[SOLVED] Explain me how op amp cascading works

[jeffrey samuel]

if i am to cascade say 6 op amps in series to compare two input signals will there be any time lag in it

or the operations of the op amp be instantaneous

must i also consider the time lag factor for my circuit design'

i am new to it so kindly explain it to me

i am struggling a lot with it so kindly be elaborate in your point

 

[godfreyl]

Re: op amp cascading problem

Each opamp has a tiny time delay. Why on earth do you want to connect six in series?

 

[mrflibble]

Re: op amp cascading problem

Well, almost by definition there will be lag. It´s a non-ideal physical object with mass. In a gravity well. At temperatures above 0 K. Without magic sauce. So there will be lag. Other than that, yeah, your typical opamp will have non-trivial delays between input and output.

 

[D.A.(Tony)Stewart]

Re: op amp cascading problem

{added) The only valid reason for cascading 6 Op Amps that I can think of is to achieve a higher gain bandwidth product at the expense of accumulated noise figure, input offset, group delay, IM distortion to mention a few. THIS IS NOT THE WAY TO DESIGN A COMPARATOR If for example you want to design a comparator, then you want the highest bandwidth,, no feedback, hence no stability issues, no noise, hence well-filtered signals to avoid lack of monotonicity causing switching noise on the output.

Read specs of parts for High Speed Comparator selection menu . These are designed differently than Op Amps as the compensation filter is not needed.
http://www.linear.com/products/high_speed_comparators

Read here to see how Propagation Delay is affected by overdrive in mV in 1st graph example of an XO.

As my signature states, what is your spec?

If you are trying to cascade 6 stages with maximum bandwidth allowing say only a gain of 10 in each stage, that may work but video amps are better for specific broadband applications that need high gain-bandwidth. Prop. delay would be improved with fewer stages.

YOu need to read some journals from respected authors such as the late Jim Williams.
http://cds.linear.com/docs/Application Note/an47fa.pdf

Absorb as much as you can.

 

[KRIS255]

Re: op amp cascading problem

The opamps better be capacitively coupled otherwise the input offset voltage on the first stage will send your output into saturation.

 

[FvM]

Re: op amp cascading problem

If have problems to see the original post as a serious question, because it's holding almost any detail in suspense.
Points that should be mentioned:
- are you designing a feedback amplifier, cascading gain blocks with individual feedback or are just a comparator
- what's the motivation for cascading 6 OPs
- any specification or other design requirements

We had a recent thread that addressed some aspects of cascaded OPs: https://www.edaboard.com/threads/261574/

 

[D.A.(Tony)Stewart]

Re: op amp cascading problem

maybe you just want a high speed comparator

4ns prop delay https://www.linear.com/product/LT1715

 

[jeffrey samuel]

Re: op amp cascading problem

maybe you just want a high speed comparator

4ns prop delay https://www.linear.com/product/LT1715

this is it i guess

but can it compare the voltages at different points on a wave with reference

- - - Updated - - -

that to there is a small prob when i went through the app it is for sampling at high speed then what really is it saying i am new so not aware of those terms

 

[FvM]

Re: op amp cascading problem

compare the voltages at different points on a wave with reference

I guess, we need to refer to another thread to understand the problem that the original poster has in mind: https://www.edaboard.com/threads/261524/

If so, it's just a helpless case.

 

[D.A.(Tony)Stewart]

Re: op amp cascading problem

The title OP Amp cascading is incorrect... try to edit the original question to say what you need,, like accurate comparator at minimum propagation delay. or "Precision high speed slicer required" or whatever you need rather than incorrectly assume your implementation needs improvement (cascaded Op Amp problem). It helps get at the best solution quicker.

In order to make any high speed circuit , such as the fastest device I suggested later with 4ns Prop Delay, you must have controlled low impedances to match the track impedance of typically 120~150 when a ground signal is on either side or underneath. Shielding, avoiding loops, and low source impedance filtered to the signal bandwidth to eliminate noise are all critical. The same is true for the reference voltage. It must be low impedance and well filtered with low ESR RF caps to prevent crosstalk from the output getting into the V ref voltage which is the threshold of the slicer along the waveform you required..

It would be more helpful to share what you are measuring and show a photo of the proposed layout.

 

[mrflibble]

Re: op amp cascading problem

1 - quantize
2 - correlate
3 - profit!

oh yeah, forgot step 0...

0 - forget about doing the signal comparison in the analog domain

 

[FvM]

Re: op amp cascading problem

An annoying point is that you didn't tell yet a rough idea of the intended time scale,

e.g. N compare points * t point time distance = T waveform length

 

[mrflibble]

Re: op amp cascading problem

If those were the only missing details things would be relatively peachy!

 

[FvM]

Re: op amp cascading problem

Of course, this are not the only open questions. I just want to promote a systematical way to present the problem. I agree that "forget about doing the signal comparison in the analog domain" is the most likely result.

Analog signal comparison, e.g by correlation can be done and has been done, when digital signal processing wasn't yet available. It's mainly a museum of electronics topic now.

The least we can learn from a comparison of analog and digital signal processing methods is a better understanding about the strong points of both ways.

 

[jeffrey samuel]

Re: op amp cascading problem

thank you all


i learnt one way of not giving a topic and many solutions to this issue i think slice the signals at a particular time
compare them and you have your comparator

thanks all
and sorry for geting you all irked out cos of the title

 

[mrflibble]

Re: op amp cascading problem

What does "slice the signals at a particular time ... compare them" mean?

 

[D.A.(Tony)Stewart]

Re: op amp cascading problem

To digitize an analog signal to binary output, you slice the signal with a stable reference voltage. Asymmetry, unequal rise fall time, jitter all introduce loss of phase margin in clock & data recovery. So we call these comparators "Slicers"

 

[mrflibble]

Re: op amp cascading problem

To digitize an analog signal to binary output, you slice the signal with a stable reference voltage. Asymmetry, unequal rise fall time, jitter all introduce loss of phase margin in clock & data recovery. So we call these comparators "Slicers"

Thanks for the clarification. Although I'm afraid I still don't get it. Then again, maybe I just misunderstand the OP's requirements. Because ...

The objective as far as I understand it is to "compare signal A and signal B to see if they are the same". If the OP means just to compare 2 instantaneous voltages (as in there is no memory element in the "compare signals" function) then 1) the OP could have been a little clearer and 2) I more or less understand the proposed use of comparators. If however the requirement is to compare two signals that are roughly the same shape, but could be spread out over time somewhat (input A is lagging behind input B for example), maybe need some scaling etc ... well then I don't quite see how it would work.

But oh well, it's marked as solved so I hope jeffrey samuel has a working solution for whatever the problem turns out to be.

 

[D.A.(Tony)Stewart]

Re: op amp cascading problem

He was trying to cascade Op Amps with low gain but high bandwidth to do what Comparators do already. So compare means output=1 if (+) is greater than (-) at any time, faster the better….. This is like slicing a signal at any level over all time.

 

[mrflibble]

Re: op amp cascading problem

So standard use of comparator.... and there I was hoping I was missing some clever analog trick. Anyways, thanks for clearing that up.