Amplifier For 600 kHz Signal

Thread Starter

yokel

Joined Apr 29, 2010
15
Hello,
I am a newbie in electronics. I need to design an AMPLIFIER to amplify

> A 600~650 kHz PWM signal
> The input PWM signal can be taken from a function generator
> The output should be 3V
> The load is equivalent to a 13nF capacitor

If you need any more info please let me know.

It would be very nice, if someone give me some instruction, sample circuit or refer me some online resources which can be helpful for me in this regard. Detail answer will be appreciated.

thanks a lot in advance
 

Thread Starter

yokel

Joined Apr 29, 2010
15
Hello,
Let me add some more information.

> The PWM duty cycle doesn't vary.
> The capacitive load I am talking about is actually a Piezoactuator. Its equivalent capacitance is 13nF and also a little resistance will be there.
> The output 600 kHz signal quality must be as good as possible and shouldn't be distorted due to capacitive load as we have to drive the actuator very precisely.
> I will use a typical function generator for initial input and the output voltage should be 3V.

Please help me to design the circuit and suggest me some online resources too for studying.

kind regards
 

ifixit

Joined Nov 20, 2008
652
Something like a 'video amplifier' might work for this application. Try google.

  1. What are the rise and fall times required of the signal?
  2. What is the function generators output impedance?
Regards,
Ifixit
 

Thread Starter

yokel

Joined Apr 29, 2010
15
Hi Ifixit,
Thanks for the response.

1. Sorry, I have no idea about the rise or fall time, please consider the usual value. But as I told before the output should be as good as possible.
2. Function generators output impedance is 50 ohm (usual)

Something like a 'video amplifier' might work for this application
Could you please tell me in a more detailed way? Actually I am totally new in this field. The best thing would be if you can kindly give me a sample circuit diagram or some useful online resource.

kind Regards.
 

Ron H

Joined Apr 14, 2005
7,063
You could try something like this push-pull emitter follower. You will have to adjust the function generator signal amplitude slightly while measuring the output voltage. If you want something more predictable, it will require a more complicated circuit.
Keep in mind this is a simulation. The actual circuit may perform differently if the component models (e.g., the piezoactuator) don't reflect reality.
 

Attachments

Thread Starter

yokel

Joined Apr 29, 2010
15
Dear Ron H,
Thanks a lot. It seems really simple and effective. I am just wondering about some issues. Please help me in this regard.

1. Very roughly, 600kHz 3V square wave into 13nF with a rise/fall time 1/40th the period (a moderately sharp square) = 42ns, so the current is I = C*3V/42ns = 1.0 A
Is the circuit you suggested able to offer around 1.5 A peak?

2. In case of the high capacitive load of this circuit, is there any chance of signal distortion while running the piezo actuator?

To others, it would be nice if anyone else can offer some more info or sample circuit?

thanks again :)
 

kingdano

Joined Apr 14, 2010
377
i just did a similar design for a high speed (not as high as yours - 50 kHz) high slew rate (200 V/us) piezo drive circuit (square pulses)

far and away your best bet is to start with Apex - Cirrus Logic.


http://apex.cirrus.com/en/products/apex/index.html

they have a product selector and FANTASTIC applications notes that literally walk you through everything you need to think about and calculate.

they have a spreadsheet which can be used to select a part as well - BUT BE WEARY because not all new models are included and you may miss the best fit by using only the spreadsheet.

their datasheets and app notes will literally guide you through the thick and thin of it all.

here is a link to their PWM amplifier section as well

http://apex.cirrus.com/en/products/pro/areas/PA139.html#PA142_open

also, please note that they do offer development kits (which i used) which are great for prototype work and come with a pre-spun PCB and some basic components as well.

good luck!


Dan


edit:

App Notes Link:

http://apex.cirrus.com/en/products/apex/documents.html


Suggested Readings

AN01 - this is a must read
AN30->35 - PWM specific
AN44 - Piezoelectric specific
 
Last edited:

kingdano

Joined Apr 14, 2010
377
You could try something like this push-pull emitter follower. You will have to adjust the function generator signal amplitude slightly while measuring the output voltage. If you want something more predictable, it will require a more complicated circuit.
Keep in mind this is a simulation. The actual circuit may perform differently if the component models (e.g., the piezoactuator) don't reflect reality.

typically you model a piezo as pure capacitance and in calculations you treat it as pure capacitance.

it is good that you added the small resistor in series with the load - as this greatly aids in stability.

generally a 1-5Ω higher wattage resistor is used.
 

kingdano

Joined Apr 14, 2010
377
Hi Ifixit,
Thanks for the response.

1. Sorry, I have no idea about the rise or fall time, please consider the usual value. But as I told before the output should be as good as possible.
2. Function generators output impedance is 50 ohm (usual)

regarding number 1:

there is no typical value or usual value for rise time - this is an application specific value based on the total voltage swing, and amount of time you have to get from low to high voltage - this should be defined by the load - specifically the load manufacturer who i assume has defined the drive specifications.

regarding number 2:

make sure you add a 49.9Ω (common value) termination resistor on your amplifier input circuit to ensure no losses due to reflections.
 

Ron H

Joined Apr 14, 2005
7,063
Dear Ron H,
Thanks a lot. It seems really simple and effective. I am just wondering about some issues. Please help me in this regard.

1. Very roughly, 600kHz 3V square wave into 13nF with a rise/fall time 1/40th the period (a moderately sharp square) = 42ns, so the current is I = C*3V/42ns = 1.0 A
Is the circuit you suggested able to offer around 1.5 A peak?

2. In case of the high capacitive load of this circuit, is there any chance of signal distortion while running the piezo actuator?

To others, it would be nice if anyone else can offer some more info or sample circuit?

thanks again :)
Peak current capability depends on the transistors you use. As I mentioned previously, simulation results may not reflect reality when it comes to details like this.
You may have to try several different types of transistors. PN2222/PN2907 might work better. The problem with going to transistors that have high beta at high current is that their Ft tends to be low, which may distort the waveform.
 

ifixit

Joined Nov 20, 2008
652
From what I read, it seems to take more than 3 Volts to actuate a piezo-actuator... more like 100V.

What is the actuators part number? What is the application?

Regards,
Ifixit
 

Thread Starter

yokel

Joined Apr 29, 2010
15
Dear Ron H,
Thanks again for your reply. I am going to find a suitable transistor with high current specifications.

The problem with going to transistors that have high beta at high current is that their Ft tends to be low, which may distort the waveform.
Sorry, I couldn't understand this point fully. Anyways, Could you please suggest me any way to get rid of this sort of distortion?


One of my friends suggested me to go with MOSFET switching where I also need to include the MOSFET gate driver. Do you think that will be a better choice instead of this simple BJT based solution?

kind regards
 

Ron H

Joined Apr 14, 2005
7,063
Dear Ron H,
Thanks again for your reply. I am going to find a suitable transistor with high current specifications.


Sorry, I couldn't understand this point fully. Anyways, Could you please suggest me any way to get rid of this sort of distortion?


One of my friends suggested me to go with MOSFET switching where I also need to include the MOSFET gate driver. Do you think that will be a better choice instead of this simple BJT based solution?

kind regards
Vishay makes a nice dual MOSFET (one NMOS and one PMOS) in a single surface mount package. The part number is Si1555DL. I've attached a simulation schematic and results, along with the .ASC file for anyone who wants to run the sim in LTspice.
It looks really good - much better than the BJT solution, and probably more believable.
 

Attachments

Thread Starter

yokel

Joined Apr 29, 2010
15
Dear Ron H,
I have gone through the datasheet of Si1555DL. But don't you think that, the output current specification of this part is not enough for my purpose? As I said before, I want to assure atleast 1.5A current-output. I might be wrong. Please correct me if I am.

I have two more questions.
>> Do I need to use a MOSFET gate driver for running this Si1555DL with
a typical function generator output?

>> How will be the output distortion due to this high capacitive load?


BTW, I am really grateful to you, I am learning a lot.

Kind Regards,
 

Ron H

Joined Apr 14, 2005
7,063
I'm pretty sure we can find MOSFETs that will work, but before I spend any more time on this, tell us what sort of piezo actuator requires only 3V to drive it?
 

kingdano

Joined Apr 14, 2010
377
i agree.

i have never seen a piezo that will actuate at that low a voltage.

generally it is anywhere from 50VDC up to 100's of VDC that will cause actuation.
 

Thread Starter

yokel

Joined Apr 29, 2010
15
It's a specially designed piezo actuator which will be used inside cell phone and you know 3/3.5 V is the highest voltage we can provide in a cellphone.

Thanks for all your cooperation. It would be nice if you kindly help me to resolve the previously mentioned questions.
 

Ron H

Joined Apr 14, 2005
7,063
OK, check this out.
The MOSFETs have an extra pin (connected to 25V). You can ignore it. The Pspice models I used have temperature effects included, represented as a voltage.
Also, ignore the inductor. It is there to simulate the effectiveness of the decoupling caps on power supply parasitics, which will be present without any effort on your part.:)
I think your terminology is the source of our confusion. What you described sounds more like a transducer than an actuator.
Function generator risetimes don't have to equal the 10nS I used. Rgen is the generator's internal resistance. Rterm should be applied on the other end of the 50 ohm coax, on the board near the MOSFETs.
You could change the 100uF to as low as 10uF with no noticeable effect.
 

Attachments

Last edited:
Top