Resonant VHF Amplifier

I played a little with the values of the inductors and it works just as well as low as 40nH and at every value above that (including even 1mH), so it looks to be primarily a constant-current effect rather than a resonant effect, although very low inductances such as 40nH do have slightly more gain due to resonance. Lower values (e.g. 10nH or even just a simple wire) have much lower gain.

Oh sorry I just noticed your reply on this. That’s interesting that changing the inductor values didn’t have much effect, I had to tweak everything so much early on. And are you saying that the inductors themselves provide gain? …or just that different size inductors result in better transistor gain?

I'm saying that the only contribution of the inductors to the circuit is to make the current constant and that no resonance is happening, so the exact value of the inductor doesn't have much effect as long as it's larger than about 40nH. But if the inductors are too small than they just act as shorts to ground completely killing any chance of signal propagation from one stage to the next. In other words they act as high-pass filters and nothing more.

Also the left-most inductor can resonate with the input capacitor at around 40nH or so which would provide an additional gain of 2.3dB but the other inductors don't have any values for which the gain increases due to resonance.

I see exactly what your saying, and I almost completely agreed with you…. But, if that were the case then I wouldn’t expect it to work so well at lower frequencies (where the parasitic B-E junction capacitance is much less, or rather the same but doesn’t have the effect of shorting that junction so much anymore). It even works at 1MHz. On the contrary to that though; you’d also expect that even if it was resonating, then it would be so frequency specific that it wouldn’t work above or below 50MHz, but on a second more thorough observation, it might not be doing either one of those things, but rather the inductors are providing some le-way to allow the alternating current to create a proportional negative voltage swing on the emitters of the common base configurations, which in turn further biases the transistors. I’m not entirely sure exactly the reason, perhaps it’s even more than one thing.

You know, I’m not even sure how the signal makes it past the first stage through the bypass capacitor at such low frequencies though…. So who knows, maybe this is even an EC error?!?

Well, if you look at the bode plot and play with the inductor values it seems pretty obvious that they act as a high-pass filter with a cutoff frequency well below the 50MHz of the signal. Assuming the inductors were arbitrarily large then the steady decrease of signal strength as the frequency drops (below 15MHz) is probably because of the input capacitor and the similar decrease but this time as the frequency rises (above 15MHz) is probably because of parasitic capacitances in the transistors which can have a very disproportionately large effect when between the collector and base due to the miller effect. Decreasing the values of the inductors reveals that they act as a fairly sharp (in terms of dB/decade) high-pass filter with no increase in gain at the cutoff frequency due to resonance. The left-most inductor on the other hand does resonate with the input capacitor with increasing Q as the inductor size decreases, but at 4μH the Q is low enough that the resonance has no appreciable effect. The increasing Q with decreasing inductor size therefore increases the gain from resonance as the inductor decreases in size but the rate at which the resonant frequency goes up with decreasing inductor size and consequent decrease in signal strength from parasitic capacitances makes it so that ultimately the signal strength actually goes down with a smaller inductor even when the signal is always tuned to the resonant frequency. So ultimately there is no inductor size for the left most inductor where the resonance has any appreciable effect because when it’s tuned to a frequency where the signal isn’t squashed by parasitics the Q is too low for much gain and when it’s tuned to a high enough frequency where the Q is high enough to be useful it’s just squashed by the parasitics.

I hope my huge block of text isn’t too hard to follow.

Honestly I guess I just gotta learn how to read a bode plot, cause I understand what your saying, but part of me is still kinda in question of some of it. Oh and no worries, I tend to send huge blocks of texts all too often myself, so I totally understand 😁

The bode plot is pretty simple, although you have to be aware of a certain assumption it makes, which is that it always assumes the input signal to be small enough that nonlinear components such as transistors are effectively linear, even if said input signal is whatever whacky value you set it to be (e.g. 1kV). This allows it to treat everything as linear making analysis much easier, but of course it won’t give accurate values if the signal is ever large enough that nonlinearities become important (but if you plan to use the amp at much smaller signal values then the bode plot will still be accurate even if you feed the bode plot higher input signal strengths). The dB (decibel) measurement is magnitude difference of the signal at the given frequency to a 1V signal. 20dB thus means 10V, 40dB 100V, 60db 1kV, -20dB 100mV, -40dB 10mV, etc. So, that means that the 1μV input signal is -120dB, so if you select that wire in the bode plot it shows a flat -120dB line because no matter the frequency it’s always -120dB relative to 1V. If you select the output wire you can see the dB value at the output (and remember it’s always relative to 1V) and as you can see at 50MHz it’s a lot higher than the -120dB of the input and at 15MHz it reaches a peak because that’s the frequency this amp amplifies best. Now, as I’ve mentioned decibels are relative. So, 20dB actually just means a generic x10, not 10V specifically, but EC always makes dB values relative to 1V. So, when giving the amplification of an amp the dB number is relative to the input voltage, so for EC it’s best to set the input voltage to 1V for the purposes of measuring amplification via the bode plot, but this isn’t necessary if you’re only interested in the relative amplification at different frequencies (e.g. seeing if 1kHz amplifies more than 10kHz without worrying about how much amplification actually happens or not at 1kHz). Of course for an amp like this 1V at the input would massively overload it, but the bode plot assumes a small signal so it’ll just report some massive 120dB at the output even if there’s no way this amp would actually output 1MV. This means that the amp will amplify by 120dB (voltage multiplication of 1,000,000) at that frequency as long as the input signal is sufficiently small.

Wow thanks for writing all that, I’m gonna have to slowly get more familiar with using/reading dB’s on the app, it seems like it shows a lot of characteristics that aren’t so obvious with traditional analysis….

Also, I got some very interesting news regarding the 220MHz oscillator design: I took my soldered pcb version of the circuit to work, and probed it with a Spectrum Analyzer, just to see if my Oscilloscope was causing problems and affecting or simply not displaying what was actually happening, and the results were very exciting and interesting…. First, I set up the Spec-An to analyze everything within 100MHz and 600MHz, and I probed it using an oscilloscope 1x probe (cause that’s all I could find at the time), and I adjusted the potentiometer to set the current on the current mirror as I was observing the results…. And what I found is that it was actually oscillating right around 420 to 450MHz !!!!! That was so cool 😮😄. So, I should mention that when the Spec-An is running with the input floating (nothing being measured) -it shows just some noise at around -40 to -45dB, but when probing the output of the circuit, it sees a positive spike of about -20 to -25dB @ 420-450MHz using a 1x 100MHz probe. So I think that just goes to show finally why I was getting very weird frustrating measurements with the oscilloscope. It was likely extremely loading the circuit and being attenuated so much that it practically looked like it wasn’t oscillating at all. Crazy right?!

Wow, that’s cool! And 420 to 450MHz is right around the 70cm ham radio band (in America at least) so there’s a chance you’re annoying some hams every time you run your circuit. Maybe you could buy a tunable VHF receiver and try to see if you pick up any signals from your circuit. That should be a good test to see whether it’s oscillating or not.

I don’t think it’s gonna put out enough power for that; not sure. Do you know anything about antenna design? Or exactly how a wave propagates through an antenna, or I guess, standing waves; would you happen to know how all that works?

Also, this is technically UHF now isn’t it?? 🤭😅

Actually yeah that is UHF. I posted a circuit recently that was a transmission line. To get a rough simulation of a monopole antenna set the left resistor in that circuit to something small (or remove it entirely) and set the right resistor to about 1-10kΩ. This will make the transmission line behave like a monopole antenna, although the characteristic impedance will probably be quite different from a real monopole antenna. But if you do add an antenna it will influence the way your circuit oscillates because it has it’s own resonance and will also radiate away a lot of power potentially killing the oscillations.

Regarding whether it’s making enough power as-is to be heard by a radio, I do believe it is because even very low transmit powers (a few watts I guess?) can reach hundreds of miles in good conditions, so with only a foot or two of air between the circuit and the radio I imagine it’ll be a very clear signal with plenty of strength even if only 1mW is being actually transmitted. The usual modulation type for that frequency range is FM so no signal will sound like loud static and a constant unmodulated signal (like might be transmitted by your circuit) will sound like silence. But you have to make sure the squelch is open, otherwise you won’t hear any static even in the absence of a signal (the purpose of the squelch is to turn off the speaker when there is no signal to prevent you from having to hear loud and annoying static). I also recommend having a radio with a tuning dial or some such because that will make scanning for your signal a lot easier than having to check individual frequencies one by one by using the buttons to type in the frequency in MHz. And if you go so far as to get your own radio you might as well get yourself a ham radio technician license too.

Really all I would like to do is eventually make my own RF wireless transmitter and Receiver to transmit/receive digital data. I think I can handle the modulation/demodulation part of it on my own, but where I need the most help is how to actually turn a sine wave into an over-the-air wave and most importantly, receiving it. Most of my problems seem to be amplifying and buffering, because any transistor configuration I try are not effective or don’t work at all when it comes to pretty much anything in the MHz range. I even have trouble using resistors in the kOhm range, the signal goes straight from a few hundred mV at RF on one side to completely flat D.C. on the other side of the resistor unless the value is no more than a few hundred ohms at most. There are a bunch of factors I can think of that might explain this, but it’s so hard to say for sure what it actually is cause I have to rely on my 200MHz oscilloscope to show me everything, so If it’s not even showing anything true at those those higher frequencies, than that could be the source of all my problems. As an example, one thing that’s really weird is, instead of measuring a signal through a resistor, I’ll try an inductor, and you would think that if a floating 100uH inductor doesn’t pass the signal, then a 1000uH definitely won’t pass anything either, but no instead it seems completely freaking backwards…. and it’s super weird stuff like that which makes it so hard for me to tell if I’m doing anything right. I’ve even made an oscillator that my oscope shows it produces a 150MHz signal a couple volts pk-pk and m, but then measured an even larger sine wave on the whole positive rail of the source itself, which makes no sense! I always use a 5v/1amp phone charger as my power supply

If you want to transmit, then I’m pretty sure you’ll need a ham radio license to be able to transmit anything legally (but it’s pretty easy to get, I know because I have one). Also, given how many problems you’re having, why not start at a lower frequency and get your radio working there before trying to go for the more difficult higher frequencies? At the real low end there’s a ham radio CW band (CW a meaning morse code only) around 137kHz.

Actually, it looks like the 136kHz is only available for general class licensees (the three classes in order from least to most privileged are technician (what I have), general, and amateur extra). But I think there’s still plenty of lower frequency stuff you can do in the HF band (3-30MHz). A quick google search tells me that technician class operators can use the 10-meter band (28MHz to 28.5MHz) for lots of things including data, so you could use that one.

Also, regarding your troubles with VHF/UHF, I think that these components have lots of inductive and capacitive parasitics that you’re not taking into account which can potentially explain all your problems. For example, a higher inductance coil will have more windings which means more contact between the windings which means more capacitance which means more VHF/UHF signals get through. And your power supply cable if coiled up will form an inductor and if stretched out will make a transmission line, both of which will cause unexpected behavior on VHF/UHF. So I recommend connecting a capacitor between the power and ground rails of your circuit if you want to keep your power supply voltage constant.

I think that’s a great idea, and I’ve thought about trying to focus on the lower frequencies before to get a different perspective on it, however i suspect proper antenna design becomes even more critical at lower frequencies, so I’ve been putting that off for a while…. But since you mention it now too, I think that means I need to just stop anticipating and try it anyway

And I’ll go ahead and look into getting a ham radio license too, thanks for all the help and input!

No problem. And also antennae aren’t hard to design as long as you don’t mind an omnidirectional monopole or dipole. In fact, lower frequencies are probably actually easier because it doesn’t have all the wacky and unpredictable problems of VHF/UHF where you have to actually work to make the non-antenna parts actually not be antennae.

For the 10 meter band a simple 2.5m wire straight up will make a nice resonant monopole antenna. Alternatively you can attach two monopoles end-to-end with the feed in the middle to make a dipole, which is typically mounted horizontally. However, that’s a little more complicated because now you have to take the ground into consideration. Also, if it doesn’t resonate perfectly you can make it resonate by adding a capacitor or inductor of the right value. This is called tuning the antenna, and in many radio setups there’s a dedicated device for it called an antenna tuner.

Ok sweet, that doesn’t sound too bad at all actually…. For the antenna, are there any ways to condense it? like for example, instead of a straight 2.5 meter wire, can I just coil 2.5m of wire to to a height of a few inches to make the device more compact? Or would that not work do you think?

That would work, but not exactly. A shorter than optimal antenna can be effectively lengthened by adding an inductor, but the length of wire used in the inductor doesn’t necessarily correlate to the extra length added by it. Only the inductance should matter I think. It’s mainly about balancing the parasitic inductance with the parasitic capacitance to make it resonate at the desired frequency. Also if you do that the characteristic impedance of the antenna will change a bit, but that’s not much of a concern I think as long as you’re not feeding it via a coax cable (or other transmission line) that’s more than a significant portion of a wavelength long. If you get the resonance right at the frequency you’re transmitting at then the antenna will look to the circuit like a pure resistor with the resistance being the characteristic impedance of the antenna, although it will have some reactance (inductance or capacitance) at other frequencies, so it might be better modeled by a damped resonator.

Hmm ok, sounds like a lot of factors to worry about, I guess I’ll just have to try a little bit of everything and see what happens

I think with an antenna it's fine to just make a thing and see how well it works and tweak it or not as necessary. The worst that can happen is a large impedance mismatch which will reflect a portion of the transmission power back into the transmitter (the technical term for that is standing wave ratio or SWR), but as long as that's not too much of a problem for your transmitter then the most that'll do is reduce how much power gets actually transmitted.

Ok cool, sounds like a good idea then. Will see what happens