Ultrasonic Analyzer

[Nigel Goodwin]

No wonder I've had a hard tiem finding ultrasonic transformers. I always assumed that audio transformers cut off at a bit above 20kHz. Most RF ones that can step up any amount of voltage start around 100kHz, or have really bad attenuation at 50kHz. Turns out that audio transformers easily go up to the 100kHz range...I guess that's just a result of physics. Neat.

I wouldn't expect audio transformers to go that high, they usually drop off before 20KHz, and are a major limitation in audio systems that use them. For ultrasonic transformers I would imagine you have to wind your own?.

 

[dknguyen]

A few do. Little ones. I think their power is a little too low though (50mW or so), and I need a tapped secondary winding and they all seem to be step-down rather than step-up with a center tap on the primary winding.

I'm not sure how to calculate the power draw either. I've just been treating it as a capacitor with a sinusoidal input of a certain voltage. With a capacitance of 500pF (at 1kHz it says, so I think I just use that for all frequencies?). With a 200Vp signal, I get 37.7j mA. I've heard of electrostatic transducers pulling a momentary 3A though, so I must be missing something. 8W...I haven't seen an ultrasonic frequency transformer that can handle that much yet, but it's momentary so it might work...but I think saturation of the core happens instaneously and is independent of duty cycle.

 

[hjames]

I'm thinking along the lines of a backlight inverter transformer - those are well above 20KHz, but they might have too many parasitics.

The fun way might be to see if it is possible to directly modulate a 400V DC source - I've occasionally come across mention of people running transistors in avalanche mode - usually really extreme pulse generators, or projects needing very sensitive photodiodes (APD's). (and no, zeners aren't that interesting...) Zetex has a pile of app notes on that sort of stuff.

 

[dknguyen]

That's pretty crazy. Isn't it even harder to bias something at 200VAC with 200VDC if you use solid-state rather than a tapped-transformer?

The best transformer I've found right now has perfect bandwidth and step-up...it's just 40-50mW power, and the I don't know how far I can push it since sonar is a pretty low duty cycle application before it is destroyed. Senscomp has some vague equally sized transformers that seem to be able to handle it.

As for the 200VDC biasing, all I need is a really really really small 5/12V-200V DC-DC converter right? It's just a bias to charge up the capacitive plates on the transducer so current draw isn't really an issue (except maybe charge up time, but that only happens once every power-up).

 

[hjames]

As long as you have a +-200V signal, you can play the same capacitive coupling trick as the polaroid schematic - the diode clamps the negative end to ground, and the positive peak goes up to 400V. Bias and signal generator all in one.

But, ah, nevermind about the avalanche thing - it would be like trying to move a car by hitting it with a hypervelocity pea-shooter. The avalanche stuff only does nanosecond wide pulses, and you'd probably tee off everyone in the RF spectrum if you tried using it like a PWM.

I'd imagine that the worst thing that would happen is that you'd short out your transformer - I doubt you'd be able to punch a hole in the speaker membrane unless the coupling capacitor is ludicrously oversized.

 

[dknguyen]

The voltage is well within the ratings (well it's rateed at 1kHz, so I'm not sure if it's worse, same or better at 60kHz). It's the current that might push the transformer beyond power limits.

 

[Nigel Goodwin]

I'm not sure how to calculate the power draw either. I've just been treating it as a capacitor with a sinusoidal input of a certain voltage.

I suggest you try measuring it? - normal transducers are completely different for transmit and receive - the transmitter is series resonant, so is it's lowest impedance at resonance, and draws a fair amount of power. The receiver is parallel resonant, so is a high impedance at resonance.

A transducer that does both?, I've no idea!.

 

[hjames]

I suggest you try measuring it? - normal transducers are completely different for transmit and receive - the transmitter is series resonant, so is it's lowest impedance at resonance, and draws a fair amount of power. The receiver is parallel resonant, so is a high impedance at resonance.

A transducer that does both?, I've no idea!.

I'm pretty sure the operational theory behind these is just it being a parallel plate capacitor - voltage pulse -> electrostatic attraction -> sound, and the reverse transformation being vibration -> capacitance change -> voltage change, just like a electret microphone.

I vaguely remember seeing nothing except a textured plate behind the gold-colored film when I accidentally took one apart (and subsequently had to order the replacement).

 

[Nigel Goodwin]

Well looking at the datasheet it's not particularly resonant like normal ultrasonic transducers.

 

[Sig239]

So...I was looking at what some other people had mentioned earlier about filtering square waves. For some reason I crossed this off my list early on since I didn't think I could get a multi-frequency bandpass filter that was accurate enough...but then I remembered about switched-cap filters. And it also seems that I do not need a bandpass filter either, just a lowpass filter right? Since there are no subharmonics in a square wave signal. Actually, I don't think I even need a switched cap filter. If my sinusoid is only ever going to be 40kHz-60kHz, I could just get a lowpass filter that has sufficient attenuation above 120kHz, and any square wave sent into the filter would exit as a sinusoid with the fundamental of the input, right?

Cutoff should be at 80kHz.

 

[dknguyen]

I've never really given impedance matching much thought since I've never had to really use it yet (I don't mess around much with audio circuits). Mostly digital, motros, and DAQs, etc. I know the mathematical proof, and I clearly see why you don't want the source impedance much higher than the load impedance...but I see absolutely no problem if the load impedance is much higher than the source impedance? THis is what I do with buffer circuits, and DAQ circuits. Why would you want this impedance matching stuff? Do I even really need it my transducer impedance is much higher than my driver impedance?

 

[Nigel Goodwin]

It depends what you're trying to do - matching impedances transfers maximum POWER (50% is the best you can do), not-matching impedances transfers maximum VOLTAGE - which is usually what you want. Mostly the impedance matching thing is to do with passive circuits, for active circuits it's not required (and usually a disadvantage).

If you're using a transformer, it's usual to match that - a transformer is a passive device, so works best matched.

 

[dknguyen]

It's just that I would like to drive the transformer with something around 5V and if I try to match the impedance it's going to requrie a much lower voltage so as not to exceed the transducer voltage. Two variables, one degree of reedom.

When impedance matching you would never add source resistance just to match the impedances right? That just introduces losses, but then satisfies the max power transfer (but is contradictory it seems since it introduces losses). That's what I'm a bit confused about. Wouldn't lack of a source resistance increase the voltage across the load as well as the current, therefore allowing more power to the load? THe thereom says otherwise though. I'm probably overthinking this.

 

[hjames]

I've never really given impedance matching much thought since I've never had to really use it yet (I don't mess around much with audio circuits). Mostly digital, motros, and DAQs, etc. I know the mathematical proof, and I clearly see why you don't want the source impedance much higher than the load impedance...but I see absolutely no problem if the load impedance is much higher than the source impedance? THis is what I do with buffer circuits, and DAQ circuits. Why would you want this impedance matching stuff? Do I even really need it my transducer impedance is much higher than my driver impedance?

For one-way driving, I don't think there is anything "wrong" with it, just that if you match efficiencies, then you can up the power output of the circuit (at the expense of efficiency and accuracy). For circuits that need to receive as well as transmit there is a problem though - if the power supply isn't matched, then the signal coming back is going to be heavily attenuated by the power supply.

In this case, that isn't that big of a deal *if* you can switch out the power supply. I'd bet (a little anyway) that the receive connection on the 6500 module is diode clamped - i.e. a low impedance, when generating the ultrasonic pulse, and a high impedance when receiving (courtesy of the coupling cap).

 

[hjames]

When impedance matching you would never add source resistance just to match the impedances right? That just introduces losses, but then satisfies the max power transfer (but is contradictory it seems since it introduces losses). That's what I'm a bit confused about. Wouldn't lack of a source resistance increase the voltage across the load as well as the current, therefore allowing more power to the load? THe thereom says otherwise though. I'm probably overthinking this.

High speed digital uses resistors heavily in order to impedance match. The worse the impedance mismatch, the worse the reflected power. Enough reflected power and you run the risk of getting new, or extended transitions when waves get reflected back from an unterminated or mismatched connection. In a perfectly matched setup, the resistors end up dissipating the power instead of shooting the power back down the wire, possibly causing grief somewhere else. (sometimes in the form of RF interference as well).

Obviously if you don't want to dissipate power, use non-resistive (i.e. L,C's) to impedance match. The drawback is that this ends up making the circuit frequency dependent. Transformers are special in that they let you scale AC impedances with relatively low losses - but they obviously don't work at DC.

The impedance matching is an engineering tradeoff - if you want to fully utilize your source's peak power(not efficiency), then you want to put a matched load on it. If you aren't matching impedances, then your source won't be fully utilized (but it will be more efficient), hence you can design a 'weaker' source and save some resources for somewhere else.

 

[dknguyen]

Well I know impedance matching is needed for transmission lines. Let's just talk about impedance matching for max power drive assuming there are no transmission lines.

I understand most of what you say. The only problem is then I think about the extra source impedance introducing a voltage drop and reducing current flow (which in my mind would reduce power to the load, but apparently not?). Or is this only true if a resistor is used? (But then it doesn't make sense since the thereom also holds for resistive loads.)

 

[Hero999]

I wouldn't have thought that impedance matching would be an issue at such low frequencies - the wave length of 60kHz is over two miles long!

 

[Nigel Goodwin]

I understand most of what you say. The only problem is then I think about the extra source impedance introducing a voltage drop and reducing current flow (which in my mind would reduce power to the load, but apparently not?). Or is this only true if a resistor is used? (But then it doesn't make sense since the thereom also holds for resistive loads.)

Try working it out, the maximum power transfer is when the impedances match - and (like I said earlier) is only 50% - half in the load, and half in the source.

 

[dknguyen]

I wouldn't have thought that impedance matching would be an issue at such low frequencies - the wave length of 60kHz is over two miles long!

It'd be a different story if it was digital . But it's not. It's good old sinewaves. So yeah. It's not a transmission line issue though. It's a power transfer one.

 

[hjames]

I only brought up the impedance matching because there are cases where you absolutely want power to be dissipated into a resistive dump. It was a nit, hence I picked it... For this situation, matching impedances with a real resistance is going to be useless unless there's some oddball ringing/parasitic resonances/just bad stuff going on.

If you look at impedance matching from the perspective of constant power, it won't make too much sense, nor does it make sense when you say that you have a power supply with zero impedance - that power supply is an infinite source of power, at any voltage... The standard assumption is that you have a power supply with some fixed voltage, and some fixed impedance. You are then tasked with trying to get as much power out of that box as possible - and you are not allowed to modify the box. Hence you have a load that matches the impedance of the power supply, and you get max power out of the box, albeit at mediocre efficiency.

Also, impedance matching just says where the max power out of the box is - it doesn't tell you where *you* want the numbers to be. If efficiency is a concern, then the you'll want to minimize the impedance of the source, but keep in mind the scales - if the transducer has some resistance R, and a 500pF parasitic capacitance, then the total impedance of the transducer is R + 1/(j(60E3*6.28*500E-12)) = R + (1/j(1.9E-4)) = R -5000jW, where R is probably some relatively small value. Handwaving the imaginary stuff aside, we've got a 5K load that we want to push some power through. If your source had an impedance of 500R, it would only be 90% as efficient as an ideal source, but it might actually be buildable.

Alternatively you can try canceling out the imaginary component with an inductor, which ends up looking like a resonant tank circuit, but that might not satisfy another of your engineering criteria.

As for the actual problem - drive the transformer with a lower voltage (using a switching regulator to efficiently transform the DC). Or use a capacitive voltage divider at the output, if you're dead-set on that particular transformer. Possibly modify the turns ratio if you're adventurous.