Active rectifier/peak detector @ up to 150kHz...

Hello Nicko,

I looked into the datasheet and it is a real current transformer (that has many secondary turns and few primary). The open load voltage will be factors above the loaded voltage. This is a property of current transformers. The 100mV/A is valid with 50 Ohms load.

This high open load voltage is sufficient to overcome the diode voltage drop of a bridge rectifier. It increases the core flux somewhat as the transformer has to supply 2*(diode drop) on top of the rectified output voltage.

Connect (in at test setup) a 4-diode bridge (low voltage small signal schottky diodes, not 1N400X series) directly at the output of the current transformer and then load the rectifier with 50 Ohms. This will probably give you useful results.

Of course, do not operate the transformer (with rectifier) without 50 Ohms load.

Best regards,

Wim PA3DJS

Inductors don't resonate. What are you really doing? Sounds similar to my induction heater.

Tim

-- Deep Friar: a very philosophical monk. Website:

Why not use two comparators, one for positive swing and one for negative?

John

to my

Its a resonant RC tank circuit that can be detuned by external influences, so I track it using a uP - similar to an induction heater where you want the circuit to resonate but where you might change your work coils and thus move the resonant frequency - either way you don't want your bridge FETs/IGBTs to over cook...

We make power amps that have a uP constantly (2000 times a second) digitizing everything of interest (voltages, currents, temperatures) and protecting our fets. We run a thermal simulation of junction temperatures and shut down at what we compute to be 140C.

Nowadays a $4 CPU could run the same model at 100K per second, if that helped any. A $4 CPU is worth it to get the maximum safe performance out of $300 worth of fets.

You might random-sample-digitize the current waveform and do math on that. Things like abs val, square, lowpass filter, whatever. That's much more versatile than things like hardware current limits.

A cheap uP with a built-in 10 or 12-bit ADC actually makes a pretty good true-RMS voltmeter.

The low-end ARMs are now going for under $1.

John

On a sunny day (Mon, 5 Apr 2010 09:34:54 -0700 (PDT)) it happened Nicko wrote in :

Not likely :-)

Really? Then why worry about SRF?

Care to qualify that statement?

-- "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." (Richard Feynman)

Pure inductances are inductive from DC to light.

Who knows what the OP actually has.

Tim

The output of the transformer can go directly to a couple of comparators. You can shut down if either one says there is too much current. This will be fewer parts than making the rectifier because the comparators come two to a package and a "wire-or" does the OR function.

There are no pure inductances. There is always stray capacitance.

If it has to be super small there's even dual comparators with built-in reference in super-tiny SOT or DFN packages. For example the LT6700 series which happens to also come in a +/- configuration for window comparators (which is what Nicko needs). Such convenience has its price, of course.

When using this one watch out during power-up, the ref is out of whack for about the first 500usec.