PopularFX
Home Help Search
Advanced search 
Login Register
Welcome,Guest. Please login or register.
2025-04-04, 20:27:14
News: If you have a suggestion or need for a new board title, please PM the Admins.
Please remember to keep topics and posts of the FE or casual nature. :)

Pages: 1 2 [3] 4 5 6 7 8 9
Author Topic: Tetra Replication  (Read 16619 times)

Group: Professor
Hero Member
*****

Posts: 3675
From my notes it should be 3.237778 MHz
Is that a frequency for ferromagnetic metallic iron or for some non-ferromagnetic iron compound ?
How did you arrive at that frequency ?
   
Group: Moderator
Sr. Member
*****

Posts: 308
I just googled iron nmr frequency.

" The absolute frequency of iron is 3.237778 MHz with respect to 100.00 MHz for TMS"
   

Group: Professor
Hero Member
*****

Posts: 3675
I just googled iron NMR frequency.

" The absolute frequency of iron is 3.237778 MHz with respect to 100.00 MHz for TMS"
That refers to the NMR frequency of individual iron nuclei subjected to an external magnetic flux of such density that it causes the single protons of the Tetramethylsilane to precess at 100MHz.  This happens at the magnetic flux density of 2.34 Tesla.

At a lower flux density, e.g. at 0.5T, the individual Iron nuclei resonate at 690kHz.

The dependence of the NMR frequency on the magnetic field is better captured by the Gyromagnetic Ratio which precisely relates how the NMR frequency of individual nuclei varies in response to the density of the externally applied magnetic flux.
For individual iron nuclei, this Gyromagnetic Ratio is 1381.56Hz/mT.  You can see the values for other metals here.

However, for the ferromagnetic bulk metallic iron, the effective Gyromagnetic Ratio and resulting NMR frequencies are wildly different because of its huge negative internal hyperfine field which affects the iron nuclei in addition to the externally applied magnetic field.

The authors of this paper experimentally measured metallic Iron's NMR frequency as 45.525MHz. ( half of it is 22.763MHz and 1/14 of it is 3.252MHz ) in the absence of an external magnetic field (and in presence of its -33.02T internal hyperfine field).

Unlike non-ferromagnetic compounds, the metallic Iron's nuclear resonance frequency is relatively independent of external magnetic fields because the internal -33.02 Tesla hyperfine field swamps any externally applied fields. 
An external magnetic field >0.75T saturates the bulk metallic Iron (i.e. coalesces and orients all its magnetic domains in one direction) and the flux density of this field does not need to be precisely controlled/correlated with the nuclear resonance frequency (unlike with non-ferromagnetic materials) because the internal -33T hyperfine field of Iron affects its resonance frequency much more than any external field, which we mere mortals could apply. 

For example, the application of a 1T external magnetic field decreases the total magnetic field to which the Iron nuclei are subjected to, to -32.7T* which decreases the Iron's nuclear resonance frequency by only ~350kHz, and for external fields well below the Iron's saturation level (< 0.6T) that frequency changes negligibly (by only -0.033%).   See the hollow squares graph line of f vs. BEXT below:



It is important to remember that even when an external magnetic field is not applied, increasing the temperature of the Iron metal significantly decreases its nuclear resonance frequency, so if the oscillator does not track the temperature then periodic cool-downs are required.



* The ferromagnetic magnetization (domain rotation and coalescence) takes 0.75T to happen. After that, the remaining external flux density directly subtracts from the internal hyperfine field.
« Last Edit: 2025-02-17, 16:10:07 by verpies »
   
Group: Moderator
Sr. Member
*****

Posts: 308
So it seems that Iron's NMR frequency decreases as the magnetic field strength decreases, which is the opposite of what spherics specified for parameters. A large magnetic field reduces the pulse rate in the device. Maybe it's not the NMR of iron I'm hunting for then...

I might need to break things down into smaller experiments to get an idea of the final setup.

Right now I'm waiting on saving up cash for circuit boards and electronics. One experiment he specified was taking a bifilar coil, delaying one winding of the bifilar while pulsing the coils at 150-300 volts. Sweep frequencies from 1khz up to 3.5Mhz, at some particular repetition rate, depending on coil parameters, you will see very high frequency, high voltage pulses begin to appear.

When I did this before the RF interference at certain frequencies does cause electronics to start going haywire. Only at certain frequencies though...

He also mentioned that applying a DC bias of at least 50 volts to a coil, then pulsing 400 volts on top of that, produces the same effect. I guess it's a place to start.

Build a small test to see the effect and then when I power up the tetra, I should be able to look for the same kind of response.

So I suppose my initial plan is, start with the bifilar, try to see the HV bursts. Then try it with a standard solenoid with the DC offset, and see if that works too. Then build the electronics for the final device.
   

Group: Professor
Hero Member
*****

Posts: 3675
So it seems that Iron's NMR frequency decreases as the magnetic field strength decreases,...
if you are referring to an externally applied magnetic field then this statement is true for individual Iron nuclei and for non-metallic Iron compounds.  For example: Iron Sulfide.

It is false for bulk ferromagnetic metallic Iron because of its internal hyperfine field's sign. BTOT = BINT + BEXT
You can observe the latter behavior on the hollow squares graph line in Fig.7 above 7kG.

He also mentioned that applying a DC bias of at least 50 volts to a coil, then pulsing 400 volts on top of that, produces the same effect.
Coils are current devices.  Referring to a voltage alone applied to coil's terminals is meaningless.
« Last Edit: 2025-02-17, 06:33:55 by verpies »
   
Group: Moderator
Sr. Member
*****

Posts: 308
Thank you for all of this great information.

I would like to test temperature now as well.
   

Group: Professor
Hero Member
*****

Posts: 3675
If you plan to generate high power nanopulses then this circuit is a good and tried method.

As a pulse frequency controller you can use the second circuit which is depicted below (with C2=22pF):

The circuit above is based on 1 chip - the 74VHC221.

Also, you might find this topic interesting.
   
Group: Moderator
Sr. Member
*****

Posts: 308
Attached is proposed timing circuit for the tetra. Dependent on gate delays, D flip flops need to be delayed slightly to avoid double pulsing. I'm building this as a separate circuit to minimize the need to redo the entire setup if a component isn't behaving well. I'm thinking of having two inputs such that more control can be gained over the delays to keep the circuit behaving as it should.

Pulse width can be controlled for all channels by altering the duty cycle of the square wave carrier input. D flip flops divide the carrier into 3 sub carriers 120 degrees out of phase. Each AND gate with drive another logic gate to buffer the output onto a 50 ohm line which will connect to a driver board. This way I can tune delays using the line length to ensure proper timing even if each driver board and timing output has slightly different delays.

I might add another output specifically to have a reference output to sync the timing to. Perhaps output the original carrier delayed by x nanoseconds and tune each driver to that delay. Then each driver will output a pulse at the exact correct timing.

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjOCmC0CcIAYB0AWMAmAzJgrO2sA7OgGywAc6YmIOCtt9OMYYAUAErgn1g8jle-eunAJxEkUhwh00xImk42YHDPQ4S3EJnRCFy1TLRa+IFCRTb6CNgHdZm8+Vkp6KF7Yfo3z10Jt7czMPc3E-LzD3FzcAxCDY7hFfPkDVGlxRaAwXdGIQbPRPWQMEszNM6yCNUy10HyrvBrN6q0LPaob22XzuyJNtAeyMeIch1lFK4dFInBGzOaywFCtZkeyVnTwC5dW2PUJZLpzwWCWihgAHAAt2A+CtafNQp5kb9H2EQ8WdzZxNjZWN7XTBlOoNFa5Bq2e4-GrPXJOYHsVRMEY+eg8XK+ESlMahVq0dHQoJY-wgEj5DGjKJHKwDQm2VFEyZ6CkIFy6GwlGSGOgs2lkxK4pikjk6NmU1mBBxC8JS2lrSYIenhTAqmmJdWqkQkhw-TDkHU6I2atWmrUayKpJK2sCELS2ACSsnIUONjIYygcNvttV1ex9-EJRXd8VhI2wWkhOkwWmBHwjogGMYGwNBLt9wfFNqYtDYLr0Lht1NzXtJ+USxHc4UiCq5IEIGobdapvhIbvJreTggpnY8Mop+U09CbQOEFdEllHkpxn2+IxIThjS-jIBuoKTtAdCO3a-eQR+JAh8r1Av+xovNLJ06HU7nsvFI7vtAnzKP7fFt5F+ffI2fZLPrivIJOEFj0iEljXuKA4vrB-ShOBApIUy-I-LBQq9j+yhbigsDRoh+FXLcYLmFBqZQf05rGihoHuOR4RXlRTCbIkx57AA9q44AKG4BAMMgkhEKuGgKHUVg0BcIicrI8AiPAoKGgoABiEDqPQsD0NkIAACIAK4AC4AJ4ADoAM4AMJGQAxgANpAbBAA
   
Group: Moderator
Sr. Member
*****

Posts: 308
I replaced the two inverters with a controllable delay chip so that gate delay can be adjusted for precisely.
   
Group: Moderator
Sr. Member
*****

Posts: 308
Top 3 outputs to X,Y, and Z coils. Bottom three will all drive coil A at 3x frequency whilst limiting the frequency the individual mosfets have to drive to 1x frequency.

Here is the FET I picked:

https://assets.nexperia.com/documents/data-sheet/GAN190-650FBE.pdf

This FET can hard switch at frequencies exceeding 10Mhz. Power dissipation is a problem even for this state of the art FET at such extreme frequencies, so I will water cool with this cold plate:

https://www.digikey.com/en/products/detail/wakefield-vette/180-10-6C/4864899

Thermal resistance of the FET is 1K/W junction to case, the cooler has a thermal performance of 0.084°C/W @ 1.5 GPM, which is extremely good, which you would expect from liquid cooling. This single plate will attach to all 6 mosfets and all 6 drivers.

Here is the driver for the FETs:

https://fscdn.rohm.com/en/products/databook/datasheet/ic/power/gate_driver/bd2311nvx-lb-e.pdf

The driver is also capable of driving the FETs at extremely high frequencies (10Mhz).

Each driver board will have a 5 volt supply input, a 300v input for Vds, and a 50 ohm terminated BNC connector to minimize transmission line effects when running at high speed.

One thing I havent found is a good way to monitor the temperature of the chips.
   

Group: Professor
Hero Member
*****

Posts: 3675
Here is the FET I picked:
https://assets.nexperia.com/documents/data-sheet/GAN190-650FBE.pdf
Did you account for the Miller effect ?

Here is the driver for the FETs:
https://fscdn.rohm.com/en/products/databook/datasheet/ic/power/gate_driver/bd2311nvx-lb-e.pdf
Just put this driver super-close to the FET and minimize the loop-area of the traces that charge and discharge the FET's gate (including the Vdd and ground paths).

One thing I havent found is a good way to monitor the temperature of the chips.
How about the venerable DS18B20
   
Group: Moderator
Sr. Member
*****

Posts: 308
Did you account for the Miller effect ?
Just put this driver super-close to the FET and minimize the loop-area of the traces that charge and discharge the FET's gate (including the Vdd and ground paths).
How about the venerable DS18B20

Haven't accounted for Miller effect...also I need a snubber circuit...
   
Group: Moderator
Sr. Member
*****

Posts: 308
Rg on = 10 ohm
Rg off = 2 ohm
Per datatsheet
Perhaps rg on 2 ohm as well, I've had instability issues in past with low r values
   

Group: Professor
Hero Member
*****

Posts: 3675
...also I need a snubber circuit...
So you are not using a push-pull circuit to drive an inductive load ?!
   
Group: Moderator
Sr. Member
*****

Posts: 308
No, goal is to apply 300v pulses to the coils to get the ether field to spin. No interest in magnetic fields for input coils, just apply voltage across copper mass and remove it as quick as possible to get the effect needed.
   

Group: Professor
Hero Member
*****

Posts: 3675
Pulse width 25ns,...
My coils are all matched to about 175mH, 144 ohms in a brooks coil configuration.
No, goal is to apply 300v pulses to the coils to get the ether field to spin. No interest in magnetic fields for input coils, just apply voltage across copper mass and remove it as quick as possible to get the effect needed.
As soon as you apply a voltage to the coils' terminals, a current starts building up in their windings and magnetic flux is generated according to this scenario.  You cannot avoid that.   ...not even for 25ns.

Inductance is a ratio of the magnetic flux to the electric current flowing through a coil.  Your coils have 175mH inductance so they are not exempt from that phenomenon.

Also, a pulsed coil accumulates energy.  What is going to happen to this accumulated energy when your single-ended FET opens ?
You cannot suddenly interrupt the current flowing through the coil without dire consequences to the FET.  You have to provide a deliberate path for that current or it will find its own path ...and it will not be a path to your liking. TVS diode can perform that function and it will dissipate the energy stored in the coil faster than any capacitor/resistor combination.

Conversely, you can suddenly interrupt the current that is charging a capacitor without consequences.
   
Group: Moderator
Sr. Member
*****

Posts: 308
I usually use a freewheel diode. I might add a snubber for the mosfet.
   

Group: Professor
Hero Member
*****

Posts: 3675
I usually use a freewheel diode. I might add a snubber for the mosfet.
Freewheeling/flyback diode across the coil provides low impedance path for coil's current and causes slow energy dissipation.  Long fall-time.

A TVS diode across the Drain and Source dissipates the energy stored in the coil much faster than the freewheeling diode or RC snubber.  This happens because the voltage drop across a freewheeling/flyback diode is 0.6V and across the TVS diode it is much higher, e.g. 400V.  That is a 666x higher voltage drop and that makes the power dissipation 666x faster, too.

The breakdown voltage of the TVS diodes must be higher than Vdd and lower than VDS_MAX of the FET.

Also, the Vdd should be applied via a second diode in order to prevent the HV spikes from back-propagating into the HV power supply and the low impedance of this supply prolonging the coil's energy dissipation and fall-time.

« Last Edit: 2025-02-18, 05:22:10 by verpies »
   
Group: Moderator
Sr. Member
*****

Posts: 308
That's a really good idea. In simulation I've noticed that the coil continuously circulates current when using freewheel diode, slowly building up over time. I will definitely try this.
   

Group: Professor
Hero Member
*****

Posts: 3675
If the coil were ideal and the freewheeling/flyback diode had a 0V voltage drop, then the current would circulate in the coil FOREVER.

To preserve the energy stored in a capacitor - keep it opened.
To preserve the energy stored in an inductor - keep it shorted.

   
Group: Experimentalist
Hero Member
*****

Posts: 2149

That's right, and that's when we realize that our capacitors are close to ideal, they can hold their charges for a long time, whereas our self-looped inductors are very, very bad, the current attenuates in a few ms.
It's a pity that the room-temperature superconductors that are supposed to solve this problem haven't been around for as long as we've been looking for them (one of my professors when I was at university was already working on the subject in the 70s!  :( )



---------------------------
"Open your mind, but not like a trash bin"
   
Group: Moderator
Sr. Member
*****

Posts: 308
Down the line when I get to testing the device I was going to plot output vs various parameters, including input power consumed. What's a good, scientific way to ensure accurate readings that aren't obscured by high frequency components? I was thinking out heating a quantity of water and plotting temperature vs time.
   

Group: Professor
Hero Member
*****

Posts: 3675
What's a good, scientific way to ensure accurate readings that aren't obscured by high frequency components? I was thinking out heating a quantity of water and plotting temperature vs time.
That's a good way to measure output power and output energy but it is a completely unsuitable way to measure the input power and energy.
However, water calorimetry is messy and requires a Devar for good accuracy.  Take a look at the photovoltaic "wattbox" as an alternative (it is described and photographed on this forum).

If your input energy is delivered as pure DC then you can just multiply the average input Volts and average input Amps to obtain average input Watts, but this calculation is invalid for anything else than pure DC.

Measuring an input power that contains AC or high frequency components fed into non-resistive DUTs/loads is very difficult for poor amateurs because it requires high frequency analog multipliers.  For example something like this:



Alternatively, you can use a 2+ channel digital oscilloscope if it is capable of multiplying two channels together on sample-by-sample basis and integrating the results, ...but there are a lot of ADC resolution and quantization gotchas with that method.

P.S.
Kill-aWatt meters cannot deal with high frequencies and rectangular pulses and high crest-factor waveforms.

   
Group: Moderator
Sr. Member
*****

Posts: 308
I guess the best way is just sample voltage and current as often as possible, maybe at 100khz, and average the results per plot point? Maybe find a way to calculate error margin. Just think out loud here. The input current can be filtered via inductor and then the adc won't have to contend with large sudden spikes of current, if any.
   

Group: Professor
Hero Member
*****

Posts: 3675
I guess the best way is just sample voltage and current as often as possible, maybe at 100khz, and average the results per plot point?
No, first you multiply the i&v samples and then you average their products.  Averaging first and multiplying later is a different mathematical operation that generates different results (results that are not indicative of the input power).
Also, 100ks/s is usually insufficient.  Fortunately even the cheap modern scopes such as the DS1054Z can sample at 1Gs/s.
The 8-bit vertical resolution of the cheap ADCs introduces significant quantization errors, especially when the ADC is underdriven or overdriven.  For this reason a 12-bit or 14-bit ADC are preferred, e.g. as in the DHO900.  See the difference below:



The input current can be filtered via inductor and then the adc won't have to contend with large sudden spikes of current, if any.
If you can filter the input so it results in pure DC then you can get away with using even a slow voltmeter and an ammeter.
   
Pages: 1 2 [3] 4 5 6 7 8 9
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2025-04-04, 20:27:14
Loading...