46 year old Voyager 1 once again sending real data

[Vanadium 50]

the magnetic field of the Sun falls with distance, at some point it has to be undetectable. But we have not been able to provide an order of magnitude.

Um...no.

One knows how a dipole falls off, and it should have been unmeasurable years ago. But it's not. There is plasma out there, and that's what is being measured. But its not a static magnetic field like the earth's or even the sun's.

 

[pines-demon]

Um...no.

One knows how a dipole falls off, and it should have been unmeasurable years ago. But it's not. There is plasma out there, and that's what is being measured. But its not a static magnetic field like the earth's or even the sun's.

At some distance it is no longer the magnetic field of the Sun, is it? If not, can we measure the magnetic field of other stars (directly without using spectroscopy)?

 

[sophiecentaur]

There are at least three factors:

1. The average magnetic field gets smaller with distance. The degree is unknown, and this is what is being measured.
2. The available bandwidth for a perfect spacecraft degrades with distance in a predictable way.
3. The spacecraft itself is degrading in an unpredictable way.

But I am still trying to understand the point being made. There are a few surviving instruments out there. Why wouldn't we look at them?

You would need to look at the minutes of the regular NASA progress meetings. Choices will have been made throughout the mission about which experiments are performing best at the time; which ones to switch off ….

 

[Vanadium 50]

can we measure the magnetic field of other stars

Sure. as soon as we get close to one. Can you wait a few million years?

which ones to switch off ….

I think the general feeling is that anything you shut off might not turn on again.

I know heat is an issue. The RTG loses power over time as the plutonium decay, and some instruments were to be warmed by waste heat from others. Nobody anticipated running this long.

 

[pines-demon]

Sure. as soon as we get close to one. Can you wait a few million years?

I cannot wait, maybe PF. Anyway the Sun magnetic field still falls off to undetectable, specially now that V1 and V2 are beyond the heliopause.

 

[sophiecentaur]

, at some point it has to be undetectable

Sorry but that is naive as a general statement. Given time, even a coarse ADC will produce a bias at zero crossover which the statistics of the noise will reveal when analysed. But we do not know what the actual magnetic data values are so we can't tell how low we are in the likely error rate of received data.

The available bandwidth for a perfect spacecraft degrades with distance in a predictable way.

The upper limit will change, agreed, but at this stage,, the significant factor is 'low pass filtering' the data to improve SNR

The inverse Square iLaw is the deep space comms engineer's friend. You only need to take four times as long to analyse a signal from twice as far away. And we have years and years for this.

 

[pines-demon]

Sorry but that is naive as a general statement. Given time, even a coarse ADC will produce a bias at zero crossover which the statistics of the noise will reveal when analysed. But we do not know what the actual magnetic data values are so we can't tell how low we are in the likely error rate of received data.

Sure but you missed the first part of that sentence, it has to dissappear at sufficiently large distance. Can we detect Alpha Centauri magnetic field from your house?

If there are things that we do not know (for example when that tails decays to noise), then that is an answer to my earlier question, but I already acknowledged that too.

 

[Vanadium 50]

You only need to take four times as long to analyse a signal from twice

I'm not sure I buy that. I think you need to slow the transmission down (or repeat it).

If you travel farther by d, your signal goes down as d^-2. If I integrate d² times as long, my signal returns to its original value, but noise goes up as d.

I think you actually lose as d^-8/3. Doubling the distance costs you a factor of 6.4.

 

[sophiecentaur]

I think you need to slow the transmission down (or repeat it).

I was assuming that the data is sent out continuously and increasing the number of samples will improve the error rate / SNR. At the speed of Voyager through the field, the mean measured values won't change over even a long integration period

but noise goes up as d.

What woud cause the noise to increase? By increasing the integration time, the noise bandwidth goes down. My reasoning comes from comms engineering and link budget calculations. I can't see that things would change at extreme distances.

There are two contributions to the noise - noise from the receiving equipment and noise from the magnetometer but why would increasing integration time not affect both in the same way?

 

[Vanadium 50]

What woud cause the noise to increase?

It grows as √t. The fractional noise goes down, but the total noise goes up with increasing t.

 

[sophiecentaur]

It grows as √t. The fractional noise goes down, but the total noise goes up with increasing t.

I was questioning your statement that noise goes up with d. d has nothing to do with it directly. Of course the rms noise goes up as √t but the signal goes up as t. I thinik that's pretty straightforward and it's how they dig signal out of the dirt - just at the expense of time taken.

To clarify in terms of noise and signal power on the link, doubling the distance will reduce the transmitted signal by 6dB but the channel noise power will be the same. There will also be the effect of instrumental noise and the SNR of the transmitted signal will be the same at each end of the link. The variation of magnetic flux will decrease as distance increases but that would depend on the specifics of the equipment as to which inverse law dominates in the received SNR.
Were you making assumptions about the accuracy of the measured H values? It could be relevant.

 

[Vanadium 50]

If you mitogate the signal attenuation with integration time (or repeating the message, which is much the same thing) you get more noise. That's why you have a 8/3 power and not a simple inverse square

 

[sophiecentaur]

If you mitogate the signal attenuation with integration time (or repeating the message, which is much the same thing) you get more noise. That's why you have a 8/3 power and not a simple inverse square

I'd like to see some derivation of that 8/3 figure. The result of adding (repeating) the same signal is linear (6dB) and adding two random signals (or halving the bandwidth) gives an increase of 3dB and an increase in SNR of 3dB. I've never questioned this before as a rule of thumb.

 

[Vanadium 50]

Distance changes by d.
Signal changes by d^-2.
Integration time changes by d^x. We will solve for x.
Therefore noise changes by d^x/2.

Integrated signal changes by d^x-2.
Signal significance goes as S²/N or d^2x-4/d^x/2.

Holding the significance constant tells you 3x/2 - 4 =0. Therefore x = 8/3. That's why a doubling of distance requires you to integrate by 6.4 times as long, not 4.

 

[sophiecentaur]

I'd need some reference about this or some more explanation of your steps.
Can you spot the difference between what you have written and the following?

Received signal power will be proportional to P_t d^-2
In terms of bandwidth, the noise power (P_n) admitted is proportional to bandwidth. Bandwidth is proportional to 1/t (t = time over which received signal is averaged) so P_n is proportional to 1/t. If the noise is all due to the receiver then SNR will be P_t d^-2 /P_n
P_s varies as 1/d. Noise power density (NPD) is the same for all link distances so P_n is NPD Δf
So SNR for the appropriate processing time (aka 1/Δf) will be P_t d^-2 /NPD Δf

Are you introducing something else that's relevant to low signall levels?

 

[sophiecentaur]

I managed to get us down a bit of rabbit hole here. My initial point was that the ISL is useful. Doubling the distance has the same effect, whether it's from 100 to 200km or from 1LY to 2LY. Compare that with the best RF feeder you can find, which will lose you 1dB for every 100m or even the losses in atmospheric links.

 

[Nik_2213]

IIRC, one of the surprises has been how much / how far the solar heliopause fluctuates. Not just tied to sun-spot cycle and flares / CMEs, but 'gusts' of galactic 'wind'. Further out, in 'genuine' interstellar environment, you get into 'Local Bubble', with conditions significantly modified by a bunch of supernovas...

IIRC, our solar system entered 'Local Bubble' about 5 million years ago, crossing the 'firewall' embedding super-nova derived iron isotopes in ancient ice...

FWIW, I've used this as a possible solution to the 'Where Are They' conundrum...
'Of Filters and Firewalls...' Post #15 in...
https://www.physicsforums.com/threads/the-short-story-thread-post-yours-here.914630/

 

[difalcojr]

The record is probably still affixed, showing earth's location, music, pictures, sounds, music. Think Johnny B. Goode was deciphered by an alien already, possibly.