Back in ancient times, before active filters got popular, we used to make oscillators and bandpass filters using slug-tuned pot cores. They were very stable, but big and expensive and were a PITA to document, wind, assemble, and mount.
--
John Larkin Highland Technology, Inc
jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro acquisition and simulation
Where is the capacitor, and how does the return current get to the source of the MOSFET?
If there is no other resistive load, the voltage does not stop there, as long as your FET does not do anything unexpected. What you have now is a Class C amplifier without load.
I have worked with similar resonance circuits for wireless power transfer u sing 5% flat wound ferrite cores and 1% Polyester film caps ( binned) and a lso with Lexmark Laser DC-DC PWM boost regulators to >5kV which also used r esonant sine and boost XFMR.
I understand there is no resistive load, yet frequency is (cast in concret e) in spec. and coil impedance already chosen @ 50. I also understand Fet d river is 100 ohm 25% d.f. to 30V This means effective source impedance is
100R/d.f.= 400R which means a Q = 400/50 = 8 so why only 60V??
I would guess there is a magnetic load involved since the effective Q is on ly=2
If you did have a Q of 8 the phase shift would be rapid around resonance. But since it is only a voltage doubler (Q=2) the phase will be more stabl e.
Vpp=60 and if Zl= @50 Ohm then I(LC)=1.2A pp
This means your RIPPLE CURRENT on the capacitor must handle around 0.5A rms , which would obviously require Film Caps only, which means these 10% CAPS must also be binned.
This requires some calculations, but I see the necessity to match binned L' s with binned C's here, which is a bit expensive on scrap.
I have seen this done to tune an antenna matching network for 200 to 400 kHz vlf transmitter running at around 100 watts. The processor adjusted a DC current through the pot core to tune the network.
Well, depends on the material. For example the Ferroxcube 3C96 which starts with an initial permeability of 1800 at 0 degrees and has a local maxima of 4500 at 70 degrees. So your inductance will be all over the place just with the temperature drift.
uction is not usable, since the inductor will drift over time and temperatu re...
ke +100
Ferrite is also susceptible to pressure or impact. Hit a core lightly and t he permeability changes. Can be tricky if you pot a ferrite in a plastic ho using and the potting compound hardens and during the process puts pressure on the ferrite and the permeability changes in the process
oduction is not usable, since the inductor will drift over time and tempera ture...
like +100
the permeability changes. Can be tricky if you pot a ferrite in a plastic housing and the potting compound hardens and during the process puts pressu re on the ferrite and the permeability changes in the process
Also, beware of the remanence hystersis. If not properly demagnetised befor e your test, the inductance will be changed due to a different operating po int on the hysteresis curve
I got some cute little Coilcraft samples, 82 uH +-20%. They measure 6% high at room temp. TC is pretty high, about +0.1% per degree C on my AADE meter. Not great but tolerable for this app.
--
John Larkin Highland Technology, Inc
jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro acquisition and simulation
Hey, I had an idea. Is the tank the source of the oscillation, or is that from some other clock? Anyway, the idea is to tune it low, then put a magnet next to the inductor. Enough flux leaks in through the shielding to reduce the inductance. When you find the sweet spot, glue the magnet in place on the board. If it is free-running, you can tune with a frequency counter, if otherwise clocked then you'd us a voltmeter to find the resonance.
The negative swing of the low side of the tank is just about to ground, so the p-p swing is close to 2*Vcc. That's just what happens.
See Spice thing below. It's interesting. The sine actually swings below ground a bit, just enough to keep the average drain current (and Vcc supply current) low. And the timing between the gate and LO is cool.
It only pulls about 5 mA from the 30 volt supply.
I can buy 2% film caps from Mouser or Digikey.
I can use a 20% inductor and trim C. Looks like one plugin cap (5 mm radial, pcb pin sockets) and the 16-step hex switch c-dac. Something like that.
Version 4 SHEET 1 880 772 WIRE -80 32 -192 32 WIRE 128 32 0 32 WIRE 240 32 128 32 WIRE 320 32 240 32 WIRE 416 32 320 32 WIRE 496 32 416 32 WIRE 416 80 416 32 WIRE 128 112 128 32 WIRE -192 144 -192 32 WIRE 240 160 240 32 WIRE 416 176 416 144 WIRE 128 224 128 192 WIRE -192 272 -192 224 WIRE -16 352 -64 352 WIRE 128 352 128 304 WIRE 128 352 64 352 WIRE 240 352 240 224 WIRE 240 352 128 352 WIRE 320 352 240 352 WIRE 352 352 320 352 WIRE -64 384 -64 352 WIRE -112 464 -192 464 WIRE -192 512 -192 464 WIRE -64 528 -64 480 WIRE -192 640 -192 592 FLAG -192 272 0 FLAG -64 528 0 FLAG -192 640 0 FLAG 320 32 HI FLAG 320 352 LO FLAG 416 176 0 SYMBOL cap 224 160 R0 WINDOW 0 62 14 Left 2 WINDOW 3 50 56 Left 2 SYMATTR InstName C1 SYMATTR Value 27n SYMBOL ind 112 96 R0 WINDOW 0 -57 27 Left 2 WINDOW 3 -66 72 Left 2 SYMATTR InstName L1 SYMATTR Value 82µ SYMBOL res 112 208 R0 WINDOW 0 -63 35 Left 2 WINDOW 3 -65 74 Left 2 SYMATTR InstName R1 SYMATTR Value 0.5 SYMBOL res -32 368 R270 WINDOW 0 -36 55 VTop 2 WINDOW 3 -44 55 VBottom 2 SYMATTR InstName R2 SYMATTR Value 100 SYMBOL voltage -192 128 R0 WINDOW 0 60 42 Left 2 WINDOW 3 61 76 Left 2 SYMATTR InstName V1 SYMATTR Value 30 SYMBOL nmos -112 384 R0 WINDOW 0 -91 -11 Left 2 WINDOW 3 -114 26 Left 2 SYMATTR InstName M1 SYMATTR Value 2N7002 SYMBOL voltage -192 496 R0 WINDOW 0 54 72 Left 2 WINDOW 3 26 117 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value PULSE(0 10 0 100n 100n 2.333u 9.333u 1000) SYMBOL res 16 16 R90 WINDOW 0 69 61 VBottom 2 WINDOW 3 79 60 VTop 2 SYMATTR InstName R3 SYMATTR Value 100 SYMBOL cap 400 80 R0 WINDOW 0 56 18 Left 2 WINDOW 3 50 57 Left 2 SYMATTR InstName C2 SYMATTR Value 10µ TEXT 216 552 Left 2 !.tran 200u TEXT 160 416 Left 2 ;Resonant Network TEXT 144 496 Left 2 ;J Larkin Nov 14, 2103 TEXT 136 456 Left 2 ;Highland Technology Inc
--
John Larkin Highland Technology, Inc
jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro acquisition and simulation
** Potcores with ground air gaps are *designed* for making precision inductors - put a threaded slug through the gap and have an adjustable one.
I have a wide band AM tuner purchased in 1974. In the audio output is a tunable pot core wired to form a passive notch filter at 9kHz exactly - it was originally set to 10kHz until AM band was reorganised to 9kHz spacings. When that happened, I changed a cap and retuned the core.
The bandwidth of the notch is about 5Hz, depth 40dB.
This is a bit superficial. For the inductance and the application, John Larkin ought to be using a heavily gapped ferrite core, where you'd gap the permeability down from about 5uH (+/-25%) per turn squared to about 160nH (+/-3%).
There are a whole range of parts relatively easily available from broad-line distributors. What John Larkin actually needs is a gapped pot core or RM core with a central hole into which you can screw a ferrite adjuster.
Farnell sells two such adjusters for EPCOS P11x7 and P26x16 pot cores - strictly legacy parts, but John Larkin's problem is a legacy application.
For these sorts of heavily gapped core, the ferrite just locates the field lines, and the inductance is pretty much determined by the geometry and the dimensions, neither of which much vary much with temperature.
John Larkins thinking on small signal inductors doesn't go much beyond what he can buy - pre-wound - off the shelf, but I would have expected that this usegroup did include a few people who could wind their own.
Not really "all" over -- if it's gapped down to, say, an equivalent 30 permeability (pretty typical, but YMMV), the difference between 1800 and
4500 is actually 30.305 to 30 (assuming the gap remains constant and the speed of light doesn't change), a 1% change, or approximately 142 ppm/C average (between peak and valley, the range of change will probably peak more like 200 ppm/C).
There are better ferrites to choose, so this is a reasonable worst-case estimate.
In practice, the gap will expand roughly at the core's expansion rate, which is about 10 ppm/C. The core area gets larger by twice this (d/dx of x^2 is 2x), while the gap stretches by this amount. Overall, inductance due to mechanical properties should rise by 10ppm/C. YMMV with various materials nearby -- a strong varnish may force the gap to expand near the rate of the bobbin. But unless you're using a 'precision' ferrite with comparable tempco, this won't be important. It's noteworthy that the copper windings shift with temperature too, and potentially humidity over long time scales depending on materials used (i.e., avoid paper!).
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.
lded inductors are mostly +-20%. So, how to tune it?
P26x16 gapped pot cores with a central hole.
fine.
mber is 2098922 - you can also get it out of Newark if nobody else in the U S stocks it.
bly more than enough to take the +/-3% tolerance on the heavily gapped core pairs.
may want to find a coil winding shop to make your coils. Hand coil winding machines are pretty cheap, and if you only need 20 parts it wouldn't be to o much of a chore to do it in house. Most places where I've worked have had a coil winding machine somewhere around the place, but the Nijmegen Univer sity workshop threw their's out a few years ago because none of the new hir es could imagine winding a transformer.
l that easy to get hold of, but I managed it - you should find it easier.
Rotating it changes the inductance by changing the effective air gap.
(threaded slug).
And you can still do it. EPCOS lists the pot cores, and Farnell/Newark stoc k the adjusters - I gave the part numbers - as Jim might have noticed if he 'd read all the text he posted, before posting it.
r using 5% flat wound ferrite cores and 1% Polyester film caps ( binned) an d also with Lexmark Laser DC-DC PWM boost regulators to >5kV which also use d resonant sine and boost XFMR.
rete) in spec. and coil impedance already chosen @ 50. I also understand Fe t driver is 100 ohm 25% d.f. to 30V This means effective source impedance is 100R/d.f.= 400R which means a Q = 400/50 = 8 so why only 60V??
o the p-p swing is close to 2*Vcc. That's just what happens.
below ground a bit, just enough to keep the average drain current (and Vcc supply current) low. And the timing between the gate and LO is cool.
only=2
e. But since it is only a voltage doubler (Q=2) the phase will be more st able.
rms, which would obviously require Film Caps only, which means these 10% CA PS must also be binned.
L's with binned C's here, which is a bit expensive on scrap.
al, pcb pin sockets) and the 16-step hex switch c-dac. Something like that.
You could wind a +/-3% inductor if you bought the right - heavily gapped - ferrite core pairs and the matching coil former. They are available off the shelf.
With only a bit more effort, you could probably buy a gapped pot core with a central hole, into which you could screw a ferrite adjuster to bridge the gap in a smoothly controllable way, giving you 78uH inductor you could tri m to exactly the inductance you needed.
I've listed the part numbers elsewhere in this thread.
I still don't follow your argument. Leave the voltage swings out of it. You have a target capacitance of X. You want to trim it over the range Y. Seems to me you would start with X-(Y/2) then add up to Y capacitance in parallel to achieve trim.
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.