tuning an LC

I need an LC tank that's resonant at 107 KHz. It will be "pumped" by a mosfet and have about 60 volts p-p across it. L is about 75 uH, so Xl is around 50 ohms. Peak inductor current is half an amp or so.

I need to hit close to 107KHz in production, and available small shielded inductors are mostly +-20%. So, how to tune it?

Trim caps and varicaps are too small, and 60 is a lot of volts.

Use a trimmable pot core, but that's a custom part PITA. RF type trimmable inductors look too wimpy for this frequency and current.

Make a capacitive DAC, maybe 4 bits, c values weighted 1-2-4-8. Switch the caps across the LC tank with a dip switch or relays or SSRs, gives 16 tuning steps. The bigger caps have to be pretty accurate to keep the steps sorta uniform.

Put pin sockets on the board for three or four caps, and manually plug in assorted film caps, like those nice WIMA parts.

--
John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    
 Click to see the full signature
Reply to
John Larkin
Loading thread data ...

Is 107kHz sacred? That is, can you shift the frequency to make the resonance rather than the other way around?

Is there any value is using a thermometer code rather than binary weighting? That is, will economies of scale allow for 16 caps of identical value C rather than binary weighting of 1 2 4 8. Thermometer codes are inherently (and obviously) monotonic. You could solder all the caps in at the board stuff then snip them out as needed.

Reply to
miso

Can you use an air-core coil, and mount it so that a brass bolt can be run into and out of it, to create a simple permeability-tuned oscillator? Or would that not meet the needs of shielding?

Two coils, end-to-end, with variable/adjustable spacing?

Reply to
David Platt

inductors are mostly +-20%. So, how to tune it?

If you measure the resonant frequency before you plug in the first cap, and keep on plugging in caps that are guaranteed to be just too small, and measuring again, you should get very close to 107kHz rather quickly.

Caps aren't easily available on finer grids than E6 (100, 150, 220, 330, 470, 680) but that should be good enough to let you get close to 107kHz fairly quickly.

It might be worth measuring the inductor's inductance fairly accurately before starting the process. IIRR Wima parts are available at +/-5%, and Farnell stocks some +/-1% (65) and +/-2% (185) film parts.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

Maybe use one of those rotary binary switches, like on an old external SCSI drive? It's easier for a production person to twist a knob (or click up/down buttons) than flick DIP switches back and forth. If the switch can't handle the juice itself, use it to drive the relay coils.

How about this: put two sockets (four pins) on the board, and make two DACs part of the test fixture, terminated with compatible pins. The "coarse" DAC has (say) 0.002, 0.004, 0.008, and 0.016 uF caps, and the "fine" DAC has (say) 0.00013, 0.00025, 0.0005, and 0.001 uF caps. Plug in the "coarse" DAC, twirl knob to find best value, replace "coarse" DAC with fixed capacitor of that value; then repeat with the "fine" DAC if you need to dial it in closer.

Yeah, the switch and test fixture will add a little capacitance that the fixed capacitor won't have, but I *think* you need something around

0.03 uF, and I *think* the fixture will only add low tens of pF, so probably not a super big deal.

There is probably some worst-case minimum capacitance that you could install as standard on the board, which would let you use finer tuning steps for the two "production select" caps.

Matt Roberds

Reply to
mroberds

--
How close?
Reply to
John Fields

On a sunny day (Tue, 12 Nov 2013 19:55:13 -0800) it happened John Larkin wrote in :

I have done it for high current high voltage by switching inductors of value 1,2,4,8 with 4 relais to tune to different frequencies in a VLF transmitter.

240 watt..
Reply to
Jan Panteltje

You need about 30nF +/-6nF. How about 22nF with a parallel 12nF PWMed at a much higher frequency, preferably phase locked to your 107kHz?

Cheers

--
Syd
Reply to
Syd Rumpo

John Larkin schrieb:

Hello,

long time ago tuneable inductors were used. But these were not small and shielded. How many windings on the used inductors, what about severals taps?

Bye

Reply to
Uwe Hercksen

Fun, Random thoughts... I guess I like the switched C idea (maybe miso's single value caps.) What about tweaking the inductance? Over wound coil and take off turns till it's close. Under wound and then measure and add some other L in series. What's the Q? You get a little freq shift as you load it. Then I started thinking about coupled oscillators...one pulling the other... but that seems much to exotic.

George H.

Reply to
George Herold

We used a 'capacitive DAC'

Have you looked at saturable core reactors, even 'home made' versions might work for you.

use phase shift to tune and sit exactly on the resonance.

Reply to
RobertMacy

PWMing a cap makes a resistor.

Cheers

Phil Hobbs

Reply to
pcdhSpamMeSenseless

How about splitting your inductor into two 39uH and rebinning the component so you can combine one higher and one lower than the nominal value to get a fairly accurate composite 78 uH using two components.

Any reason why it has to be 107kHz? How accurate does it have to be?

A tunable core inductor of about 1mH in parallel should give you some tuning leeway and continuous adjustment. It must still be good for the full voltage but only has to carry about 10% of the current.

--
Regards, 
Martin Brown
Reply to
Martin Brown

I used to place inductors of smaller and smaller values in series with shorting PCB traces around each, then cut a trace(s) during manufacturing to roughly tune stuff. Actually pretty cheap, even throwing away all those unused inductors. Service was a bitch, though.

The key here is how to maintain high Q at 107kHz. One Amp is not negligible.

Yeah, far easier to use the 'capacitor DAC' approach and stick with high Q audio caps.

Reply to
RobertMacy
2C-C DAC (you know, R-2R ladder, but flipped to give the same response in capacitance).

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
 Click to see the full signature
Reply to
Tim Williams

Scratch, scratch.. Phil you are both too smart and to cryptic for me to understand what that means.

George H.

Reply to
George Herold

If you PWM a cap at a frequency f that is well above the signal frequency, it dispenses charge at the rate of I = V*f*C. It's in phase with the voltage, so it looks like a resistance of 1/(fC).

Of course that's if you discharge it on the other half cycle. If you don't, it just looks like a non-PWMed cap. Either way, you don't get variable capacitance by PWMing a cap.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

How about one of Joerg's famous Y5V varactors? You're way up in the nanofarads, where you can still get Y5V, so you could put a 100-nF 50V cap in series with the main tank cap, with some highish DC voltage applied to the midpoint.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

Good idea! I've made on-chip thermometer-code DAC's... trivial to get many bits with no trimming. I even designed one once that was logarithmic output to control the bias/power-output of a WiFi repeater :-) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

Note that the resistance is due to switching losses. Charging the cap through the (resistive) switch, discharging it inbetween (or whatever's going on).

Presumably, one could pump the charge into a reactance and dissipate an arbitrarily low energy (as Larkin is so proud of claiming when "shorting" two capacitors together), but coming up with that... shall be left as an exercise for the reader.

No one mentioned if the PWM is in phase or not. If you switch it, cycle by cycle, synchronous with zero crossing, you can stretch a few cycles, then run a few without, and so on. Using a too-large capacitor, one could sigma-delta to an average frequency. Maintaining a phase lock within one cycle would probably be difficult with such tolerances, but can be reduced arbitrarily by adding more levels to the DAC. That might be especially handy if it's being used as an oscillator or reference or something.

Amplitude variation, particularly if the S-D produces a constant state for a bunch of cycles, might not be ideal.

If it's just being used as a resonator, the filter spectrum ideally is a single peak at the center frequency, but the S-D will split that into two peaks adjacent to the center, being an FSK spectrum. The actual response will be along the valley between the two, so it still won't be at resonance, as such, though it will average out to it. It would be interesting to play with a tracking resonator: it's always at resonance, until you move beyond the passband. Counting the S-D codes should provide low bandwidth frequency-ADC action, since it's ultimately just a PLL. Locking is fast (indeed, immediate), because it's a phase lock, not a frequency lock (the resonator is always excited by the input frequency).

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
 Click to see the full signature
Reply to
Tim Williams

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.