tuning an LC

"John Larkin" wrote in message news: snipped-for-privacy@4ax.com...

That's more info, but didn't answer the question... what happens when you anneal them?

Tim

JUST like that. The government can't do anything well. Some things it must do and it doesn't matter if there's waste. It's gotta be done. Those are damned few, though.

You're right. AAMOF, Murata doesn't even show the C/V curves for C0Gs. Too bad they're such small capacitances (at such high prices).

It's hard to know what the real object here is... being a Larkineque bloviation with no exact specification/definition... only the usual vagaries.

However, if the object is to get an accurate 60V P-P sine wave at

Any kiddy just out of school knows how >:-}

I'll post the trivial solution when the bloviating asshole stops spouting.

(Simulation results posted to ABSE)

("Asshole" IS a transitive verb... asserted for those of you considering criticizing my use of "pluperfect" :-) ...Jim Thompson

High Q does reduce the power required, though--power dissipation per radian of phase is E0/Q, where E0 is the instantaneous energy stored in the tank. Any of us can hook up some big ugly power op amp, but John's tank has a stored energy E0 = 1/2*75uH * 0.5A**2 = 9 uJ, so the drive power is

P_drive = (2 pi f/Q) E0 = 6.3W / Q.

Q=50 is dramatically easier to drive than Q=2, no?

And high Q also improves the THD, unless (as in the Y5V varactor proposal) the distortion comes mostly from the reactive components.

Cheers

Phil Hobbs

How do your comments add to the information content? Sounds like mostly sickophant support.

I said high Q was not required. The exhibition was Q~1. It's be nice if we could return to real engineering on this group instead a BS after BS.

As long as Larkin wants to name-call I will simply bring him down technologically. He really is quite incompetent. ...Jim Thompson

My LC simulates a transducer of some sort, possibly an LVDT sort of thing; I'm not allowed to know, exactly. The point is that I do need to simulate the transducer impedance and its waveform, distortion and all, when driven by the given circuit. My Q will be close to the actual device's Q. And I'm required to hit the drive resonant frequency pretty close.

Bloviate! Bloviate! What a dork ;-) ...Jim Thompson

Didn't try that. But we found the problem on a production board that had been reflow soldered. The caps on the reel, in stock, were similarly low in value.

Dip switches are non-volatile!

I have measured many C0G (aka NP0) chip capacitors that lost 40 % C or more at rated voltage, but still have very low tempco. That includes several batches each from multiple manufacturers. Alright that was 30 years ago, but still. Not better than films for bias voltage, only for temperature.

But it's 60

floating.

too

processing will be

with the

be a bunch

Maybe only John Larkin really feels the box, it seems pretty insubstantial to me compared to other methods offered.

A non-volitale 2C-1C ladder divider using magnetic latching relays for trim also seems reasonable to me.

?-)

On a sunny day (Mon, 18 Nov 2013 20:36:14 -0500) it happened Phil Hobbs wrote in :

Well, it is clear now, after the flux-capacitor we now have the flux-inductor. No wonder he keeps it a secret ;-)

y a mosfet and have about 60 volts p-p across it. L is about 75 uH, so Xl i s around 50 ohms. Peak inductor current is half an amp or so.

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

d P26x16 gapped pot cores with a central hole.

e fine.

number is 2098922 - you can also get it out of Newark if nobody else in the US stocks it.

bably more than enough to take the +/-3% tolerance on the heavily gapped co re pairs.

ou may want to find a coil winding shop to make your coils. Hand coil windi ng machines are pretty cheap, and if you only need 20 parts it wouldn't be too much of a chore to do it in house. Most places where I've worked have h ad a coil winding machine somewhere around the place, but the Nijmegen Univ ersity workshop threw their's out a few years ago because none of the new h ires could imagine winding a transformer.

all that easy to get hold of, but I managed it - you should find it easier.

re.

ded slug).

oviation with no exact specification/definition... only the usual vagaries.

z with good side-bands, the solution is trivial and _does-not_require_ prec ision tuning, nor high Q.

In fact John Larkin has spelled out that it is to substitute for LVDT coil (or something like it) in a test bed for one of his customers.

an of phase is E0/Q, where E0 is the instantaneous energy stored in the tan k. Any of us can hook up some big ugly power op amp, but John's tank has a stored energy E0 = 1/2*75uH * 0.5A**2 = 9 uJ, so the drive power is

l) the distortion comes mostly from the reactive components.

He seems to be planning on driving it in class-C, so the energy going in is almost a Dirac spike, with equal levels of every harmonic up to the width of the spike.

Obviously the spike gets smaller if the Q of the tank is higher. With Litz wire he might get a Q of about 150.

Baxandall's class-D driver feeds in what is nominally a square wave, where the harmonics decrease in proportion to the harmonic number, but the feed i nductor inject a load-independent second harmonic component, which means t hat there's a constant harmonic content for sufficiently low loads.

My current mirror variation of that can push the harmonic content down by 3

I've modelled that in LTSpice using the John Chan model - we got a third ha rmonic content about 80dB below the funadamental with a tolerably heavily g apped core, and confirmed it on a test winding (hand wound and shipped off to London for measurement on a friend's rig). The design then moved to the core with the biggest gap that I could buy, where the harmonic content simu lated at 104dB down.

That needed a better drive, which I've designed, bought the parts for, and should eventually build.

We had a ceramic 100uF/10V TAIYO YUDEN LMK325BJ107MM-T, farnell: 1838744.

I tested it both with LCR meters and the old fashion way with a scope and function generator:

(bias versus capacitance)

Cheers

Klaus

C(f,v), or is that C(t,v)?

Now you have me curious as to EXACTLY what the function is?

You've just lost your temper, is all. You claimed that Q didn't matter, and I showed that it reduced the drive power by a lot, from watts to tens of milliwatts. You like math, remember?

Of course high Q isn't required, if you're driving it with a Kepco BOP, but (at Q=1) six watts of gate drive is a bit ridiculous. You could roast hot dogs on the driver stage.

Today's feature is like Godzilla vs Rodan. Both quite capable of ravaging Tokyo as long as the other one isn't around. ;)

Pass the popcorn, Mabel.

Cheers

Phil Hobbs

On a sunny day (Tue, 19 Nov 2013 10:11:32 -0500) it happened Phil Hobbs wrote in :

I already guessed 'transducer', and the capacitance value points to piezzo. So what does 'frequency' tell us?

He wants to weld plastic.

I sort of doubt that. John's customers are usually a bit higher tech than that. It could be an inductive transducer that has some built-in capacitance, e.g. a resonant magnetostrictive drive or something driving air or a fluid. Matching acoustic impedances is a lot easier with a resonant transducer.

Cheers

Phil Hobbs