Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Tank resonance locator

Status
Not open for further replies.
For giggles and kicks I tried to make the tank circuit with an oscillator with <10mΩ source impedance but not consume much power driving the series tank circuit.

It was a bit problematic but it oscillated at resonant frequency. The challenge was to get Barkausen criteria with positive feedback and minimal DC offset into a 50mΩ current shunt with negative feedback for DC operating point and not suck much current. I used +/- 3.7V to DC couple and prevent DC current thru the inductor.

But in the time I spent on it, it wasn't good enough to be reliable for a wide range of LC values.

View attachment 94455

To make a good sample and hold , you can use CMOS analog switches but low bias FET input OP Amps and low leakage caps ( Plastic film)
I understand but this looks good enough.
I have a question. Let's say I have a system which sweeping so fast that it is going through the resonant frequency. If I want to make it more sluggish or let us say I want it to lock some KHz above the resonant frequency (so that it is more practical), how would I go about it?
 
...Let's say I have a system which sweeping so fast that it is going through the resonant frequency. If I want to make it more sluggish or let us say I want it to lock some KHz above the resonant frequency (so that it is more practical), how would I go about it?
How about designing it to do what you want? What kind of question is that?
 
How about designing it to do what you want? What kind of question is that?
My question was that is it practical to stop way before it hits resonance to avoid frequency sweeping to the capacitive side, as a precaution. Or would it always lock at resonance.
Is it just the capacitance and resistance connected to VCO that I can change.
 
Last edited:
Just like Spectrum Analyzers, the sweep rate is controlled by the RBW ( resolution Bandwidth)

If the RBW is 1kHz then the sweep rate must be < 0.2kHz/s but to sweep 150kHz, that would be too slow.
EDITED
One way to do it is to differentiate the Peak voltage across the sense R and use the inverted and amplified voltage to slow down the +Ramp voltage vs f. or the reverse if visa versa. Then the sweep rate can be done 50x faster then reduce to 1x or 20% of RBW (-3dB)( where RBW=f_res/Q)

To make the response time of the peak detector fast, use a fast S&H with low sag, and compare an input vs output V S&H to drive the CMOS Analog switch.

This is far faster than an RC filter on a peak detector, but prone to spike noise, so the current signal must be prefiltered.

Then adjust the mixer gain in an inverting Op Amp, so the ramp slew rate is fixed V_target but can be modulated by a current source the Precision Op Amp peak detector voltage ( boosted to a few volts)


An RLC meter should be able to display RLC, Rs, Cp, , D, Q and have auto-cal with open cct and short circuit on fly leads.
 
Last edited:
100kHz would be nice and offered in some.

This is what bought.
http://www.extech.com/instruments/product.asp?catid=56&prodid=349
~$210 http://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_198678_-1
1mΩ resolution
with 0.1uH & 5 digit resolution. More expensive than than 1uH resolution, less than 0.01uH.
120 HZ and 1kHz std.
I doubt any Chinese Clones< $100 can match performance.

I would suggest for you , the next model up
**broken link removed**
5 frequencies up to 100kHz

Inductance 20μH, 200μH, 2000.0μH, 20.0000mH, 200.00mH ±(0.5%rdg + 5 digits)
2000.0H, 20.000H, 200.00H, 2000.0H (DF<0.1)
Capacitance 20pF, 200pF, 2000pF, 20.000nF, 200.00nF, 2000.0nF ±(0.5%rdg + 5 digits)
20.000μF, 200.00μF, 2.0000mF, 20.00mF (DF<0.1)
Resistance 20.00Ω, 200.00Ω, 2.0000kΩ, 20.000kΩ, ±(0.5%rdg + 5 digits)
200.00kΩ, 2.0000MΩ, 20.000MΩ, 200.0MΩ
DF (with C) 0.000 to 999
Q 0.000 to 999
Phase ±90°
Test Frequency 100Hz/120Hz/1kHz/10kHz/100kHz


but only 0.01Ω resolution.
 
Last edited:
100kHz would be nice and offered in some.

This is what bought.
http://www.extech.com/instruments/product.asp?catid=56&prodid=349
~$210 http://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_198678_-1
1mΩ resolution
with 0.1uH & 5 digit resolution. More expensive than than 1uH resolution, less than 0.01uH.
120 HZ and 1kHz std.
I doubt any Chinese Clones< $100 can match performance.

I would suggest for you , the next model up
**broken link removed**
5 frequencies up to 100kHz

Inductance 20μH, 200μH, 2000.0μH, 20.0000mH, 200.00mH ±(0.5%rdg + 5 digits)
2000.0H, 20.000H, 200.00H, 2000.0H (DF<0.1)
Capacitance 20pF, 200pF, 2000pF, 20.000nF, 200.00nF, 2000.0nF ±(0.5%rdg + 5 digits)
20.000μF, 200.00μF, 2.0000mF, 20.00mF (DF<0.1)
Resistance 20.00Ω, 200.00Ω, 2.0000kΩ, 20.000kΩ, ±(0.5%rdg + 5 digits)
200.00kΩ, 2.0000MΩ, 20.000MΩ, 200.0MΩ
DF (with C) 0.000 to 999
Q 0.000 to 999
Phase ±90°
Test Frequency 100Hz/120Hz/1kHz/10kHz/100kHz


but only 0.01Ω resolution.
Thank you. I have decided to go with the latest model.
 
Wise decision.

Then save all part data to PC
and make a database or design a spreadsheet to mix and match all parts, computing f, Q, D, efficiency , power loss, impedance vs f. harmonic content etc.
Drive levels ...

For very high power coils they need to be copper water tubes for cooling , high f, then use Polyestor or Polyurethane (huge industrial caps) for things like very efficient localized red hot steel for industrial use.
 
then you don't have to be clairvoyant with dynamic VCO that slows down as it approaches resonance at high Q>10 with fast slew rate and low power.

Another simple idea is use a 3 terminal LDO as a CC source such as 100mA and variable Nch FET on low side with limited current on high side LDO and invert Current Sense feedback from 75mV shunt at max current and gain > x 50k to make Series Resonant Oscillator. Negative feedback controls DC Vds to V/2 and positive feedback causes oscillation instantly at 0 deg phase shift. Using 10x wide BW Current V_Amp to result in low phase shift. THus at 100kHz with Av=50k >> GBW= 0.1MHz*50k*10= 500MHz may need two stages. Much simpler than sweep method and direct F reading within 1 second, as long as it doesn't resonate at spurious harmonics where phase shift is 360 deg and fundamental at 0 deg in series.

**broken link removed**

High value of current at resonance produces very high values of voltage across the inductor and capacitor. So limiting current to <<1A is desirable.

But at a minimum you can expect peak Vd or Vc to be 2x Vcc due to inductor with low Rs.

The Series Resonator Oscillator method is faster since rise time to full oscillation is only a function of k * 1/BW-3dB = Q/f * k, whereas sweeping almost three (3) decades takes 100x to 1000x longer for Q~10, where k depends on full scale signal compared to seed impulse noise that starts oscillator.
 
Last edited:
Eureka. I think I have a working micropower clock generator that starts instantly from a parallel tank circuit

Here's is my interactive simulation.
Try this

Negative feedback is for self Bias with small hysteresis reduction.
Positive feedback is to satisfy Barkhausen criteria with Tank circuit.
Try click switch to change frequency 3x ,
Press Reset to see startup operation in slow motion.

Thus with gain >>1e6 in a high speed comparator, the logic level feedback is integrated to a sinewave current due to parallel resonance is much higher than resonant reactance ~< 1 Ohm so the Q should be high and current is limited by 3.7V / 100K iuf using a LipPo battery. Noise abatement is critical, as AM radio will be stronger than the injected signal using shielded or twist pair with large CM choke ( ferrite torroid). Of cource add RF cap across comparator close to chip. Wires cannot be shown like in schematic to DUT with 100 uVpp for example.

upload_2015-10-2_0-47-25.png
 
Eureka. I think I have a working micropower clock generator that starts instantly from a parallel tank circuit

Here's is my interactive simulation.
Try this

Negative feedback is for self Bias with small hysteresis reduction.
Positive feedback is to satisfy Barkhausen criteria with Tank circuit.
Try click switch to change frequency 3x ,
Press Reset to see startup operation in slow motion.

Thus with gain >>1e6 in a high speed comparator, the logic level feedback is integrated to a sinewave current due to parallel resonance is much higher than resonant reactance ~< 1 Ohm so the Q should be high and current is limited by 3.7V / 100K iuf using a LipPo battery. Noise abatement is critical, as AM radio will be stronger than the injected signal using shielded or twist pair with large CM choke ( ferrite torroid). Of cource add RF cap across comparator close to chip. Wires cannot be shown like in schematic to DUT with 100 uVpp for example.

View attachment 94479
Wow Tony Stewart
This works like a million dollars on simulation. I'll try it doing it practically today itself.
Thanks a lot. You're a genius.
 
Wow Tony Stewart
This works like a million dollars on simulation. I'll try it doing it practically today itself.
Thanks a lot. You're a genius.
There is still a challenge to choose the right comparator. They have high speed , low latency but also low gain compared to Op AMPs.

I don't know recall these days which is the best high gain comparator or high GBW Op AMP used as a comparator.
if BW is 7x150kHz for reasonable square edges, with a minimum gain of 1e5, the GBW product would be extremely high or 1e6Hz*1e5 = 1e11.
Cascading two comparators is tricky to prevent spurious oscillations.

Background.
I estimate from 100uV to 3.7V square wave, it would be like 100uV sine to <100Vpp sine clipped to 3.7 and similar to a gain of 110~120dB depending on slew rate out, which isn't critical. 100dB gain would be 1e5 or 10Vpp sine clipped to 3.7 might be the worst case large signal gain.

The LM339 only has a gain of 200V/mV typ ( 50min) which is 2e5 compared to Op Amps with a gain of 3e6 but lower slew rate for large signal swing. The LM339 may be marginal for some LC combinations of Q which also gives gain with the 100K feedback R/X(f) ratio.

CORRECTION(edit)
It may be possible to use two CMOS buffered inverters each self biased with 1M Feedback and 10uF coupling which has 10s time constant or so. then use the resistor feedback to input with the Parallel LC load. However it may have spurious resonances from stray reactance and required good Vcc cap. My experience is each buffered CMOS inverters have a analog signal gain of 1000 ( 3stages x10) or 60dB so 2 inverters self biased could be 120dB with high slew rate or transition frequency much greater than you need. THe value of C must be slightly higher than the DUT capacitance, in the comparator mode where self bias (-input) RC= T and (+input) feed impedance RC =T
creates a relaxation effect but resonates when the external load reactance is in phase at resonance.
So good luck.
 
Last edited:
Note the comparator Vcm must be able to work near Vdd (V-) or 0V and out must swing near rail to rail even if open collector with Pullup R, unlike many BJT Op Amps.
 
If you want a dead simple oscillator that runs off an LC tank, then a negative resistance oscillator using a Lambda diode is about as simple and reliable as you can get. For example:
**broken link removed**
The oscillator circuit is at the bottom of the page.
I've used essentially the same circuit for measuring resonance of LC tanks.
 
If you want a dead simple oscillator that runs off an LC tank, then a negative resistance oscillator using a Lambda diode is about as simple and reliable as you can get. For example:
**broken link removed**
The oscillator circuit is at the bottom of the page.
I've used essentially the same circuit for measuring resonance of LC tanks.

Unfortunately BobW, ( I believe that circuit can't work here, because the negative resistance is too gm*Id is too low.

( I recall) In order for it to oscillate, the net resistance in series must be negative or equal to zero with the current sources ( or regerative gain at 0 deg phase= 1 or more.)

The reactance at resonance here is < 1Ohm.
 
Good point. I just had a look back through the discussion and hadn't realized the huge C/L ratio. I have the negative resistance test oscillator in my shop and could test it with the proposed L and C values, except that I'm on the road for a few days. I'm thinking that if it doesn't work as is, then a broadband impedance matching transformer could bring the resonant impedance into a suitable range. But that would make the negative resistance circuit somewhat less appealing. Will report back later.
 
any luck with my cct?
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top