Author Topic: What determines the minimum/maximum frequency an AC transformer will work at?  (Read 3354 times)

0 Members and 1 Guest are viewing this topic.

Offline BeaminTopic starter

  • Super Contributor
  • ***
  • Posts: 1567
  • Country: us
  • If you think my Boobs are big you should see my ba
I was quite surprised to learn that after shutting off a transformer there is still measurable current that can be seen on "human" time scales, I believe 10's or 100's of milliseconds. Is this actually capacitance and not the same principle as "stepping up/down current" and it just seems like it is storing charge? If so what's the lowest frequency you can transfer current between two coils and what determines this speed? Size/distance? What happens as you make the frequency higher and higher is there an equivalent transformer that uses mm waves, IR, or even light somehow?
Max characters: 300; characters remaining: 191
Images in your signature must be no greater than 500x25 pixels
 

Offline bjbb

  • Contributor
  • Posts: 44
  • Country: us
All hail the search lords of duckduck/bing/google.

https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19660001049.pdf

Core construction, core and wire material, and voltage determine applicable frequency.

The excitation energy takes much time to decay, depending on core size and material.
 

Offline bob91343

  • Super Contributor
  • ***
  • Posts: 2675
  • Country: us
On the low frequency side, the limiting factor is the core saturation.  In order not to draw excessive current, the source must have a minimum frequency and voltage combination.  If the frequency is lowered, the voltage must also be lowered in proportion.  So a 120 V 60 Hz transformer primary can be operated safely at 60 V and 30 Hz.

On the high frequency side, the situation is entirely different.  Basically there is no upper limit other than that posed by the distributed capacitance of the windings and the inductance of the leads.  However, the hyteresis and eddy current losses in the core increase wirh frequency so there is heating just from the excitation.  At some point it becomes impractical to use the device.  Typically the upper frequency limit is very roughly ten times the design frequency.

Transformer design is nearly a black art.  Without decades of experience, it's simple to screw up a design.  That's why there are experts to consult and established companies to manufacture.  Even the relatively mundane factors such as winding patterns become important for certain applications.
 

Offline BeaminTopic starter

  • Super Contributor
  • ***
  • Posts: 1567
  • Country: us
  • If you think my Boobs are big you should see my ba
So my question (like most of my questions) is more theoretical then practical. I'm not interested at say 30hz, or building one, but rather what happens at 1Hz or a gigantic transformer at less then 1hz. When does AC cease to become AC and just turn into pulsed DC with no transfer of energy? What kind of magic happens at the extreme ends of the spectrum. IR/Light is just high frequency why can't it work in a transformer (If you think this is a stupid /easy question then you are not thinking about it hard enough i.e. why is glass an insulator of EMF at lower frequencies but a good one at high frequencies if we go past the simple answer of: refraction down fiber optics make it work, or what happens between light and RF)?   

Imagine a magic DC to daylight signal gen what would the outputs on the back look like assuming you started at 1hz and cranked it up to 1000THz and noted what happens to your transmission lines as you went up in small 10hz steps.
« Last Edit: November 30, 2019, 06:09:39 pm by Beamin »
Max characters: 300; characters remaining: 191
Images in your signature must be no greater than 500x25 pixels
 

Offline Gyro

  • Super Contributor
  • ***
  • Posts: 10174
  • Country: gb
Any time you have a change in current (no matter how small) then you have a change in magnetic flux - translate small change as a larger change over a longer period. It is a matter of whether the core is large enough to couple it to the secondary.
Best Regards, Chris
 

Offline larry42

  • Regular Contributor
  • *
  • Posts: 174
  • Country: 00
So my question (like most of my questions) is more theoretical then practical. I'm not interested at say 30hz, or building one, but rather what happens at 1Hz or a gigantic transformer at less then 1hz. When does AC cease to become AC and just turn into pulsed DC with no transfer of energy? What kind of magic happens at the extreme ends of the spectrum. IR/Light is just high frequency why can't it work in a transformer (If you think this is a stupid /easy question then you are not thinking about it hard enough i.e. why is glass an insulator of EMF at lower frequencies but a good one at high frequencies if we go past the simple answer of: refraction down fiber optics make it work, or what happens between light and RF)?   

Imagine a magic DC to daylight signal gen what would the outputs on the back look like assuming you started at 1hz and cranked it up to 1000THz and noted what happens to your transmission lines as you went up in small 10hz steps.

There appears to be a fundamental misunderstanding of physics here.

J Fourier showed that real signals can be written as the sumation of sinusoids. Hence "pulsed" DC is, in fact, a summation of sinusoids and will therefore be transferable through a transformer.

In order for a transformer to operate in the usual fashion, as a near ideal coupled inductor, there must be a conversion of a field, generated by the wire into a flux, and vice versa at the secondary.
For an efficient conversion of field to flux, a material with a non-unity relative permeability may be used, depending on the freq. range of interest. Similarly, multiple turns may be used in practice instead of a single turn in order to have a greater field for a given current.

I have made transformers at hundreds of MHz - and have bought ones that work even higher. These don't work at low frequencies, as their inductance is insufficient.
At the other end - around VHF to some point at GHz, the ferrite materials become too lossy, the coils self resonate and the currents radiate (== act as an antenna) rather than pass through the wire. The transfomer can not longer be regarded as a monolithic ideal entity, but as a collection of R's, L's, G's and C's, each with a transfer function that is a function of freq - and your transformer equation (and operation) no longer holds.

With the example of glass it can be regarded as a question of boundary conditions (see Maxwell's equations). I believe that you could make a waveguide (== fiber) at 10MHz and use it as a transmission line - but it's just not particularly practical to have a 10m diameter glass rod, instead of a roll of RG-223 coax.

NB glass remains an insulator at light frequencies as well - it is not "conducting", but acts as a dielectric waveguide - the same way as glass can be used as a dielectric medium for making capacitors at low/RF frequencies...

If you have an animated GIF in your avatar or signature then I reserve the right to think you're a dolt.
 

Offline bob91343

  • Super Contributor
  • ***
  • Posts: 2675
  • Country: us
Operation at very low frequency is certainly possible.  What happens is that when a voltage is applied to a winding, the current begins to increase, since of course it's an inductance.  The current continues to increase as long as the voltage is applied.  At some point, the flux density, which has been increasing with time, reaches saturation of the core and no more field can be generated.  Then the inductance drops sharply and the current rises abruptly.

To avoid the latter situation, the time has come for the voltage to be removed or reversed in order for the core to respond.  Then the second half of the cycle can begin.

So for very low frequency, one needs more winding turns and a larger core cross section.  As the frequency drops, so the core and winding must be such that they can stay in normal conditions.  Bigger cross section, more turns of wire.

Any given core can support as low a frequency as you like, as long as the amplitude of the applied voltage is small enough.  More turns on the core makes saturation take longer to occur, but this is limited by the practical maximum number of turns you can wind.  Smaller diameter wire allows for more turns but you can only go so far before the wire gets too small.  And a giant core will help but again, you will have to have a big factory to house such a core.

These are legitimate questions and it's good that someone asks them.  I have designed thousands of transformers and know how to balance the parameters to obtain desired results.  I once made a calculation to learn how much power a particular core can handle.  For a given material of core and wire, the ruling factor is window area.  Window is the core opening in which the winding wire resides.  Smaller wire means more turns possible but the same amount of power.  And some practical factors steal window area, such as multiple secondaries, high voltage insulation, temperature capability, and more.
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5571
  • Country: us
And the question hiding behind your question is:  What is the problem you are trying to solve and what resources are available to solve it?

You can make a transformer to work at frequencies a few orders of magnitude below the ones we typically think of, but may need a transformer the size of a shipping container or larger.  Can your application afford that space, weight and cost?  Engineering is the art of the practical.  These thought experiments are important and useful for exploring the edges of practicality, and practical varies from application to application.  What is practical for a government working on a project of national importance is very different from what is practical for a starving college student working on a hobby.
 

Online ejeffrey

  • Super Contributor
  • ***
  • Posts: 4034
  • Country: us
So my question (like most of my questions) is more theoretical then practical. I'm not interested at say 30hz, or building one, but rather what happens at 1Hz or a gigantic transformer at less then 1hz. When does AC cease to become AC and just turn into pulsed DC with no transfer of energy?

There isn't really any limit at the low end as long as you avoid core saturation, but eventually the primary reactance becomes lower than the resistance which limits your power transfer.

Quote
i.e. why is glass an insulator of EMF at lower frequencies but a good one at high frequencies if we go past the simple answer of: refraction down fiber optics make it work, or what happens between light and RF)?   

It isn't that glass is an insulator at RF and a conductor at DC.  It actually behaves exactly the same for RF and light -- a low loss dielectric with eps_r ~ 2.  The difference is that with light we are always talking about free space waves whereas with RF we are interested in both signals carried by wires (i.e., PCB traces or coax) as well as free space waves (radio).  A large enough diameter glass rod would be a perfectly good "fiber" for RF signals!

Quote
Imagine a magic DC to daylight signal gen what would the outputs on the back look like assuming you started at 1hz and cranked it up to 1000THz and noted what happens to your transmission lines as you went up in small 10hz steps.

At 10 Hz you would need wires/coax for your signal as a waveguide would be prohibitively large.  By around 100 GHz you would want to switch from coax to waveguides when at all possible as the center conductor losses on coax would be very high.  At 200 THz even a metal waveguide would have too much loss and you would need to use an optical fiber made of only dielectric such as glass.
 

Offline JustMeHere

  • Frequent Contributor
  • **
  • Posts: 827
  • Country: us
J Fourier showed that real signals can be written as the sumation of sinusoids. Hence "pulsed" DC is, in fact, a summation of sinusoids and will therefore be transferable through a transformer.

Common example: Ethernet uses transformers on each end to isolate the grounds.  (Thus no ground loop problems.) 
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
There's no such thing as "DC" to a transformer, because a DC voltage implies a linearly rising flux density with respect to time.  The highest flux densities reached on Earth are around 60T (sustained) to 500T (explosively pumped).

Unless you're planning a trip to a neutron star, I wouldn't worry about anything even approaching DC.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline BeaminTopic starter

  • Super Contributor
  • ***
  • Posts: 1567
  • Country: us
  • If you think my Boobs are big you should see my ba
So my question (like most of my questions) is more theoretical then practical. I'm not interested at say 30hz, or building one, but rather what happens at 1Hz or a gigantic transformer at less then 1hz. When does AC cease to become AC and just turn into pulsed DC with no transfer of energy?

There isn't really any limit at the low end as long as you avoid core saturation, but eventually the primary reactance becomes lower than the resistance which limits your power transfer.

Quote
i.e. why is glass an insulator of EMF at lower frequencies but a good one at high frequencies if we go past the simple answer of: refraction down fiber optics make it work, or what happens between light and RF)?   

It isn't that glass is an insulator at RF and a conductor at DC.  It actually behaves exactly the same for RF and light -- a low loss dielectric with eps_r ~ 2.  The difference is that with light we are always talking about free space waves whereas with RF we are interested in both signals carried by wires (i.e., PCB traces or coax) as well as free space waves (radio).  A large enough diameter glass rod would be a perfectly good "fiber" for RF signals!

So how would the large glass rod work if the radio freq. RF energy (Long photons) don't have a metallic "sea of electrons" skin to travel down? My understanding of why glass is transparent is that it is really complicated but has to do with photons being absorbed and reemitted at the same wave length,  that's much different then from an AC signal traveling through a metal antenna.

Quote
Quote
Imagine a magic DC to daylight signal gen what would the outputs on the back look like assuming you started at 1hz and cranked it up to 1000THz and noted what happens to your transmission lines as you went up in small 10hz steps.
Quote
At 10 Hz you would need wires/coax for your signal as a waveguide would be prohibitively large.  By around 100 GHz you would want to switch from coax to waveguides when at all possible as the center conductor losses on coax would be very high.  At 200 THz even a metal waveguide would have too much loss and you would need to use an optical fiber made of only dielectric such as glass.
So what happens in the gray area at 100 thz or 10thz or where ever things begin to change from really small microwaves/mm waves and long IR, imagine you are cranking the freq. dial up and noting the changes; its easy to say glass is for light metal for RF but whats the transition like? Is it sharp or more intriguingly gradual as is usually the case.
Max characters: 300; characters remaining: 191
Images in your signature must be no greater than 500x25 pixels
 

Online ejeffrey

  • Super Contributor
  • ***
  • Posts: 4034
  • Country: us
So my question (like most of my questions) is more theoretical then practical. I'm not interested at say 30hz, or building one, but rather what happens at 1Hz or a gigantic transformer at less then 1hz. When does AC cease to become AC and just turn into pulsed DC with no transfer of energy?

There isn't really any limit at the low end as long as you avoid core saturation, but eventually the primary reactance becomes lower than the resistance which limits your power transfer.

Quote
i.e. why is glass an insulator of EMF at lower frequencies but a good one at high frequencies if we go past the simple answer of: refraction down fiber optics make it work, or what happens between light and RF)?   

It isn't that glass is an insulator at RF and a conductor at DC.  It actually behaves exactly the same for RF and light -- a low loss dielectric with eps_r ~ 2.  The difference is that with light we are always talking about free space waves whereas with RF we are interested in both signals carried by wires (i.e., PCB traces or coax) as well as free space waves (radio).  A large enough diameter glass rod would be a perfectly good "fiber" for RF signals!

So how would the large glass rod work if the radio freq. RF energy (Long photons) don't have a metallic "sea of electrons" skin to travel down?

Same way RF travels through air or vacuum. There is no need to have a conductor to propagate em waves of any frequency.  Metal conductors mostly help you do it in a smaller space.

Quote
Quote
Quote
Imagine a magic DC to daylight signal gen what would the outputs on the back look like assuming you started at 1hz and cranked it up to 1000THz and noted what happens to your transmission lines as you went up in small 10hz steps.
Quote
At 10 Hz you would need wires/coax for your signal as a waveguide would be prohibitively large.  By around 100 GHz you would want to switch from coax to waveguides when at all possible as the center conductor losses on coax would be very high.  At 200 THz even a metal waveguide would have too much loss and you would need to use an optical fiber made of only dielectric such as glass.
So what happens in the gray area at 100 thz or 10thz or where ever things begin to change from really small microwaves/mm waves and long IR, imagine you are cranking the freq. dial up and noting the changes; its easy to say glass is for light metal for RF but whats the transition like? Is it sharp or more intriguingly gradual as is usually the case.

No abrubt change.  You can use a dielectric waveguide at ordinary rf if it is big enough.  The point wasn't "metal for microwaves glass for light".  The point is the waves are the same the difference is how we humans interact with them.  That is limited because we are about 2 meters tall and only have materials on the periodic table
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5571
  • Country: us
There's no such thing as "DC" to a transformer, because a DC voltage implies a linearly rising flux density with respect to time.  The highest flux densities reached on Earth are around 60T (sustained) to 500T (explosively pumped).

Unless you're planning a trip to a neutron star, I wouldn't worry about anything even approaching DC.

Tim

No such thing as DC to an ideal transformer.  Resistance limits current to a constant at DC in a real transformer, which means no flux change and no output.  But this the essence of the physics making low frequency transformers hard.  The secondary current is proportional to the field rate of change, so drops linearly with frequency.  The asymptote is zero, which can only be conquered by an infinite number of turns in the secondary. 

High frequency limits on transformers are set by parasitics as stated previously.  You have to get into the interactions of waves with atomic level structure to understand the transitions with frequency in fibers and other materials.  I am not aware of any one sentence, or even one paragraph explanations that don't depend on much much background in the area.  The following link provides some overview of the behavior.

https://www.google.com/imgres?imgurl=https%3A%2F%2Fars.els-cdn.com%2Fcontent%2Fimage%2F3-s2.0-B9781437778175000043-f04-12-9781437778175.jpg&imgrefurl=https%3A%2F%2Fwww.sciencedirect.com%2Ftopics%2Fchemistry%2Fdielectric-constant&tbnid=P-EQ1ho6oMaUyM&vet=12ahUKEwj79MKA75TmAhU1IH0KHeTHCRUQMygFegUIARD3AQ..i&docid=lkphjELth_8z1M&w=470&h=314&q=dielectric%20function%20vs%20frequency&ved=2ahUKEwj79MKA75TmAhU1IH0KHeTHCRUQMygFegUIARD3AQ

 

Offline BeaminTopic starter

  • Super Contributor
  • ***
  • Posts: 1567
  • Country: us
  • If you think my Boobs are big you should see my ba
There's no such thing as "DC" to a transformer, because a DC voltage implies a linearly rising flux density with respect to time.  The highest flux densities reached on Earth are around 60T (sustained) to 500T (explosively pumped).

Unless you're planning a trip to a neutron star, I wouldn't worry about anything even approaching DC.

Tim

No such thing as DC to an ideal transformer.  Resistance limits current to a constant at DC in a real transformer, which means no flux change and no output.  But this the essence of the physics making low frequency transformers hard.  The secondary current is proportional to the field rate of change, so drops linearly with frequency.  The asymptote is zero, which can only be conquered by an infinite number of turns in the secondary. 

High frequency limits on transformers are set by parasitics as stated previously.  You have to get into the interactions of waves with atomic level structure to understand the transitions with frequency in fibers and other materials.  I am not aware of any one sentence, or even one paragraph explanations that don't depend on much much background in the area.  The following link provides some overview of the behavior.

https://www.google.com/imgres?imgurl=https%3A%2F%2Fars.els-cdn.com%2Fcontent%2Fimage%2F3-s2.0-B9781437778175000043-f04-12-9781437778175.jpg&imgrefurl=https%3A%2F%2Fwww.sciencedirect.com%2Ftopics%2Fchemistry%2Fdielectric-constant&tbnid=P-EQ1ho6oMaUyM&vet=12ahUKEwj79MKA75TmAhU1IH0KHeTHCRUQMygFegUIARD3AQ..i&docid=lkphjELth_8z1M&w=470&h=314&q=dielectric%20function%20vs%20frequency&ved=2ahUKEwj79MKA75TmAhU1IH0KHeTHCRUQMygFegUIARD3AQ


This kind of explains what I was getting at:
https://www.researchgate.net/figure/Frequency-dependence-of-complex-dielectric-constant_fig2_305116040

What does the sigma represent? The scale on the bottom seems to be Hz

You are on to something to answer the question
Max characters: 300; characters remaining: 191
Images in your signature must be no greater than 500x25 pixels
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Sigma?  ...Epsilon?

That's a representative diagram, showing how different polarization and loss mechanisms manifest as flat regions, or time constants and resonances, respectively.  The dimensions are arbitrary, since the mechanisms and magnitudes vary by material.  The horizontal axis is ballpark Hz, yes.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5571
  • Country: us
I have forgotten most of what I knew about this subject so can't spout a quick explanation.  But the text that introduced me to the subject was "Introduction to Solid State Physics" by Charles Kittel.  You could od worse than obtaining and absorbing a copy.
 

Offline BeaminTopic starter

  • Super Contributor
  • ***
  • Posts: 1567
  • Country: us
  • If you think my Boobs are big you should see my ba
I have forgotten most of what I knew about this subject so can't spout a quick explanation.  But the text that introduced me to the subject was "Introduction to Solid State Physics" by Charles Kittel.  You could od worse than obtaining and absorbing a copy.

Those peaks explain why there isn't a smooth transition from say a metallic wave guide to glass. I knew I was on to something. Sometimes when you approach a problem with less info you can come at it more abstractly with the stigma of how its supposed to work attached.
Max characters: 300; characters remaining: 191
Images in your signature must be no greater than 500x25 pixels
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf