Andrew Holme
Guest
Thu Mar 11, 2010 10:45 pm
Does such a chip exist? I can't find it.
MAX9375 or PTN3310 = Single LVDS to LVPECL
tb5d1m = Quad LVTTL to LVPECL
MC10H352-D = Quad CMOS to PECL
sn65lvelt22 = Dual LVTTL to LVPECL
MC100EPT622-D (obsolete) = 10-bit LVTTL to single-ended LVPECL and I only
want single-ended LVPECL to drive 4 data inputs on a MC100EP142
Any suggestions?
TIA
Jim Thompson
Guest
Thu Mar 11, 2010 11:05 pm
On Thu, 11 Mar 2010 20:45:17 -0000, "Andrew Holme" <ah_at_nospam.com>
wrote:
Quote:
Does such a chip exist? I can't find it.
MAX9375 or PTN3310 = Single LVDS to LVPECL
tb5d1m = Quad LVTTL to LVPECL
MC10H352-D = Quad CMOS to PECL
sn65lvelt22 = Dual LVTTL to LVPECL
MC100EPT622-D (obsolete) = 10-bit LVTTL to single-ended LVPECL and I only
want single-ended LVPECL to drive 4 data inputs on a MC100EP142
Any suggestions?
TIA
See page 3...
www.pericom.com/pdf/applications/AN047.pdf
Looks like all it takes is an R-network.
Also see page 3...
www.fairchildsemi.com/an/AN/AN-5029.pdf
...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at
http://www.analog-innovations.com | 1962 |
The only thing bipartisan in this country is hypocrisy
John Larkin
Guest
Fri Mar 12, 2010 12:51 am
On Thu, 11 Mar 2010 20:45:17 -0000, "Andrew Holme" <ah_at_nospam.com>
wrote:
Quote:
Does such a chip exist? I can't find it.
MAX9375 or PTN3310 = Single LVDS to LVPECL
tb5d1m = Quad LVTTL to LVPECL
MC10H352-D = Quad CMOS to PECL
sn65lvelt22 = Dual LVTTL to LVPECL
MC100EPT622-D (obsolete) = 10-bit LVTTL to single-ended LVPECL and I only
want single-ended LVPECL to drive 4 data inputs on a MC100EP142
Any suggestions?
TIA
You could use something like a 10EP17, a quad ecl diff-in, diff-out
buffer. Its inputs are (probably...check me on this) LVDS compatible,
and its outputs are ECL.
If your LVDS common-mode voltage is too low to meet the common-mode
input specs of the EP17, you can power it from +3.3 and -2 or
something like that. It's good to 5.5 volts.
John
Andrew Holme
Guest
Fri Mar 12, 2010 1:47 am
"John Larkin" <jjlarkin_at_highNOTlandTHIStechnologyPART.com> wrote in message
news:1vrip5hjima4kfimk9lpefq85iflqdl94t_at_4ax.com...
Quote:
On Thu, 11 Mar 2010 20:45:17 -0000, "Andrew Holme" <ah_at_nospam.com
wrote:
Does such a chip exist? I can't find it.
MAX9375 or PTN3310 = Single LVDS to LVPECL
tb5d1m = Quad LVTTL to LVPECL
MC10H352-D = Quad CMOS to PECL
sn65lvelt22 = Dual LVTTL to LVPECL
MC100EPT622-D (obsolete) = 10-bit LVTTL to single-ended LVPECL and I only
want single-ended LVPECL to drive 4 data inputs on a MC100EP142
Any suggestions?
TIA
You could use something like a 10EP17, a quad ecl diff-in, diff-out
buffer. Its inputs are (probably...check me on this) LVDS compatible,
and its outputs are ECL.
If your LVDS common-mode voltage is too low to meet the common-mode
input specs of the EP17, you can power it from +3.3 and -2 or
something like that. It's good to 5.5 volts.
John
OnSemi do a MC100LVEL17 which is 3V3 PECL levels.
Their PECL data sheets show internal input pull downs to VEE and talk about
inputs having default states if left open. This sounds like it's OK for the
input to go all the way down to VEE; but then the DC characteristics give
min/max VIH/VIL that are strictly PECL spec. They also have an app note
AN1503-D which gives Spice info including input / output internal schematics
which look like they'd be totally screwed by rail-to-rail inputs.
Jim Thompson
Guest
Fri Mar 12, 2010 2:10 am
On Thu, 11 Mar 2010 23:47:39 -0000, "Andrew Holme" <ah_at_nospam.com>
wrote:
Quote:
"John Larkin" <jjlarkin_at_highNOTlandTHIStechnologyPART.com> wrote in message
news:1vrip5hjima4kfimk9lpefq85iflqdl94t_at_4ax.com...
On Thu, 11 Mar 2010 20:45:17 -0000, "Andrew Holme" <ah_at_nospam.com
wrote:
Does such a chip exist? I can't find it.
MAX9375 or PTN3310 = Single LVDS to LVPECL
tb5d1m = Quad LVTTL to LVPECL
MC10H352-D = Quad CMOS to PECL
sn65lvelt22 = Dual LVTTL to LVPECL
MC100EPT622-D (obsolete) = 10-bit LVTTL to single-ended LVPECL and I only
want single-ended LVPECL to drive 4 data inputs on a MC100EP142
Any suggestions?
TIA
You could use something like a 10EP17, a quad ecl diff-in, diff-out
buffer. Its inputs are (probably...check me on this) LVDS compatible,
and its outputs are ECL.
If your LVDS common-mode voltage is too low to meet the common-mode
input specs of the EP17, you can power it from +3.3 and -2 or
something like that. It's good to 5.5 volts.
John
OnSemi do a MC100LVEL17 which is 3V3 PECL levels.
Their PECL data sheets show internal input pull downs to VEE and talk about
inputs having default states if left open. This sounds like it's OK for the
input to go all the way down to VEE; but then the DC characteristics give
min/max VIH/VIL that are strictly PECL spec. They also have an app note
AN1503-D which gives Spice info including input / output internal schematics
which look like they'd be totally screwed by rail-to-rail inputs.
Just because an input is _allowed_ to go all the way to VEE doesn't
mean it will necessarily produce a legitimate output.
Since what I've previously posted showed resistive level shifting I
doubt that you can directly make a connection from LVDS-out to
PECL-in.
Study what I've posted AND the LVDS and PECL data sheets.
...Jim Thompson
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at
http://www.analog-innovations.com | 1962 |
The only thing bipartisan in this country is hypocrisy
John Larkin
Guest
Fri Mar 12, 2010 3:12 am
On Thu, 11 Mar 2010 23:47:39 -0000, "Andrew Holme" <ah_at_nospam.com>
wrote:
Quote:
"John Larkin" <jjlarkin_at_highNOTlandTHIStechnologyPART.com> wrote in message
news:1vrip5hjima4kfimk9lpefq85iflqdl94t_at_4ax.com...
On Thu, 11 Mar 2010 20:45:17 -0000, "Andrew Holme" <ah_at_nospam.com
wrote:
Does such a chip exist? I can't find it.
MAX9375 or PTN3310 = Single LVDS to LVPECL
tb5d1m = Quad LVTTL to LVPECL
MC10H352-D = Quad CMOS to PECL
sn65lvelt22 = Dual LVTTL to LVPECL
MC100EPT622-D (obsolete) = 10-bit LVTTL to single-ended LVPECL and I only
want single-ended LVPECL to drive 4 data inputs on a MC100EP142
Any suggestions?
TIA
You could use something like a 10EP17, a quad ecl diff-in, diff-out
buffer. Its inputs are (probably...check me on this) LVDS compatible,
and its outputs are ECL.
If your LVDS common-mode voltage is too low to meet the common-mode
input specs of the EP17, you can power it from +3.3 and -2 or
something like that. It's good to 5.5 volts.
John
OnSemi do a MC100LVEL17 which is 3V3 PECL levels.
Their PECL data sheets show internal input pull downs to VEE and talk about
inputs having default states if left open. This sounds like it's OK for the
input to go all the way down to VEE; but then the DC characteristics give
min/max VIH/VIL that are strictly PECL spec. They also have an app note
AN1503-D which gives Spice info including input / output internal schematics
which look like they'd be totally screwed by rail-to-rail inputs.
The EP input is just an NPN differential pair. The min/max specs are
for single-ended inputs working against Vbb, which isn't what I
suggest. LVDS is differential and doesn't go rail-to-rail.
Assuming a 3.3 volt supply:
The LVDS common-mode voltage will be around +1.2 and the lines will
wiggle a couple of hundred millivolts. The minimum common-mode voltage
on the EP17 is spec'd as VEE+2, which is +2 if you run the EP from
+3.3 and ground. So direct LVDS drive is out of spec.
But if you run the EP from +3.3 and -2, then the input common-mode
extends down to VEE+2, namely ground. Then it should be happy with
direct LVDS diferential inputs.
The EP17 outputs will still be ECL referenced to +3.3, namely PECL.
John
