Question for the Oscilloscope Experts

[KeepItSimpleStupid]

I like your pics in #137. Look here https://www.boatdesign.net/forums/electrical-systems/din-rail-wiring-opinions-39529.html at post #3.

I wish I had some photos of the 6 or so wall mounted stuff I did. I even used the technique to build custom rack mounted instrumentation. I used this **broken link removed** stuff for my PCB's in my rack instruments.

Winford https://www.winford.com/products/ has a lot of DIN rail stuff. Particularly look at their brackets.

Some of my network stuff at home is mounted on DIN rail, I have some stuff (home automation) ready to be installed mounted on DIN rail.

The gray stuff you see is wiring duct. When I did my wiring, I used 18 AWG, stranded, but with few strands. This meant it stayed in place when bent. I have a low-cost source (direct from the manufacturer) for wire. Problem: You have to wait until they make it and they MAY have the same $50.00 USD minimum order.

Here, **broken link removed** isn't a bad place to get wire.

DIN terminals are fun to use. It's basically an erector set and you can re-use the parts. It might be bit difficult to grasp the concepts. In some simple terms, there are end brackets which fix a cluster of terminals.
There are terminals which have to have at least one end cap per group because one side is totally exposed unless the cap is present. Then there are jumper bars to bus terminals together. You can use end caps to group the terminals. There are separate ground terminals. I mostly used two level terminals and the TS-35 rail.

What's important, is to use a set of terminals for the wires that enter and exit the box.

Generally, in a plant the box is steel with a 1/8" aluminum plate mounted in the bottom of the enclosure. Generally, you only have to tap a few holes for the rail and the wiring duct.

 

[Mikebits]

I understand, but I am one of those that if he needs a special car starts to think how to make even if it is available in the market and even if it means reinventing the wheel or discovering the black yarn or inventing the hot water again

There is a huge difference between you and me. You are a studied and qualified Electronic Engineer, at least i read that out of your answer, and I am an interested person in learning a little bit trough out my life and as a young boy I once built a radio during my high school time using bulbs under the guidance of the teacher and since them I never forgot that experience and it never did really let me go.

By the way I developed the whole wiring and mechanical logic for my observatory for being remote controlled. Look at the images below. I even thought how to measure the current, (peanuts for you but not for me until I found the logic how to do it), of the connected devices in order to know if they are working or not, I went from the point that measuring voltage was a nonsense as the device could be connected but if it is not drawing current it is not working. So I made manually a simple current loop detector or whatever it is called.

Yes right but that is no fun. I once made a amplifier using the INA122P for building a cloud detector and that worked. I searched the internet how to do it and it is working. The thing at the end of the red pipe

View attachment 94784

BTW the grey case is a custom made 180° All Sky camera also made by me.

Below you can see some of the wiring working in my Observatory

View attachment 94785 View attachment 94786 View attachment 94787

I am very impressed with the level of quality, time, and money you have put into your passion of observing the world beyond this blue marble. I think this shows your love of the sciences, the unknown, and your quest to understand how the world of things work. It seems to me that you were/are making good progress, and setbacks due to learning curve are to be expected, and part of the fun in the exploration process. How many tries did Edison make to get the lightbulb? I think you would be making a mistake to give up on your vision, and to quit would be a dent in your self esteem. Keep going brotha

 

[rsfoto]

Hi KISS and Mikebits,

Thanks for the head ups.

Somebody here asked about if my Oscilloscope has a FFT function. Well yes it has one.

I made some Trial & Error experiments and I get now the following FFT curves when analyzing input at the LM358 and as well the Output of the LM358.

Below the FFT curve sampled at 10ms at the input

The blue dotted line is sitting at ~ 60Hz which was the main noise I had before. Now I see no relevant spike there so the 60Hz noise is gone ¿ right ? but I see a relevant spike at ~ 15Hz which is one quarter of 60Hz. Maybe due to now correctly chosen values for the RC component , but Ok I think I am making progress but still using the LM358. I bought some TL082CP, based on a recommendation in this thread, but still have to fiddle out how to use them.



Now measuring the output at pin1 from the LM358 and analyzing it with FFT I see no relevant spike so I guess the output is more or less Ok, but what I have the feeling that my sensitivity has gone down.



One problem I still have is to get at the output of the LM358 real 0V output and not an amplified signal too.

¿?

 

[KeepItSimpleStupid]

Cool, you figured out FFT.

Back, a bit to the OEM schematic. The 500K to Vref basically forces U1b at 1/2 the supply voltage, so it actually sees above and below an artificial zero of 2.5V of the waveform. It's initially DC coupled, so the sensor and the OP amp offset is initially buffered too.

Don't use the TL082. The DC offset is too high.

Then they remove the offset with C10. C10 and R10 for a filter with a -3db frequency of )1/(2*Pi*10M*2.2e-9) or 72Hz, I think. =3db is the same as the signal has lost amplitude of sqrt(2) or about 70%. The point is called a "pole".

 

[audioguru]

Your 'scope is set to 10ms per division and there is one division between the huge blue spikes so their frequency is 1/10ms= 100Hz, not 60Hz.
The opamp has some input offset voltage and is amplifying it so of course its output is not 0V. If you add a capacitor in series with the feedback ground resistor then it will not amplify the DC offset voltage.

 

[rsfoto]

Your 'scope is set to 10ms per division and there is one division between the huge blue spikes so their frequency is 1/10ms= 100Hz, not 60Hz.

Hi Ag,

I referred to the vertical blue dotted line which I positioned at the left side. If you look at the Oscilloscope screen pn the left lower corner there are the values in Yellow Letters

Cursor: Δ Time 16.4ms Δ Frequency 60.80Hz

When Moving the cursor that part showed me continuos different frequencies, and I left it at 60Hz in order to see if there was a relevant spike.

Thanks a lot for the hint with the capacitor in series with the Resistor at Pin 2. That did it.

I still have a max. voltage of 31.4mV when nothing moves but that voltage I also have it coming from the sensor directly.

When taking away the capacitor C1 the sensitivity increases but immediately the 60Hz noise comes back.

Below how the circuit looks right now before feeding the LM3914

 

[KeepItSimpleStupid]

Yea, primarily because you have a Hi Z blowing in the wind. High Z picks up everything. 60 Hz is everywhere. In the walls, ceiling etc.

 

[rsfoto]

Yea, primarily because you have a Hi Z blowing in the wind. High Z picks up everything. 60 Hz is everywhere. In the walls, ceiling etc.

Hi KISS,

From what I see in the FFT curve I have no 60Hz noise anymore because there is no relevant peak. The relevant peak is at 15Hz. From where comes that ?

My RC is ceramic capacitor 104 and carbon pot at 27KΩ

Look at the diagram below

 

[KeepItSimpleStupid]

Seek and you shall find. See: https://en.wikipedia.org/wiki/Schumann_resonances

Is that what you observe?

 

[audioguru]

A ceramic capacitor is a microphone and picks up vibrations and low frequency sound. Use a film capacitor instead. 104 is 0.1uF or 100nF.

 

[Tony Stewart]

rsfoto
You ought to change your question to.
How to monitor ground vibrations for telescope imaging?
Then define exactly what you wish to accomplish.

I would suggest your band of interest is 1 to 100Hz. You must convert g to v to x which is 2 integrations. ( like a low pass filter but infinite gain at DC)
Many cheap seismic sensors are limited >10Hz.

Seismic monitors vary from $10 to $10k.

High end seismic sensors using double integral charge amplifiers can range from DC -10kHz. I
I have used only these types which can easily be calibrated for g by free fall = -1g and ac displacement in μm with exceptional care.
Roof resonances can range all over the place above 1Hz.

Basically you want to buy a seismic sensor with a response down to 1Hz the $50 seismic geophones are good for 10 Hz. This will be far cleaner signal and more sensitive than your piezo . ( check amazon or ebay.) But you will be better off with amplifier included.

MEMs technology is good but requires good electronics and calibration. Maybe you can find a iPhone with one and a good app.
HP MEMs sensors have a response down to 0.01 Hz

Some work down to DC ( so they say)
https://www.colibrys.com/product/vs1000-vibration-sensor/

Most structures of any type have many resonances and Q or gain > 10 including your telescope . Obviously due to 2nd integral of acceleration , the higher the resonant frequency , the better.

 

[Tony Stewart]

These are cheap but remember displacement 4x at 1/2 the frequency must be mentally calculated unless you find an instrument to convert to displacement with peak to peak position swing as the important measurement to compare image wobble in peak to peak.

An open loop Op Amp with a 0.1 Hz High Pass filter will get you velocity . Two stages gets you position. All Op Amps have an input offset and after a DC gain of great than 1e6 ( million) a few mV offset should be a few kV out but saturates near the supply rail. So the DC offset must be nulled using special methods such as choppers and auto zero, which is beyond your scope ( no pun intended)

For something, far less perfect yet a 1000x better than what you have, you could use any old phonograph input to a quiet stereo ( no hum) which supports very low frequency response but then output to a meter or rectify to an LED indicator.

Then get a 10Hz $50 Geophone or try this $2 MEMs with a quiet battery operated pre-amp mounted solidly with beeswax to PCB bottom. ( toilet ring is pure beeswax) and low noise battery power
https://www.ebay.com/itm/ADXL335-3-...-angular-transducer-for-Arduino-/400985206886
https://www.sparkfun.com/products/9269 ( Similar but has datasheets )
Single-supply operation: 1.8 V to 3.6 V

Making an amplifier with a gain of a million with a DC input response would be too difficult without experience, so I suggest this as a starting step. Then you can find a VCO or voltage controlled oscillator to hear the vibration near 1Hz or at least under 100Hz or just use an AC analog volt meter. Then you can pick up knee flexing on the roof induced scope wobble or see a car going buy on the meter but not be true position.

If you do want to make it portable with battery operation only for no hum, and need an Op Amp..... But remember long jumpers are like antenna to noise. The smaller the better.

0.1Hz to 10Hz Noise Voltage: 45nVP-P means with a gain of 1 million and a DC null adjustment pot the noise of 450nVpp x 1e6= 250mV with suitable filters between 0.1 and 10Hz Input offset is 10uV max, which is very good. but x 1million is 10V so DC gain must be low using large capacitor for DC blocking such that RC>>10 seconds.

Mouser and Digikey probably deliver next day to Mexico.

THis is not a complete solution, just ideas. Remember you have to inegreate twice to get position so micro-g's gets converted to micro-meters as a function of frequency properly.

$400 product **broken link removed**
Lower the noise and frequency response.... the more expensive but better it is.

 

[KeepItSimpleStupid]

I'll try to explain integrals and derivatives quickly.

If s is distance ds/dt is velocity

If v is velocity then dv/dt is acceleration. It can also be written as the second derivative of s with respect to time.

dx/dt is called the derivative to x with respect to time.

So what is a derivative. It's an equation of the instantaneous slope of a line.

There is a similar anti-derivative called an integral. One simple explanation is it's the area under a cure, It's also the inversee of a derivative, sort of (multiplication and division) and (addition) and subtraction. Since there are an infinate number of lines that have the same slope, integral is not precise. So, the integral of a derivative is called an indefinite integral.

When you put boundaries or initial conditions, you can find the definite integral.

Remember y- mx+b; Say y=10x+3; Well dy/dx( y=10x+3) = 10 , or y=10 which is the same as the slope.

If we integrate dy=10dx, we get y=10x+C, note it's missing the 3. C is an arbitrary constant. So, C can be anything.

If we knew a value that satisfied y-mx+b, we can actually find C to be equal to 3 for our original function.

The derivative and integral thing is easy for polynomials.

There are electronic differentiators and integrators and they were the basis for analog computers.

This is hopefully a seat of the pants explanation with little explanation. It's Calculus.

 

[rsfoto]

I'll try to explain integrals and derivatives quickly.

If s is distance ds/dt is velocity

If v is velocity then dv/dt is acceleration. It can also be written as the second derivative of s with respect to time.

dx/dt is called the derivative to x with respect to time.

So what is a derivative. It's an equation of the instantaneous slope of a line.

There is a similar anti-derivative called an integral. One simple explanation is it's the area under a cure, It's also the inversee of a derivative, sort of (multiplication and division) and (addition) and subtraction. Since there are an infinate number of lines that have the same slope, integral is not precise. So, the integral of a derivative is called an indefinite integral.

When you put boundaries or initial conditions, you can find the definite integral.

Remember y- mx+b; Say y=10x+3; Well dy/dx( y=10x+3) = 10 , or y=10 which is the same as the slope.

If we integrate dy=10dx, we get y=10x+C, note it's missing the 3. C is an arbitrary constant. So, C can be anything.

If we knew a value that satisfied y-mx+b, we can actually find C to be equal to 3 for our original function.

The derivative and integral thing is easy for polynomials.

There are electronic differentiators and integrators and they were the basis for analog computers.

This is hopefully a seat of the pants explanation with little explanation. It's Calculus.

 

[rsfoto]

Seek and you shall find. See: https://en.wikipedia.org/wiki/Schumann_resonances

Is that what you observe?

Hi KISS,

No, I have no peaks at the frequencies described in the Schumann resonances article.

 

[KeepItSimpleStupid]

OK, OK.

Your in a car. The distance (displacement or s) to the next town is 60 km.

It takes you 1 hr to get there. Your average speed is 60km/hr. Speed is also called velocity Velocity is therefore a function of t. s=v(t) or s is "some function of time.

OK, You floor the accelerator and you feel a force on your back. That's acceleration.

We feel a force on our feet when we stand. It's due to the acceleration due to gravity of F = m * a or the Force i mass * acceleration.

The point is, if you know the acceleration and some other stuff, you can find the speed. If you know the speed ant the the time it took to get there you know the distance.

So 60 km/hr * 1 hr = 60 km traveled.

There is just some complicated math (Calculus) that does the same thing, but with time varying quantities.

So, if you have a function of the distance traveled as a function of time, you can create a function v(t) that would have your instantaneous speed with respect to time.

If we do the magical process called integration on v(t) and we know that s=0 when t=0, we can actually find s(t).

 

[rsfoto]

OK, OK.

Your in a car. The distance (displacement or s) to the next town is 60 km.

It takes you 1 hr to get there. Your average speed is 60km/hr. Speed is also called velocity Velocity is therefore a function of t. s=v(t) or s is "some function of time.

OK, You floor the accelerator and you feel a force on your back. That's acceleration.

We feel a force on our feet when we stand. It's due to the acceleration due to gravity of F = m * a or the Force i mass * acceleration.

The point is, if you know the acceleration and some other stuff, you can find the speed. If you know the speed ant the the time it took to get there you know the distance.

So 60 km/hr * 1 hr = 60 km traveled.

There is just some complicated math (Calculus) that does the same thing, but with time varying quantities.

So, if you have a function of the distance traveled as a function of time, you can create a function v(t) that would have your instantaneous speed with respect to time.

If we do the magical process called integration on v(t) and we know that s=0 when t=0, we can actually find s(t).

Hi KISS,

What I do not understand is what relation has your calculation to do with what I am trying to achieve ?

That is what I do not understand

 

[rsfoto]

$400 product **broken link removed**

Thanks Tony,

¿ But where is the fun ?



and how do I learn a bit about electronics ?

which is beyond your scope ( no pun intended)

Got the indirect ... Thanks, you are right

 

[KeepItSimpleStupid]

Hi KISS,

What I do not understand is what relation has your calculation to do with what I am trying to achieve ? That is what I do not understand

Just TRYING REALLY HARD to try to explain Tony's second integral of g (m/s^2) is displacement (m) WITHOUT actually doing the math.

The similarity that you SHOULD be able to relate to is the car example.

G (acceleration due to gravity) is generally a constant for most problems, but in your case the change in g (now a small letter) is related to vibration (displacement).

 

[Tony Stewart]

Unless your telescope has some slack in the moving parts which can cause high frequency buzz motion from low frequency vibration or a resonance, the only frequencies of interest you should be looking at are the seismic frequencies I posted. These of course are not sounds but vibrations you can feel and see.

10% or 1/10 of the signal in log scale is 20 dB

This requires a two stage low pass filter so that frequencies above 10 Hz are attenuated , so anything at 100Hz will be 1% of input or 40 dB down and at 1kHz 0.01% of input.

Since DC offsets with high gain will saturate or clip the signal, DC must be balanced or DC blocked = a high pass filter HPF above 0.1Hz which is a series RC product= time constant of >2/f = 20 seconds. e.g. 20uF in series into 1MOhm load. But recall sensors must be high end to detect down to this level, since cheap geophones are only 4.5Hz (worse 10Hz geoph. and much worse piezo on glass) thus a HPF of 1Hz or 2 second time constant approx.

edit
So your primary range of interest is 0.1Hz to 10Hz, but there is slack or hysteresis or stiction then abrupt motion stimulates any mechanical resonances in the telescope and base as a step contains some energy for all frequencies ( limited by rise time and step size.) ( like driving over a road bump stimulates the resonant frequency of the car mass on springs dampened by shocks but a loose tie rod creates a click sound )

A 2nd order LPF slope effect is the same as converting acceleration, g to motion, x but only over the range where the filter slope is 2nd order LPF meaning attenuates higher frequencies by two x 20 db per decade .e.g. 40 dB attenuation as f acceleration rises from 10Hz to 100Hz to give a flat displacement response.

everything else you might see is just interference or noise and may not move the imager, but low frequency vibration will be most significant. end edit

the initial assumption is critical however... no slack in moving parts, like the focus parts or scope mounts, bearings etc.