high voltage amplification

Discussion in 'The Projects Forum' started by nimaajbphs, Jun 1, 2011.

  1. nimaajbphs

    Thread Starter New Member

    Jun 1, 2011
    14
    0
    First off, Hi to all in the community! =)
    this is my first posting. I'm an undergrad physics student and I need to build a high voltage amplifier that is, I need to turn an input signal(sawtooth 0 to 10 V pk to pk) from a function generator into an output signal : 0 to 100 V peak to peak at maximum gain.I also need a gain adjustment knob. I should also mention that output current needs are minimal as in they are in the fractions of a microampere range.

    I'm sort of a noob in circuit development and I'm in desperate need of suggestions and advice! Currently I'm planning on achieving this by using an HV opamp such as http://www.cirrus.com/en/pubs/proDatasheet/PA240U_G.pdf but i'm not sure if that is the best choice. I'm planning on setting a basic non inverting configuration and again I'm not sure if thats the best design. I have attached a copy of the schematic. Please let me know if what I've drawn is correct. Any suggestions or advice is greatly appreciated..
    [​IMG]
     
  2. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    If your slope magnitudes are equal, your waveform is more properly called a triangle wave.
    What is the frequency range?
    What is the load? Are you sure it's not capacitive?
     
  3. nimaajbphs

    Thread Starter New Member

    Jun 1, 2011
    14
    0
    sorry i should have specified. As you said the signal is a triangle wave. the frequency range is from 10 Hz to 5 kHz and the load is a piezoelectric transducer with a capacitance of 0.09 micro farads (http://www.thorlabs.com/Thorcat/15600/15639-S01.pdf)

    Edit: I should have better thought this through, It seems to me that at 5 kHz and 100 volts I will be drawing 0.283 amperes.

    thx again
     
    Last edited: Jun 2, 2011
  4. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    The data sheet seems to imply that this chip requires a spit (+-) supply.
     
  5. nimaajbphs

    Thread Starter New Member

    Jun 1, 2011
    14
    0
    Does that mean that I need to supply -100V to 100V, Or can I supply -50V to 50V and then somehow add a DC offset to my resulting output?
     
  6. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    I think they spec'd the minimum supply voltage @ +-50V, so you want to be above this. A +-100V supply would give you lots of head room to play with but +-60V should be OK but +-70V would be better. I looked to see if the output of this chip is capable of a rail to rail swing but I didn't see that spec. I may have missed it though.

    I need to hit the sack.;)
     
  7. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    You need to be aware that a .09uF load can be a challenge for op amp stability, depending on the op amp. The usual solution is to use a low-value series resistor to isolate the cap from the output.
    Also, your transducer will not draw DC current, but the current required to charge and discharge the capacitance is I=C*dV/dt.
    In this case, at 5kHz, dV/dt will be ±100V/100usec=±1e6 v/sec, so current will be I=±9e-8*1e6=±90mA.
    90mA is not insignificant, and you (we) need to be sure your op amp can handle both the current and the resultant power dissipation.
     
  8. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Datasheet says common mode range is -Vs+12v to +Vs-14v; so you would need around -20v for the negative rail, and around 120 for the positive rail - but +70 and -70 would work, too.
    Output is rated 120mA peak, 60mA continuous absolute maximum - the latter is a problem; you need at least 90mA continuous, preferably more (you don't want to operate a component at maximum capacity, as it will have a much shorter life than expected.)

    I did a very abbreviated search through their parametric selection list (somewhat frustrating) and found this:
    http://www.cirrus.com/en/products/pa107.html
    Maximum overkill, and probably will cost a good bit more than you'd want to pay for such a thing - however it's something that might be viable. However, perhaps the less expensive opamp and some HV transistors wired as emitter followers would do the trick; they would need to be well heat-sinked of course.
     
    Last edited: Jun 2, 2011
  9. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    Have you considered driving the Piezo with a transformer? There are a number of 12V circuits out utilizing one.
     
  10. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    It might work, but trying to get 10:1 voltage increase into a .09uF cap, and maintain a triangle wave at 5kHz, seems problematic to me.
     
  11. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    nimaajbphs has not clarified the problem enough to explain whether the wave shape is important. For all we know, he could be a newbie that just wants to get a squeak out of a piezo and doesn't know the easiest way to do it.

    nimaajbphs, please point out the important aspects of what you need.
     
  12. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    I would like to see an app note for the part. I can't envision what it is for. It generates heat that is a function of frequency, but I don't know if this is a feature or a liability. There are other ways to generate frequency-dependent heat.
    It also has a displacement of 3 microns at 100V, which I suppose is useful for something, but it seems awfully small.
     
  13. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    I have a confession to make. As a general rule I've been subconsciously ignoring some aspects of an OP's requirements. At least..I think it's subconsciously! :rolleyes:

    The OP's first post indicated that he wanted to drive a Piezo. I saw no logical reason for the need of a triangle waveform, since the Piezo will reshape it into a sine wave anyway. ;)
     
  14. nimaajbphs

    Thread Starter New Member

    Jun 1, 2011
    14
    0
    thank you very much everyone for taking the time to reply it has been very useful. =)
    sorry I should have been more explicit on the application. I am currently working on a new diode laser system design, the output frequency of the laser is dependent on the length of the external cavity and here is where the piezo comes in: the piezo is attached to a semireflective mirror whose position determines the length of the external cavity. By feeding the triangle wave to the piezo I can scan across a very fine range of output frequencies. Specifically frequencies corresponding to electronic transition structures in rubidium atoms whose spectrum is being used for other applications such a frequency locking and stabilaztion etc.

    The shape of the input signal is important since any irregular deviations such as non linearity in the gain will translate directly a shift in the laser output light frequency, so I need to preserve the shape of the triangle wave in the output.


    It seems to me that high driving frequency is problematic for this circuit so after collaborations I've decided to reduce this frequency by a factor of 5 bringing it down to 1 kHz.
    At this point I'm beginning to think an op-amp is not the best choice for this application. Specially with the cost of over 200 just for the Op-amp.

    After some searching I found this : http://www.emcohighvoltage.com/pdfs/caseries.pdf
    I'm not sure if it could be used for my application. If anyone has had any experience using these your advice is much appreciated.

    I also found this : http://www.mmech.com/images/stories..._Converter/PPT04P2020XB1_MMech_Data_Sheet.pdf
    but they dont specify their operating frequency range, I will call them tomorrow to see if they meet my requirements.


    ------------------------------------------------
    I tried looking online for such a transformer but i couldn't find what you were refering to =(. It would be awesome if you could point me in the right direction
    -------------------------------------------------

    I'm still open to any suggestions you guys may have. Thanks.
     
  15. #12

    Expert

    Nov 30, 2010
    16,248
    6,745
    Now we're cooking. The more specific you get, the less we try to use universal answers that are not optimized to the application.
     
  16. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    Please post the particular Piezo you plan to use.

    EDIT: Ooops, I see that you posted it. I'm not used to this. :D
     
  17. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    OK, I read through your Piezo data. You can disregard my earlier statements stating that the Piezo will convert a triangle to a sine wave. Your Piezo is a totally different animal and is not the typical models that we see here often. In fact, it's polarized and can't see a reverse polarity without being damaged. This means that your output must never go negative. Transformers are definitely out.
     
  18. CDRIVE

    Senior Member

    Jul 1, 2008
    2,223
    99
    Did a bit more reading of the Piezo data sheet and I'm getting a bit confused. In one breath they warn about reverse polarity and in another breath they provide a graph of heat produced when a sine wave is sourced! Now I'm not sure if the reverse bias warning pertains to a steady state DC or not.
     
  19. nimaajbphs

    Thread Starter New Member

    Jun 1, 2011
    14
    0
    I would assume that the sine wave has an offset so that the lowest end of the peak is at zero volts.
     
  20. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    The piezo device datasheet indicates that there is an offset; the positive input (red lead) must never go negative with regards to the negative input.

    The devices you linked to in http://forum.allaboutcircuits.com/showpost.php?p=368199&postcount=14 are high voltage DC-DC converters with variable outputs. They are not designed to output triangle waves. They would have a rather long response time to the control input compared to an opamp. Besides that, the output current those converters would not be adequate for charging/discharging the Piezo beyond a few Hz.

    If you decreased your requirement to below 2kHz, you might be able to use the original opamp that you found; as the drive current requirement will be considerably less than at 5kHz. However, you will still need cooling for the opamp; heat sinking will be required.
     
Loading...