Increasing speed of Opamp used as Comparator

[John Larkin]

And how would you clamp a follower anyway? And feedback won't work
without the input being carefully bounded.

Follower? I thought he needed a comparator. Depending on the situation
(like, is the comparison level constant?) a negative-feedback clamp
might be arranged, but it will certainly be a nuisance.

Hey, I'm not an LM324 fan!

John

 

[Spehro Pefhany]

LM324 is a rotten opmap anyhow. Don't even think about pulling any of
the inputs even a little below ground.

John

You won't find a better quad for the price.


Best regards,
Spehro Pefhany

 

[Jim Thompson]

Follower? I thought he needed a comparator. Depending on the situation
(like, is the comparison level constant?) a negative-feedback clamp
might be arranged, but it will certainly be a nuisance.

Hey, I'm not an LM324 fan!

John

The output may be considered as followers.

As I said before, negative feedback clamping would be generally less
than worthless.

...Jim Thompson

 

[John Larkin]

You won't find a better quad for the price.

Another way of saying that is that it's hard to find a worse opamp at
any price.

John

 

[John Larkin]

The output may be considered as followers.

As I said before, negative feedback clamping would be generally less
than worthless.

...Jim Thompson

Wouldn't it keep the internal nodes from winding up? I've seen opamps
that took *seconds* to come off a rail.

John

 

[Jim Thompson]

Wouldn't it keep the internal nodes from winding up? I've seen opamps
that took *seconds* to come off a rail.

John

That's why I mentioned input *bounding*. If I knew the input range
there might be a solution... of course feedback limits output swing.

...Jim Thompson

 

[[email protected]]

John Larkin skrev:


The datasheet for 5 different manufactors I looked up didn't mention
this and I have never heard of this. Is this some soft of undocumented
feature that is constrained to us westerneers so the terrorist bombs
explode prematurely due to the quirks of the LM324s? (bad joke, sorry)

Look at the Linear Technology data sheets for the LT1006. and for the
LT1013 dual and LM1014 quad versions of the same amplifier at

http://www.linear.com/

The LT1006 is much better behaved than the LM324 (and much more
expensive) so the data sheets include a small gloat - see sheet 9 in
both data sheets.

They also give some information about the performance of the parts as
comparators.

Pity John Fields won't read this - the creep is claiming that I post
"Basically nothing but opinions formed long ago and 'You're not doing
it the way I would, so you must be wrong '"

 

[Michael A. Terrell]

Bill is so full of himself that he had to sign it twice!


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

 

[John Larkin]

Bill is so full of himself that he had to sign it twice!

Yeah, too bad a technical post couldn't stay a technical post. I guess
we know what really matters here.

John

 

[Klaus Kragelund]

[email protected] skrev:

Look at the Linear Technology data sheets for the LT1006. and for the
LT1013 dual and LM1014 quad versions of the same amplifier at

http://www.linear.com/

The LT1006 is much better behaved than the LM324 (and much more
expensive) so the data sheets include a small gloat - see sheet 9 in
both data sheets.

They also give some information about the performance of the parts as
comparators.

Pity John Fields won't read this - the creep is claiming that I post
"Basically nothing but opinions formed long ago and 'You're not doing
it the way I would, so you must be wrong '"

--

I have tried the idea by Jon:

www.microdesign.dk/tmp/LM324Simulation_Sch.pdf
www.microdesign.dk/tmp/LM324Simulation_Plt.pdf

LM324 with no tricks, in saturation has about 15us of delay before
output reaches VCC/2

LM324 with the clamping trick has about 6us before VCC/2

LM324 with a BJT connected like this:

www.microdesign.dk/tmp/LM324_plusBJT.pdf

has 7us of delay until output shifts to VCC/2

The nice thing about the last one is that it is cheap and has logic
level output for use with microcontrollers and so on

Comments?

Regards

Klaus

 

[Klaus Kragelund]

Klaus Kragelund skrev:

[email protected] skrev:


I have tried the idea by Jon:

www.microdesign.dk/tmp/LM324Simulation_Sch.pdf
www.microdesign.dk/tmp/LM324Simulation_Plt.pdf

LM324 with no tricks, in saturation has about 15us of delay before
output reaches VCC/2

LM324 with the clamping trick has about 6us before VCC/2

LM324 with a BJT connected like this:

www.microdesign.dk/tmp/LM324_plusBJT.pdf

has 7us of delay until output shifts to VCC/2

The nice thing about the last one is that it is cheap and has logic
level output for use with microcontrollers and so on

Another circuit:

www.microdesign.dk/tmp/LM324_plusBJTClamp.pdf

This circuit has a response of 5us

Regards

Klaus

 

[Fred Bloggs]

Klaus Kragelund skrev:



Another circuit:

www.microdesign.dk/tmp/LM324_plusBJTClamp.pdf

This circuit has a response of 5us

Regards

Klaus

Well if you're going to use all that then go discrete all the way and
get 50ns response.

 

[John Larkin]

Klaus Kragelund skrev:

Another circuit:

www.microdesign.dk/tmp/LM324_plusBJTClamp.pdf

This circuit has a response of 5us

Regards

Klaus

OK for an intellectual exercize, but in real life, why add 11 or 12
additional discrete parts to turn a bad opamp into a mediocre
comparator? I doubt this saves space or money.

John

 

[[email protected]]

Yeah, too bad a technical post couldn't stay a technical post. I guess
we know what really matters here.

My apologies. I should know better than to post a reply at four in the
morning, when I really ought to be in bed. but the LT1006/LT1013/LT1014
datsheets were just too tempting to let me hit the sack. Bad judgement.

 

[Spehro Pefhany]

OK for an intellectual exercize, but in real life, why add 11 or 12
additional discrete parts to turn a bad opamp into a mediocre
comparator? I doubt this saves space or money.

John

I count 7 or so *additional* parts (over using an additional
comparator chip with pullup resistor), with a parts cost and power
consumption less than adding the cheapest IC comparator, but I agree
it looks a bit messy. It's not completely insane for huge volume.


Best regards,
Spehro Pefhany

 

[John Larkin]

I count 7 or so *additional* parts (over using an additional
comparator chip with pullup resistor), with a parts cost and power
consumption less than adding the cheapest IC comparator, but I agree
it looks a bit messy. It's not completely insane for huge volume.

I like using LVDS receivers as comparators. They're cheap, have
infinite Zin, brutal cmos outputs, and are blindingly fast.

Or how about using current-steering diodes into an opamp that stays
linear? That's more like 4 additional parts.


John

 

[Spehro Pefhany]

I like using LVDS receivers as comparators. They're cheap, have
infinite Zin, brutal cmos outputs, and are blindingly fast.

Or how about using current-steering diodes into an opamp that stays
linear? That's more like 4 additional parts.

John

According to the data sheet, even if you keep it linear, and hit the
inputs with sledgehammer (a 2.5V step), the output slews at a
leisurely ~0.5V/usec.

The simple RTL inverter on the output of an unadulterated
comparator-connected amplifier looks like the best bang for the buck
to me (7usec, the OP says), and only 1 or 2 additional very cheap &
small jellybean parts. It works because he only needs fast response in
only one direction, and the single supply output clamps itself at
ground with no additional parts. If he feels lucky he could try
biasing the base more positive by a couple hundred mV with another
resistor. ;-)


Best regards,
Spehro Pefhany

 

[[email protected]]

Klaus Kragelund skrev:

Another circuit:

www.microdesign.dk/tmp/LM324_plusBJTClamp.pdf

This circuit has a response of 5us

Regards

Klaus

Another possibility is to go the other way: depending on what you're
doing, LM339 comparator sections can used as nasty, slow, horrible op
amps[1], and the left-over section makes a dandy comparator.

Or use the LM392, which combines an LM324 op-amp and LM339 comparator
in one package (any quads out there?)

Cheers,
James Arthur

[1] Not that an LM324 sets much of a standard[2].
[2] Though I'm still a fan of the nasty little beggars.

 

[[email protected]]

I like using LVDS receivers as comparators. They're cheap, have
infinite Zin, brutal cmos outputs, and are blindingly fast.

Any particular standouts?

Regards,
James Arthur

 

[John Larkin]

Klaus Kragelund skrev:

Another circuit:

www.microdesign.dk/tmp/LM324_plusBJTClamp.pdf

This circuit has a response of 5us

Regards

Klaus

Another possibility is to go the other way: depending on what you're
doing, LM339 comparator sections can used as nasty, slow, horrible op
amps[1], and the left-over section makes a dandy comparator.

Or use the LM392, which combines an LM324 op-amp and LM339 comparator
in one package (any quads out there?)

Cheers,
James Arthur

[1] Not that an LM324 sets much of a standard[2].
[2] Though I'm still a fan of the nasty little beggars.

May you step on an LM324N, pins up, barefoot. [3]

John

[3] I have. Maybe that's why I dislike them so much.