Author Topic: Broadcast (Wideband) FM Receiver Design  (Read 1317 times)

0 Members and 1 Guest are viewing this topic.

Offline ignilux

  • Supporter
  • ****
  • Posts: 29
  • Country: us
Broadcast (Wideband) FM Receiver Design
« on: March 14, 2019, 04:32:17 pm »
Hi, all-

I'm an EE student in my late twenties, having returned to school after becoming deeply passionate about hobby electronics over the last 5 years or so. Recently, I've been bitten by the RF bug. There's something so intriguing about knowing the secrets of what other, smarter people refer to as "black magic". Anyway, I've done the typical crystal radio, amplified crystal, litz antenna coil with inductive coupling to an amplified 1N5711-based detector, RF amplifier, etc etc. That was fun and all, but AM radio is only so interesting. SO! Without further ado-

My next/current project is building a wideband FM receiver for the broadcast band here in the States, i.e., 88 to 108 MHz. I realize that there are all-in-one IC's that can do everything for me, but I'd like to stick to basic building blocks. PLLs, mixer ICs, amplifiers, etc. I also want to keep the signal path as analog as possible. My initial thoughts were to use a superhet architecture, mixing the RF input with a VCO (for tuning) to get a low IF, say, 5 MHz. Feed that to a PLL with a center frequency of 5 MHz, and the error signal becomes the baseband. But... then I learned about image frequencies.

After that came the direct conversion, zero IF route. Bandpass-filtered RF straight in to a Tayloe detector to obtain baseband quadrature signals. As an aside, I understand that the Tayloe detector essentially digitizes the signal, but the low pass filter afterwards filters out the higher harmonics. Anyway, here's where I'm stuck. Most direct conversion designs seem to feed the I/Q signals directly into DSP / SDR / etc. I'd like to use analog demodulation, but any references I've seen seem to be designed for SSB transmission. Is there an analog way to recover the full, wideband FM from the baseband quadrature signals, or am I way off here? If it's as simple as phase detection, can I simply feed the signals through the phase detector block of a PLL as I would have in the superhet? I really can't tell if I'm overthinking this, or vastly underthinking it.

If there's another architecture you think I should consider, please let me know. I like the direct conversion design because it teaches me a whole hell of a lot about mixing, quadrature signals and modulation, high speed signals, etc. I'd like to make my career in RF/MW electronics, so this is as much educational as hobby/fun.

Thanks in advance, everyone.
 

Offline radiolistener

  • Frequent Contributor
  • **
  • Posts: 928
  • Country: ua
Re: Broadcast (Wideband) FM Receiver Design
« Reply #1 on: March 14, 2019, 05:23:51 pm »
Analog receivers and digital SDR works in the same way, but in digital domain it is more easy to do these things.

For example in order to demodulate FM from quadrature input in digital domain, there is just need to calculate atan2 from input IQ stream. The result will be the current phase angle. Next, just substract phase angle obtained for previous IQ sample and you will get delta phase. Next just scale it and you will get FM demodulated output :)

In analog world it works the same, but it is more complicated to implement.

 

Offline ignilux

  • Supporter
  • ****
  • Posts: 29
  • Country: us
Re: Broadcast (Wideband) FM Receiver Design
« Reply #2 on: March 14, 2019, 06:22:46 pm »
I realize that it's much simpler to do the last step digitally, but I'm wondering how it's done with analog components, or at least jellybean logic. Most projects would be made easier by throwing a microcontroller at it, but I'm hoping to learn more this way by struggling with the difficulty.

So what I gleaned from your reply is that magnitude of the I/Q phase difference is not what matters, but dΦ/dt . Is that correct? Something like a phase detector followed by an opamp differentiator? Analog scaling is easy enough.
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 2667
  • Country: lv
Re: Broadcast (Wideband) FM Receiver Design
« Reply #3 on: March 14, 2019, 07:09:42 pm »
I realize that it's much simpler to do the last step digitally, but I'm wondering how it's done with analog components, or at least jellybean logic.

Analog superheterodyne FM receiver:



You may consider using SA605 and disregard that it is single chip IC. It's analog after all.
« Last Edit: March 14, 2019, 07:14:44 pm by ogden »
 

Offline radiolistener

  • Frequent Contributor
  • **
  • Posts: 928
  • Country: ua
Re: Broadcast (Wideband) FM Receiver Design
« Reply #4 on: March 14, 2019, 08:41:17 pm »
but dΦ/dt . Is that correct?

yeah, it is known as "arctan demodulator":

atan2(Q, I) => dΦ/dt => scale => result

There is also needs to unwrap phase angle.

This is my working code for FM demodulator:
Code: [Select]
                var i = iq[i].Re;
                var q = iq[i].Im;
                if (i == 0 && q == 0)
                {
                    i = 1;
                    q = 1;
                }
                var phase = atan2(q, i);
                var deltaPhase = phase - last_phase;
                if (deltaPhase < -M_PI) deltaPhase += TWO_PI;
                if (deltaPhase > M_PI) deltaPhase -= TWO_PI;
                deltaPhase /= M_PI;
                last_phase = phase;

                audio[i] = deltaPhase * gain;

Here is some theory about quadrature detection: http://kom.aau.dk/group/05gr506/report/node29.html


here is alternative FM demodulator in digital domain, similar to analog one from picture above:
Code: [Select]
                // Polar discriminator
                var f = iq[i] * conjugate(_iqState); 
                _iqState = iq[i];

                // Limiting
                var m = sqrt(f.Re*f.Re + f.Im*f.Im);
                if (m > 0.0) f /= m;

                var phase = atan2(f.Im, f.Re);

                audio[i] = phase * gain;

conjugate means just sign inversion for imaginary part (Q value): conjugate(v) = complex(v.Re, -v.Im);

Multiply operation can be done with mixer in analog domain.


« Last Edit: March 14, 2019, 10:30:20 pm by radiolistener »
 

Online vk6zgo

  • Super Contributor
  • ***
  • Posts: 4978
  • Country: au
Re: Broadcast (Wideband) FM Receiver Design
« Reply #5 on: March 14, 2019, 11:40:30 pm »
I would suggest sticking with the superhet design.
FM receiver IFs are usually quite a bit higher, at 10.7MHz.
Image interference is not a big problem, as a Station 10.7 MHz away is a fair way down the skirt of the RF input tuned circuit.
This also works for "the other way round" where the interfering Station is 10.7 MHz away from the local oscillator at two points on the dial.

Everybody loves Direct Conversion, but the IF is at base bound audio, & there is often no RF selectivity prior to the detector
Strong unwanted signals can force their way in & be envelope demodulated, producing "in band" noise.

SDRs would seem to able to solve the problems with Direct Conversion, but a superhet doesn't have those problems to begin with.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 9946
  • Country: us
  • DavidH
Re: Broadcast (Wideband) FM Receiver Design
« Reply #6 on: March 15, 2019, 01:46:30 am »
I think for analog FM demodulation in a direct conversion receiver you can use a Costas loop which is completely feasible in the analog domain.  (1) A PLL FM demodulator relies on the content being within the bandwidth of the PLL but a Costas loop relies on the content to be above and the loop only tracks the carrier.  So a PLL FM demodulator acts as a low pass filter and a Costas loop acts as a high pass filter.

But for simplicity, I think you would be better off with a single conversion superheterodyne with as high an IF frequency as you can get away with while still getting sufficient gain and limiting.  Dual conversion is likely necessary for maximum sensitivity. (2)

(1) I have considering designing a discrete analog WWVB kit like this (not direct conversion) but I suspect it is patented.

(2) Too much gain at a single frequency tends to result in instability because of coupling between the input and output so additional conversions and IFs are used to divide the gain up.
 

Offline ignilux

  • Supporter
  • ****
  • Posts: 29
  • Country: us
Re: Broadcast (Wideband) FM Receiver Design
« Reply #7 on: March 15, 2019, 02:01:45 pm »
Thanks for your input, everyone. At some point last night I realized that I had never designed a superheterodyne receiver of any kind, and that seems like a rite of passage in the RF world. My plan from here is to do so with an IF of 10.7 MHz, as suggested by vk6zgo. That does indeed seem to be the standard broadcast FM IF in the US. Makes IF filtering easier, too, with crystal filters available at that frequency at a decent price.

I breadboarded a 74HCT9046 PLL last night as an FM demodulator, and successfully recovered a baseband signal from a 10.7 MHz FM input from my function generator. If I can get that to work on a breadboard I feel pretty good about it working well on a real PCB. My tentative mixer will be the SA612A, which, while somewhat dated is still the "classic". My only real remaining question is the VCO for tuning the LO frequency. Is it feasible to obtain a reasonable degree of stability from a varactor-tuned LC oscillator in the VHF range? I have a Vackar oscillator working in LTSpice, but we all know how that goes once it's implemented in the real world.
 

Offline radiolistener

  • Frequent Contributor
  • **
  • Posts: 928
  • Country: ua
Re: Broadcast (Wideband) FM Receiver Design
« Reply #8 on: March 15, 2019, 02:17:50 pm »
My only real remaining question is the VCO for tuning the LO frequency. Is it feasible to obtain a reasonable degree of stability from a varactor-tuned LC oscillator in the VHF range? I have a Vackar oscillator working in LTSpice, but we all know how that goes once it's implemented in the real world.

you can use cheap si5351 module with cheap Chinese STM32F103C8 arduino controller. It is cheap, simple and get very clean and stable output with low jitter from 2 kHz to 200 MHz. It covers all FM bands and VHF band. Also you can use it for SW, MW, LV and even VLF band receiver.

Here is si5351 module:  https://www.aliexpress.com/item/CJMCU-5351-Si5351A-Si5351-I2C-Clock-Generator-Breakout-Board-Module-Signal-Generator-Clock-8KHz-160MHz-For/32845925389.html

It has 3 channels and 2 PLL's and has much more clean output than DDS chips.

You can use cheap STM32F103C8 arduino module to control it:
https://www.aliexpress.com/store/product/STM32F103C8T6-ARM-STM32-Minimum-System-Development-Board-Module-For-Arduino-TE435/3105031_32832983782.html
 

Online TheUnnamedNewbie

  • Frequent Contributor
  • **
  • Posts: 828
  • Country: 00
Re: Broadcast (Wideband) FM Receiver Design
« Reply #9 on: March 15, 2019, 02:49:48 pm »
Thanks for your input, everyone. At some point last night I realized that I had never designed a superheterodyne receiver of any kind, and that seems like a rite of passage in the RF world. My plan from here is to do so with an IF of 10.7 MHz, as suggested by vk6zgo. That does indeed seem to be the standard broadcast FM IF in the US. Makes IF filtering easier, too, with crystal filters available at that frequency at a decent price.

I breadboarded a 74HCT9046 PLL last night as an FM demodulator, and successfully recovered a baseband signal from a 10.7 MHz FM input from my function generator. If I can get that to work on a breadboard I feel pretty good about it working well on a real PCB. My tentative mixer will be the SA612A, which, while somewhat dated is still the "classic". My only real remaining question is the VCO for tuning the LO frequency. Is it feasible to obtain a reasonable degree of stability from a varactor-tuned LC oscillator in the VHF range? I have a Vackar oscillator working in LTSpice, but we all know how that goes once it's implemented in the real world.

Do you need high stability? I think it should be fine for audio. Keep in mind that since there is no information at DC, you don't care about any offsets - so long you are 'close' you are good.

A second project in the future could be to build a PLL-based synth? Or you can generate your LO with a XTAL and a multiplier, and then mix it with a lower frequency VCO. This is what I love about RF, so many ways to solve your problem of generating just the right sinewave.
The best part about magic is when it stops being magic and becomes science instead
 

Offline radar_macgyver

  • Frequent Contributor
  • **
  • Posts: 385
  • Country: us
Re: Broadcast (Wideband) FM Receiver Design
« Reply #10 on: March 15, 2019, 03:06:38 pm »
I realize that it's much simpler to do the last step digitally, but I'm wondering how it's done with analog components, or at least jellybean logic. Most projects would be made easier by throwing a microcontroller at it, but I'm hoping to learn more this way by struggling with the difficulty.

So what I gleaned from your reply is that magnitude of the I/Q phase difference is not what matters, but dΦ/dt . Is that correct? Something like a phase detector followed by an opamp differentiator? Analog scaling is easy enough.

Another method (besides the PLL) is to use a 'slope detector', which is a narrow-band circuit tuned just off the IF center frequency with a known frequency response slope. By choosing the frequency response so that the middle of the slope coincides with the IF, the frequency deviations of the IF are turned into amplitude modulation, which you can then pass to an AM detector.

The slope detector method is prone to distortions if the TRF bandpass slope is not linear. The PLL method suffers the same issue if the VCO's tuning response is not linear. Not sure how VCO linearity affects Costas loop receivers.
 

Offline ignilux

  • Supporter
  • ****
  • Posts: 29
  • Country: us
Re: Broadcast (Wideband) FM Receiver Design
« Reply #11 on: March 15, 2019, 06:34:24 pm »
Radiolistener, I admire your willingness to help and your enthusiasm for newer technology, but that's not what this project is about!  ;D

I had considered a PLL synth route, but I realized that that, too, would be much easier with a microcontroller. Changing divider ratios and a LO frequency simultaneously without a micro would be laborious and error-prone, would it not? I'll run with the LC oscillator idea-- I have a prototype board all layed-out to test later on. Just gotta dust the cobwebs off my etch tank (because of course I'm one of those guys, too).

Thanks again, and I'll be sure to post the end result here when it's done!
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 2667
  • Country: lv
Re: Broadcast (Wideband) FM Receiver Design
« Reply #12 on: March 15, 2019, 08:11:11 pm »
Note that FM receiver shown was 45MHz (?) narrowband FM cordless phone (prehistoric). Those have no image problems with single IF stage because channels are not overcrowded. This is not the case with FM broadcast station frequencies. Broadcast FM receiver requires two IF stages
 

Offline biastee

  • Contributor
  • Posts: 18
  • Country: my
    • 9W2LC
Re: Broadcast (Wideband) FM Receiver Design
« Reply #13 on: March 16, 2019, 02:19:13 am »
Makes IF filtering easier, too, with crystal filters available at that frequency at a decent price.

My only real remaining question is the VCO for tuning the LO frequency. Is it feasible to obtain a reasonable degree of stability from a varactor-tuned LC oscillator in the VHF range? I have a Vackar oscillator working in LTSpice, but we all know how that goes once it's implemented in the real world.

You won't need to worry about oscillator stability in a FM broadcast receiver.  Due to the +/-75 kHz deviation, the receiver's IF bandwidth is typically 180 kHz. So, the LO is unlikely to drift out of the channel during normal operation. To support the wide IF bandwidth, the filter is usually ceramic (not crystal).
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 13947
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Broadcast (Wideband) FM Receiver Design
« Reply #14 on: March 16, 2019, 03:05:30 am »
Funny thing is, once you understand RF, it's really just 1. being careful with where you allow signals to go (transmission line theory, impedance matching), and 2. solutions of polynomials (filter functions, impedances).  Not easy by any means, but hardly the insurmountable monolith some would have you imagine. :)

In keeping with a perhaps overly-classic analog approach, my radios have been tube based, like this topical one:





Which uses plug-in tuning coils to select RF, IF and LO.  6688 RF amp, 6J6 balanced mixer, 6C4 LO.

Not pictured unfortunately, the IF strip.  I made that for 33MHz (as a best case for image rejection), with 3 x 5702 (submini pentode) and 6AL5 detector.  This gives an AM output, which is wired for AGC (sort of serving as FM limiting as well), and for FM detection I added a 28.5MHz LO2 and 6BN6 (gated beam discriminator) at 4.5MHz.  (No mixer needed, I diplexed the LO2 and IF together, and let the 6BN6 grid do the mixing; as janky as that sounds, it works nicely!)

Haven't made 120 or 144MHz coils yet; think 144 was the last thing I was working on.

Not much else to look at in the low VHF range I think, basically FMBCB, ATC and amateur?

On a related note, the harmonics from the 6C4 are impressive, if weak.  They extend out to 800MHz, that I noticed.

There's very little fundamentally limiting about most ("modern"*) tubes, it's just that, for a lot of practical cases, the transconductance to capacitance ratio is rather poor, and the output impedance is rather high (the input impedance is usually on the high side, too).  The result: a useful bandwidth of maybe 10s of MHz for most types.  (For baseband, that's the LPF cutoff -- e.g. the bandwidth of a DC coupled oscilloscope amplifier.  That's also the bandpass width of AC coupled or tuned RF amplifiers.  Same thing, give or take filter order, and heroic efforts like distributed amplifiers, which only offer incremental advantage over the basic (first order) performance of the amplifier.)

*Since the 50s, say.  As designs moved away from stem bases (= excessive lead length), and optimized for lower voltages (100-300V) and higher performance, bandwidths went up, from ~MHz (e.g., an early FM radio might've needed tuned RF, or no RF preamp at all, just straight into the mixer), to 10s MHz (like the 6J6 and 6688 above, and all the types Tektronix was fond of: 6DK6, 6DJ8, a few somewhat-specialty low capacitance dual tetrodes and such), to the low 100s of MHz with the best (still more or less conventional) types, the planar triodes.  (Traveling wave tubes continue -- to this day -- to surpass whole GHz of bandwidth, but they are quite specialty types indeed.)

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Online TheUnnamedNewbie

  • Frequent Contributor
  • **
  • Posts: 828
  • Country: 00
Re: Broadcast (Wideband) FM Receiver Design
« Reply #15 on: March 16, 2019, 06:31:51 am »
Funny thing is, once you understand RF, it's really just 1. being careful with where you allow signals to go (transmission line theory, impedance matching), and 2. solutions of polynomials (filter functions, impedances).  Not easy by any means, but hardly the insurmountable monolith some would have you imagine. :)


I think the insurmountable monolith comes in when you try to turn your solutions of polynomials into bits of copper on a PCB or chip or holes in a block of metal or bits of plastic that you then call a waveguide.  ;)


  (Traveling wave tubes continue -- to this day -- to surpass whole GHz of bandwidth, but they are quite specialty types indeed.)



On the note of TWT - they are seeing a resurgence in research. With interest in communications at high millimeter wave to low THz frequencies (D-band, 110-170 GHz, 300+ GHz short range), and micro-machining TWT being an interesting possibility they could solve the lack of output power (getting 0 dBm at 200+ GHz in CMOS is impressive). Some people report 40 dBm (10 watt) of output power from a TWT at D-band and up, which, as someone who works on circuits at those frequencies, just blows my mind.
The best part about magic is when it stops being magic and becomes science instead
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 13947
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Broadcast (Wideband) FM Receiver Design
« Reply #16 on: March 16, 2019, 09:25:02 am »
I think the insurmountable monolith comes in when you try to turn your solutions of polynomials into bits of copper on a PCB or chip or holes in a block of metal or bits of plastic that you then call a waveguide.  ;)

Heh, various aspects of pesky reality do come into it, yes.  At least most of these things can be simulated quite accurately.  (Can, not necessarily on a budget, mind... ;D )


Quote
On the note of TWT - they are seeing a resurgence in research. With interest in communications at high millimeter wave to low THz frequencies (D-band, 110-170 GHz, 300+ GHz short range), and micro-machining TWT being an interesting possibility they could solve the lack of output power (getting 0 dBm at 200+ GHz in CMOS is impressive). Some people report 40 dBm (10 watt) of output power from a TWT at D-band and up, which, as someone who works on circuits at those frequencies, just blows my mind.

Other physicists simply said "hold our beers" and invented free-electron lasers, too. :-DD

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline vk3yedotcom

  • Frequent Contributor
  • **
  • Posts: 535
  • Country: au
    • vk3ye dot com (radio articles and projects)
Re: Broadcast (Wideband) FM Receiver Design
« Reply #17 on: March 18, 2019, 09:08:14 am »

My next/current project is building a wideband FM receiver for the broadcast band here in the States, i.e., 88 to 108 MHz. I realize that there are all-in-one IC's that can do everything for me, but I'd like to stick to basic building blocks. PLLs, mixer ICs, amplifiers, etc. I also want to keep the signal path as analog as possible.

Maybe try something super simple.

Eg a 1 or 2 transistor superregen.

Incredibly sensitive.

There's a bit of distortion but there's lots you can experiment with them eg using external quenching.

Two of my examples.



If you're into amateur radio you might enjoy my books. Choice of 6. Electronic or paperback. Details here: http://home.alphalink.com.au/~parkerp/vk3yebooks.htm
 

Offline ignilux

  • Supporter
  • ****
  • Posts: 29
  • Country: us
Re: Broadcast (Wideband) FM Receiver Design
« Reply #18 on: March 29, 2019, 03:25:26 pm »
Progress so far:

Built a prototype varactor-tuned LC oscillator for use as an LO. With 2 to 12 V bias on the varactor I'm able to tune from 80 to 120 MHz with a nice -10 dBm output level in to 50 ohms. If biased from 0.5 to 20 V I'm almost able to tune across an entire octave! The signal is a sufficiently clean sinusoid, too, with the 2nd harmonic coming in 35 dB down, and the 3rd+ way down in the noise of my 8568B.

Audio frequency stage breadboarded and working well, too. I won't bore anyone with the details as it's a fairly run-of-the-mill design: two stage JFET preamp using J310's that feed in to a NE5532A-based class AB amplifier.



Currently working on the RF section, which will consist of a third order Chebyshev bandpass filter and BGU6102 RF amplifier. I did have one question that I can't find the answer to, and I suspect it's because it's one of those things that everyone is supposed to know. You know, like the first time you read Sedra and Smith, and they mention something once in passing on page 3, and they expect you to remember it for the entirety of the book. Anyway, when looking at RF designs/app notes/etc I very often see a series resistor on the input, and shunt resistor on the output. Is this simply a statement of the inherent input and output impedance of the system, or am I physically supposed to insert resistors in my design?  Try not to laugh as you answer this  8)
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 2667
  • Country: lv
Re: Broadcast (Wideband) FM Receiver Design
« Reply #19 on: March 29, 2019, 05:40:51 pm »
Anyway, when looking at RF designs/app notes/etc I very often see a series resistor on the input, and shunt resistor on the output. Is this simply a statement of the inherent input and output impedance of the system, or am I physically supposed to insert resistors in my design?  Try not to laugh as you answer this  8)

No worries, quite valid question. Are you talking about input/output of RF/IF filters? Any example to show what you are talking about? Those most likely are impedance matching resistors. When you do not match your filter input/output impedance properly, it's frequency characteristics will be distorted.
 

Offline ignilux

  • Supporter
  • ****
  • Posts: 29
  • Country: us
Re: Broadcast (Wideband) FM Receiver Design
« Reply #20 on: March 29, 2019, 06:26:58 pm »
The attached picture is a screenshot from Elsie, showing my RF bandpass filter.

 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 2667
  • Country: lv
Re: Broadcast (Wideband) FM Receiver Design
« Reply #21 on: March 29, 2019, 07:15:27 pm »
The attached picture is a screenshot from Elsie, showing my RF bandpass filter.

Those resistors are 50 Ohm impedance matching resistors for simulation - signal source of simulator have impedance of 0 Ohms so you add series 50 Ohms, and filter output needs 50 Ohm load as well. When you put your filter into RF circuit, it (circuit) supposedly shall have output and input impedance 50 Ohms so no impedance correction in form of resistors is necessary.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 13947
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Broadcast (Wideband) FM Receiver Design
« Reply #22 on: March 29, 2019, 09:55:36 pm »
Note that a 440pF cap with less than 6nH ESL is hard to come by (an 0805 ceramic chip is already about three), and more to the point, you won't be able to tap onto the "top" and "bottom" of that 6nH as suggested by the schematic!

The series branch is a bit easier to deal with, as 260nH inductors with much less than 10pF stray (end-to-end and to ground) are common enough.

What effect does this have?  Basically you'd be connecting to a tapped inductor, which means impedance transformation -- whether you meant to or not.

Consider transforming the entire network, so that for example it's three LCs chained, with a bit of mutual inductance linking them, and the two end tanks using tapped feedpoints (tapped L, or C divider).

Or just raising the impedance, so that the end tanks are reasonable (say, 200 ohms), but then the middle tank will be unreasonable (you won't find a ~uH inductor with <1pF strays), so again, consider a series-parallel transformation.

Or use the strays to your advantage as impedance transformation.  This requires a good model of them, but it is solvable. 

I made this 100-130MHz bandpass,





with this [simulated] response,



This can be described as a:
- One port unterminated filter (this is a perfectly normal type -- at least one port must be terminated, and the other can be whatever impedance, including open or short, it just uses different coefficients; see tables in Zverev)
- 2nd order highpass, from the bias choke L1 and coupling cap C2
- 5th order lowpass, approx. Chebyshev 0.5dB or so I think
- Impedance transformer from the capacitor divider (C2, C5, C6)
- Two zeroes from the inductor stray capacitance
- Terminated into a complex load (Q2's base input impedance is around 500 ohms and a few pF, adjustable by R13)
- Capacitors adjusted with trimmers and gimmicks (bits of wire/PCB tacked on)

Overall, the impedance matching gains several dB over a more cookbook approach, and more than 6dB over a naive double terminated filter.

So, it's a neat curiosity, but hardly necessary with gain-bandwidth being as affordable as it is!

This is more to say -- you can pull off an oddball circuit, which behaves like a certain amalgam of prototypical filters and well understood transformations, and which is motivated by real parasitics that are missing from the standard ladder (alternating parallel-series) prototype.  But choose your transformations wisely, because some are easier than others, and an optimal (lead component count?) approach can incur a lot more design and tuning time.

So, impedance transformations are easy, and series-parallel transformations are modest.  Keep things symmetrical, so the same thing happens at both ends (and you maintain 50 ohm system impedance).

Series-parallel transformation is where you transform one branch into another.  The transition band response is preserved, but the asymptotes may shift, depending on what reactances are used to perform the transformation.  The most common example is a quartz crystal, which is usually series resonant between two pins (actually it's both, due to internal capacitance), but a parallel resonant response is required.  By making a pi network with two capacitors and the crystal between them, a transformation is performed and the desired response is obtained.

Similarly, we can transform a parallel-series-parallel LC bandpass into an all-parallel bandpass, where the parallel tanks are identical and linked with series capacitors.

Or more generally, any coupling will do, not just series capacitors -- series inductors could be used, or mutual inductance between the tank inductors, or many other topologies as well.

If series caps are used, the asymptotic response tends towards high-pass (the HF asymptote is flat, a constant -whatever dB, corresponding to the C ratios; while the LF is the full -20*(filter order) dB); if inductors, low-pass; if mutual, then symmetrical band-pass.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf