Start From Crystal (II)
2/12 2019
Crystal Set
Well, In episode Experience Crystal Set, I know several important things about the Crystal Set. We must had a good antenna, and what we received is what we received. It's not by design, especially for a newbie. Middle wave broad cast is even worse, mainly depend where you are.
I decide use 5 cm diameter coil, taped as band selection, not limited it to MW band. Calculation indicate 60 Turns coil should enough cover Medium wave and several Shortwave bands.
This time still begin from Crystal Set. Taped inductor is critical for searing the crystal available band. My location is not good spot for receiving MW broadcasting.
LC tank not function?
Well, connecting antenna to Top of tank is recommended for beginner.
I had upgraded my antenna to a Invert V, each arm had 15 meters long wire. I suppose this definitely make me easy to get some positive result.
And it does. I notice some unusual phenomenon. No matter which tap i use, the station i got is SAME, definitely SAME. And the signal copied take whole scales, no selectivity at all.
This phenomenon has explained by W0BTU.
Directly connecting antenna to TOP of the tank is the WORST choice ever! My antenna is 15 meter for each arm, this is the minimal length for crystal radio to recepte medium wave broadcasting per W0BTU's opinion.
Directly connecting the antenna to Top tank, ruined everything. The antenna is short ( 50 feet, 15 meter) , electronically speaking, so it's capacitance is huge for Medium wave. This capacitance might domain the hole tank resonant frequency.
It's complex than i thought, this explained why only that strong station emerge on each tap. But still can not explain why that station disappear if don't connecting the variable capacitor to inductor.
Connecting the Antenna to top is also an interesting radio demo to children. I want more than that. One of the Invert V arms use as crystal antenna, and another arm use as ground connection.
Solution for Invert V coupling:
Make another winding on the tank as Antenna input. By this way, crystal set might take full capability of the antenna, i suppose. Obviously, medium wave need more couple turns than shortwave band. This seems works. The selectivity comes back, different tap now resonant in different band. I even received one MW station. I hook up a TA2822 Audio amplifier which had 40dB gain, then I got several other weak Medium wave station.
This is working well for HF band, but for Medium Wave this seem not enough.
Detector Module
Two diode soldering on the supporting bracket. 1N34 and 1N60. The red 2 mm banana socket connect to inductor Tap. Other 2 2mm sockets use as output.
Use quality component and carefully building make the module seems valuable, then i might like to keep it for later experiment.
This whole module mount to a steel chassis which use as GND.
Reference for crystal radio:
Couple Antenna For Medium Wave Band
There is more important notes form W0BTU's blog. What important to antenna, is make it resonant somewhere. For medium wave receiving, it's better as long as possible. But we can not do that for so may different reason. My invert V had 15 meters Arm, to make it efficiency, a LOADING COIL is suggested by W0BTU. It might need as much turns as tune winding, even more, a taped inductor is a good thing to play with.
For the medium wave use invert V antenna's 2 arm is not good enough. I gonna take this chance to make a good ground. I buried a 50 cm M8 screw vertically, on it's body, twisting full of tinned copper wire.
I expecting this would working far more better than the Invert V. the result disappoint me pretty much. I actually found this is just as better as use the power cord ground. The better ground for me is my working tables's steel shelves.
The 30T on PCV coil forms an antenna tuner, the Ferrite rod could adjust for best sensitivity.
Another tricks is light capacitive couple to top tank. series a small capacitor, around 1 to 10 pF for electronically short antenna. For electronically long antenna, might be 100 pF, not my case though. In this way the antenna capacitance is killed. I tried it, it useful, but for WM receiving use another winding is better than this one.
Crystal radio experiment probably end here. There is more tricks to play, like wave-trap for strong local radio( Here do have one strong shortwave station ).
More on crystal radio:
Summary
Searching MW and SW band make crystal set easy to sound.
There is a Local SW broadcasting very powerful, wave trap is prefer.
Inductance couple the Invert V to tune tank help SW reception.
Medium Wave need absolutely good ground.
0V1 TRF
Audio Amplifier
Testing crystal radio in house is kind of torture, the signal is weak no matter whatever you do. Adding Stages of Audio amplifier is the simplest and straight way to improvement the crystal set any way and so well known.
How about the simplest audio amplifier?
You better not connecting any low impendance earphone on this little Audio PA. 300 Ohm and higher is better. The old telephone speaker is perfect for this amplifier.
The input impedance typically 1~2 Kohm whereas the quiescent current is several mA. 47K resistor is only a DC loading for crystal detector.
While the signal became strong, the voice is audible in quiet room on night.
For a different type transistor, it's possible oscillating. For SS8050, it gonna to be lots of high frequency hiss. Add a degenerate emitter resistor would be better.
TL431 Audio Amplifier
I saw this novel TL431 amplifier on Teclib's website. And this little audio amplifier been used wildly in theradioboard.com project. The gain of this little beast probably 40 dB and more. This is a simple class A amplifier, and it's input impendance is very high, so very suitable for crystal audio amplifier.
1.6 Mohm bias, on 12V rail got about 30 mA quiescent current, roughly estimate by (vcc-2.5)/300 .
330pF feedback killed most of high frequency response(too much high frequency been killed, try 100p please). Headset is far more sensitive than speaker, high frequency hiss is really annoying, feedback capacitor help a lot. Try 30 pF to 330 pF as your wish.
47 k loading resistor set the impendance of this little amplifier.
I try to setup a test bed with modules. The modules i ever built are almost garbage, been thrown away while done with it. There is nothing left while i begin to exploring new stuff. I decide make the module pretty, and keep them for later use.
Summary
TL431 amplifier is simple, high gain and high input /low output impendance, perfect crystal radio audio stage.
Limited by small quiescent current, it's easily over-driven and cause audio distortion, especially with a low impedance speaker. Choose current limited resistor from 68 ohm to 1000K as your wish.
Crystal radio to room volume ! beating LM386 (If you don't care power efficient).
Wild 1V1 TRF
Un-tuned wild RF Amplifier
I wonder what would happen if adding a RF amplifier before couple to Tuner without preselect filter.
I suspect no band pass filter would be a totally disaster , tremendous noise would be amplified. Ltspice suggest the amplifier would have 15~25 dB voltage gain with a resistor collector load, use RF Choke increase the gain to 30 dB. Gain roll down about 15-25 Mhz.
These two little piece amplifier is built. The result is amazing, signal was strengthened enough to overloading the TL431 audio amplifier. Even without the audio amplifier, the crystal detector itself could drive the earphone into room audible volume at HF band.
If dear to connect the RF amplifier to any tap, the strongest HF signal gonna take over whole tuner scale.
Magnetically couple the signal by another winding is much better.
(NOTE: better to use one or two turns as couple winding.)
DO NOT directly couple the signal Directly to the crystal tune tank ! The First rule for RF amplifier: extremely light couple the crystal tank! (Gain is too high and the tap had too much turns.)
Connect a wire to RF output, it would act like a antenna. The other end of wire need NOT connect to GND, just put that wire around to the crystal tune.
For HF band, connected the other end to GND help to reduce the gain hence make light couple easily. This RF amplifier had too much gain, you could add 15-51 ohm degeneration resistor to improve the stability .
TRF Choices
Strong couple from RF stage would ruined the detector's selectivity
Strong signal overwhelm the audio amplifier
High rf Gain might cause oscillating
Prefer inductance couple, which is far more better than tapped couple. RF stage before tune circuits is wild.
Options available for curing these problem:
Audio. TL431 could not handle large signal, it's class A. Low Gain LM386 or similar chip could be alternative. Personally prefer to use a simple push-pull or a Class-A amplifier match the impedance by a transformer.
RF pre-amplifier
An RF amplifier with too much gain is hard to tame. It easily overload next stages. Spreading the gain in different stages is better.
Another tradeoff is tuned amplifier or not tuned. The RF amplifier act as an active antenna instead of LAN. Such a RF stage amplify too much out-band signal, overwhelm the later stage. The HF station is extremely strong ... To cure the problem, couple between the RF stage to Tune Tank should be very light, the high gain RF stage is no good.
Simple radio receiver have only one tuned stage, prefer use After Tune RF Amplifier. Selectivity and out band rejection done in same tuned circuits.
An advanced receiver most likely use a before tuner Rf Amplifier with high dynamic range capability and out band signal filter. And there might also have antenna tuner /resonant antenna/ preselector.
Handle strong Audio Signal
MW signal might only several mV after detector, but the HF signal reached to around 200 mV, this gonna overloading any simple Audio stage.
One of solutions might be feedback amplifier. But simulating did not suggest any audio amplifier can handle such a strong signal and also keep sensitive to small signal.
Looking for an audio stages could:
Handle large signal and keep relatively small distortion
Handle Small signal in full power of Gain
Possible solution: AGC, Compressor and peaking-limit
Start with a simple 2 stages transformer output class A PA.
60 mW maxiom output. 40mA quiescent current
overdrive generating noise and severe distortion
input impendance is low
First Stage input impendance is several Kohm, not very suitable for directly connected to crystal style detector. Second stages input Independence is only ~500 ohm.
Audio transformer has 3:1 turns ratio, only 50 mW available. 40mA quiescent current generated about o.5W heat dissipation, half of SS8050 power capacity, quiet hot.
Adding a input filter which knee point about 4-7 khz
A 1N60 voltage limiter is added to limitation overdrive the PA
2 100 pF capacitors at the C-E junction of transistor kill the high frequency instabilities.
1N60 use to limit the maximum input signal delivery to the PA. 200 mV peak signal happens to be exactly the overdrive point. Output is obviously distortion. Hearing such a round top signal is not that bad, much better than overdrive.
Efficiency of Class A Audio Amplifier
I thought use an audio transformer with ratio 5:1 or 6:1 could delivery more peak power on same quiescent current. Such a PA has at least 2 factors limited maximum available power:
available current: quiescent current, multiply the turns ratio is the maximum current could delivery to the load.
available voltage swing: transistor needs 2v bargain, the peak voltage is (vcc-2), then divided by turns-ratio delivery to the load.
The 3:1 ratio easily drain the quiescent current while the voltage swing still had potential. I replace the transformer with a high ratio one, 5:1. The output still distorted at almost same voltage level, nothing improved!! Stuned!
I tried several different audio transformer, Nothing improved !! Go for Ltspice... After play with everything, just can not make breakthrough: 1 Vpp output swing is kind of hard limit! (with the same quiescent current)
Changing the impendance ratio would run into other limiting factor. The input signal could not exceeding 0.7 Vpp, it gonna reverse bias the transistor anyway. To overcome this problem, changing the bias to resistor divider, add a degenerate emitter resistor, and then it gonna use a higher base voltage to provide same quiescent current. Thus reduce the available voltage swing, this is a endless loop.
Suddenly It hit me, i actually trying to improve the efficiency instead of max output power. 25% maximum, i use a so-so transformer and not carefully design, so 10% efficiency! ( the transformer primary had 10 ohm DC resistance)
From power point of view, everything explained. with the fixed quiescent current, max available power is not easy to improve. My 12V 40mA quiescent for a very good efficiency, let say 20%, we definitely could not reach the theory maximum efficiency(50%), the output voltage only 1.2V if keep it not so distortion!!!! And for a 6:1 ratio transformer the voltage available about only 1.6V peak (12Vcc - 2V margin ), this indicate even increasing the quiescent current, maximum 100 mW (1.3 Vpeak) power is almost the hard limit.
While transformer's current drain out (highest voltage swing), the output became distortion, for a better signal, gonna need few more mA current, efficiency become even lower.
Back to Handle Large signal
Both amplitude limiter and audio compressor tested, none of them could handle 0.x mV to exceeding 200 mV audio signal.
AGC is last straw. Reduce the gain could make small signal disappeared, and then reduce noise while received strong signal. Shunt type AGC could be deployed easily and improve dynamic range.
Typical shunt AGC.
2N7000 is much cheaper than JFET
MOSFET/JFET has good linearity than BJT
input impendance reduced while large signal applied
a buffer stage is prefer while connected to crystal style detector
This little amplifier could provide 40 mW to 60 mW power. If it been over driven, tremendous noise is generated, voltage spike on the transformer might be high, tens volt even hundred. The Audio AGC prevent it from heavily over driving.
Notes and results:
30 dB input signal signal dynamic range without severe distortion, sound good.
AGC time set long to prevent oscillation, change 82k to larger value in case it is unstable.
To improve AGC control, change the output transformer to 4:1 to 6:1 version.
Overall gain 35 dB -40 dB, frequency bandwidth from 500 to 5k kHz (simulation result)
AGC Control Method
Use BJT as AGC Control device is rather common back to transistor ages. Use negative voltage reduce the transistor current then reduce gain.
Use another transistor control the quiescent current is much better than one transistor. In this way,wider control range is achieved with small AGC current.
But control the quiescent current impact the amplifier. The input impendance is changed a lot, the very biggest problem is that a very strong signal gonna make the AGC pull out all current from transistor, the transistor then became distorted. The result is worst sound under strongest signal.
The shunt type AGC is much better. Shunt type AGC gonna change the input impendance rapidly, the lowest impendance defined by preceding divider resistor. Use with a pre-amplifier or buffer is suggested.
Enhanced MOS-FET shunt type AGC control is very easy to use and have wider control range,and easily setting the time constant. JEFT is another favor way but JFET is expensive than MOS-FET nowadays.
Try to receiver MW signal from Strong HF background
Bigger antenna better signal, but with most strongest HF signal instead of MW signal. The simple receiver's tuned tank easily been destroyed, occasionally, the strongest Shortwave signal flooded everywhere.
The strong out band signal ruined everything.... literally everything. Out Band Signal must be eliminated in fist places, a AM band LPF or BPF help a lot.
Try this TRF Receiver
Out Band reject filter remove the Excessive HF signal, this is the key for receive the weak MW signal.
one RF stage might not enough to recovery the weak signal. So another RF amplifier added to output of the LC tank.
2Mhz Butterworth LPF
The terminator resistor is not that critical, for my simple RF amplifier, it's input impendance around 500 R @ 1Mhz, and 100R @5Mhz.
It's suppose use a band-pass filter. But below 500kHz, there is no stations. Simple LPF did the tricks.
MW stations emerged, and soon, I found the gain is insufficient receive weak MW stations. I need sensitive detector or add one more RF amplifier.
Sensitive envelop detector
V. Polyakov suggested a small detector circuits. which bias the transistor on it's starting up ramp, works well for detecting the AM signal. Sensitive Envelope Detector.
V. Polyakov AM detector
quiescent current about 0.5mA.
Only 10-15 uA flow through the diode, dynamic resistance is quite high (maybe 10k - 100k range).
Output DC component could provide AGC voltage from 1.5V to 0.55V, DC voltage is high while there is no signal.
22k and 680p capacitor set the knee frequency about 10 kHz
V. Polyakov said this detector has good efficiency works well and voltage gain is around 60. Input impedance is not too high, and output impedance is not low as expected, thus need a audio amplifier better have dozens kilo ohm.
The second diode bias by a 330k , double the sensitivity compare to the one diode version.
Double diode version works from hundreds μV to 10 mV , more than 0.5V output will get clipped. Using a LED set collector voltage should be much better for dynamic range.
After tank RF amplifier
Only thing need to mention about this is that very small couple capacitor from tank to this RF amplifier.
Because the RF amplifier is BJT, it's input impendance gonna ruined the High Q tank. The small couple capacitor do the trick, but compromised the gain.
With is configuration, the receiver now received the AM signal well . One stage tuned circuits is enough to separated adjacency stations, peak one signal is hard without speed reducer.
This amplifier terminated the 2Mhz LPF. Too much Gain, maybe 40dB.
Tend to oscillation, bad input impendance 500 R @ 1Mhz, and 100R @5Mhz.
Passband could be bumpy. Signal above 4Mhz is well filter out. ( by Spectrum Analyzer).
Insert An audio LPF after the Detector
A simple 500R 5khz audio LPF inserted.
The terminator impendance is not that critical as source impendance. 300 ohm to 1k won't impact too much, only little bit group delay change. Instead, the source impendance is critical. The L6 value choice is 'random', from 15 - 25 mH, the band pass changing not too much.
AM tuner tank
The final audio amplifier with AGC , reuse the module built weeks ago.
This Audio amplifier already had a RC type LPF, which cut off frequency is near 5Khz, and it might overkill, lost too much high pitch sound.
TRF with after-tuner RF amplifier
It's time say goodbye to crazy mad TRF. The most problem due to the uncontrollable RF amplifier. Too much gain and tend to oscillate, cause tremendous noise.
RF amplifier after tuner is good choice, avoid troubles to make a wide bandpass filter , avoid much more trouble to terminated the filter.
Stable the RF amplifier
RF amplifier had too much gain is disaster. I dip into that wild amplifier several months, so impressed . RF voltage Gain > 25 dB (17x voltage gain) is almost a hard limit before it's tending to oscillate. Of course, this Gain limit depending what transistor you use.
Per experiment, one stage RF Amplifier deploy 10dB to 20 dB voltage gain, i.e, voltage 3x to 10x, is easy and stable. Use RFC as load and capacitor couple. In this way, RF voltage gain is around 20 dB @1 Mhz/per stage.
To cure the unstable RF amplifier, 2 methods deployed to RF amplifier:
Insert a 15 Ohm to 33 Ohm degeneration resistor. low overall voltage gain, while coupled to next stages, it's gain varied from 20 dB to 30dB (around 1Mhz), input impendance around 200 Ohm, and decrease quickly while frequency increase.
Couple the stages with a small enough capacitor, so the low frequency gain decreased. NOTE: this is just pushing the self resonant frequency higher, and if luck enough, there might no gain anymore.
It's now stable while coupled to next stage.
Couple Huge antenna to small TRF receiver
An 15 meter wire antenna use to testing the receiver in door, for a long time, I suspect the MW signal been flooded by the very strong HF stations.
No matter how light i coupled the wire to the Rod, even only one turn, or even just put wire around the rod antenna. The HF station always dominated the whole band. LC tuner did not work at all.
Finally, i notice this: connect the wire antenna via a 22pF capacitor to the top of the LC tank, everything works very well. 22pf isolated the huge capacitance introduced by the long wire.
Connect the wire to tap or another winding not a good method for this antenna. This probably because the wire capacitance and the tap inductance resonant at HF.
Another possible reaseason, might be the shield of the rod antenna is poor, rod itself is antenna after all. I winded another Toroid inductor, use a standalone one turn winding as antenna couple point, that one works well.
TRF, especially this rough one, selectivity is poor, any stray couple from that HF friendly antenna is gonna make the TRF slide to HF band.
Capacitive ground plane
Recently, a good laboratory breadboard builted. It's not intend to making RF modules. That won't scare me, I built a TRF receiver on it. Yeah, a very noise TRF, it does not oscillating, though.
I try to cure the problem, there is no ground plane, and the GND is routed everywhere, totally not conform any RF rule.
But, while i connected one GND joint to the bellowing copper clapboard, magical thing happened, the TRF silent as it built on a ground plane. I guess this clapboard make a capacitance with corresponding wire on the breadboard, make it partly ground plane, a capacitance ground plane.
It's works well on MW, HF? don't know yet.
How about this sensitive detector
Sound much more noisy than diode detector. Diode detector gonna clipp the weak signal, let only the strong signal pass, but transistor amplified weak signal. Transistor detector's threshold is much lower than diode.
Remove one stage RF amplifier help a lot. BTW: use 15m long wire attenna.
Outdoor Testing
To take the receiver outdoor, i need duplicate a smaller platform for the modules. The 2V386 version is lack of sensitivity, only one station received. but lately i realized the 365pF capacitor is problematic, it's rotator soldering pin is missing, contact resistance is high and then, selectivity is poor.
With a new small testing receiver, this 3V386 receiver is under testing. 3 RF amplifier crowed in a small board , it's disaster, and tending to oscillate. Adjust the ferrite rod antenna help a little, at least most of the AM band is quite.
3V386 could received 3-5 station in downtown. the loudspeaker volume is acceptable and audible in a relatively quiet environment.
Unstable RF amplifier
Put 3 RF amplifiers on the small clapp board is disaster. One step forward, deploy Polyakov transistor detector leads to unavoidable oscillate. The overall feedback might not enough to make it became a stable oscillator, but enough to make a regenerative receiver.
60 ~ 80 dB RF amplifier gain is 'hard' limitation to the simple, rough TRF. And with 3xRf amplifier and a diode detector, 3V386 could deliver a audible TRF receiver with perceptible selectivity.