GHz ac frequency- Maximum Practical generation - Solenoid question for the experts.

GHz ac frequency- Maximum "Practical" generation - Solenoid question for the experts.

It's me again. Big ideas, big questions.

(A quick special thanks to everyone who has ever taken a moment to answer my questions:
(Bob S, bipolar, Phrak, f95toli, jambaugh, waht, Pumblechook, ZapperZ, Astronuc, mheslep, TallDave, granpa, olgranpappy, weejee, Lurch, JesseM, pervect, )
You've all really been a big help in forming my ideas and bringin them to reality.)

Ok, end of "shoutout"; now for the question!

Ok, I've been asking about THz frequency generation and the likes, but I'm begining to understand the difficulty generating these frequencies, especially from your responses, so for the experts:

1. What's the highest ac frequency, GHz range, that we can generate reasonably "easy" with our current technology? "Reasonably" low noise and "decent" efficiency. I hate using these qualitative descriptions rather than quantitative, but I'm hoping you'll understand. Just think "practical" on a fairly well funded DIY budget.

To give you an idea of the experiment being performed, and to lead me to the next question:

It involves Two solenoids fed by high frequency ac... This being the case-

2. If you have a ferromagnetic core (Iron for simplicity) in a solenoid, what is the time lag between turning on the magnet and the core being full aligned with the field? Maybe there's a formula that factor's in size, length, radius, B field generated by the wire, etc. I'm sure it's in nanoseconds.

With question 2, I'm hoping to find the frequency where the change in freqeuncy outpaces the alignment of the core. Though, this is not the point of the experiment, merely the determining factor of whether the magnets I'm using will have a core or not.

and last but not least:

3. Have any of you ever worked with ac electromagnets? I havn't really been able to find too much info on this.

Thanks!

What's the highest ac frequency, GHz range, that we can generate reasonably "easy" with our current technology? "Reasonably" low noise and "decent" efficiency. I hate using these qualitative descriptions rather than quantitative, but I'm hoping you'll understand. Just think "practical" on a fairly well funded DIY budget.
Frequency multiplication will get you that high quickly. Start with a submillimeter source such as a 100 GHz gunn oscillator, or a GPS harmonically synthesized multiplier, then begin applying frequency doubler or tripler stages as needed.

If you have a ferromagnetic core (Iron for simplicity) in a solenoid, what is the time lag between turning on the magnet and the core being full aligned with the field? Maybe there's a formula that factor's in size, length, radius, B field generated by the wire, etc. I'm sure it's in nanoseconds.
This is negligible. The solenoid will increase inductance to further impede AC current. Also ferromagnetic materials don't respond well at high frequencies, generally a few MHz.

I had been browsing around for frequency multipliers now that you mention it.

You recommended a 1.1-1.7 THz multiplier with .025% efficiency.

a few posts back.

Is the .025% effiency the percent of input power that makes it through the conversion?

Thanks for the iron core/high frequency advice btw.

Last edited by a moderator:

Four concerns about solenoids with iron or ferrites:
1) Intraturn capacitance of solenoids makes them become capacitive at higher frequencies, and may induce circuit resonances. Test them for inductance vs. frequency on a vector impedance meter or network analyzer.
2) Ferrites are lossy. They have a parameter called a loss tangent that essentially makes them look like frequency dependent resistors.
3) The speed of light is 1 / sqrt (e0 u0) (about 30 cm per nanosec). The speed of light (speed of magnetization) in a ferrite is possibly 30 times slower; 1 sqrt( e0 u u0) where the first u is the ferrite's permeability. This means that the magnetization cannot penetrate more than several millimeters per nanosecond.
4 Iron solenoids suffer from eddy current losses, which is why transformer iron is laminated. At higher frequencies, oxidized iron wire is useful (like in automobile ignition coils). Beyond this, ferrites are used.

I had been browsing around for frequency multipliers now that you mention it.

Is the .025% effiency the percent of input power that makes it through the conversion?

Yes. Frequency multiplication is very inefficient, and especially in the sub-millimeter range. Yet, it is the only way to produce a stable source at sufficient power levels. You might get a stronger signal from a FIR laser, but the output is a pulsed, not a true CW wave.

For more power, you would have go with a clinotron tube, but they are bulky, require high voltage for tuning, up to 8,000 volts in some cases, and require alot of cooling. But clinotrons are not a stable signal sources.

The next promising technique, for THz generation is photomixing, but currently power levels are well below what you could achieve with a multiplier.

Different viewpoint:

Solenoids are inductive.
The applied voltage should be know and the solenoid inductance can be calculated. Thus it is fairly simple to calculate the current.

Exactly what are you attempting to do?

All the firrites that I am familiar with are brittle and easily broken.

Two solenoids fed by high frequency sounds like a radio frequency transformer??

Now the juices are flowing...

Ok, I've definately been convinced that a solenoid core is not what I need here. Thanks for the formulas and explanations Bob, that really cleared some things up for me. I had an inclination that once the frequencies were up there that the "mag lag" would become more and more pronounced.

So, with these formulas, is it safe to say that the core of a high frequency ac magnet will have multiple N-S alignments progressing from the outside to the center?

Looking at my calculations, I'm feeling pretty comfortable with something in the range of 74-300 GHz. Maybe even a little higher.

What am I trying to do? Well, let's just say I want to break the laws of physics... or maybe redefine them, because laws are meant to be broken right?

I do have to say though, I wouldn't be able to do this without everyone's help I wouldn't be as close as I was now.

Oh, I almost forgot my other questions.

How do I shift the sinosoidal frequency to cosinosoidal - frequency independant. (+Pi/2 if the frequency is 2 Pi.

Can this be done accurately at the high frequency level, or is it best done early with two seperate multiplier chains?(I'm thinking the later, and the chains would be identical... both sides would be the same frequency, just offset like a sin and cosine wave)

Is it best to amplify the signal at several points in the multiplier chain, early in the chain, or as high as possible in the chain? Keeping noise to a minimum is important here.

For the 500 GHz range, should I stay with fiber optic cable?

I think those are all my remaining questions... this way I won't need a ton of seperate posts.

Building a better flux capacitor? I seem to recall mist around the flux capacitor in Back to the Future...

Oh, I almost forgot my other questions.

How do I shift the sinosoidal frequency to cosinosoidal - frequency independant. (+Pi/2 if the frequency is 2 Pi.

Can this be done accurately at the high frequency level, or is it best done early with two seperate multiplier chains?(I'm thinking the later, and the chains would be identical... both sides would be the same frequency, just offset like a sin and cosine wave)
Phase shifting before the multiplier chain, or inbetween stages will be unpredictable, as the whole process is very non-linear. So if you need a precision phase shift, then it would be better done at the output.

The highest phase shifter I saw was in the W band, which goes from 75 to 110 GHz, it works by changing length of a dielectric slug inside a waveguide, others phase shifters work by adjusting ferrite material in waveguides, or you can simply adjust the physical length of the transmission line. I'm sure its possible to make this work at 500 GHz, but keep it mind you are working with sub-millimeter lengths now, precision might be difficult.

Is it best to amplify the signal at several points in the multiplier chain, early in the chain, or as high as possible in the chain? Keeping noise to a minimum is important here.
That depends on the components you get, it will usually say in the datasheet what the maximum input power rating is for a doubler, or a tripler stage. I think VA in the link already supplies assembled high multiplier chains for different frequencies.

For the 500 GHz range, should I stay with fiber optic cable?
Transmission lines for THz is still an issue. There are waveguide that go up 500 GHz, but they have a such small hole you could barely see it with naked eye.

I'm not aware of any fibers that can directly pass 500 GHz, but the standard way of sending RF through fiber is to modulate a laser with your signal of interest, send the laser over fiber, then with a special photodiode demodulate the laser and extract the RF. I'm not sure if this can be done at 500 GHz yet.

f95toli
Gold Member

At 500 GHz your best bet would probably be to use a dielectric waveguide. Or, alternatively just use quasi-optics (silicon lenses etc)

Btw, you DO realize that you can't feed a solenoid 500 GHz, right? It is very difficult to "bend" signals at that frequency range even with a waveguide and the whole concept of inductors/capacitors simply do not work anymore.

fairly well funded DIY budget
It would be good to know what you mean by "well funded". Even equipment that works in the "normal" microwave range (up to say 65 GHz) can be extremely expensive (a simple 2.4mm connector can easily cost \$100) and the prices go up exponentially as you enter the sub-mm range since most of it is made to order and much of it is custom made since there is virtually no commerical market for this technology yet (unless perhaps if you happen to be using a frequency that is also used in e.g. military installations and can buy surplus equipment).
You need several thousand dollars just for passive components (waveguides, connectors and so on), and that still leaves the source etc.

Bipolar, do you have any idea what I'd give right now for a Delorean? hahahahaha. Man... Stainless steel, Mr. Fusion generator, Hover Tires...

Waht, I think when this is over I'm going to have to put you on the patent. After all our talks, I've reformulated to just a 100 GHz oscillator gunn like you mentioned and skip the multipliers. I really want to just build a proof of concept prototype, rather than a full scale working item. Plus it seems 100GHz is a lot easier to work with, especially concerning that Phase shifter you just mentioned.

Plus I think this goes with what you're saying f95... My budget at most could be a couple grand or maybe a little more... Especially when you get into the customization and expensive technology of the THz range.

But would you say the 100GHz range (waveguides, etc.) is more affordable for the serious inventor? I think this W band is alot easier to manage.

Waht, taking this range into account, what kind of fibers would be recomended? Also, with the W band phase shifter, is it most practical to split the signal and run one "loop" aligned sinosoidally and one "loop" cosinosoidally or take one "loop" and shift half of it with on dielectric "in" and one dielectric "out" (if that second arrangement is even possible)?

If that arrangement is possible, would it be better than the split? And what are the consequences of trying a split at 100 GHz?

Thanks for all the great answers and explainations! I've tried asking my teachers, but they seem to be only focused on the lesson at hand and always divert their answers to that. ( GOD BLESS EM! :) )

Thanks

Waht, I think when this is over I'm going to have to put you on the patent. After all our talks, I've reformulated to just a 100 GHz oscillator gunn like you mentioned and skip the multipliers. I really want to just build a proof of concept prototype, rather than a full scale working item. Plus it seems 100GHz is a lot easier to work with, especially concerning that Phase shifter you just mentioned.

Plus I think this goes with what you're saying f95... My budget at most could be a couple grand or maybe a little more... Especially when you get into the customization and expensive technology of the THz range.

But would you say the 100GHz range (waveguides, etc.) is more affordable for the serious inventor? I think this W band is alot easier to manage.
I'm flattered, but would you mind sharing what you are working on so we can help you better. Because you are planning on spending literally thousands of bucks on parts that may or may not achieve what you intend to do.

And now since you decided to lower the frequency, here's a tip. Cost of microwave components rises exponentially with the frequency, mostly because of industry supply and demand. So if you were to cut the frequency again to 40 GHz, you could reduce cost to easily fit your budget. I brought up 40 GHz because it's the frequency widely used in telecommunications, hence parts are easily available.

Waht, taking this range into account, what kind of fibers would be recomended? Also, with the W band phase shifter, is it most practical to split the signal and run one "loop" aligned sinosoidally and one "loop" cosinosoidally or take one "loop" and shift half of it with on dielectric "in" and one dielectric "out" (if that second arrangement is even possible)?

If that arrangement is possible, would it be better than the split? And what are the consequences of trying a split at 100 GHz?
Since you mention fiber optics, I assume you need to transport your signal over long distance? It's possible to send millimeter wave over fiber via laser modulation, but a cost of such of a link is comparable to a new car. Alternatively there are waveguides, but they are used only to interconnect various subsystem. That's why you need to generate the signal right on the spot.

So if you need two phase shifted signals, you can split the source with a waveguide power splitter, and attach the phase shifter on one output, while the other output will be the reference. That way you would have two independent signals, and like with any power splitter, you will lose half the power at each output. The phase shifters are adjustable from 0 to 360 degrees, so you would have to tweak it carefully to compensate for any lead or lag as needed.