200kHz HF Transformer needed

[acinfo32]

hf transformer design

Dear all,

I'm searching for a special transformer with the following specifications:

I have a block pulse of 200kHz and 12Vpp (transformer input). The transformer output is connected to a transducer of 1200ohm and must have pulses of about 500Vpp.

Does someone know a transformer with comes close to this specification and which can be bought on the market (without winding myself).

With kind regards,

 

[FvM]

hf power transformer design book

Hello,

I don't know of a particular transformer, but in general small SMPS (switched mode power supply) trafos for high primary voltage (off-line transformer) could be usable with reversed windings. The other possibility are EL backlight inverter trafos, but they are probably too low in power handling respectively too high in impedance.

You can find some transformers of the said types from magnetic device manufacturers as coilcraft. For final selection, the specification should be supplemented. I understand, that the load should regarded resistive, if not, it would change a lot. You should know, if the power capacity implied by your specification (≈ 50 W) is CW or burst-only, the latter allowing for a smaller device if some losses are acceptable. Furthermore, a transformer can be expected to have some stray inductivity. You should take it into account in your specification.

Finally, winding the transformer yourself isn't an issue for high frequency because of the low winding count. I have a design kit from Wurth Electronics Midcom for this purpose.

Regards,
Frank

 

[acinfo32]

resonance transformer

Dear Frank,

Thank you very much for your help. Here a supplement.

Hello,

I don't know of a particular transformer, but in general small SMPS (switched mode power supply) trafos for high primary voltage (off-line transformer) could be usable with reversed windings. The other possibility are EL backlight inverter trafos, but they are probably too low in power handling respectively too high in impedance.

These transformers are probably hard to buy.

You can find some transformers of the said types from magnetic device manufacturers as coilcraft. For final selection, the specification should be supplemented. I understand, that the load should regarded resistive, if not, it would change a lot. You should know, if the power capacity implied by your specification (≈ 50 W) is CW or burst-only, the latter allowing for a smaller device if some losses are acceptable. Furthermore, a transformer can be expected to have some stray inductivity. You should take it into account in your specification.

The device is a piezo ceramic transducer. I don't know if it is resistive but I know that it gives a power of about 300W (round 500V no need for exact values). This is done with a burst of +/- 3 mS and is given 10 times per second.

Finally, winding the transformer yourself isn't an issue for high frequency because of the low winding count. I have a design kit from Wurth Electronics Midcom for this purpose.

Regards,
Frank

With kind regards,

acinfo32

 

[FvM]

% diodes wurth pdf

Hello,

without detailed analysis, I would guess that for a piezo transducer the situation is somewhat different. Basically, you can build a resonance transformer. The transducer has a resistive or real input impedance, but ony at the transducer center frequency.

This is something like a sonar, right? Then time resolution also depends on transducer damping und should no be degraded by the transformer. The frequency dependend transducer impedance should be known to calculate optimal transformer parameter, also driver the properties.

Regards,
Frank

 

[echo47]

resonant transformer design

500V peak-to-peak into 1200 ohms (assuming sine wave in resistive load) is 26 watts, not 300 watts. There's a wrong number somewhere.

If it's 26 watts, Coilcraft may have a suitable part for you. They stock lots of parts, and have easy on-line ordering.

If it's 300 watts, then I built a similar transformer some years ago, but not with such a low input voltage. Going from 12Vpp to 500Vpp at 300 watts --- wow, that's a lot of input current. Maybe an RM6 or RM8 core with one turn of 8 strands of 28 AWG wire on the primary (strands spread out to cover the bobbin, or use an equivalent wide copper strip), and 42 turns of 32 AWG wire on the secondary. Unless I bungled the math! That seems like an unusual transformer to buy off-the-shelf someplace.

I assume you meant 300 watts of electrical power, and not 300 watts of acoustic power.

Your project description resembles a small sonar transmitter.

 

[FvM]

nick inverter , sf peninsula

Hello,

an additional remark:
As the application implies pulsed operation with high peak power and low duty cycle, it seems meaningful not to use a transformer for the transducer drive voltage but a high voltage operated drive circuit together with a continuously operated DC/DC converter with a sufficient energy storage (a capacitor).

Furthermore, if bandwith isn't an issue, the driver must not necessarily operate from 500V DC (or 250 in case of a full bridge) but may also utilize impedance transformation in a CLC PI-circuit. Probably not from 12V DC, but maybe from 50 or 100V driver supply

If driver circuit bandwith counts, then only a high voltage direct drive without a transformer could give maximum performance.

Regards,
Frank

 

[kender]

200 khz air transducer

Guys, you seem well-versed in sonar design. Could you recommend and good introductory books that cover underwater ultrasound instrumentation design?

- Nick

 

[FvM]

40khz 300 vpp ultrasonic transducer driver

Hello,

as for me, I unfortunately can't. I have an 35 years old Valvo (Philips) application on piezo transducers and once did an utrasonic level meter project. Also, I'm working on projects connected to piezo transducers today. But not related to sonar at all.

Regards,
Frank

 

[echo47]

hf power transformer

I haven't seen any books with much sonar electronics details. Most systems are specialized and company confidential, although some were built by academic institutions, so you may find them in research papers or engineering journals or web pages.

Here's a nice book describing the acoustic theory, but very little electronics:
"Principles of Underwater Sound" by Urick
https://www.amazon.com/Principles-Underwater-Sound-Robert-Urick/dp/0932146627/

I've been curious about this book, but haven't seen it yet. I wonder if it's useful, or just a broad introduction:
"Sonar for Practising Engineers" by Waite
https://www.amazon.com/Sonar-Practising-Engineers-D-Waite/dp/0471497509/

 

[kender]

airmar sonar circuit

echo, thanks for the tips on the books! I'll look the Waite book up. I think I've read the 1st ed. of the Urick's book. He says explicitly in the beginning of the book that he's not goind to touch transducer design. (Come to think of it, there could be references to transducer books from that place.)

- Nick

 

[acinfo32]

design of hf switching transformers pdf

Dear All,

Thank you all for your reactions. Because I had no internet the last days hereby my reaction a little bit late.

My design is indeed for a sonar transducer. Of this transducer I know for sure that is works only on 200KHz. At this frequency it has a load of 1200 ohms and a power consumption of max 380W. I heard that these devices are piezo electronic but the consequences of this is unknown to me, e.g. if these devices always has an inductance or resistance (but anyway 1200ohms at 200KHz)?

At 500V and 1200ohms it gives I think P=U^2 / R = 500^2 / 1200 = 208W. This is indeed not 300W but I do not care at working at the exact max power. Or do I miss something?

How can I deal the best with the transducer damping? Is this a trail and error or is it always given by the manufacture of the transducer.

I have contacted Coilcraft and give them the specifications but they said that they didn’t have any products for that (hopefully they didn’t misunderstand my specs as I gave to them because they have indeed a lot of nice products).

What I also don’t understand is that the spec says that it has a power of 380W but seen the coil size, of a sonar which uses these transducers, it’s illogic to me, see the attached picture of the coil. Maybe it’s 380W acoustic power and is the electric power consumption much less. But that is not what I understood from the specs. What I now is that it is driven with only a small burst.

With kind regards

 

[echo47]

hf hv diode

Hello,

P=U^2/R ------- That's correct if U is volts RMS, but not if U is volts peak-to-peak (Vpp).

Here's how I calculated the power from your parameters:
You mentioned 500 Vpp and 1200 ohms. I assumed the waveform was a sinewave.
A sinewave of 500 volts peak-to-peak = 250 volts peak = 177 volts RMS.
Power is volts RMS squared divided by resistance: 177^2 / 1200 = 26 watts.

That's electrical power. Acoustic power will be less depending on transducer efficiency. Acoustic output power can't be greater than electrical input power because that would violate physics.

The voltage required to pump 380 watts into a 1200 ohm load is sqrt(P*R) = sqrt(380*1200) = 675 Vrms = 955 Vp = 1910 Vpp (sinewave).

In what way would you like to control damping? If it's a high-Q (narrow bandwidth) transducer, then it's going to ring like a bell. That can be a good thing or a bad thing, depending on your application. If it's low-Q, then you can transmit a wider bandwidth signal to carry more information or sharpen the range resolution or whatever. Sometimes you can buy a transducer that fits your need, but many customers need to specify a unique custom transducer, and then they build specialized electronics to drive it.

Can you trace out the schematic to see what that 12mm diameter coil does in the circuit? The coil appears to have an air core (no ferrite), so maybe it's a tuning inductor. The circuit may be some sort of resonator that rings a high-Q transducer at its resonant frequency. That would be a simpler system than the power amplifier output transformer that I described earlier. The 12mm coil is a little smaller than the RM6 or RM8 core that I mentioned. An RM6 is about 15mm square.

Possibly helpful or interesting, here's a company that sells a variety of rather expensive transducers:
http://www.itc-transducers.com/

 

[FvM]

why are transformers needed

Hello,

I think, it could be misleading to calculate transducer peak power from given voltage and impedance values, as long as you don't know, how the manufacturer determined these parameters. A reason could be, that a piezo transducer is often operated with a series or parallel inductance in a resonant circuit. In series tuning, the 500 Vpp could be the driver voltage and 1200 ohms the effective series resistance formed by acoustic radiation. The transducer voltage would be much higher than 500 V in this case.

However, without knowing the full transducer specification, I'm only guessing. Obviously, it's a big difference between 26 and 380 W peak power, and I would try to clear this point before defining the hardware in detail. The other point is, if the transducer should be operated in an resonant circuit, it would be meaningful to make the transformer part of this circuit.

When the 380 W peak power value turns out correct, this would require a somewhat bigger core, although wire gauge could be designed for the lower mean rms current. The most unpleasant thing with a 380 W design is the high primary peak current for the low voltage driver to my opinion.

Thus I would like to emphasize my previous suggestion of using a DC/DC converter with a secondary energy storage and a high voltage driver. Do you already know burst duration and duty cycle for the circuit?

I append a piezo transducers equivalent circuit from the said old Valvo/Philips application. You will notice that it's, execpt for the transmitted energy resistance, similar to a crystal, allowing series and parallel resonant operation. Assuming linearity, the parameters could be measured also at low voltage with a dived transducer (the tank should have a sufficiant extension to limit disturbance due to reflections).

Regards,
Frank

P.S.: An additionally question. Is maximum peak power an application essential or is the 380 W just a value from the transducer specification? In this case, it could be meaningful to analyze which drive voltage respectively power is actually needed in terms of signal-to-noise ratio and what's usual with similar existing designs.

 

[acinfo32]

transformer frequency definemax power

Probably I'm a little bit misleading by the 380W max power. This is according the specs the voltage was not given.

I have traced the original sonar and it is just as I thought and as I had seen in another project. The transformer in the picture above is not only an inductor. This transformer is driven by a MOSFET, shown beside it. Just like in the HR of a television. The output of the transformer is almost directly fed to the transducer. With a 470pF / 3KV capacitor parallel with the transducer and transformer output. All three with one side connected to GND.

Seen the size of the coil and MOSFET, I think the input power is much less than the 380W, but for now I don't know which driving voltage should be enough. The duty cycle is probably around 10%.

The reason why I didn't look at a DC/DC converter is because it is probably hard to make one with 12 to many hundreds of volts. I also think that it consumes more power and the devices work on relatively small 12V batteries and seen the transformer above, probably overkill. But maybe I see it wrong, I'm not a expert at DC/DC converters.

I will try to make a transformer with the RM6 or RM8 core and do some test. The physical size it not a big problem. @echo47 can you tell me how you calculated the windings (probably I need some 1000V output to test but I will try to get more detail about this voltage)? Is there also a program for these calculations with these cores?

Maybe I can get some more specifications. Also I will keep in mind the DC/DC solution. Although I have only seen the use of transformers, this will not say for me that it is the only solution.

With kind regards,

 

[FvM]

tuning sonar transducer to look resistive

Hello,

if the design can work with one MOSFET at low voltage and a small transformer, than the DC/DC and high voltage driver solution ist not necessary, it shouldn't consume more power but would be more complex anyway.

As the transformer in the reference circuit is apparently operated in a resonant circuit, ist isn't sufficiant to know the primary and secondary winding count, also the core dimension and properties should be known, or better, the transformers ectrical parameters.

The other option would be an empirical optimization (the designer of the original circuit most likely did it, or he had another reference circuit...). It would be helpful to know the primary and secondary waveform in operation, also the drive waveform of the MOSFET.

Regards,
Frank

 

[echo47]

transformer cw specification mean

The original design does sound like some sort of resonant circuit. Those things can look like a big mystery until you measure the component values (including the transducer), and simulate its operation in something like spice.

I couldn't imagine building a suitable amplifier without knowing more transducer specs. Try sweeping it with an impedance analyzer. Try to somehow determine how much voltage you need to get your desired acoustic source level. Consider building a simple low-power version of the amplifier, and try driving the transformer/transducer combination at 1% power (so nothing blows up), to get a feel for its response. After you've convinced yourself that you understand everything that's happening, go for high power.

Sometimes I need a tightly-coupled transformer (like the one I described earlier). Sometimes I need a loosely-coupled transformer (so I can resonate the secondary). That's a big design decision influenced by the transducer characteristics. The air-core transformer in the original system suggests it is loosely-coupled.

One nasty problem you may encounter is the high peak voltage. You need careful wiring and transformer construction to avoid arcing. I've heard too many pulse amplifiers go "tick tick tick tick SNAP SNAP SNAP SNAP silence", followed by a puff of smoke. Sometimes you get flying FETs too. Wear safety glasses.

I wish I could share this awesome little transformer design program, but it's company proprietary. Most of the math (especially proximity effect) came from chapter 11 of the classic out-of-print book "Soft Ferrites" by Snelling:


You'll need a carefully designed PC board to reliably drive high switching currents into the transformer. If you've never built anything like that before, you'll be amazed how much unwanted voltage can appear across a few millimeters of PCB trace inductance.

 

[kender]

air cored transformer maths

Transducer equivalent circuit

Frank,

Are there a formulae that can give a rough estimate of these parameters from the transducer geometry, parameters of the piezo material, surrounding environment, etc? Could you give an idea for the values on the transducer equivalent circuit?

- Nick

 

[FvM]

piezoceramic 300w

Hello,

Are there a formulae that can give a rough estimate of these parameters from the transducer geometry

It should be considered, that the quoted equivalent circuit describes transducer behaviour near an operating frequency. Like most mechanical resonators, the transducer has several modes with respective resonance frequencies, they are not covered by the model. Furthermore, coupling to the enviromental media may also introduce additional resonant circuits.

Piezo transducers can be designed in a wide range from a few 10 kHz to several MHz operating frequency and power levels from mW to kW. Thus parameters change over several orders of magnitude. It's possible however, to define equivalent circuit parameters for a given transducer (with given enviromental coupling) either from transducer data or from a few basic measurements. This circuit parameters can be used for drive circuit simulation.

It's also possible, to introduce a strong damping with transducer design. Then the resonant equivalent circuit becomes unsuitable. It could be replaced e. g. by transmission line models that are coming nearer to the actual acoustical wave behaviour.

With sonar application, damping would reduce acoustical power output and is thus almost unwanted, but it is necessary to achieve a high timing resolution in time-of-flight flowmeter or material testing applications.

Regards,
Frank

 

[acinfo32]

design ultrasonic transformers mosfet

Dear all,

Finally I managed to find someone with an impedance analyzer (LCR meter). In the attachment you can find my measurements. The pulses at the transducer are 1000V top-top and 200kHz.

Can you give some ideas and formulas seen these specs. I have for example the questions:

"loosely-coupled transformer (so I can resonate the secondary)"

What do I need (which calculations) to get the secondary resonate? I think a transducer parallel at the secondary output of the (air-core) transformer should do the job. Maybe an extra capacitor to get it resonate?

I have a RM6 and RM8 air-core and cupper wire (0.14mm², 0.08mm², 0.05mm², 0.03mm².

Do you have some tips on how to avoid arcing?
Which calculations of ch11 of "soft ferrites" are important? I don't need a feedback of the output so phase shift and so on are not important for me.

Or maybe some of you have better ideas.

With kind regards

 

[acinfo32]

parameter hf transformer

Tank you for your reactions.

The analyzer was not able to measure more frequencies.

I didn't acquire the data of the original transformer because.

- I could not get it out of the circuit without a high risk of damaging the print.
- Measuring the coil in the circuit did not work.
- The person who measured it was a stranger for me who did it for free and I didn't want to use too much of his time.

And last but not least was that I thought:

1 – First calculate the inductance of the secondary winding to get the resonance frequency.
2 – Make the number of windings of the secondary part of the transformer with these specifications (know RM8 or RM6 air core etc).
3 – Calculate the number of windings for the primary part so that, at 200kHz and 12V, the voltage at secondary output gives 1000V tt.

But I don't know if this is a good way to think and which formals are important.

Furthermore I do not understand the reaction with the sitemaps.

With kind regards