# Revisit amps input impedance

• posted

I got this high input impedance amp out again and redid my input capacitor. I have it at about 0.3pf.

The writer says the input impedance is near 5 gigaohms but, I put in a 1 Mhz 4Vpp signal and got out a 4Vpp signal. Then I put a ten Meg in series with the input and my output dropped to about 0.5Vpp. That certainly doesn't seem like it is anywhere close to 5 Gigaohm. So can someone calculate the input impedance, I want to figure out if I have something wrong or if the writer got it wrong.

Thanks, Mikek

• posted

Are you serious ?:-) ...Jim Thompson

```--
| James E.Thompson, CTO                            |    mens     |
• posted

"amdx"

** He does not say that at all.
** The writes says there is 1.4pF of capacitance at the input.

THAT alone has an impedance of just over 100 kohms at 1MHz.

Radio frequencies and impedances in the megohms do not go together.

FFS learn some math.

... Phil

• posted

Ok, I missed that, he's figuring additional capacitance from his input wires and input connection. Just the 1.4pf has an impedance of about

115K ohms. Then the 0.3pf in series with 20 Megaohm has an impedance not much more than 20 Megaohms, so we have 115k parallel 20 meg, so we are still about 115K. Then throw in the gate impedance and we are slightly less than 115k. So where did the writer come up with:

"But the input resistance of the amplifier is much higher then 20 M.Ohm, in theorie even 17² times higher (so, 5780 M.Ohm), this is because over the 20 M.Ohm resistor is only 1/17th part of the input voltage. In practice the input resistance will be lower then 5780 M.Ohm because of dielectric losses e.g. in the gate of the FET."

145k is certainly lower than 5780 Megaohms, but I would not call it high impedance.

???

I'm not sure how the 1.4 pf should be treated, the circuit is used to measure Q of LC circuits, so the 1.4pf could just lower the resonant frequency slightly and the Q would be affected only by the quality of the 1.4pf of capacitance. So maybe this isn't as bad as the 115 kohm impedance suggests, but I'm still stuck with my 10 megaohm experiment that started my question.

Yes, that would be very helpful, maybe when I retire. If I was starting over I would have studied math and physics. Thanks, Mikek

• posted

It's a silly circuit. A bootstrapped jfet or phemt could have unity gain (no need for all that other junk) and a fraction of a pF of actual terminal capacitance; he has 1.4 pF, mostly parasitic.

John

• posted

The *magnitude* of the input impedance is 94.6k.

• posted

I wonder what he measured that with.

That's verging on electrometer-type input resistance.

```--
"For a successful technology, reality must take precedence
over public relations, for nature cannot be fooled."```
• posted

He didn't, that's a calculation. It's also been called goobledegook.

I'm thinking the capacitance just adds to the resonating capacitor of the LC being measured and should not be called part of the input impedance when measuring the Q of an LC circuit. The loss in the capacitor will show as a large parallel resistor.

Mikek

• posted

First, the author is claiming that the circuit looks like 0.3pF in parallel with 5Gohm, not 5Gohm bare. That 0.3pF has a reactance of around 500kOhm, which is almost sorta-kinda consistent with what you're seeing.

Second, the author is using theory that doesn't match the part he's using. He needs to sit down, make his little FET all nice and comfy, and explain to the little critter how it is supposed to work and hope that it changes its ways. Or, he needs to understand what the hell he's talking about.

I'm not inclined to go through the math right now, but the effective input impedance at the gate of the FET is going to look like the FET gate capacitance in series with some resistance that's going to be in the kilo- ohms. I'm not sure _how many_ kilo-ohms, but it's going to be 100k or less.

Keep in mind, though, that the circuit ought to be doing pretty much what the author claims for it: making measurements on LC circuits without significantly affecting their Q. I do _not_ see him claiming that the resonant frequency won't be affected. So when you clip that thing onto a tank circuit you should expect that the resonant frequency _will_ change (probably more from the parasitic capacitance of the leads than the 0.3pF cap), but it shouldn't get loaded down to any significant degree.

Whether the rest of the amplifier maintains calibration close enough for even remotely useful measurements is left as an exercise for the reader

-- if you're using it for narrowband ratiometric stuff, like finding the bandwidth of the tank, then it'll probably work fine.

```--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.```
• posted

hmm... His writing is not real clear on that, (as I read it).

I have modified John S's LTspice file to include an LC driven through a

100Meg ohm. I then connect the (input circuit) .3pf in series with (paralleled) 20meg and 4.8pf to ground. This looks like an input impedance of 950K according to LTspice. If I add 0.7pf as my input stray capacitance the input impedance drops to 275k.

Here's where I want to get some understanding.

I think the .7pf becomes part of the resonating LC, so it should not be considered as part of the load on the LC. It is just additional capacitance that will lower the resonant frequency. Yes there will be losses in the stray capacitance.

That's what he designed it for.

I cannot find info on how to generate an LTspice netlist that I can post. Will someone walk me through it? Thanks, Mikek

• posted

Something tells me he has much more parallel capacitance at the input than he thinks.

Jamie

• posted

It could be, here's what I did. I measured the different in the output voltage of my fet with and without a 10meg resistor in series with the input. I plugged the circuit into LTspice and juggled the stray capacitance number until it matched my measurements. .7pf is the number that matched. I will post the netlist when I find out how to. Thanks, Mikek

• posted

That sounds about what a connector may generate at the input.

You also must remember you could be getting some SWR inside the unit and not enough drive over come that.

Jamie

• posted

True. That wouldn't keep the circuit from being useful for measuring the effective Q of a tank circuit -- it just wouldn't be good for much else.

```--
Tim Wescott
Control system and signal processing consulting```
• posted

LTSpice files are all text. Just open your asc file in a text editor, select all, then cut & paste into your posting.

```--
Tim Wescott
Control system and signal processing consulting```
• posted

That's not a netlist, .asc files are the schematic itself. You can't import that into another simulator. In the LTspice _View menu, select Netlist, the ctrl-A (copy all), then open your usual text editor and ctrl-V to paste the netlist into a file. ...Jim Thompson

```--
| James E.Thompson, CTO                            |    mens     |
• posted

Are you using the BFR92A he shows?

• posted

He didn't say it was a netlist. It is actually everything it takes to draw the schematic and define the components.

• posted

You may not be able to import it, but you can sure do it the way I said

-- that's what gets handed around here as "LTSpice" files.

Try pasting the following into a text file with the "asc" extension, and see if it doesn't work:

Version 4 SHEET 1 880 680 WIRE -96 -32 -144 -32 WIRE 16 -32 -96 -32 WIRE 96 -32 16 -32 WIRE 320 -32 192 -32 WIRE 432 -32 320 -32 WIRE 544 -32 432 -32 WIRE 704 -32 544 -32 WIRE 768 -32 704 -32 WIRE 544 0 544 -32 WIRE 16 32 16 -32 WIRE 432 48 432 -32 WIRE 544 96 544 64 WIRE 144 144 144 32 WIRE 320 144 320 -32 WIRE 704 144 704 -32 WIRE -144 176 -144 -32 WIRE 16 192 16 112 WIRE 80 192 16 192 WIRE 432 192 432 128 WIRE 432 192 384 192 WIRE 144 256 144 240 WIRE 240 256 144 256 WIRE 320 256 320 240 WIRE 320 256 240 256 WIRE 16 272 16 192 WIRE 240 272 240 256 WIRE 432 272 432 192 WIRE -144 384 -144 256 WIRE 16 384 16 336 WIRE 240 384 240 352 WIRE 432 384 432 352 WIRE 704 384 704 224 FLAG 16 384 0 FLAG 240 384 0 FLAG 432 384 0 FLAG -144 384 0 FLAG 544 96 0 FLAG 768 -32 Vout FLAG -96 -32 Vin FLAG 704 384 0 SYMBOL npn 80 144 R0 SYMATTR InstName Q1 SYMATTR Value 2N5550 SYMBOL npn 384 144 M0 SYMATTR InstName Q2 SYMATTR Value 2N5550 SYMBOL res 224 256 R0 SYMATTR InstName R1 SYMATTR Value 10k SYMBOL zener 32 336 R180 WINDOW 0 24 72 Left 0 WINDOW 3 24 0 Left 0 SYMATTR InstName D1 SYMATTR Value BZX84C6V2L SYMATTR Description Diode SYMATTR Type diode SYMBOL res 416 256 R0 SYMATTR InstName R2 SYMATTR Value 10k SYMBOL res 416 32 R0 SYMATTR InstName R3 SYMATTR Value 100k SYMBOL pnp 192 32 M270 SYMATTR InstName Q3 SYMATTR Value 2N5401 SYMBOL voltage -144 160 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value SINE(90 20 120) SYMBOL res 0 16 R0 SYMATTR InstName R4 SYMATTR Value 100k SYMBOL cap 528 0 R0 SYMATTR InstName C1 SYMATTR Value 1µ SYMBOL current 704 144 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName I1 SYMATTR Value 10m TEXT -178 408 Left 0 !.tran 10m TEXT -184 440 Left 0 !.ic V(Vout)=0 TEXT -184 472 Left 0 !;dc I1 0 100m

```--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.```
• posted

Post a netlist and I can run on PSpice.

Post a .asc and anyone can run on LTspice.

Different animals, different uses. ...Jim Thompson

```--
| James E.Thompson, CTO                            |    mens     |