Alternatives to creeping contacts for high frequencies

[_maxim_]

I anticipate immediately that I am not sure if this is the right place to ask my question, in case I had made a mistake I would migrate willingly into the relevant area.

I have an RF circuit on small PCB board (about 3x2 cm) that moves along a vertical axis in a measuring probe (it serves essentially to capture RF data in the presence of magnetic fields).

Now the point is that for each "quote" reached by the circuit for the point of measurement, it would be advisable to connect an RLC circuit sized appropriately; the moving part remains fixed and that contains a linear coil, the resonant circuits (capacitors and capacitors) would have to be changed.
Given that at each position the frequency response of the entire circuit is fixed, and that can change from a few hundred MHz to a few tens of MHz, it is not conceivable to use a single broadband circuit for the whole excursion.
My idea would be to connect in some way (yes, but which one ??) the PCB to different circuits, cut in frequency based on the chosen odds.
I have no idea how it can be done given the frequencies involved. The power transmitted along the line (adapted to 50 ohms) is of the order of watts or at most a few tens of watts in transmission and microwatt or less in reception; the problem for which I can not see solutions is: how to connect the PCB to the various circuits?
Is it conceivable a kind of controlled switch able to select only one RF channel at a time, without introducing substantial losses?

I had thought about creeping contacts, but here too: how to make them?

Thank you for your attention and I will try to provide further details if my request was not clear.

Best

 

[phinds]

...I would ask for venereal services

Ah ... you may want to look that word up

 

[_maxim_]

Ah ah sorry you're right... Just corrected, my keyboard has been faster than my mind. Hope nobody considered offensive the mistyped

 

[phinds]

Ah ah sorry you're right... Just corrected, my keyboard has been faster than my mind. Hope nobody considered offensive the mistyped

No, not offensive. It was clearly a typo and just amusing.

 

[gneill]

Can you place different LC circuits along the path of the probe and have them inductively couple to an inductor in the probe? As in a transformer, impedance in the load is reflected to the primary. This might allow you to "tune" your probe without physical connections being required.

 

[_maxim_]

No, it can not be realized in such way: coils as in transformers generate magnetic fields which interact with the static field to be measured.
The coil must be connected electrically with the RF circuits.

 

[gneill]

Can you change the tuning instead by varying the capacitance? A Varicap comes to mind whereby the capacitance can be selected by altering the DC bias across it. I don't know off hand if there are varicaps with a suitable range for your application, but I do know that they are used in RF tuners.

 

[gneill]

This varicap might fill the bill:

https://www.nxp.com/docs/en/data-sheet/BB135.pdf?

It's apparently used in UHF tuners.

 

[_maxim_]

It looks like I can only use capacitors, as in the following general circuit

Either the Matching Capacitor and the Tuning Capacitor are elements which I must select depending on the Frequency range.
The L is the coil which moves in the magnetic field which must be connected with different Cm/Ct capacitors.

 

[gneill]

It looks like I can only use capacitors, as in the following general circuit

Well, varicaps are capacitors. They're variable by applied DC reverse bias. Wikipedia shows a few example circuits.

https://en.wikipedia.org/wiki/Varicap

 

[Averagesupernova]

What do you have as a receiver? I am a bit more confused about this setup than maybe I should be.

 

[_maxim_]

The receiver is something like the following: http://www.conspiracyoflight.com/NMR/NMR.html

 

[Averagesupernova]

Am I missing something here? It seems the receiver is only interested in 18 MHz. Why concern yourself with other parts of the spectrum?

 

[_maxim_]

No no.. As I wrote, the schematic and the receiving stage is something like what I have, it is a very simplified version of my instrument.
But the principle is the same.

 

[Averagesupernova]

So you have a receiver already and just need to couple the signal into it and this is what is in question or do you need help with the whole design, receiver and all?

 

[_maxim_]

I just need a way to connect the coil to several RLC circuits which resonate at different residencies.

 

[Averagesupernova]

Then I'd say you need to group up some bandpass and bandstop filters.

 

[Baluncore]

Maybe if you make and position all your LCR circuits, you can then connect them in series.
Only those near resonance will provide an output so the output of the string will be the sum of all.
No matter what the frequency is, it will produce an output without moving connections.

 

[Tom.G]

You could use CMOS switches to switch in the appropriate tuning caps at the receiver. Here is an Application Note on them: https://www.maximintegrated.com/en/app-notes/index.mvp/id/5299
And a 23pg tutorial. https://www.analog.com/media/en/training-seminars/tutorials/MT-088.pdf
above found with: https://www.google.com/search?&q=cmos+switch+design

Also look at T/R switches (Transmit/Receive switches), they can handle the transmit power.

For further ideas, look at schematics for Amateur Radio (Ham Radio) transceivers, you are close to the 15 meter band.

Probably one of the Ham radio operators here can supply many more details. Paging @berkemanCheers,
Tom

 

[_maxim_]

You could use CMOS switches to switch in the appropriate tuning caps at the receiver. Here is an Application Note on them: https://www.maximintegrated.com/en/app-notes/index.mvp/id/5299
And a 23pg tutorial. https://www.analog.com/media/en/training-seminars/tutorials/MT-088.pdf
above found with: https://www.google.com/search?&q=cmos+switch+design

Also look at T/R switches (Transmit/Receive switches), they can handle the transmit power.

For further ideas, look at schematics for Amateur Radio (Ham Radio) transceivers, you are close to the 15 meter band.

Probably one of the Ham radio operators here can supply many more details. Paging @berkemanCheers,
Tom

Thank you Tom.
About the Analog Switches, I think the main problem arises from the frequency: can 850 MHz @ tens of watts be handled ?
Frequencies and powers delivered are in the range of radio/antenna applications...
T/R switches are interesting...
I will have a look around.

Probably one of the Ham radio operators here can supply many more details. Paging @berkeman

Do you mean can I ask directly to Mr. Berkeman?

 

[Tom.G]

About the Analog Switches, I think the main problem arises from the frequency: can 850 MHz @ tens of watts be handled ?
Frequencies and powers delivered are in the range of radio/antenna applications...
T/R switches are interesting...
I will have a look around.

From you brief description, I had the impression the signal of interest was the 18MHz low power receive signal.
Moderately available Analog Switches are available for use up to 1GHz, I don't know their power rating but I suspect tens of watts would be hard to find.

T/R switches are used to switch an antenna from Transmitter to Receiver. Although more costly than a simple Analog Switch, they have the power rating needed and could be used in place of the Analog Switches if needed.

Do you mean can I ask directly to Mr. Berkeman?

You could, by finding his home page here on PF and sending him a Private Message. However our preferred method is to have most technical discussions in public so others can benefit too. The way I referenced berkemen in my post triggered an automatic notification to him that he was mentioned. (you noticed his reference was in Blue? that indicates that a notification was triggered. you can contact him by clicking on that link, click on 'Information', click on 'Start a Conversation'.)

It often takes people a few days to respond to these references, life gets in the way sometimes; or they may be gathering information.

Cheers,
Tom

 

[Averagesupernova]

Are there specific multiple frequencies of interest or would it be anything across the spectrum you have defined and anything in between?

 

[Baluncore]

Given that at each position the frequency response of the entire circuit is fixed, and that can change from a few hundred MHz to a few tens of MHz, it is not conceivable to use a single broadband circuit for the whole excursion.

The vague nature of the description suggests to me that you do not understand the physics of what you are trying to do, or the available solutions.

The dimensions of the equipment is small in wavelengths so moving the detector will probably not make much difference in signal strength. Do you use a coupling loop to sense the RF?

If the RF probe you use is a log-amplifier, then you will need to tune the front-end or think again. But if it was a direct digital down-converter, based on quadrature mixing with a low-pass filter, you could select the frequency band of interest without any need to tune the RF, then use a log converter on the output.

Any tuned circuits that remain in the field will ring like bells if/when struck by a transmit pulse.

Without a better understanding of the system we will not be able to see the obvious solution.
So what exactly is special about the line you move the detector along?

 

[_maxim_]

From you brief description, I had the impression the signal of interest was the 18MHz low power receive signal.

18 MHz is an intermediate frequency which is used for subsequent elaboration of signal. The response picked up by the coil is the same of the frequency transmitter (that's why resonance).

Moderately available Analog Switches are available for use up to 1GHz, I don't know their power rating but I suspect tens of watts would be hard to find.
T/R switches are used to switch an antenna from Transmitter to Receiver. Although more costly than a simple Analog Switch, they have the power rating needed and could be used in place of the Analog Switches if needed.

I see

You could, by finding his home page here on PF and sending him a Private Message. However our preferred method is to have most technical discussions in public so others can benefit too. The way I referenced berkemen in my post triggered an automatic notification to him that he was mentioned. (you noticed his reference was in Blue? that indicates that a notification was triggered. you can contact him by clicking on that link, click on 'Information', click on 'Start a Conversation'.)

It often takes people a few days to respond to these references, life gets in the way sometimes; or they may be gathering information.

Cheers,
Tom

Thank you again, Tom!

 

[_maxim_]

The vague nature of the description suggests to me that you do not understand the physics of what you are trying to do, or the available solutions.

Thank you for your comments but for the purposes of defining the problem it makes no sense to describe all the devices involved in the project.
To make this discussion productive, it is sufficient to focus on the type of problem posed: how to make an efficient electrical connection given the frequencies involved for a coil that moves in a magnetic field and requires resonant RLC circuits cut on specific frequencies.

Here they are a fairly simple lessons even to those who do not know NMR / MRI

The dimensions of the equipment is small in wavelengths so moving the detector will probably not make much difference in signal strength. Do you use a coupling loop to sense the RF?

The Larmor Frequency is dependent from the magnetic field.

According with the theory, when placed in a magnetic field of strength B, a particle with a net spin can absorb a photon, of frequency ν.
That frequency depends on the gyromagnetic ratio γ of the particle. As known, ν = γB , being B the magnetic field.
For ¹H nucleus ν = 42.58 MHz / Tesla. Since the field strength B varies with the position, the frequency varies accordingly, even for a small movement of the detector.

If the RF probe you use is a log-amplifier, then you will need to tune the front-end or think again. But if it was a direct digital down-converter, based on quadrature mixing with a low-pass filter, you could select the frequency band of interest without any need to tune the RF, then use a log converter on the output. Any tuned circuits that remain in the field will ring like bells if/when struck by a transmit pulse.
Without a better understanding of the system we will not be able to see the obvious solution.

The AMP is a wide-band RF amplifier operating in linear class AB (N-MOSFET) that provides hundreds peak RF power over a frequency range 6-500 MHz for any combination of pulse width and duty cycle up to 100ms and 10%.

So what exactly is special about the line you move the detector along?

Sorry, I don't understand clearly your question.
I don't know what add more...

 

[_maxim_]

Are there specific multiple frequencies of interest or would it be anything across the spectrum you have defined and anything in between?

Every frequency at which the RLC circuit resonate (coil/resonator + selected tank filter Ct+Cm) is interesting to detect.

Obviously it is unthinkable to read them all, I would just read only 4 or 5, from the highest (700-800 MHz) to the lowest (1-30 MHz), stepped.

 

[Baluncore]

To make this discussion productive, it is sufficient to focus on the type of problem posed: how to make an efficient electrical connection given the frequencies involved for a coil that moves in a magnetic field and requires resonant RLC circuits cut on specific frequencies.

You have got stuck trying to use, or to avoid the use of sliding contacts. I am trying to identify why you need them in the first place. Others do not seem to have your problem.
“The only interesting answers are those which destroy the question”. Susan Sontag.

 

[_maxim_]

I did not refuse the idea of using sliding contacts, simply I have no idea on how to implement that on my project...
Can you share with me some papers about sliding contacts?
Think that the coil/resonator will move by at least 50-80 cm from one side to the opposite side in the magnetic field along the vertical axis.

 

[Baluncore]

Think that the coil/resonator will move by at least 50-80 cm from one side to the opposite side in the magnetic field along the vertical axis.

You are confusing up and down in the vertical axis with horizontal side to side movement.
1. What do you gain by having irrational coordinates in multiple reference frames?
2. Is the “coil/resonator” the probe?
3. Is that probe a coupling loop antenna?
4. Is the sample held in and moved with the probe?
5. Would your spectrum of CCLR tuned circuits only be part of the receive circuit, or will the transmit pulse also use that path? Even with isolation switches, the presence of tuned circuits in the signal path will ring and produce artefacts that I expect will hide or distort the signal from your sample.

Can you share with me some papers about sliding contacts?

But you do not need sliding contacts.
I might digitise probe position and then use that to set the centre frequency of my receiver, based on the previously mapped field strength. I would find it too difficult to calibrate a whole spectrum of tuned circuits, then to re-adjust their physical position as the magnetic field drifted.

 

[_maxim_]

I want to thank you for giving me the opportunity to clarify some aspects:

You are confusing up and down in the vertical axis with horizontal side to side movement.

Never talked about horizontal movements. I talked ONLY about movement along the magnet axis, which is VERTICAL.

1. What do you gain by having irrational coordinates in multiple reference frames?

?

2. Is the “coil/resonator” the probe?

The coil (or resonator) is a part of the probe, it is inside the probe. By probe we mean the device in its entirety that allows one to make a measurement when connected to a properly hardware.

3. Is that probe a coupling loop antenna?

A probe is a probe. Are you talking about the coil?

4. Is the sample held in and moved with the probe?

Yes, the sample is inside the coil and it moves together with the coil.

5. Would your spectrum of CCLR tuned circuits only be part of the receive circuit, or will the transmit pulse also use that path? Even with isolation switches, the presence of tuned circuits in the signal path will ring and produce artefacts that I expect will hide or distort the signal from your sample.

The probe is involved in both TX/RX process. Power from linear transmitters/amplifiers is delivered to the probe for "exciting" the sample (TX), thus the small NMR signal is detected by the same probe (RX), then preamplfied, detected, digitized, filtered and elaborated by PC.

But you do not need sliding contacts.
I might digitise probe position and then use that to set the centre frequency of my receiver, based on the previously mapped field strength. I would find it too difficult to calibrate a whole spectrum of tuned circuits, then to re-adjust their physical position as the magnetic field drifted.

But it is not a question of calibrating the whole spectrum of tunable frequencies; rather, it is a matter of creating n discrete tuning circuits suitable for frequencies predefined by the field/position: it is much simpler!

Cheers

 

[Baluncore]

Your diagram in post #26 shows the capacitors and inductor that form the resonator. The capacitors Cm and Ct are continuously adjustable so the probe antenna, L, can be matched to Zo of the transmission line. You have just confirmed that “Yes, the sample is inside the coil and it moves together with the coil”. But you seem to be wanting many pre-tuned CCL resonators along the magnetic gradient axis. That would require transfer of the sample through all those resonator inductors as you move the sample along that axis, which seems somehow illogical to me. What have I misunderstood or got wrong?

 

[Tom.G]

That would require transfer of the sample through all those resonator inductors as you move the sample along that axis, which seems somehow illogical to me. What have I misunderstood or got wrong?

I think the OP is trying to vary the tuning of ONE coil (which contains the sample) depending on its physical position relative to the magnetic field. Sounds do-able depending on the required tuning bandwidth.

@_maxim_ , can you tell us the frequency range of the exciting pulses and of the received signal? And, what is the frequency correlation, if any, between the exciting and received signals.

For instance if the exciting frequency is 200Mhz and 1GHz is expected back, the architecture would be different than if the frequencies were identical.
Also do you:

• Require continuously adjustable tuning or are discrete frequencies acceptable?
• Need rejection or suppression of nearby frequencies? If so, how close and how much suppresion?

Cheers,
Tom

 

[_maxim_]

Dear Baluncore

Your diagram in post #26 shows the capacitors and inductor that form the resonator. The capacitors Cm and Ct are continuously adjustable so the probe antenna, L, can be matched to Zo of the transmission line. You have just confirmed that “Yes, the sample is inside the coil and it moves together with the coil”. But you seem to be wanting many pre-tuned CCL resonators along the magnetic gradient axis. That would require transfer of the sample through all those resonator inductors as you move the sample along that axis, which seems somehow illogical to me. What have I misunderstood or got wrong?

The diagram in post #26 is a very basic thank circuit.
If the coil L does not move in the magnetic field (with its sample inside), only one thank RF circuit is required.
If the coil L moves from one position to another experiencing a different magnetic field, as in my case, a different RF thank circuit for is required.
The sample remains inside the coil all the time.

I think the OP is trying to vary the tuning of ONE coil (which contains the sample) depending on its physical position relative to the magnetic field. Sounds do-able depending on the required tuning bandwidth.

Hi Tom,

Yes, it is exactly this.

can you tell us the frequency range of the exciting pulses and of the received signal? And, what is the frequency correlation, if any, between the exciting and received signals.

Exciting pulse and receiving frequency are exactly the same since the sample will emits a tiny wave at the same frequency of the transmitting pulse: that's why it is called Resonance.

For instance if the exciting frequency is 200Mhz and 1GHz is expected back, the architecture would be different than if the frequencies were identical.

In a magnetic field of 1.75 Tesla the frequency for 1H isotope is 75 MHz, at 3.52 Tesla is 150 MHz, ..., at 14.09 Tesla is 600.13 MHz and so on. The highest field I can handle is 19.97 Tesla for a maximum frequency of about 850 MHz, with transmitting and receiving identical frequencies for the same field.

Also do you:

• Require continuously adjustable tuning or are discrete frequencies acceptable?
• Need rejection or suppression of nearby frequencies? If so, how close and how much suppresion?

Cheers,
Tom

Discrete frequencies are fine, the final tuning of caps (maybe through small variable capacities) can be done directly when the sample is inserted in laboratory outiside the magnetic field, since the dielectric will change a bit depending of the nature of the sample observed.
About rejection, I would expect that there is no interference from one frequency to another due their relative distance, let say 800 MHz, 500 MHz, 300 MHz and 50 MHz.
The point is how to switch electrically and mechanically from one thank circuit to another.

Thank to all for your interesting investigation.

 

[Baluncore]

If the coil L moves from one position to another experiencing a different magnetic field, as in my case, a different RF thank circuit for is required.
The sample remains inside the coil all the time.

The magnetic field for NMR needs to be strong, flat and stable throughout the sample volume so as to give clean sharp spectra. The reason for having a gradient field in MRI is to identify the positional source of sample resonance along an axis by resonant frequency, which makes 3D imaging possible.

With a single sample, if you have a steep magnetic gradient, your spectra will be blurred by the size of your sample. With a gentle magnetic gradient, there can be no advantage in moving the sample as it will only change the resonant frequency slightly.

So I still cannot see why you need to measure the same small sample at different magnetic field strengths between the poles of the same magnet, as all it does is upset your matching network for no additional information.

What exactly are you trying to achieve?

 

[Tom.G]

That is going to be... CHALLENGING. I'm sure there are a few RF experts around here with better ideas, but here is a semi-informed first pass.
Overall it looks like separate transmit and receive coils are a good bet. I'll be concentrating on the receive part.

At 800MHz a coil of 22nH and a capacitor of 1.8pF will resonant. To give an idea of size, a 1 turn coil, 3/8in. dia., of 40AWG 30AWG (0.010 dia) wire yields 22nH. Keeping stray capacitance below 1.8pF requires rather small circuit construction. That would indicate putting a pre-amp directly adjacent to the coil. A Mini-Circuits LHA-13LN+ is a possibility for a pre-amp https://www.minicircuits.com/WebStore/dashboard.html?model=LHA-13LN+.

Since you are talking tens of watts, the pre-amp will need input protection. A pair of PIN RF-Microwave diodes in anti-parallel across the input may suffice. Here is a link to one source of PIN diodes. https://www.microsemi.com/document-...icrowave-diode-and-transistor-product-catalog

For higher power and isolation, another design for diode switching is a bridge circuit. As used in a fullwave bridge rectifier circuit, but the RF comes in and exits where the power transformer would be, and the control voltage is applied at what would normally be the output. When the control voltage biases the diodes ON there is an RF path thru the bridge. At higher powers, the control voltage polarity can be reversed to ensure the diodes stay OFF.

The pre-amp would be wideband with the receive tuned circuits at the RF receiver, away from the head. Tuned circuit selection could be done using a diode switching approach. For examples see:
https://www.radio-electronics.com/info/circuits/diode-rf-attenuator/pin-diode-switch.php
https://www.electroschematics.com/3002/pin-diode-rf-switch/

The Transmit is "Left as an exercise for others."
edit: corrected wire size
Hope this helps.
Tom