Certainly. But does temperature change to a dielectric behave the same way as applying mechanical force to a vacuum cap?
-- John Larkin Highland Technology, Inc trk The cork popped merrily, and Lord Peter rose to his feet. "Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"
I was wondering if changing the temperature (hence dimensions and dielectric constant) of a polar dielectric has any special electrical effects. It does take work to change the temperature of a mass.
-- John Larkin Highland Technology, Inc trk The cork popped merrily, and Lord Peter rose to his feet. "Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"
It gradually degrades the noise floor, but it takes more than two decades for that 1/f contribution to get up to the flatband of Phil A's noiseless TL084.
3.5 Hz / (16 nv /1 nV)**2 = 0.014 Hz.If you have a squint at the battery paper I linked to, you'll find a discussion of the two-channel correlation technique. With CMOS amplifiers, you can get the noise floor down way below the amplifier noise, though of course you have to average for awhile if you want it to work well at very low frequency.
Sure. But there are a lot more bogies down in the very low baseband than just flicker noise.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
I have repeated these measurements with batteries I could actually buy and not some undisclosed stuff that F.Walls was not allowed to endorse.
<I did not use cross correlation but my amplifier had < 220 pV/rtHz noise floor. The good traces are limited by the 30 dB/decade rise below 50 Hz. (The short/60 Ohm measurements in that text were taken after the cap at the op amp inputs. I had not yet identified that as a problem.)
The amplifier at that time was 20* ADA-4898 parallel, the problem was
10 * 10uF Wima PP foil and 10K bias resistor. You can clearly see the 30 dB/decade floor rise below 50 Hz. It levels out at full 10K voltage noise.That went away with a ?100 4700 uF wet slug tantalum but the amplifier did not survive for long.
<The disaster happened when I investigated some dBs of noise rise above
500 KHz. The tantalum (already recovered in the picture) was not the culprit but the thin U-shaped routing to the 2 * 10 op amps. Skin effect. The wire grid healed that.But then I forgot about that bridged tantalum some weekends later when I tried to measure the noise of some 18650 Lithium batteries. :-(
Building such amplifiers is like peeling an onion. It takes some iterations. The current one is 16 * CPH3910 without feedback, intended for cross correlation.
Gerhard
In part. The CTE of your average plastic is of the same order as its TC of epsilon, and both contribute to the TC of capacitance.
For instance, the CTE of PTFE gets up to 2900 ppm/K near its glass transition temperature. Thermal expansion increases the area quadratically and the thickness linearly.
Neglecting fringing, the capacitance goes as area/thickness, so if there were no TCepsilon, the capacitance tempco of a stacked-film PTFE capacitor would follow the CTE. (A rolled film-and-foil cap might behave a bit differently on account of the smaller CTE of the metal and the softness of PTFE.)
However, the measured TC of metallized PTFE capacitors is a fairly steady -150 ppm/K with no big whoopdedoo at T_G. So the TCepsilon is following the CTE with an additional thermal contribution. I could wave my arms about thermal excitations opposing alignment of the microscopic dipoles, but that would be about as reliable as the Friday stock market report. ;)
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
Am 26.02.21 um 17:35 schrieb Phil Hobbs:
PTFE has that unfortunate phase change at room temperature.
Gerhard
Yes, and I know how to use it, thank you. Most of his output is foul language, but sometimes there's a gem.
Jeroen Belleman
Unless we have some weird non-linear dielectric, I'd assume so, but some dielectrics /are/ weird. I have to think this over.
Jeroen Belleman
Oof, that hurts all the way to here.
Yup. Peel off a layer, cry a bit, peel off another....
What drain currents are you running?
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
Right, but it doesn't make Teflon capacitors misbehave at all.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
Am 26.02.21 um 18:08 schrieb Phil Hobbs:
abt 55 mA @2 to 2.7V for the 16 FETs. Not sure if this is really final. Spice says the gain is constant vs. temperature in spite of the removed feedback. Spice also says 200 pV/rtHz, but I get only 320.
Gerhard
I don't understand your calculation.
The typical LT1028 hits 1nV/root(Hz) at 10 Hz, and 0.8nV/root(Hz) at 100 Hz.
The preamp starts to roll off around 1 Hz, so 0.014 Hz is well below the cutoff frequency.
Fred Walls has published an amazing number of papers on low noise techniques.
The system illustrated in Fig. 1 requires duplicate identical noise sources, and duplicate amplifiers with positive and negative inputs to take the difference between the two noise sources.
The low noise preamp described here has only a single input. It cannot take the difference between two inputs.
The preamp starts to roll off around 1 Hz. It does not respond to very low frequencies.
-- The best designs occur in the theta state. - sw
I let others find the gems. They are too rare to waste time on.
-- The best designs occur in the theta state. - sw
r
f
Batteries.pdf
Nice, thanks.
1.) Is the first take away that NiCad's are good. 2.) Re: pg. 3 NiMH (with no load) is the 'stuff' between 100 -1 kHz acoustic/ vibrations perhaps? The vibration problems I found were first in 9V transistor batteries, but persisted at a lower level with a series string of alkalines.George H.
ifier
lic/
Am 26.02.21 um 19:37 schrieb George Herold:
No, it is harmonics of the 50 Hz grid. It stops at 1 KHz because the FFT analyzer switches bandwidth at decade boundaries and the wider bandwidth adds a smoothing effect.
The Agilent 89441A cannot do a logarithmic sweep; the plot is combined from a linear measurement for each decade. My control program corrects for filter BW before sending the results to gnuplot. It assumes that I measure noise.
I don't have an idea how to recognize and handle discrete spurs in such a noise density plot. They scale differently than noise with regard to bandwidth. A 10 nV carrier is still 10 nV in 10 times the BW. Noise gets worse in 10 times the BW.
And remember that 0 dB in the plot is 1 nV/rtHz.
If you try to measure the noise of a say 5 Vdc power supply, it will typically have only a few millivolts riding on the DC voltage. Feed that DC voltage directly to the varactor through an inductance. The rate of capacitance change is still quite linear in the few millivolt range of interest.
The frequency variation might be quite small, perhaps a Hz or two, so you may need a few frequency multiplier stages before mixing down with a clean crystal signal.
I have used this method to check the quality of a crystal oscillator. The oscillator was multiplied to the GHz range and then mixed down by an other, clean, signal 1 kHz apart producing a 1 kHz audio beat tone Listening to the audio tone, it was easy to check e.g. circuit microphonics.
I wondered why the audio beat tone varied sometimes every few seconds (200-500 mHz), until I noticed that my breathing caused a small temperature variations on the oscillator when doing the tests :-)
No doubt, adding a varactor and feeding it with a noisy power supply voltage would have also caused to generate beat frequency variations detectable by ear. No doubt measuring down to 15 mHz (1 minute cycle time) would be detectable.
Steve Wilson wrote: ==================
** We should all plonk a nut case like you...... Phil
====================
** I have many times seen *very* noisy batteries - particularly 9V types.They were *not* fresh and may have had internal contact issues. The 6 cells inside such a 9V battery are held in pressure contact.
None of your lab test data covers any such thing - cos it was all carefully eliminated from the test procedures.
Almost all the published data on batteries has similar flaws. Lab tests do NOT correlate with actual use.
Ever managed to get the advertised 500 cycles from a pack of NiCd cells ? Easily done in a lab - with *careful* automatic charging and discharging at some favorable rate. No way any user can match that set up.
...... Phil
I get around 0.7 nV for one fet running at 12 mA or so--they're about 1 dB better than the BF862 for my uses. They also parallel better closer to I_DSS. BF862s run very happily at I_DSS, so you can parallel them directly and be sure they're all pulling their weight.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
The 1/f noise goes up as, well, 1/f. That's in terms of power, so to get N times more noise _voltage_, you go N**2 times lower in frequency to get N**2 times more PSD.
That's just on account of your circuit topology and parts values, not the limitations of the op amp.
Again, that just requires moving a few wires or changing some component values.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
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.