Hoorah for the patent system

Have you looked at a portable Josephson junction calibration standard?

At the moment they require liquid helium, but if somebody can make suitable SQUIDs with high temperature superconductors you might be able to get away with liquid nitrogen.

-- Bill Sloman, Nijmegen

You can always lowpass filter the reference output, at variously heroic levels. The real killer is the below 10 Hz noise. ADR440 is spec'd at 1 uV p-p from 0.1 to 10 Hz, the best I know of. The Intersils are around 3 uV p-p.

The next step would be to lowpass filter then average a bunch of references.

Temperature control, or at least isolation, is important. Millikelvin thermal noise, from air currents and such, whacks the chips and their interconnects and adds lf noise. RF can hurt, too.

John

I did that but those 5uF film capacitors are not exactly small. I also floated a large thermal mass inside the box to mount all the parts onto. This lowered the low frequency stuff I saw on the output of the filter a whole bunch.

I wonder how noisy the forward voltage of an LED or Laser is. The tempco of their forward voltage is the nearly same as that of a normal silicon diode. I could even string several in series if needed.

Yes and in my current prototype the 0.001 to 0.1Hz noise translates to noise at the system output. I am looking to move the operating frequency up so that DC folds to up above the band of interest.

Or the other order so you only need about 5 of those big capacitors. I wonder if taking the middle 1/3rd or some other nonlinear average would work better.

So can mechanical stress. You never want the PCB flex to be large at the reference. This can be helped by an extra mounting hole or even isolating cuts in the PCB.

From my experience, the high temperature SQUIDs are noisy. There seems to be a temperature effect even in superconductors. The noise was about 100 time higher than reported for the low temperature ones.

The 0.1 on the output to the adjust makes the impedance there low at RF frequencies. The LM317 likes to demodulate RF frequency current ripple.

No, that is the op-amp.

I thought of that, but the lowpass filters must include at least one opamp per, and those opamps will have ballpark 1 nV/rthz noise too. So filtering then summing (passively!) has an advantage.

Yeah, the next time you build some high-precision thing, push on the tops of the ic's with a pencil or something. Fun!

John

Not really. Congress pressured the USPTO to move applications through the system quickly. And then they haven't funded it to support the workload.

The basic policy there is: If an examiner can't find grounds to deny an application, issue it. And examiners only have a matter of hours to review each one. If there's a problem, it can be hashed out in court by the affected parties.

No doubt, this was pushed through Congress by patent trolls and other assorted special interests. But it was sold as necessary to protect the 'little guy', the lone inventor working in his basement.

...the guy who can't afford the 'hashing it out in court' - yeah, sure, it's definitely being done 'in his/her interest' - in the interest of stealing anything he/she invents and handing it on a platter to anyone with deep pockets. Yup...

Speaking as someone who once considered patenting something, and then discovered that there was no way in hell to even bother starting without a few million in legal offense/defense funds, so I let it rot.

I hope you at least published it, so it couldn't be patented by another.

If you have semi-deep pockets, you can attempt to depend on trade secrets and non-disclosure agreements.

.. or just yell at the PCB and watch the scope.

Some circuits make good microphones.

On Jan 11, 7:49 am, John Larkin wrote: [... multiple references ...

You can still sum in front of the filter:

REF1 ---/\\/\\---- REF2 ---/\\/\\----+ REF3 ---/\\/\\----+ " " REFN ---/\\/\\----+-----+------!+\\ ! ! >-----+----More filter if needed === --!-/ ! ! ! ! GND ----------

In an extreme low pass filter it is the op-amps noise current that hits you the hardest.

But THAT opamp ^^^^^^^

has 1 nv/rthz noise!

John

Yes it does but the number of references would have to be:

before the filter before the op-amp doesn't help.

The "more filter if needed" can further reduce the noise from that point. Since there aren't many op-amps below 1nV/sqrt(Hz), for most practical situations the extra filter won't be the determining noise level of the whole system.

I suspect that the circuit could be made better if each input was soemthing like this:

Others REFN ----------!+\\ ! ! >--+---\\/\\/------++-------!+\\ --!-/ ! ! ! >---+--- To more filter ! ! === ---!-/ ! R2 +--/\\/\\--+ ! ! ! ! ! GND ----------+ C1 +---!!---+ ! ! ! ! D2 +--!!---- ! ! ! ---/\\/\\------------------------------- R1

The outliers would be given less weight when the diodes D1 and D2 conduct. The circuit as drawn assumes an exact match between the references. Since this is never the case, a more complex circuit would be needed but I think this clearly shows the idea. The noise from a reference is not perfectly uniformly distributed. There tend to be noise events. These could be ct off.

Fund it? The USPTO is a *profit* center. Why do you think you have to pay "maintenance" on patents?

Which is pretty bad since the courts look at the USPTO as the arbiter of "novel" and "useful".