fast comparator, inputs operate to the negative rail?

[Winfield Hill]

What do you like for fast (better than 50ns if possible)
5V-supply comparators with inputs that operate to the
negative rail? Surface-mount package. Finally, would
a rail-to-rail logic output be too much to ask? :>)

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Jeroen Vriesman]

EL5181C from elantec (intersil), 8 pin SO package. 8ns prop. delay.

input to 0.1 V below ground, rail to rail output (well, almost)

Cheers,
Jeroen.

joenix at gmx dot net

 

[Jeroen Vriesman]

MAX919 also, but that might be a bit of an overkill, don't know the price
but sounds expensive.

 

[Jeroen Vriesman]

oops, not the MAX 919, got microseconds confused with nanaseconds, sorry

 

[Winfield Hill]

Jeroen Vriesman wrote...

EL5181C from elantec (intersil), 8 pin SO package. 8ns prop. delay.
input to 0.1 V below ground, rail to rail output (well, almost)

Thanks Jeroen, a good one. I have the LT1720 and LT1116 in stock.
TI's TL3116 is an improved LT1116. But I hadn't found the EL5181,
and it has everything I've been looking for. Thanks!

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Glenn Gundlach]

What do you like for fast (better than 50ns if possible)
5V-supply comparators with inputs that operate to the
negative rail? Surface-mount package. Finally, would
a rail-to-rail logic output be too much to ask? :>)

Thanks,
- Win

whill_at_picovolt-dot-com

Would this device be close enough? I've used lots of them on dual 12v,
dual 5v single 5v.

http://www.analog.com/UploadedFiles/Data_Sheets/258917749AD790_d.pdf

GG

 

[Winfield Hill]

Jeroen Vriesman wrote...

oops, not the MAX 919, got microseconds confused with nanoseconds,
sorry

Maybe not the MAX919 Jeroen, but Maxim's handy app note 886, on
selecting the right comparator, reveals the MAX9011 (dual) and
MAX9013 (LT1116 replacement), which are nice 4.5ns parts running
on a miserly 900uA supply. But they don't have rail-rail outputs.

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Winfield Hill]

Glenn Gundlach wrote...

http://www.analog.com/UploadedFiles/Data_Sheets/258917749AD790_d.pdf

Yes, thanks very much, the AD790 is slower than the others we've
tossed around here, but it's a nice rail-to-rail-output part with
a very low 250uV voltage offset version available and with 400uV
of built-in hysteresis, which is attractive. Truly a _precision_
comparator!

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Phil Hobbs]

Maybe not the MAX919 Jeroen, but Maxim's handy app note 886, on
selecting the right comparator, reveals the MAX9011 (dual) and
MAX9013 (LT1116 replacement), which are nice 4.5ns parts running
on a miserly 900uA supply. But they don't have rail-rail outputs.

Thanks,
- Win

I'm a big fan of the MAX900 and brethren. Besides being a quick and
well-behaved part, it has a latch input that you can drive from its own
output. What's useful about this is that it gives the effect of AC
hysteresis but without putting junk back out the inputs. Bipolar
thresholds, window comparators, and so on benefit.

Most such latches are in the output structure, which means that looping
the output back will cause oscillations and other bad behaviour. The
MAX900's latch is in its input structure, so there's no setup time
requirement.

Cheers,

Phil Hobbs

 

[Jim Thompson]

I'm a big fan of the MAX900 and brethren. Besides being a quick and
well-behaved part, it has a latch input that you can drive from its own
output. What's useful about this is that it gives the effect of AC
hysteresis but without putting junk back out the inputs. Bipolar
thresholds, window comparators, and so on benefit.

Most such latches are in the output structure, which means that looping
the output back will cause oscillations and other bad behaviour. The
MAX900's latch is in its input structure, so there's no setup time
requirement.

Cheers,

Phil Hobbs

"MAX900's latch is in its input structure", ala MC1650/51 ;-)


...Jim Thompson

 

[Winfield Hill]

Phil Hobbs wrote...

I'm a big fan of the MAX900 and brethren. Besides being a quick and
well-behaved part, it has a latch input that you can drive from its
own output. What's useful about this is that it gives the effect of
AC hysteresis but without putting junk back out the inputs. Bipolar
thresholds, window comparators, and so on benefit.

I agree, and have been a long-time user of the MAX903. In fact it
was the 3.5V TTL output of that '903 and most other LT1016 / LT1116
variants that set me off on a search for a different part. :>)

OK, so far we have several suggested comparators, all available in
8-pin surface-mount packages. All have input ranges that includes
the negative supply rail and can operate from a single 5V supply.
Most parts have a pair of complementary outputs, which reduces
supply noise, although two ICs address this issue with a separate
Vcc pin for the logic output stage. Two types are dual comparators.
Three types have rail-to-rail output swings. :>)

part speed Vo+ I_b Vo P_d latch hyst comp other
number (ns) (V) (uA) (mV) (mW) ? ? out? features
------ ---- --- ---- --- --- --- --- --- --------
MAX903 8 3.5 4 4 18 X = X self-latch
MAX9013 5 3.3 + 5 4.5 X X X
MAX9012 5 3.3 + 5 2.4 - - - dual comparators
LT1116 10 3.4 10 3 135 X - X
TL3116 8 3.9 0.7 3 63 X - X
AD790 40 5 2.5 1* 50 X X - sep. logic Vcc
EL5181 8 5 3.5 3 40/55 X - - sep. logic Vcc
LT1720 4.5 5 <6 4 20 - X - dual comparators

notes
+ bias I_B = 2uA ?, or = 0 if Vin well above Vee, say > 1V...
* AD709K version available with Vos (max) = 0.25mV (miniDIP only)

Any other contenders for the list?

Thanks,
- Win

whill_at_picovolt-dot-com