went closer to -80dBc. This was a long time ago so I can't really remember the specs...
-
My first DDS was a simple thing for a telecom chip (a gate array, probably 2 micron, back in the late 1980s). This was for a telecom T1 board, and I needed a 4 x 1.544MHz clock, locked to the system 8.192 MHz clock. I ended up designing a 63/256 NCO to generate a 2.016 MHz output, using just two bits of the output word to drive weighted resistors giving me a four-level DAC output. The first alias of this freq was 6.176 MHz (8.192 - 2.016), which I ran through a simple LC filter giving me a sufficiently clean (low jitter) 6.176 MHz clock. This took up a tiny corner of my gate array and saved a ton of money / space / power when compared to the competition who used PLLs and VCXOs to do the same thing. All done with schematic capture, as Verilog hadn't yet been invented. This started my love affair with synthesis, various divider structures, digital filters, etc., and led to a nice string of patents and some designs I am still quite proud of.
I was a bit surprised to discover that clock frequency relationship, because the underlying 1.544 MHz is 193 x 8KHz (193 being a prime number). It wasn't obvious until I stared at it for a while. Telecom is full of those oddball, "Hey, let's add one bit for framing and multiplex it for signaling too!" frequencies. Bandwidth used to be precious.
-
Remember that the -70dB requirement was just the OPs WAG (Wild-Ass Guess), and perhaps not what is actually necessary for his actual needs.
But this remains a very interesting discussion and demonstration of how much component characteristics and construction / measurement techniques matter (they do, a lot).
What I wild-ass-guess the OP meant, is that he needs the undesired frequencies to be below -70dBc. I assume that the other spectral components are already 30-something dB below the carrier, that would relax requirements for the filter considerably.
-
FWIW here is the schematic for the 7th order elliptic filter. I'm not sure this qualifies if the parts have to come from Aliexpress as it requires parts with fairly tight tolerance and also the inductors are modelled using the data for the Coilcraft 1206CS series. I'm not sure these Coilcraft parts can be ordered from Aliexpress.
However, purchases from Digikey have been considered, so maybe have a look at the circuit below. The most critical part of the filter is the centre resonator as this is resonant just above 30MHz. To try and reduce the issues with component tolerances the 49.5pF resonator cap has been made from three 33pF caps. The reasoning behind this is that three 33pF 5% caps should usually have a tighter tolerance (when configured as a 49.5pF cap) compared to a single 49.5pF 5% tolerance cap. This is fairly desperate stuff, but every little bit helps here.
The filter only needs three cap values, 10pF, 33pF and 100pF and only one inductor value of 560nH. If you can buy 33pF 1% tolerance caps from Digikey and maybe 2% 1206CS 560nH Coilcraft parts (or order them as free samples from Coilcraft) then this risk reduces the design still further.
The resonator on the right also uses a 33pF cap and so if you buy tight tolerance 33pF caps then this filter ought to give fairly similar results to the simulation below. I'll try and get some 1206CS 560nH parts at work tomorrow as there are dozens of Coilcraft sample kits at work.
Realistically, this filter should be built with 0603 SMD caps for the shunt capacitors to keep the package inductance down. Even 1nH package inductance will be enough to cause a very poor stopband up in the 700MHz region because this filter has series caps that bypass the inductors up at UHF. So the only filtering up at UHF is due to the capacitive dividing in the circuit. Any package inductance in the shunt 100pF and 33pF capacitors will destroy the capacitive division up at UHF meaning the stopband performance will also be destroyed up at UHF. I'd recommend fitting a basic pi roofing SMD filter after this filter if you need fairly good stopband performance above about 400MHz.
-
I rebuilt the breadboard after T3sl4co1l noticed one of the caps was not connected and repeated the test. That post was then updated.
While everything was setup, I swapped the discs for mica, double checking they were plugged in the right holes this time.
-
OK, but please bear in mind that I only tacked this filter together quickly to prove that getting 70dB rejection at 30MHz wasn't 'impossible'. I aimed for over 90dB to try and show this. This filter isn't meant to be the solution for the 70dB requirement. It offers about 90dB rejection and not 70dB rejection. I tacked a quick and dirty LPF together on a bit of spare PCB material. The construction method I used is only suitable for prototypes. I stuck microstrip pads onto a plain copper PCB to mimic a microstrip circuit. It looks horrible but is quicker than making a microstrip PCB.
I've attached the schematic and simulation below. I used 120pF and 270pF caps to keep the number of cap parts low. Each cap value costs £1 per 100 from Aliexpress so I designed it this way to keep the parts cost down. I'm not sure what you mean by mathematical analysis but here's my initial simulation below and also a copy of my earlier VNA plot. The agreement is quite good. If I press the toroids together so the windings touch within each other the stopband gets worse at 30MHz but it is still about 90dB.
The reason each cap value has 1nH next to it is because this is the package inductance of the SMD part. A real PCB construction would need lots of via holes very close to each cap if the stopband performance is to stay below about 70dB up through VHF and into UHF.
That looks very interesting. Why do you say he cant use it though if the cut is 90db and he only needs 70db i would think 90db would be even better. That should get the clock feedthrough factor even lower.
The math analysis would be something along the lines of a Nodal analysis for both AC and transient responses.
OK but I didn't actually say it couldn't be used, I jut clarified that it was a demo filter to show what could be achieved with a filter built using just two different component types that are readily available from Aliexpress. 0.7dB insertion loss at 15MHz and over 90dB rejection at 30MHz.
The problem with using SMD inductors will be the the huge reduction in Q compared to the toroids. The Q would collapse from about 150 to something under 30 if Coilcraft 1206 series inductors were used. SMD inductors are generally fixed in value as well. So this is a big disadvantage compared to the toroids. The toroids are fiddly to wind and can look untidy on a PCB though.
I didn't think these 1206 Coilcraft inductors were available from Aliexpress but maybe I'm not looking and searching correctly. If SMD inductors were used with my filter then the insertion loss at 15MHz would probably increase to over 3dB. There's also going to be some sinx/x droop in the DDS output by 15MHz which will make the droop appear even worse. In this case though, it might be possible to correct the droop in the digital domain as the DDS design looks to use an external DAC.
Otherwise, if the 3dB droop is a problem, and SMD parts are mandatory then it might be worth re-visiting the elliptic filter design. This should give a lower loss compared to a 9th or 11th order cheby filter. The risk here is from the component tolerances. I think it would be OK to do this for a one-off filter, especially if some fine tuning of component values is needed.
I did a quick simulation to show what I mean. The filter is a 7th order design and it only just achieves 70dB rejection at 30MHz. The other issue with this type of filter is that it reduces to a ladder network of capacitors up at UHF and just the package inductance of the caps will cause stopband problems up at UHF. When the via hole inductance is added to this, things will get even worse. If the filter was built using 1206 SMD caps I think it would be even worse again. The stopband would probably reduce to 20-30dB by 500MHz.
The simulation below assumes the filter is built using three different 0603 cap values and one Coilcraft SMD inductor value. This will help with costs and also with tolerancing and maybe help with the UHF stopband issues if several 0603 caps are used in parallel for the shunt caps. However, if good stopband performance is needed above about 500MHz a roofing filter would be needed.
The filter response is as below. I don't think I'd get this through a design review at work as the component tolerances would cause the stopband to sometimes fail to meet -70dB at 30MHz. So no good for mass production. However, it would be OK for a one off filter design I think. You can see that the insertion loss is predicted to be just over 1dB at 15MHz. This is fairly reasonable I think.
Hello again,
That's one of the things i was initially worried about for such a sharp design, and more to the point, do we have any temperature data yet?
I'm wondering what the sensitivity is with respect to component change especially in the presence of a temperature gradient. Maybe it would be necessary to stuff it into an insulated, metalized, insulated project box.
-
The Coilcraft ceramic core SMD inductors are extremely stable over temperature. Not as stable as a good C0G SMD cap but not far off. I'm not sure what temperature range you have in mind, but think in terms of about 1% change in inductance across a -30degC to +70degC temp range for the Coilcraft inductors. The C0G SMD caps should be slightly better than this over this temperature range. This assumes genuine parts and not fake parts.
-
The Coilcraft ceramic core SMD inductors are extremely stable over temperature. Not as stable as a good C0G SMD cap but not far off. I'm not sure what temperature range you have in mind, but think in terms of about 1% change in inductance across a -30degC to +70degC temp range for the Coilcraft inductors. The C0G SMD caps should be slightly better than this over this temperature range. This assumes genuine parts and not fake parts.
Hello again,
Ok great. So what do the sensitivity figures look like for the critical frequencies and sharpness, or has this not been looked at yet. That's the thing i was worried about. I didnt get to do anything mathematically yet with the circuit i have a bunch of other things going on at the moment. Maybe later this week i hope. Im involved in a pretty important project at the moment i want to see through to the very end (nothing to do with filters).
-
If it helps inspire some confidence in the use of SMD parts, I found some Coilcraft 1206CS 560nH inductors and built up the 7th order elliptic filter using these parts and some 0603 SMD caps. I used 100pF, 33pF and 10pF caps in the filter.
Remember, this filter was designed to get a low passband insertion loss at 15MHz, despite the low inductor Q of about 26 at 15MHz. This is why this filter only just achieves 70dB rejection at 30MHz. It was designed to do this.
See the simulation plot and the plot of the real filter. The result is quite good considering I used 5% parts. The iserion loss is just under 1.2dB.
-
I also designed a 9th order elliptic filter using 560nH inductors and 10pF, 33pF and 82pF caps. See the simulation below. You can see that this filter offers a significant improvement in stopband performance, but the insertion loss at 15MHz is about 1.4dB. If SMD parts are mandatory and Coilcraft 1206CS inductors are available (and you definitely want less than 2dB insertion loss at 15MHz and >70dB rejection at 70MHz) then this 9th order filter could be a contender.
-
Can you point a hot-air gun or hairdryer at the filter and see if there is any appreciable change in performance? I doubt that there will be much if any measurable effect given the ceramic-core inductors and the COG capacitors, but apparently MrAl is concerned.
Even with a mix-6 iron powder toroid, the temp stability is 35 ppm / deg C, so with a 0-50 temp swing that's 0.175% change in permeability. A ceramic core should be even better.
-
It's a good suggestion and I did the hairdryer test on the toroid filter yesterday after reading MrAl's post and it was fine. Any change was less than a trace width on the VNA and I heated the board quite aggressively with the hairdryer.
The other classic test is to put it in a freezer for a few hours. I don't think this would show any issues with drift either.
-
Another surface mount lowpass.
***
Attempted to measure the same filter with the LiteVNA using 100 averages and an IFBW of 200Hz. Cheap filter, cheap VNA made on the last attenuator circuit board.
-
If it helps inspire some confidence in the use of SMD parts, I found some Coilcraft 1206CS 560nH inductors and built up the 7th order elliptic filter using these parts and some 0603 SMD caps. I used 100pF, 33pF and 10pF caps in the filter.
Remember, this filter was designed to get a low passband insertion loss at 15MHz, despite the low inductor Q of about 26 at 15MHz. This is why this filter only just achieves 70dB rejection at 30MHz. It was designed to do this.
See the simulation plot and the plot of the real filter. The result is quite good considering I used 5% parts. The iserion loss is just under 1.2dB.
Exactly what I need. Thank for the effort.
I am still waiting for the samples from coilcraft 1206. I also put a free sample request to Murata for caps - if they do not approve, I will go ahead with the digikey ones. In either case, this gives me confidence that I will be able to get the performance I need.
I also changed my build techniques and measurement setup and now getting pretty good results - results which are in agreement with simulation. I will post a new reply with that info soon.
-
It's a good suggestion and I did the hairdryer test on the toroid filter yesterday after reading MrAl's post and it was fine. Any change was less than a trace width on the VNA and I heated the board quite aggressively with the hairdryer.
The other classic test is to put it in a freezer for a few hours. I don't think this would show any issues with drift either.
Hi,
That sounds pretty good. I think it might be a different story with a bandpass or notch filter. The depth of the notch of a sharp notch filter can vary quite a bit even 2nd order. Maybe many of us where thinking about that during this discussion and spoke too soon. The cutoff of a LP can change quite a bit and still be a decent LP filter. The schematic i provided would probably work in a similar manner.
-
I also changed my build techniques and measurement setup and now getting pretty good results - results which are in agreement with simulation. I will post a new reply with that info soon.
Good job. Looking forward to seeing your progress.
Soldered up the case and made some fancy labels. The attenuator I think was around $10 for the board with 4pcs on it. The connectors are total crap but cheap. Uses 5 CS inductors and some AVX 0603 NPO caps. Maybe there's $8 + labor...
The cooler was from another experiment and there's a lot of loss. Not sealed well and conduction. Managed to get it to -15C - 60C with an hour hold. Using the LiteVNA and min/max it's hard see anything. Making a Touchstone file for the two extremes then normalizing we can get some idea of the drift. This could be normal drift of the VNA. Would need more testing.
-
I couldn't find the attenuator boards I made these from. They were very poor attenuators and this was a good use for them. I did get this board as well which actually worked fairly well and has much nicer connectors.
If you want a cheap and dirty way to build up your filter, this may be one option. Bandsaw and X-acto knife and you're good to go.
https://www.aliexpress.us/item/3256803767890537.html?spm=a2g0o.productlist.main.7.68d43680KC6gpN&algo_pvid=d9edcc59-aa20-4bda-b619-8f4a286bda81&algo_exp_id=d9edcc59-aa20-4bda-b619-8f4a286bda81-3&pdp_ext_f=%7B%22sku_id%22%3A%2212000027556541338%22%7D&pdp_npi=2%40dis%21USD%213.24%212.75%21%21%21%21%21%40211bea7b16728782680287924d06d7%2112000027556541338%21sea&curPageLogUid=QRnny8YvC2KD (https://www.aliexpress.us/item/3256803767890537.html?spm=a2g0o.productlist.main.7.68d43680KC6gpN&algo_pvid=d9edcc59-aa20-4bda-b619-8f4a286bda81&algo_exp_id=d9edcc59-aa20-4bda-b619-8f4a286bda81-3&pdp_ext_f=%7B%22sku_id%22%3A%2212000027556541338%22%7D&pdp_npi=2%40dis%21USD%213.24%212.75%21%21%21%21%21%40211bea7b16728782680287924d06d7%2112000027556541338%21sea&curPageLogUid=QRnny8YvC2KD)
-
Comparing the LiteVNA with 100 averages compared with the HP3589A 100k - 30MHz.
-
FWIW here is the schematic for the 7th order elliptic filter. I'm not sure this qualifies if the parts have to come from Aliexpress as it requires parts with fairly tight tolerance and also the inductors are modelled using the data for the Coilcraft 1206CS series. I'm not sure these Coilcraft parts can be ordered from Aliexpress.
However, purchases from Digikey have been considered, so maybe have a look at the circuit below. The most critical part of the filter is the centre resonator as this is resonant just above 30MHz. To try and reduce the issues with component tolerances the 49.5pF resonator cap has been made from three 33pF caps. The reasoning behind this is that three 33pF 5% caps should usually have a tighter tolerance (when configured as a 49.5pF cap) compared to a single 49.5pF 5% tolerance cap. This is fairly desperate stuff, but every little bit helps here.
The filter only needs three cap values, 10pF, 33pF and 100pF and only one inductor value of 560nH. If you can buy 33pF 1% tolerance caps from Digikey and maybe 2% 1206CS 560nH Coilcraft parts (or order them as free samples from Coilcraft) then this risk reduces the design still further.
The resonator on the right also uses a 33pF cap and so if you buy tight tolerance 33pF caps then this filter ought to give fairly similar results to the simulation below. I'll try and get some 1206CS 560nH parts at work tomorrow as there are dozens of Coilcraft sample kits at work.
Realistically, this filter should be built with 0603 SMD caps for the shunt capacitors to keep the package inductance down. Even 1nH package inductance will be enough to cause a very poor stopband up in the 700MHz region because this filter has series caps that bypass the inductors up at UHF. So the only filtering up at UHF is due to the capacitive dividing in the circuit. Any package inductance in the shunt 100pF and 33pF capacitors will destroy the capacitive division up at UHF meaning the stopband performance will also be destroyed up at UHF. I'd recommend fitting a basic pi roofing SMD filter after this filter if you need fairly good stopband performance above about 400MHz.
Hi G0HZU,
Please can you kindly contribute to the thread :
https://www.eevblog.com/forum/beginners/433-bandpass-filter/ (https://www.eevblog.com/forum/beginners/433-bandpass-filter/)
Thanks!
-
How about using a notch filter to remove the 2nd harmonic (30MHz), thus only requiring the low pass filter to remove the 3rd and higher harmonics (>=45MHz)