Is Feedback Loop possible using Stepup Voltage Transformers in place of Voltage Amplifiers?

Discussion in 'General Electronics Chat' started by Matjelo Naleli, Nov 23, 2015.

  1. Matjelo Naleli

    Thread Starter Member

    Oct 7, 2015
    39
    0
    Recently I have been thinking of a possible way of implementing a feedback control loop without using active components. It's quite common to use amplifiers in feedback loop perhaps for implementing summing circuits, forward gain or even compensation networks.

    Imagine a simple case whereby a signal (voltage) is fed into an amplifier and the output of this amplifier is fed back and added/subtracted from the input signal at the summing point to complete the loop. One can imagine that if the output voltage is less than the voltage at the summing point then current won't flow back from output back to the summing point to fascillitate feedback. So some form of amplification or buffering would be necessary and we can use amplifiers to get around this.

    But what if we use a transformer to step up the voltage and then feed back this higher voltage back to the summing network at the input? Is it a viable way of achieving feedback or am I overlooking certain barriers with regard to the dynamics of the transformer itself maybe?

    If this is viable, regardless of how complicated the design might be, I wish to invest a bit on studying models of transformers and see if I can demonstrate simple feedback loop with them.

    Thanks in advance for your thoughts on this.
     
  2. crutschow

    Expert

    Mar 14, 2008
    13,009
    3,233
    You can adjust the input resistors of standard op amp inverting amplifier to get whatever sensitivity you need for the feedback at the summing junction, so you generally don't need additional amplification.

    And an op amp is much cheaper than a transformer if you do need voltage gain.

    Finally, transformers are AC only, so won't work if you need DC or very low frequency feedback.

    So you can study this if you like, as an academic exercise, but I see few, if any, practical applications.
     
  3. Matjelo Naleli

    Thread Starter Member

    Oct 7, 2015
    39
    0
    Thanks crutschow for your response.
    Yeah sure, the summer can be made to also offer amplification so I agree with you.

    One application which I was thinking of is in ion trapping (with voltages around 500V and more) and I am not sure if there are op amps that can work in those regimes. But if there are then ofcourse I wouldn't have even thought of attempting feedback using passive components only, except for mere curiousity. I guess there are transistors that have Vce in the range of kilovolts but I am not sure if their base-emitter voltage range can go beyond 100V.
    In any case I would suppose it's more expensive to use active components since they require DC power supplies with high stable voltage supply in my case ofcourse.

    As for the frequency, I will be operating around 13.5MHz with no need for DC offsets so an AC only device won't be a problem.

    So all that I wanted to hear is whether a transformer sounds like a viable way of implementing feedback. From your comment you don't sound like you totally find it as an exercise in futility so perhaps there is hope.

    Thanks again for your comments
     
  4. crutschow

    Expert

    Mar 14, 2008
    13,009
    3,233
    What is the impedance of this 500V source?
    Can't you just use a resistive voltage divider to reduce the voltage to something that can be handled by an op amp?

    Edit: Getting a transformer to operate at 13.5MHz is not a trivial task.
    I avoid transformers as much as possible, due to their cost and generally non-ideal behavior.
     
  5. Matjelo Naleli

    Thread Starter Member

    Oct 7, 2015
    39
    0
    Ok let me explain our setup a bit.
    We have a signal generator (50 Ohms) which drives a home-made class E amplifier. The amplified voltage gets filtered (and further amplified) by the resonator. This amplifier couples inductively (with a transformer) to the function generator at its input as well as to the resonator at its output and this achieves impedence matching as well. The resonator output then feeds the ion trap (load) which is essentially a capacitor (15pF) with some parasitic resistivity perhaps.

    So what is required in our expperiments is that the voltage across the load be on the order of 500V peak to peak or more and be narrowband (Quality factor greater or equal to 1000). I guess we could cascade amplifiers to get the voltage we want and feed the output straight to the load but then it won't be narrowband and hence why a resonator is still required as a filter. The resonator also offers an amplification at resonance which can help reach those high voltages without much voltage demand from the amplifier stage. Infact in the Ion Trapping community some people don't use the amplifer stage but rather drive the resonator (hellical resonator mostly) directly by the RF source.

    What I am focusing on is the implementation of the resonator itself. Previously we connected our ion trap (a capacitor if you like) to an inductor in series resulting in an RLC resonator which gived resonance at a desired/operating frequency. However the quality factor of this RLC connection was terribly low (around 40) on its own. However it increased to about 130 when I included the inductive coupling (for impedence matching) to the amplifier.

    A hellical resonator is very common in ion trapping community, infact every thesis that I have come across has a chapter or atleast a section on hellical resonator. I am in the process of building one for the lab but I thought it would be interesting to try other alternatives and make an evaluation study of those at the end. So if my ideas sound a bit crazy, that's because I am desperate to find an alternative that I can compare with the hellical resonator.

    There are two approaches that I was thinking of when I first attempted an alternative resonator. One was to cascade multiple resonators all with the same resonance frequency. In this way the overall transfer function (voltage to voltage) becomes the product of the individual transfer functions of each resonator and therefore the quality factor becomes the product of individual quality factors. I was hoping to do this using only passive components (RLC circuits) because it seems a cheaper way to go but I met a problem with impedance loading which required buffering between adjacent resonators. I used ladder networks but I am open to any suggestion on other topologies that might give narrow band filtering.

    Using opamps it becomes easy to improve the quality factor but the problem is how to get output voltage in the order of 500Vpp without clipping with op amps.

    The second approach that I thought of was using feedback control. In this way I can push the poles of my resonator towards the marginal stability inorder to increase the quality factor by simply varying the forward voltage gain of the open loop. It's quite easy to do this with opamps, only if they could give output voltage that high. I got seduced by IGBT transistors which seemed to have pretty high Vce voltage range however the Vbe voltage range is not that high. But I am still trying to figure how I can realize feedback with them perhaps by attenuating the voltage that is fed back to the base. I am not so well-informed with small signal analysis of transistors.

    It was during these struggles when I asked myself if there is a feasible way of implementing feedback with just passive components. Noticing that gain variation would be required, the transformer was the first and perhaps the only thing that I could think of. But I am getting scared now if you say it's not trivial to model it at those frequencies. I was hoping to pick a small transformer (from radios or chargers) or make a simple one, treat it as a black box and get its frequency response from which I can attempt to make a simplified model of it around that frequency so that I can include the dynamics of the model when designing feedback.

    Thanks again crutschow for your critiques & suggestions.
     
  6. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,797
    1,103
    Rather than a transformer, could you use a capacitive voltage-divider for the feedback? Suitable caps able to handle 500V should be easily obtainable.
     
  7. Matjelo Naleli

    Thread Starter Member

    Oct 7, 2015
    39
    0
    I haven't thought of that and I don't know how it will work out. But if you think it can provide feedback then that's great. I will just need to understand how it works.

    Thanks Alec_t
     
  8. alfacliff

    Well-Known Member

    Dec 13, 2013
    2,449
    428
    tubes or high voltage mosfets are a possability. I agree with reducing the voltage with a restive devider, it would work on both ac and dc, and should have no frequrncy response problems. special voltage sensing relays are used on the power lines to operate motors in regulating transformers, but even they use either transformers or voltage deviders to get the voltage down to a useable level for feedback.
     
  9. Matjelo Naleli

    Thread Starter Member

    Oct 7, 2015
    39
    0
    I haven't looked at voltage ratings of high voltage mosfets but it would be a good alternative too. Tubes, I am not familiar with them at all but I will check them out too, thanks..
     
  10. DickCappels

    Moderator

    Aug 21, 2008
    2,653
    632
    Tubes are like high voltage N-chanel JFETs.

    Sanyo used to make high voltage NPN bipolar transistors such as the 2SC4637 (2KV/50ma) and maybe they are still around.

    Also, see the High Voltage Stack in this Wikipedia article:
    https://en.wikipedia.org/wiki/Cascode
     
  11. Matjelo Naleli

    Thread Starter Member

    Oct 7, 2015
    39
    0
    I found the high voltage stack quite interesting. I will have a look at the tubes and 2SC4637 as well.

    Thanks DickCappels
     
  12. MrAl

    Well-Known Member

    Jun 17, 2014
    2,433
    490

    Hi,

    It is not only possible but i had posted this exact circuit on the web either in this forum or on the other one many of us here visit too. It's a little tricky but you just have to figure out the right arrangement of transformers in order to get a difference as well as an inversion and gain.
    I dont remember the exact circuit offhand but it goes something like this...
    If the two outputs of two transformers are arranged such that they subtract, you have the differencer. Just add a third transformer for the gain (or maybe you can work that into the original two transformers somehow) and you should have the AC equivalent of an op amp.
    As others have noted, there will be some limitations, but then again there are some advantages too like higher voltage if needed using the same basic setup as for lower voltages.

    See if you can figure out the exact arrangement for your needs given the above information, then maybe connect two resistors to form a regular inverting amplifier.
     
  13. Matjelo Naleli

    Thread Starter Member

    Oct 7, 2015
    39
    0
    Thanks MrAI, I am busy trying to figure it out and will report back if on my findings..cheers
     
Loading...