I have 3 "Emitter-Coupled" Oscillator, the difference is that I don't use a resistor on the tank circuit and I do vary the Inductor inductance. They're close on the PCB and they got some disturbs, if only one is running everythings is great, but if the others starts to oscillate, then everything got messed up. What can I do to fix this?
Everything got messed up: when more than 1 oscillator are working, their frequency isn't as stable as before, take for an example, running alone Osc 1: 110kHz, Osc2: 75kHz, Osc3: 45kHz, they oscilllate about 500Hz, but when more than 1 are working together, it seems like they frequency connect to each other.
I'm using this circuit as a "Metal Detector", so when I got some metal stuff near my inductor a LED should turn on, but when something get near to Ind from Osc1 if more than one is active, multiple leds will turn on.
My Oscilator are close, my inductors not much.
UPDATE: Tuned them on the same frequency, used 500ns on my oscilloscope and now my wave is looking like a hysteresis.
Edit:
Trying to clarify myself.
My board is 3 oscillators, like the one I said on first line, my tank circuit is capacitors, a coil 1:1 and a inductive loop, this one is kinda far from my board. I have them on a 8051 which I'm measuring their frequency and checking if there's a metal near.
About AC analysis I'm trying to do the things the way I know, I don't know much about oscillator nor Control theory, so I'm just observing things slowly. About the capacitor I can try, another notable thing is, a circuit like mine that works have 2 differences, their oscillators are far apart from each other and they use Neon Lamp to protect the loop while I use a bourne gas discharger. I read that gas discharge have a high capacitance, would that interfer this much?
Tried to couple each oscillater voltage supply(5V) using a 100nF ceramic capacitor, still the same problem.
I guess I have a problem with the stored energy on inductor, how could I eliminate "odd" current, I mean eliminate the current that was brought by other oscillator or the metal messing up with the magnetic field? Also, tried using a 1M resistor parallel to LC Tank, nothing changed, should it be minor?
\$\begingroup\$
\$\endgroup\$
3
-
\$\begingroup\$ Did you perform already a loop gain analysis (ac analysis) ? Be aware that excessive loop gain (magnititude much larger than unity) may cause such problems because the active devices are driven deep into saturation. More than that, a resistor in parallel to the tank can reduce the gain and makes the circuit less sensitive to parasitic effects. \$\endgroup\$– LvWCommented Jun 8, 2018 at 16:13
-
\$\begingroup\$ Added some edits on main post, if there's anything confusing or that I didn't explained please correct me. \$\endgroup\$– Marlon Franco SantosCommented Jun 8, 2018 at 18:08
-
\$\begingroup\$ Impossible to see what you have done \$\endgroup\$– D.A.S.Commented Jun 9, 2018 at 23:03
Add a comment
|
1 Answer
\$\begingroup\$
\$\endgroup\$
6
It sounds like you have more than one high Q tank oscillators in close proximity and close frequencies. If the coils are in the same orientation inline or near fringes then they couple weakly like a transformer and will track the one with the higher current flux if within the bandwidth.
The physical layout and stray capacitance also affect everything so we are blind to the real issues.
The Q is simply the impedance ratio at resonance.
Q=Req/Zc(f) Since there are two R’s as shown you can have collector only with Q=100 in this example or both for Q=50. Q1b=Q2c are on the same node.
Q is the effective gain above Q=1 and AT crystals have a Q in the 10k value.
Upon reading your edits of purpose it sounds like a lot of improvements are needed with details missing.
Added
GDT’s go from low pF to low Ohms thus over dampening the Q like an SCR crowbar. Not good. A linear TVS may be better with heat managed somehow.
-
\$\begingroup\$ Added some updates, hope it can clarify things a bit. \$\endgroup\$ Commented Jun 8, 2018 at 19:47
-
\$\begingroup\$ Add 1uF capacitors to the VDD power input of each oscillator; use 10 ohms or 100 ohms from each oscillator VDD/capacitor to your global VDD/battery node, and place 1uF/10uF/100uF across the battery. \$\endgroup\$ Commented Jun 9, 2018 at 5:02
-
\$\begingroup\$ Tony, besides decoupling capacitors on each oscillator VDD, is there a way that I can isolate them, so they can be next to each other but don't "track the same thing". \$\endgroup\$ Commented Jul 20, 2018 at 19:24
-
1\$\begingroup\$ Coils must be orthogonal and perhaps a shield between or gap >2x bigger than coil \$\endgroup\$– D.A.S.Commented Jul 20, 2018 at 19:28
-
\$\begingroup\$ Those are my three oscillators, each board have them, but they interfere between other boards. The board behind this one can't use the same frequencies as the one at front. \$\endgroup\$ Commented Jul 20, 2018 at 19:33