Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Rail to Rail High Bandwidth and High slew rate opamp

Status
Not open for further replies.
Hi MrAl
Yes I did try the Ron's circuit. But when I go at frequencies in MHz, the rectification is not smooth. I am actually trying to rectify IEEE 8/20 micro second surge so that I can feed it to A/D converter
 
Hi,

Actually i meant take Ron's advice and adjust the gain. Did you adjust the gain? The wrong gain would contribute to uneven full wave rectification. I found the gain of the first stage had to be set to a much lower value in order to get equal 'bumps' on the output.
If on the other hand you are seeing distortion due to the slew delay (that's what i call it but i dont know the official name for this) then you need a faster amp. If the fastest amp you can find still doesnt work, then perhaps you need a different design. If we could find a way to keep both amps in the linear range all the time that would at least help.

Alternately, how about a synchronous rectifier using an analog switch(s) ? That should work up to maybe 20MHz. The op amps, if needed, would always be operating in their linear range.
 
Last edited:
Hello again,

You do realize that you have to also filter it too right, unless you have an extremely fast ADC, which is also a possibility.

To rectify it synchronously, one idea would go like this...
First, an op amp to invert the signal. That gives you a two phase signal, phase A and phase B, where the two signals are 180 degrees out of phase.
Next, use a very fast comparator to detect when the input signal is positive. The output of the comparator controls the single-pole-double-throw analog switch, switching phase A to the very output when the input is positive, and when the input goes negative it switches phase B to the output. Since phase B is inverted and gets switched to the output during the input phase of 180 to 360 degrees, the output is therefore always positive.
This technique also keeps the op amp(s) always working in their linear range so the output will be clean except for a tiny transition period around the zero crossings of the output.

Just to recap...
For the input signal from 0 to 180 the non inverted signal (which may be just the input signal itself) gets switched to the output, and during the input signal phase 180 to 360 the inverted signal is switched to the output. The output will therefore be a full wave rectified sine.

Understand this now?
 
Last edited:
Thanks MrAl
I am using pic24 mplab starter kit for ADC. Your idea about sync rectifier was good, but I could not understand it. Are there any circuits available that I can try and understand?
 
Thanks MrAl
I am using pic24 mplab starter kit for ADC. Your idea about sync rectifier was good, but I could not understand it. Are there any circuits available that I can try and understand?

Hi,

Ok then that wont be fast enough so you'll also need to filter the output with at least a resistor and capacitor. That will give you the average from which you can calculate the peak value provided you always have a sine wave input.

Here is a block diagram (not showing the filter)...
 

Attachments

  • SyncRectifier-01.gif
    SyncRectifier-01.gif
    10.4 KB · Views: 150
Hi
Thank you MrAl for the block Diagram. I found something on precision rectifiers using 2nd generation current conveyors. Do you know what they are and how do the work? also can you suggest me some if they are available in market. I am attaching the paper that I found out
 

Attachments

  • precision rectifier.pdf
    1 MB · Views: 198
Hi there,

As you probably know, a regular op amp can be modeled as a voltage controlled voltage source. That means it senses an input voltage (difference) and outputs another voltage based on that input. There are also current mode op amps that sense current.

Apparently the difference is that the CCII+ is capable of sensing BOTH voltage and current simultaneously. For example, a voltage input at the Y terminal will produce a voltage at X terminal of equal value, and then the current out of the X terminal
is mirrored out of the Z terminal, and the current out of the X terminal is based on the voltage at X and the external circuit impedance connected to X.
In the simplest case, a resistor R1 from X to ground will cause a current Ix=Vy/R1 to flow out of X which in turn will cause that same current Ix to flow out of the Z terminal unless there is gain B, in which case the current out of Z would be Iz=B*Ix and so if there was another resistor R2 connected from Z to ground the output voltage would be simply Vz=Iz*R2 which again would be Vz=B*Ix*R2 which can also be written as Vz=B*R2*Vy/R1 which tells us this is a voltage amplifier and the gain is B*R2/R1.
The difference in the output is that since you are controlling a current and not a voltage it is going to be more compliant.

So it is an interesting device, that is like a (simultaneously both) voltage and current controlled current source which for one way we might abbreviate VCCCS. This is in contrast to VCCS and CCCS which only sense either voltage or current but not both in the same device.

We could probably come up with a simple model for spice, but yeah i think it would be wise to look around for the types of devices available on the market to make sure one will reach up to your desired bandwidth.
I dont know any off hand, but we could look. I'll take a look a little later today or early tomorrow morning.
 
Hello again,

Apparently the current conveyor is used as part of anther device, such as a current mode op amp. That means it might not be available as a separate device in itself. For example, Maxim makes some fast current mode op amps which have a BW of 1GHz. That should be able to handle 8MHz.
Aside from that, maybe you'd want to build one yourself, but it may be hard to do with all the transistors, depending on accuracy needed.
National Semiconductor, when they were around, used to make a integrated circuit transistor that i think would behave in a similar manner, but i dont remember the part number. It may not be fast enough anyway.

Another way to view the current conveyor in a simpler way is to think of it as an NPN transistor with zero base emitter voltage drop, infinite input impedance, and Beta of only 1. That would mean that any voltage that is applied to the base shows up at the emitter, and any resistor from emitter will draw a current, and that current will be mirrored in the collector given a Beta of exactly 1 and zero input base current. Any collector resistance would drop Ic*Rc and thus there would be a voltage gain.
 
I recall NS had one chip qualified as "dam fast" in the very first line of the datasheet. Not sure if an op amp or just a buffer.
 
The LH0063 known as the "Dam Fast Buffer Amplifier", represents an extreme slew rate limit in commercially available buffers of 6000 volts/microsecond.

no gain just buffer

I can guess and extreme price as well...
 
A precision 8MHz wave rectifier does not have any useful purposes, that I can think of and you would be better off to define the reason or specs. for all signal ranges and interference.

This leads to the better solution already mentioned by Crutchow , being a S&H detector.
By using a Comparator to enable a fast Analog switch and small cap such that the switch is active only when the peak signal is detected, which can be done by several simple means depending on nosie immunity requirements.

The result is a precision envelope detector with a LPF defined bandwidth.

If you were doing QAM demodulation for example, the design might be different.

But if you just want a pretty waveform,.... I would use a simple unity gain inverting amp and a CMOS Analog switch configured as a SPDT and toggle between the two sources depending on signal polarity with a CMOS logic inverter acting as a limiter.

So IMHO always start with good specs or requirements.... than just a how-to question.
 
Hi,

From what i read they are discontinued (LH0063).
 
Hi MrAl
Yes I did try the Ron's circuit. But when I go at frequencies in MHz, the rectification is not smooth. I am actually trying to rectify IEEE 8/20 micro second surge so that I can feed it to A/D converter

Why not use a DSO? Do you know how to probe it? Compensated for cable impedance? e.g. 10:1 with tuning cap? or 50 Ohm terminations.

Why rectify it? There are two standard surge signals.

1. Combination wave: a unipolar pulse that occurs most often outside a facility (e.g., a lightning strike)
2. 100 kHz Ringwave: an oscillating waveform that occurs most often inside a facility
 
Why Not just use a Peak Detector, Like This?

View attachment 96863

this 1000:1 capacitive divider uses diodes to clamp the most negative peak wave then rectify the most positive peak and the attack time depends on the source R value and RC result.
 
Why Not just use a Peak Detector, Like This?

View attachment 96863

Hi,

Chemelec:
We used a peak detector like that to test for ripple on the output of a Zetex 300 or 310 LED driver chip way back when the Zetex chips just came out around year 2002. I cant believe that was some 14 years ago already.

The attachment shows that circuit, although it was tailored to fit the application with an added high frequency filter to filter out the short duration peaks but keep the main bulk of the 300kHz ripple peaks. Of course removing the filter and using a signal diode would be better.

The main point is though that there is some calibration required, using a scope. The one in the diagram was calibrated using a scope. If the signal shape is known then the output can be correlated to the input, and for sine waves this should not be too hard to do by measuring several points and then extrapolating between points.

Tony:
I think the output cap voltage will 'pump' up to the input to some degree because there is a rectifier so it is not 'just' about the voltage divider ratio of cap values, however some calibration would be required because there will still be loss in the diode(s). How much loss would have to be compared to the scope measurements and may or may not affect the overall measurement if the input signal is large. The ratio of caps does seem pretty large however, that seems a little too extreme.
 

Attachments

  • RippleDetect&LS.gif
    RippleDetect&LS.gif
    8.4 KB · Views: 137
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top