Of course, I am not proposing voodoo! I have just noticed the other thread, so maybe we should continue there, my post above would fit there better than here.
Sy,
25 nanohenries is the inductance of about 3" of straight wire. You could get the same results as shown with an LC filter but it would have more L C & R.
No this is not an LC filter. Some of the technique is really a trade secret, although a bit of an open one.
Yes the time scale is important. I did not hide it, you certainly are capable of calculating the loss from the LC values given. They do not account for the results.
Jan,
When I bought it it was direct from China, not through distribution. It was really cheap. When I got it I realized it is a copy of a Tektronix, but not quite as good.
ES
And some are fans.
An externally hosted image should be here but it was not working when we last tested it.
The inductance of a wire is from two parts.
The internal inductance of a cylindrical cross section conductor is 15 nanohenries per foot. This drops as a consequence of skinning, the frequencies you examine are certainly reducing the internal inductance. 3 inches at dc would be about 4 nanohenries.
The external inductance of any wire is 100% dependent on the proximity and geometry of the return conductor. Since you cannot measure the inductance without the return conductor, you must take this into account.
When I saw your mention of 25 nH, I woldered how you measured it accurately. It is not easy, most cannot perform this test at this level with any accuracy. I am quite confident you did not accurately measure L.
I tried to measure my 60 picohenry load resistor, but was unable to get any measurement below 250 picohenries as a result of the current path geometry forced upon me by the dissipation requirement.
Your estimate of the wire inductance is of question. If you chose a larger inductance in your model, what value would be required to duplicate your result?
Cheers, John
John,
You are right, I have the value too high. I went with that value because I reasonably believe I am below it.
Even with the probe the load is less than 100pf. That would require a filter pass resistance above 1000 ohms for the filter effect shown. That is not the case.
The return path is a cast AL ground plane. I reused a CATV case to hold the parts.
ES
Last edited:
Sy,
If you measure it without an RF field and I do the same with one, we get different results. But what if I don't know there is the RF field? If I don't have the equipment to detect it how do we know why we differ?
Yes, I once had a football field where a few of the amplifiers did not work properly due to RF as show by a wide bandwidth (500mhz) scope on the output terminals. That is where the EMI was entering. The other 100 or so of the same make and model did not have problems. Could have been the slight difference in the speaker cable length or some manufacturing variance.
When we do experiments and the results agree exactly then we should worry!
ES
Actually, you have the value too low.
Eddy currents, path geometry, and skinning are all going to conspire to throw you off as well, both in L, Q, and distributed C.
The biggest challenge you are going to face is..how to measure it correctly. Be wary of drawing incorrect conclusions as a result of uncontrolled testing. You are not the first to be hit on the head by bad test design, you won't be the last.
Parasitics are a b....
Cheers, John
Price, it's OK for tinkering like my little handheld Wavetek VOM. If I had to look again there is a Chinese brand that looks a bit better. If I was to spend >$1500 I would just get a Tek.
The OWON software is surprisingly better than expected.
Scott,
I wish the Tek were only $1500! Last one I bought was closer to 5K!
The Owon was $200 ish and allows me to produce publishable pictures. Tek wanted more than that for the added software!
John,
You can question the inductance value but it does not rise to 300uh as would be required to produce the results as shown.
I am pretty sure if you think about it you can figure out what is going on. But circuit theory is not the best approach.
ES
At those frequencies, one would think striplines and cavities (or so says my 1960s era memories of microwave experiments).
Tek "lunchbox" scopes start at $990. I'm making the assumption of course they perform better. I came across a "reverse" knockoff. A Tiwanese company makes USB microscopes for checking for fake money (great toy BTW for SMT work) that they sell for $25 direct. I've seen the same thing for as much as $399!!! at PCGEAR with a slightly nicer stand. In this case from the pics it IS the same thing.
Tektronix : Oscilloscopes
Last edited:
That’s a sanity test Anatech, and thanks for putting it on the table.
Metrology involvement taken aside (top notch), they (NIST) are into every far edged scientific knowledge-seeking investigation themselves.*PS
They are aware of every possible spin-offs that worth dealing with.
Needless to say, they are able to evaluate the "worth" of any such spin-off.
Regards
George
*PS
“To know a thing's nature is to know the reason why it is.”
“We possess scientific knowledge of a thing only when we know its cause.”
(Aristotle, Posterior Analytics)
500kHz is the period between bursts. What's the ringing frequency in the burst? Can't make that out in your screenshot but it looks pretty high.
Actually, this is not as simple as you think.
When trying things like this, you must ask;
1. What is the return path for a dc current?
2. What is the return path for 20 hz to 20Khz?
3. What is the return path for the 500 Khz signal?
And, what are the dissipative losses and path resistivity for those three conditions?
They will be different for all three cases. So when you perform a measurement, what exactly are you measuring?
Remember, exceptional claims require exceptional proof. I'm not here to bust your chops Ed, I like what you do..
Um, my idea of circuit theory ain't exactly what others think... If you choose an explanation which falls outside established physics, then I repeat...exceptional claims require exceptional proof. I do not believe you have considered the return current path.
They are unfortunately not without their foibles as well. But those are few and far between.
That frequency exceeds the level where current path consideration must be dealt with.
Cheers, John
It seems the device not only attenuates the signal but also shifts its frequency by about 9kHz! Is this noise, or a nearby AM transmitter?
I believe it is 1/f noise causing the choke in the noise source to ring. Notice the noise floor also drops.
The issue as I understand it is that the Bybee device is supposed to attenuate noise not signal. So the issue is how does a passive device tell the two apart. There are two possibilities, one is that noise is bidirectional and the device acts as a one way signal pass or stop. The other is that 1/f noise is quite distinct from signal among other features approaching in some theory infinite amplitude with a fast rise time typical.
So if I can load just the 1/f noise that might to a limited extent do what is possible with a passive device. I am not trying to use some esoteric technology to quench the spike but rather well known mechanisms.
That is why my noise source is a lamp not a pseudo random noise generator. I get very high peak noise spikes from a lamp running at a few volts when the noise average is below 10mv. If I were to look at the noise with an FFT it would be quite tricky to separate the 1/f noise from the pack. So I used a scope with sampling. My two screen shots are similar spikes but try not to read more into them than should be there.
Perhaps, if you are a physicist.
If you are a meta-physicist, or a para-physicist it may be the beginning of an entirely new set of principles.
- Status
- Not open for further replies.