precision full wave rectifier

Discussion in 'General Electronics Chat' started by qlireslan, Nov 4, 2011.

  1. qlireslan

    Thread Starter New Member

    Nov 4, 2011
    5
    0
    ' like to make a Hi Speed Full wave rectifier circuit. Input frequency from 1Khz to 1MHZ sin wave. my circuit is Attached But this
    circuit can rectify up to 200 Khz and after that all kind of problems start to appear from oscillation to distortion as i go above KHz .i tried to change the diodes with transistor or schottky diode and for oscillation i tried capacitors wherever i can use but it really didn't help so much ...

    How do I improve my circuit? I wish It works well up to 1MHz or some where around that range

    Many thanks
    Regards

    Ali Reslan
     
  2. kkazem

    Active Member

    Jul 23, 2009
    160
    26
    I have 2 suggestions for you. First, since you didn't say what you application is, you might try using germanium diodes (like 1N270, I think) as they have the lowest drop of any semi and would should work well at 1MHz with about a 0.2V drop in the milliamp range, perhaps even the tens of milliamps range. However, if you truly need close to a zero drop on the precision rectifier, I would recommend that you add small feedback caps in the range of about 10pF to eliminate oscillation. Also, you have a 15K from the + input to gnd on your last stage, a differential amp and it probably should be 10K, otherwise, make the last stage (differential amplifier) from 3 separate op-amps in the standard fashion as you're asking too much from a single-stage diff-amp. Also, I don't know if you're talking about a real ckt or a simulation only. Even if you're only now just simulating it and havn't built-it, troubleshoot it by checking if the two rectifier stages oscillate and distort or not. All your problems may be in the final stage and could be eliminated by my suggestion of the 3-stage diff-amp instead of 1-stage diff-amp. Use the same LT318A. If you're talking about a real ckt, it may be layout related as you would need a very carefully thought-out and placed layout to keep out oscillations and distortion.

    That's all for now, feel free to reply if you have additional questions about what I've said.

    Good luck,
    Kamran Kazem
     
  3. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    The main problem is the bandwidth (and slew rate) of the amplifier limits you. I designed an AC rectifier circuit for my desktop DMM to extend the AC bandwidth. You can use it if you want. It uses a wideband full wave rectifier and peak detector. The voltages are negative because that's what the meter uses internally. The op amps were the fastest ones available back then, maybe faster ones now but they will be pricey.
     
    Last edited: Nov 4, 2011
  4. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    Here is the frequency response. It's good to well over 200 kHz (about 1.5% error at 500 kHz).
     
    Last edited: Nov 4, 2011
  5. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    That looks like a precision AC-DC converter. I don't think it will work as a true full wave rectifier.
     
  6. crutschow

    Expert

    Mar 14, 2008
    13,003
    3,232
    Post you LTspice .asc file and I'll try to simulate it.
     
  7. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    It's a precision rectifier followed by a peak detector so the sine wave peak value is detected. After that, there is a resistive divider so it will read the RMS value of the sine wave. It is not a true RMS detector, it is a precision sine wave rectifier. It's designed to detect the negative peak of the sine wave, the positive detecting version would have all the diodes flipped and the caps polarities reversed.

    BTW: all "rectifiers" of any type convert AC to DC, hence are "AC-DC converters".
     
    Last edited: Nov 4, 2011
  8. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    No offense, but trying to build wide band AC circuits from simulations is like trying to get a monkey to do brain surgery. Take a look at the compensation on my schematic used for peaking. That is how you tune for maximum bandwidth and flatness. You actually have to build circuits sometimes to get them to work.
     
  9. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Yeah, I know. It looks like a great circuit. If he wants to measure the amplitude of a steady-state sine wave, it will probably work for him, at least to a few hundred kHz. I'm just saying that he may actually need the output to be a FW rectified sine wave, with no filtering.
     
  10. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Here's one I've been playing with - not perfect, but not too bad either.
     
  11. crutschow

    Expert

    Mar 14, 2008
    13,003
    3,232
    Well, you're welcome to your opinion (which I disagree with). That simulation monkey can give quite accurate results but you have to have the correct models for the parts, including various parasitics. And those parasitics are not that difficult to estimate for a 1MHz circuit. I would never build any circuit without first simulating it. Of course they may require some tweaking after you build them but the simulation will usually get you well into the ball park.

    I don't see any compensation in your schematic that likely couldn't be simulated with fairly good accuracy.
     
  12. t_n_k

    AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    The sadly missed guru of electronics Jim Williams designed a 2.5MHz precision rectifier.
     
  13. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    I played with it too...
     
  14. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    Yeah, thanks. I spent the last 20 years at National Semiconductor so I know a ton about the accuracy of sims for AC performance and parasitics. Like I said: if you actually want to see how it works, build it.

    BTW: if you had ever been involved in the development of the sims that semiconductor makers come up with to release with their products (I have been on dozens) you would know how worthless they are for all but basic use. But they make people happy... and then I get to take the irate calls when they go to first gen board without testing anything and they have problems.
     
  15. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    I warned him that carpooling with Bob Pease wasn't a good idea........:eek:
     
  16. t_n_k

    AAC Fanatic!

    Mar 6, 2009
    5,448
    782
    Ouch! You're treading on thin ice there.

    I guess 'acceptable' humor varies from country to country.
     
  17. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    I worked with Bob for 20 years. He spent his whole life poking sacred cows. He would be laughing louder than anybody....

    Bob was a genuine kick. He's one of very few things I miss about working at that crap factory.
     
    Last edited: Nov 5, 2011
  18. qlireslan

    Thread Starter New Member

    Nov 4, 2011
    5
    0
    i wanted to thank everyone who shared their knowledge and their put time into this.i ended up building Ron H circuit.it worked fine, although i had few changes to the circuit due to the limitation of the components that I am allowed to use for this particular project.
    thanks a lot guys
     
  19. Adjuster

    Well-Known Member

    Dec 26, 2010
    2,147
    300
    Clearly, any simulation is not reality, and its results can never be regarded as infallible predictions of what will happen in practice. I cannot however agree with the conclusion that simulation is always worthless.

    Except that, if the likes of Bob Pease thought simulation to be a mere foolishness, should anyone dispute it?
     
    Last edited: Nov 9, 2011
  20. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    Look, I know how much people like sims but they have created a monster we will be living with for decades: abunch of engineers who can't think. A computer is only a viable tool if the user is smart enough to know when it's lying. And based upon what I read in posted questions, most of the of the current crop of engineers in training don't. That's a serious problem.

    There was another thread where the poster displayed a "computer sim schematic" for a linear regulator that would be dissipating 50W and said he needed to build it using 2N3904 transistors.

    At national semi, we had tons of sims. When I left, they had adopted a policy that every new product had to have a spice model to release. Most of them were only accurate over a very narrow range of conditions, and the problem is the users would not know that.

    But it was all about the message and the message to the customer was "don't worry, you don't have to know anything or do anything, the computer does the design for you".

    The bottom line is that anybody who ever designed a power circuit knows that there are only two possible development cycles that will have a high probability of a positive outcome:

    1) Come up with a design idea, build it up, get it working and ship it.

    2) Come up with a design, run a simulation, build it up and get the bugs out and ship it.

    I used #1 in my career, and I never saw the added step in #2 doing me any good except to shut my boss up.

    The real problem we saw in the last decade was that the "new" design cycle looks like this:

    1) Run a sim using the model.

    2) Give the schematic to the PC layout guy who knows less about power design than a donkey knows about brain surgery.

    3) Build the thing

    4) If anything goes wrong, call the IC vendor support line and have them tell you why it went wrong and how to fix it.

    5) Sue the IC maker for damages because their "documentation" or software support mislead the user.

    Seriously. We used to do free power supply design for HP, Cisco, Apple and the rest just to make sure we avoided #4 and #5.


    The problem is people are lazy. Professors are lazy and students are lazy. As long as they have a tool that will spit out circuits they won't learn what they need to learn by actually (gasp!) building things and testing them.
     
Loading...