Choke input Filters

It does in that topology. A choke-input filter is usually driven from a full-wave rectifier. Half-wave, a high-value choke keeps the diode on essentially all the time, so the diode doesn't rectify.

You can fix that by adding a catch diode to ground.

But a choke-input filter is probably not a good RF detector.

Interesting, but a diode in series with an inductor is a red flag. A diode in series with a capacitor is a super-dooper red flag.

John

After the point in time where the diode would conduct if there were no choke, the choke limits the rate of current rise. Rectification still occurs at this point in time, but current is limited. Once V_load exceeds V_in, the choke shows its flywheel effect, and current flow continues past that point in time. Current flow stops before the time its ready to begin conducting again, and the cycle repeats. I'm struggling to think of a use for such a circuit though.

In the real world detectors do have a little inductance.

NT

John,

Thanks for your reply, greatly appreciated..

Well, some texts do show a choke input with a single diode, I think the idea being that the flywheel action of the diode keeps the diode conducting. But I do agree... I'll go experiment with catch diodes..

Yes, but then a cap in series with a diode was very common in early circuits. Again, I think it was the leakage current of the diode which allows it to work.

Thanks ....... Zim

Well,

I added a full-wave diode arrangement, but the circuit still refuses to conduct over a certain value choke.

Now I'm really confused...

........ Zim

in that circuit, your friend is correct.

Just think of the input of the half wave acting like AC and do the math against the load R you have on the output.

Btw, that type of circuit is normally used to help remove the higher frequencies in the signal, to a point where you can get a DC flat like given enough loading capacitance to maintain it between cycles.

But I am guessing here, that isn't what you're looking for?

Jamie

And it's not going to make any difference with a full wave because the time cycle is still the same.

Maybe you don't fully understand the need for the L in that circuit?

One of the problems with DIODE detectors is the band gap in the diode which gives you an abrupt on/off in the Vf (Voltage forward), and this can generate unwanted noise in your signal which generally is in the higher frequency spectrum. Using that L reduces this effect.

Also, because diodes are not linear, you can also get some over tones generated in the results. L1 reduces this, too.

Keep in mind that this type of detector isn't linear and may require you to use a n-log amp or some sort of compensation if that is an issue.

Basically, you calculate L for spectrum you're trying to remove, but not any lower. If you look up "Shot noise in DIODES" on the net, you'll find a lot of crap out there on the subject.

Jamie

Makes no sense. What's the waveform at the diode/inductor junction?

What kind of diodes?

John

Hi John,

I agree. Here it is again with input and output

The graphs show two values of inductor which bracket the critical value.

I think they are generic signal diodes (0.7V forward). But I will check.

Tomorrow I'm going to make a breadboard model and vary the input frequency either side of the critical values and see what happens.

later ........ Zim

Thanks Jamie, but no, the idea was to build a higher impedance detector, analogous to using a Choke Input filter in a Power Supply to give Average voltage out, rather than letting the Cap charge to Peak.

But the original idea is pretty much irrelevant. My interest now is to find out why largish values of choke seem to make a choke-input filter go wacky.

I could sort of see why it didn't like a single diode, and was sure it would work with a full-wave input, but it doesn't.

Thanks ....... Zim

I think that a small inductance, well into the discontinuous-conduction regime might reduce the video impedance of the detector by increasing the conduction angle.

Cheers

Phil Hobbs

Phil,

Yes, that was the original idea.

......... Zim

You need a _small_ inductor for that.

Cheers

Phil Hobbs

How closely are you looking for voltage at the resistor? You're going to consume most of your 1V in the diode drop.

What simulator?

With LTSpice I get signal at the resistor. Here's my LTSpice file.

Version 4 SHEET 1 880 680 WIRE -16 48 -96 48 WIRE 160 48 -16 48 WIRE 256 48 224 48 WIRE 304 48 256 48 WIRE 416 48 384 48 WIRE 416 64 416 48 WIRE -16 80 -16 48 WIRE 416 176 416 144 WIRE -16 192 -16 160 WIRE 80 192 -16 192 WIRE -16 208 -16 192 WIRE 80 208 80 192 WIRE -96 224 -96 48 WIRE -64 224 -96 224 WIRE -64 272 -96 272 WIRE -96 304 -96 272 WIRE -16 304 -16 288 WIRE 144 304 -16 304 WIRE 256 304 256 48 WIRE 256 304 208 304 FLAG 80 208 0 FLAG -96 304 0 FLAG 416 176 0 SYMBOL voltage -16 64 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value SINE(0 1 1meg) SYMBOL e -16 192 R0 SYMATTR InstName E1 SYMATTR Value 1 SYMBOL diode 224 32 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName D1 SYMATTR Value 1N914 SYMBOL diode 208 288 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName D2 SYMATTR Value 1N914 SYMBOL ind 288 64 R270 WINDOW 0 32 56 VTop 0 WINDOW 3 5 56 VBottom 0 SYMATTR InstName L1 SYMATTR Value 1m SYMBOL res 400 48 R0 SYMATTR InstName R1 SYMATTR Value 1k TEXT -116 354 Left 0 !.tran 1m

Hi Tim,

Well there is 0.7V lost across the diode and 0.2V at the resistor. But yes, it's good point. I'll crack up the generator a bit.

I'm using QUCs at present.

I must admit I'm not a Simulator person. I'm much happier with a Soldering Iron and a CRO.

I don't know if I can use your SPICE script with QUCs I'll look into it.

Thanks for your suggestions.......... Zim

Your results seem to agree; about 120 millivolts detected. Seems reasonable. The choke makes the diode node swing negative, which lowers the average. The L isn't a good idea here.

John

John,

Yes, I quite agree. But I'm trying to understand just why it dies above a certain value of Inductance.

I would have thought that with a full wave input, there would be a net DC to cause current to flow through the choke.

And for that mater, the same with the single diode version.

Obviously I'm missing something, or doing something wrong.

.......... Zim

The inductor tends to keep current flowing, which keeps diodes on longer, even after the input voltage has swung through zero. So the diodes stop acting like diodes.

The single diode case is especially interesting. With a really big L, the diode almost never stops conducting. So it doesn't work like a diode should. The intermediate-L value cases are interesting; simulate them to see what's going on, observing all the nodes.

John

After looking at that image, I would say you should start using a different simulator.

1. The diodes need to have some sort of basic specs like a 1N914. Otherwise, a basic diode will not have the switching speed and low capacitance needed for 1MHZ. This will result in not seeing the small (+)voltage swing that should be there at the input of L. How ever, the flat bottom at 0 volts indicates that it knows it should be there? strange behavior of the sim.
2. Does your sim allow you to set the R value of L? and if so does it happen to be very high ?

A lot looks wrong there.

Jamie

Try LTspice, That simulator isn't doing you any favors.

It's making errors, especially on the output, somehow?

Jamie