Topics Topics Edit Profile Profile Help/Instructions Help Member List Member List  
Search Last 1|3|7 Days Search Search Tree View Tree View  

Speed up an op-amp?

:: EPE Chat Zone ­:: ­Radio Bygones Message Board :: » EPE Chat Zone » Archive through 14 April, 2015 » Speed up an op-amp? « Previous Next »

  Thread Last Poster Posts Pages Last Post
  ClosedClosed: New threads not accepted on this page        

Author Message
Top of pagePrevious messageNext messageBottom of page Link to this message

741
Frequent Contributor
Username: 741

Post Number: 651
Registered: 08-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Saturday, 14 March, 2015 - 12:36 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

The circumstances where "slew-rate limiting" occurs is related to both frequency and amplitude of output: Things have to move faster to output 1MHz 5V than 1MHz 0.5V.

Given that the op-amp designers have already done their best, is there any point in deliberately reducing output amplitude and then externally amplifying? If yes, what is a good, approximately rail to rail, design?

I'm driving a small MOSFET, with small gate charge so I think the MOSFET input capacitance is not much of a problem.

(One gotcha is extra loop-gain promotes instability, but that can be addressed).
Top of pagePrevious messageNext messageBottom of page Link to this message

alec_t
Frequent Contributor
Username: alec_t

Post Number: 817
Registered: 03-2009

Rating: N/A
Votes: 0 (Vote!)

Posted on Saturday, 14 March, 2015 - 08:41 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

"is there any point in deliberately reducing output amplitude and then externally amplifying?"
Yes, according to simulation with LTspice (using an LM324 model plus an ideal external amp).
text/plainSim file
GainCompare.asc (2.0 k)
Top of pagePrevious messageNext messageBottom of page Link to this message

741
Frequent Contributor
Username: 741

Post Number: 652
Registered: 08-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Saturday, 14 March, 2015 - 10:01 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi,

Thanks - I grabbed a model from the TI site.

I should perhaps have phrased my question as "is there any point in adding a home-made external voltage gain stage, given that the op-amp designers have already done their best to design one for me?".

I can see an add-on stage can incease power handling; maybe that means I can charge stray capacitance faster or whatever. Basically here, I'd like to improve slew rate. I guess the slightly defeatist answer is buy a faster amplifier.

Still, I wonder...?
Top of pagePrevious messageNext messageBottom of page Link to this message

zeitghost
Frequent Contributor
Username: zeitghost

Post Number: 1956
Registered: 01-2006

Rating: N/A
Votes: 0 (Vote!)

Posted on Monday, 16 March, 2015 - 09:37 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Surely you can get faster snails than the lm324?
Top of pagePrevious messageNext messageBottom of page Link to this message

741
Frequent Contributor
Username: 741

Post Number: 653
Registered: 08-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Monday, 16 March, 2015 - 06:23 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Yes, I have a veritable greyhound, the OPA2322.

But the question is a general one really.

Given some op-amp, can I add external circuitry to keep the low-amplitude frequency performance, but raise the allowable output level?

Its seems its not 'just' GBW contributing limits here.

Maybe 100mV in: 100mV out (gain = 1) on some op-amp lets us run to X MHz, but 3V in: 3V out (same gain) gets slew-rate limited at the same frequency and the same gain.
Top of pagePrevious messageNext messageBottom of page Link to this message

james
Frequent Contributor
Username: james

Post Number: 498
Registered: 02-2007

Rating: N/A
Votes: 0 (Vote!)

Posted on Monday, 16 March, 2015 - 08:12 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi Stephen,

Yes, for a given gain, GBW puts a limit on frequency at small signal amplitudes. At larger signal amplitudes slew rate limiting comes into play, as you say.

Maximum undistorted frequency,f = Slew Rate/(2*p*Vpk)

Where Vpk = Peak amplitude of signal.

Or swapped around,

Max amplitude without slewing = Slew Rate/(2*p*f)

This is where the opamp's Full Power Bandwidth specification can come in.
Full Power Bandwidth is the bandwidth over which the amp's output can produce an undistorted output (ie. without slewing) with a max amplitude signal (ie. an amplitude just below clipping).
This bandwidth will be less than that calculated by GBW.

Sometimes the Full Power Bandwidth figure is quoted at a smaller amplitude than the clipping amplitude to give a more realistic figure which is not to pessimistic.
Having said that FPBW is not often quoted in my experience possibly because it varies with magnitude of supply voltage.

Cheers

James
Top of pagePrevious messageNext messageBottom of page Link to this message

james
Frequent Contributor
Username: james

Post Number: 499
Registered: 02-2007

Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 17 March, 2015 - 12:50 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

As an example...

Consider the LM358.
GBW = 1MHz
SR = 0.3V/us = 300,000V/s

Consider a situation with 1mV in and 100mV out.
Therefore Av = 100.

fmax due to GBW limitation = GBW/Av = 1MHz/100 = 10KHz.

fmax due to slewrate limitation = SR/(2*p*Vpk)
= 300,000/(2*p*100mV)
=477KHz.

Clearly the bandwidth is limited by GBW rather than slewrate at this low 100mV amplitude.

So, 10KHz is the max frequency due to the GBW limitation with Av = 100.

If the magnitude of the output signal was to increase due to an increase in the input signal it would be useful to know how large the output signal could be (at 10KHz) before the Slewrate started to distort it.

Max undistorted output signal,Vpk = SR/(2*p*f)
= 300,000/(2*p*10,000)
= 4.8V

And at 8V amplitude...

max undistorted frequency = SR/(2*p*Vpk)
= 300,000/(2*p*8)
=6KHz

So at 8V amplitude the Slewrate limits the bandwidth to 6KHz which is significantly lower than the 10KHz GBW limited bandwidth.

Cheers

James
Top of pagePrevious messageNext messageBottom of page Link to this message

741
Frequent Contributor
Username: 741

Post Number: 654
Registered: 08-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 17 March, 2015 - 10:08 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

That's a nice summary James - thank you.

Your formula reminds me of something I read once, and need to look up again.

From shaky memory:

Max slope of sin wave being output is max slew needed to avoid distortion.

The slope of sin(x) is simply cos(x).

So the max of the sin slope is max of cos, i.e. signal amplitude. Thus amplitude almost directly gives the max slew required.

Now the op-amp designers might allow for Av of 1 to about 200 (say), and the output stage caters for that. Maybe for some fixed gain, say 10, I can overcome the SR limit with an external stage?
Top of pagePrevious messageNext messageBottom of page Link to this message

james
Frequent Contributor
Username: james

Post Number: 500
Registered: 02-2007

Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 17 March, 2015 - 10:41 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi 741

Your memory serves you well.

Vo = VmSinwt

dVo/dt = wVmcoswt

(dVo/dt)max = wVm

Slewrate = wVm where w = 2*p*f

Cheers

James
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1637
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 17 March, 2015 - 10:52 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hola Stephen,

Started to consider a project where I would use a MOSFET for the first time. Nothing seems simple in this life!

An oportunity to say my part, being sure you are not going to learn anything new.

Relationship, or better, limitations to GBW impossed by slew rate I could really understand it by watching this:

https://youtu.be/UooUGC7tNRg

I am not really into watching videos for learning but this EE has some which are always short and to the point. A bonus: I understand all what he says ).
Agustín Tomás - Buenos Aires - Argentina
Top of pagePrevious messageNext messageBottom of page Link to this message

741
Frequent Contributor
Username: 741

Post Number: 657
Registered: 08-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 17 March, 2015 - 04:18 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi Agustin,

Yes, its a nicely put together (and 'delivered') tutorial. I never thought about differing slew rate limits (rising v falling).

So for my op-amp (OPA2322), GBW = 20MHz = 2E7 in scientific notation. Slew rate S = 10V/uS = 1E7.

 
f < S/(2.p.Vpk)
f < 1E7/(6.28*Vpk)
If output amplitude 'Vpk' is 1 V, then
f < 1E7/6.28 i.e. 1.6 MHz!


On the other hand, for small circuit ('closed loop') gain, such as a gain of x1, I can only approach GBW limits if the output level is very small indeed. Suppose Av = 1. Then "small signal" bandwidth is GBW = 20 MHz... At 20 MHz, how does slew rate limit amplitude, Vpk?

 
Vpk < S/(2.p.f)
Vpk < 1E7/(6.28*2E7)
Vpk < 0.5/(6.28) = 0.079V ???

An "eye-opener" as they say!

(Have I missed some factor out, or made some error, in these calculations?)

P.S.
Boy, has that guy got a nice oscilloscope!
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1638
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 17 March, 2015 - 04:46 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Well, he works for Tek as a Senior Engineer for some years now...
Agustín Tomás - Buenos Aires - Argentina
Top of pagePrevious messageNext messageBottom of page Link to this message

james
Frequent Contributor
Username: james

Post Number: 501
Registered: 02-2007

Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 17 March, 2015 - 04:54 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi Stephen

Your calculations look fine to me.

But it is an "eye opener" as you say, such a small maximum output amplitude (79mV) if slewing distortion is to be avoided, even when you consider that your trying to get the maximum GBW limited frequency at a gain of 1.

I suppose this situation will get worse (lower maximum amplitude) for op amps with an especially high GBW specification if the Slew Rate specification is not increased in proportion.

Cheers

James
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1639
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 17 March, 2015 - 05:13 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hola Alec

I could not simulate your .asc file. LTSpice complains about reversed (or inverted?) something.

Bringing it to an understandable format by LTSpice seems the point. Could you help?

Previous ones from you had no problem on my side.

Gracias.
Agustín Tomás - Buenos Aires - Argentina
Top of pagePrevious messageNext messageBottom of page Link to this message

741
Frequent Contributor
Username: 741

Post Number: 658
Registered: 08-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 18 March, 2015 - 06:59 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

This site is excellent for LTSpice guidance.

1. Obtain the model from Texas Instruments (file name is "LM324.5_1")

At right is
"Featured tools and software
LM324, LM324A, LM324X2 PSpice Model (Simulation Models)"

Place this file in the same directory as your .asc file.

2. Add SPICE .inc directive
.include LM324.5_1

3. Open the file in Notepad, look for ".SUBCKT"
--> .SUBCKT LM324

4. Add a "UniversalOpamp2" symbol

5. Right-Click, edit "Value" to LM324

application/x-zip-compressedSlew.zip
Slew.zip (1.9 k)
Top of pagePrevious messageNext messageBottom of page Link to this message

alec_t
Frequent Contributor
Username: alec_t

Post Number: 818
Registered: 03-2009

Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 18 March, 2015 - 11:48 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hola Agustin,
If the LM324 model is what stops the sim working then just use another low GBW opamp model instead, to demonstrate the effect of signal amplitude on frequency response.
There seem to be several LM324 spice models. I got mine from the Yahoo LTS User Group. It does seem flaky, though. Perhaps the TI version that 741 uses is better.

Administration Administration Log Out Log Out   Previous Page Previous Page Next Page Next Page