Creating an oscillating waveform.

[Cincinnati10]

I want to create a waveform that resembles a human heartbeat. (At this point I don't need the little "ticks" before and after the beat.) My idea was to start with a square wave and "process it" so that on the rising edge of the pulse, my voltage would rise to a positive voltage and then quickly taper back to zero V. Then with the falling edge of the pulse, the voltage would fall to a negative voltage, quickly tapering back to the baseline of zero V.

As far as parameter values: I'm thinking the Voltage spikes would be +/- 5V on the low side to maybe +/- 12V on the high side.

Is there a simple method of producing this — with discrete components on a circuit board? …computer generated?

If it can be done with digital/analog components: Is there a simulation program that I can use to experiment with the circuit design so that I can see the output waveform result for various inputs and various circuit modifications?

 

[VEBill]

Don't miss his post of '4 Feb 08 22:09' where he mentioned +/- 5000 (!) volts and 50-100 watts (!). Also frequency between 40 kHz and 1.5 MHz (?).

Confusing.

 

[Cincinnati10]

I'm not sure what is confusing about my design parameters.

I want to create a positive voltage spike immediately followed by a negative voltage spike—Have a period of "rest", then repeat the cycle. I want relatively little power dissipated.

From time t=0 until the two voltage spikes are complete, I anticipate that being about 10% of the Period of the waveform. But I want to be able to vary the "rest" time to end up with a waveform frequency that is adjustable between 40KHz and 1.5 MHz.

 

[MacGyverS2000]

Considering the frequency and voltage requirements, I think we can safely assume this is NOT designed to simulate a human... or any other biological entity known to man.


Dan - Owner

http://www.Hi-TecDesigns.com

 

[ScottyUK]

The confusing thing is that you want a power of about 50 - 100W in a pulse which lasts, according to your spec, zero time ('immediately'), something which requires a Dirac delta function or something close to it). What you're asking for up to now requires a very high instantaneous energy level during the pulse followed by a quiescent period. Normally a pulse is specified in Joules rather than Watts, and the rise time, pulse duration, and fall time of the pulse are defined. For a 1.5MHz repetition frequency your are looking at a very fast rise and fall time and to achieve 5kV peak voltage the dv/dt is going to be remarkably fast. Start putting values to the standard parameters for a pulse and you'll get more specific responses.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[VEBill]

It's one thing to create an arbitrary waveform at 'signal' amplitudes as per the original post. There's hundreds of methods to accomplish that. You can choose any method you like. Not quite trivial, but no big deal either.

It's another thing entirely to produce these incredibly high voltages and moderately high power level and at such high frequencies while maintaining an arbitrary waveform. It seems that you have no idea what you're asking for.

If my boss gave me those initial requirements, and then added those voltage and power and frequency specs a month later, I'd simply go home for the rest of the day for a nap. I probably wouldn't come back to work until the following week.


This thread reminds me of the old story about the bidders' conference for a comlicated military system to be installed on a submarine. Just as everyone was getting up to leave, draft specs and SOW in hand, the Admiral suddenly remember one more requirement that he'd forgotten to mention.

"Oh yeah, sorry folks, one more thing. The entire system has to enter the ship through a 3-inch diameter pipe, four feet long, with a 90-degree elbow about halfway along. Is that going to be a problem?"

 

[Cincinnati10]

VE1BLL — From your UserName, I'm guessing that you served our nation at some point in your career. I am always thankful for people who sacrificed their time for the freedoms we all enjoy. I also appreciate your taking time to consider my design challenge and your time in responding.

Since I interpret your responses having a bit of bite to them, let me respond about the design parameters. In New Process Development, one often must make a best guess of requirements and go until we find a reason to vary the course. Sometimes it's just a small correction, other times it's a hard turn on the path. If you were leading men into battle and the conditions required a hard change to be successful, would you really tell your superiors you were going home for a nap and taking a week off? I somehow believe you possess much more character than to actually do this. However, I remember the frustrations of working in an engineering group when I too thought my clients didn't know spit from shineola — especially in how to accomplish what they were asking me to do.

It seems from other's responses that the best way to accomplish my waveform is with a lookup table and a micro, outputting a voltage signal through a DAC or video card, and then boosting the signal. Of course the biggest hurdle as you point out is the voltage spike and the frequency. My thought is that if I can product the shape of the waveform, I could amplify it or use a transformer (auto ignitions system coil or amplifier) to get the 5 kV.

I didn't intend on concealing details, just sharing what I discovered when it impacted the design spec. I also did not see value in writing two pages of requirements that I do know how to accomplish. Therefore I only asked about the portion on which I needed assistance. If anyone can direct me to a source that can lead me through the process to accomplish this, I would be grateful if you would share it with me.

 

[VEBill]

No, I work for a company that supports the military. But I'm fulltime on military projects for many years. It was the closest fit on the pick list. And YCMV (your country may vary).

The 'bite' was only intended to ensure that the severity of your new requirement was made clear to everyone. It seemed that some were still answering the original question and hadn't noticed the 5kV etc.

The problem with creating a precisely-shaped spike, and then running it through a step-up coil or transformer, is that the precise shape is likely to be affected.

Personally, I don't see any (cheap and efficient) way to separate the two parts of the circuit. Of course, if you're willing to throw money at it, then you simply make the waveshape and run it through a monsterously-expensive wide-band amplifier. Even 50-watt wide band amplifiers can be a two-man lift and cost a small fortune.

If you're a bit flexible on the shape of the spike, then the voltage and power become much easier. But I don't think it will be quite as simple as an ignition coil.

As someone already suggested, start at the load and work backwards. Making the arbitrary waveform isn't the hardest part of this project. Making it 5kV is.

 

[Cincinnati10]

AH! So you have had to endure our government bureaucrats. Your bite was rather mild for that scenario.

I do not have a lot of money to throw at this project. But it's a personal idea and I have to pursue it for a while.

At this stage, I need to keep energy consumption low to prevent process heat build-up. I believe using an alternating Voltage spike at low current and high voltage may accomplish this. I want to run as high a frequency as I can achieve but I don't know what the process may require - hence the 40KHz to 1.5MHz range. My thoughts are to vary the "rest time" between spikes to vary the frequency.

As far as the process, I won't have the hardware for several more weeks. At that point I will know the exact system load resistance.

I don't believe the spike shape is critical. Nor am I concerned about a perfectly clean signal. A bit of random noise riding on the waveform should not be harmful in the process, and it may even help.

I got the idea about the auto ignition coil when I realized that the 12V battery system is used to create 25-30 kV pulses at the spark plug. So I thought why not run a +/- 2 or 3 V signal through the same coil, generating the approximate +/- 5kV pulse.

Am I overlooking something obvious?

Would anyone have a different way of accomplishing this?

 

[VEBill]

If you run an ignition coil at a reduced voltage, then you're not likely to get 50-100 watts at the output terminal.

 

[jimkirk]

Is your 50 to 100 watts instantaneous peak or average power?

 

[OperaHouse]

The primary of an ignition coil is approximately 300V or 1:400 ratio. This is why CD ignition systems operate at about 400V. You might look at charging a cap and duping the charge with a FET instead of an SCR. Google some CD ignition systems for some ideas.

 

[VEBill]

"...primary of an ignition coil is approximately 300V..."

The primaries of traditional ignition coils are obviously a nominal 12 volts.

 

[OperaHouse]

The collapsing magnetic field creates 300V on the primary, that is boosted up to spark voltage by 1:400 ratio. Therefore, most 12V waveforms fed into an ignition coil will not produce a voltage capable of sparking. You know this and you time would be better spent explaining to the OP how to use an ignition coil for his purpose. Personally, I don't believe the OP has any chance of success no matter how long this thread goes on.

 

[VEBill]

I'd forgotten that an ignition coil (even a traditional one) only sparks on field collapse. I appreciate the correction.

 

[OperaHouse]

25 years ago when I was last playing with this stuff, all ignition coils were really 6V with internal or external resistors to drop the voltage and kep inductance as low as possible.