Please help -- SPICE simulation and imolementation mismatch

Could some electronics guru please provide some hints about what might be yhe problem in the following setup ?

First I SPICE simulated a simple sertes circuit consisting of a signal source (referenced to ground) connected to a capacitor, follwed by a series resistor and finally a capacitor to ground. Both capacitors are non-ideal(includes ESL, ESR). This is a very simple model of capacitive coupling. With a signal of 6V

60Hz sine wave, transient analysis provides very good results. The capacitance values are approximately 10.0pF and the resistor is 5K.

Now, as an experiment, I hand constructed the two capacitors with single layer copper clad board, separated by a phenolic sheet

3,0 mm thick. The dimensions of the copper area of Each of the single layer copper clad pieces is 2 cm x 2 cm and four 2 mm diameter stainless steel screws hold the three pieces(2 capaciror plates and insulator) firmly together. Using a relative permittivity of 8 for phenolic, the capacitance value is 9.44pF for each capacitor.

I connected a signal generator set at 100 Hz,

6V sine wave to the circuit described for the SPICE simulation, with my homebuilt capacitors and a 5K resistor, and read 0.0 Volt AC at the resistor terminal -- the SPICE simulation showed 5.094 Volts. Given that a capacitor is a short to AC, what could be the problem ? I understand that polysttrene capacitors are ideal for AC operations, but this setup ought to show some results. I repeated the experiment with a 6-0-6 500 mA output transformer, but there was no chnage.

Any hints/suggestions as to what the problem might be would be greatly appreciated. Thanks in advance for your help.

Reply to
dakupoto
Loading thread data ...

and a 5K resistor, and read 0.0 Volt AC at the resistor terminal -- the SPICE simulation showed 5.094 Volts. Given that a capacitor is a short to AC, what could be the problem ? I understand that polysttrene capacitors are ideal for AC operations, but this setup ought to show some results. I repeated the experiment with a 6-0-6 500 mA output transformer, but there was no chnage.

This is really a question for the basics newsgroup, but you are making an assumption that a tiny capacitor is a short to a low frequency AC signal, when it actually presents an impedance (or reactance) of X(c) =

1/(2*pi*f*C). You must be doing something wrong in LTSpice. You could copy your ASC text file for LTSpice if you want someone to try it.

Paul

Reply to
P E Schoen

When you typoed "imolementatian" I thought maybe you'd burnt something up (immolation, y'know). Boringly, that appears to not be the case.

a 10pF capacitor has a capacitive impedance of around 160 megohm -- that's 160 * 10^6. When we liken a capacitor to "a short circuit at AC" we mean AC of infinite frequency. At any real frequency the capacitive impedance is non-zero, and it is inversely proportional to frequency.

Either you did not use 10pF caps in your SPICE model, or you used a much higher frequency than 100Hz.

The rule of thumb for using coupling caps (which, since it's a rule of thumb, is only usually good) is to choose a capacitive impedance which, at the lowest frequency of interest, is 10 times the impedance it's "looking into". You've got a series pair of 10pF caps looking into a 5k resistor, so that means you need to do your test with a 25MHz source.

--
www.wescottdesign.com
Reply to
Tim Wescott

For all the critiques you're getting -- it's GOOD to experiment. Your experiment has just presented you with wildly unexpected results, which you're going to remember _forever_. Even though an experienced engineer may say "of course that'll fail", a lot of the good ones (myself included) have just that sort of "failed" experiment in their past.

--
www.wescottdesign.com
Reply to
Tim Wescott

It'll be even more memorable and useful if the OP calculates it for himself. Experoments are good for developing intuition fast, and algebra is even faster.

Cheers

Phil Hobbs

Reply to
Phil Hobbs

If your voltmeter had infinite impedance (like the spice probe tool) then you would see something but the 1Meg or 10Meg input resistance of your real-world voltmeter is simply too big a load for such a tiny current. Use spice to measure the current flow in your series 5K resistor and tell us what the current is.

piglet

Reply to
piglet

*That* is your problem. A capacitor is *not* a short for AC. It's an impedance Z=1/(j*2*pi*f*C). At 100 Hz, a 10pF capacitor would be j160 Mohm!

Another thing you may be unaware of is that Spice uses base units throughout. Capacitor values are thus given in Farads! Fortunately, it understands (most) multiplier prefixes, so a

10pF capacitor value can be written as 10p.

Finally, LTspice .asc files are simple text files. Next time you want to discuss simulation results, append the .asc file to your message, so that other people can see what you did.

Jeroen Belleman

Reply to
jeroen Belleman

I do not think he is doing anything wrong in spice. He made a (albeit high impedance) voltage divider which in Spice works as such. As soon however as you construct the circuit and connect a real voltage meter to the mid-point, the internal impedance of the capacitive voltage divider 1/(2*2*3.14*60*5*10^-12) = 1/(0.36*10^4*10^-12) = 300 MOhm (capacitive) is so high that the voltmeter acts almost like a short circuit, hence the very tiny, if not 0,measured voltage.

Maybe a capacitor of 1000uF would be a better choice, but is not so easy made at home.

joe

Reply to
joe hey

Apparently he still has to learn that, and this was a good lesson.

joe

Reply to
joe hey

On 19 Sep 2015 12:36:24 GMT, joe hey Gave us:

Exper-o-mentally speaking, of course. :-) Hhehehehe

Reply to
DecadentLinuxUserNumeroUno

Phone keyboard and fat fingers.

Cheers

Phil Hobbs

Reply to
Phil Hobbs

Thanks to each of you for pointing out the capacitor reactance, that I overlooked, and has caused much misery. The Ngspice file is listed below:

.PARAMS PERMFS=8.854187817p RELPERM=8 .PARAMS LEN=0.02 THK=0.005 ; METERs

.SUBCKT NICAP 1 2

  • 1 IN
  • 2 OUT C0 1 3 '(PERMFS*RELPERM*LEN*LEN)/THK' L_ESL 3 4 4.5nH R_ESR 4 2 0.5 .ENDS

.SUBCKT CC 1 2

  • 1 IN
  • 2 OUT XC0 1 3 NICAP XC1 4 2 NICAP R0 3 4 5K

R0 1 2 1.0 VSIG 1 0 DC 0.005 SIN(0, 6.0, 60.0, 0, 0, 0) VTEST 3 0 DC 0.0 AC 0.0 XJ0 2 3 CC

.OPTIONS NOPAGE METHOD=GEAR .IC V(2)=0.0 .PROBE .TRAN 50ms 500ms 10ms UIC .PRINT TRAN V(2) I(VTEST) .END

Reply to
dakupoto

On Sat, 19 Sep 2015 12:06:25 -0400, Phil Hobbs Gave us:

Expeditiously speaking, of course. ;-)

Reply to
DecadentLinuxUserNumeroUno

There's an .ENDS missing in there.

Didn't you mention that you were doing this with LTspice?

Jeroen Belleman

Reply to
jeroen Belleman

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.