ecl comparator hysteresis

I'm going to use an ADCMP565 ecl-out comparator. The latch enable pins can be used to program hysteresis.

formatting link

On the A section, the hyst pins are sometimes called LEA and \LEA, and in fig 6, LE and LEB.

Has anyone used these with hysteresis? If I ground LEA pin 5, should I apply a positive or negative voltage to \LEA pin 7 to get hysteresis?

This is tricky enough that I can imagine getting it wrong. Fig 10 could have helped, but doesn't.

ADI publishes horrible data sheets. Fig 7 is typical.

Does fig 8 make any sense? Wired-OR of differential ECL?

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin
Loading thread data ...

Looks straight forward to me. The LEB in Fig.6 mean LE "BAR"or /LE. To adju st hysteresis Fig. 10 shows LE pulled hard high by connecting to GND. Then the resistor divider voltage is applied to /LEB ==LEB for the adjustmen t. LE-LEB negative means /LEB is pulled so many tens of mV above GND. Not m uch hysteresis to speak of pulling LEB below GND.

Makes perfect sense for a window comparator. It's not an OR, it's an AND. I nput abouve lower thresh AND below upper thresh gets a logic "1" output, et c.

Reply to
bloggs.fredbloggs.fred

Assuming the polarity is right. The data sheet sure doesn't inspire trust.

Differential ECL is interesting when both lines are pulled high. Treating the outputs as single-ended, I guess it might work.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

djust hysteresis Fig. 10 shows LE pulled hard high by connecting to GND. Th en the resistor divider voltage is applied to /LEB ==LEB for the adjust ment. LE-LEB negative means /LEB is pulled so many tens of mV above GND. No t much hysteresis to speak of pulling LEB below GND.

. Input abouve lower thresh AND below upper thresh gets a logic "1" output, etc.

ECL shouldn't be "puling" the lines high, it is pushing them. And even thou gh you don't need the complementary output, it's still a good idea to load it to maintain current balance on the power supply distribution in the chip and eliminate inductive ripple you can't decouple.

Reply to
bloggs.fredbloggs.fred

OK, from now on we'll all rename "pullup resistor" to "pushup resistor."

I fixed the hysteresis problem.

formatting link

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

ECL load resistors are pull-downs. The standard ECL termination is 50R (or 75R) to -2V.

The ECL output is the emitter of a (fast) NPN transistor. "1" is about -0.8V, "0" about -1.6V. The transistor is always carrying some current, and it never saturates.

Open collector TTL uses pull-ups. The TTL output is a transistor used as saturating switch, to get a "low" of about 0.4V, by pulling down at resistor returned to the +5V rail.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

ins

and

d I

s?

o adjust hysteresis Fig. 10 shows LE pulled hard high by connecting to GND. Then the resistor divider voltage is applied to /LEB ==LEB for the adj ustment. LE-LEB negative means /LEB is pulled so many tens of mV above GND. Not much hysteresis to speak of pulling LEB below GND.

AND. Input abouve lower thresh AND below upper thresh gets a logic "1" outp ut, etc.

Fine if it works for you, but overcomplicated compared to the figure 10 sch eme.

What kind of FET is driven by that minuscule ECL output swing?

I can't believe this super high speed gizmo is available as a dual.

Reply to
bloggs.fredbloggs.fred

Just one more resistor, in case the polarity is wrong, ideally not stuffed.

It's a Mini-Circuits enhancemet mode PHEMT. They turn on pretty good around half a volt, almost like a bipolar NPN.

Most of the good ECL comparators are duals. I'm wasting the other half.

Not many people make fast (sub 1 ns) comparators any more. Maxim gets close, but theirs are weird.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

e pins

A, and

ould I

esis?

0

. To adjust hysteresis Fig. 10 shows LE pulled hard high by connecting to G ND. Then the resistor divider voltage is applied to /LEB ==LEB for the adjustment. LE-LEB negative means /LEB is pulled so many tens of mV above G ND. Not much hysteresis to speak of pulling LEB below GND.

an AND. Input abouve lower thresh AND below upper thresh gets a logic "1" o utput, etc.

scheme.

Okay, was wondering about that. Don't believe the hype about independence. Bias the inputs to the unused comparator so everything is in a solid logic state. You never know what could be "shared."

Reply to
bloggs.fredbloggs.fred

I can't go skiing; too much snow to drive up. So I designed this little gadget to stay amused. The problem now is that all those fast analog and ECL parts and power supplies are going to get very hot.

This board will slide into a little Hammond extruded box.

formatting link

That will be horrible, thermally. 150C rise maybe.

I could use a chunk of gap-pad between the bottom of the board and the box, but that would be hard to install or remove. The gap-pad stuff is soft and sticky and needs to be compressed.

Hmmm, the stackup

PCB bottom - sticky GapPad - Aluminum Plate

would slide in and out pretty easily, and not waste pcb surface area.

formatting link

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

Right, I'm applying a healthy diff voltage to the unused input pair. Outputs are not connected, shouldn't matter.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

Huh? ECL's output is an NPN emitter. It doesn't "push" anything.

Reply to
krw

70mA on -5.2V, 430mW total, Ouch. What're you doing that needs a 300ps propagation delay?
--
 Thanks, 
    - Win
Reply to
Winfield Hill

e pins

A, and

ould I

esis?

0

. To adjust hysteresis Fig. 10 shows LE pulled hard high by connecting to G ND. Then the resistor divider voltage is applied to /LEB ==LEB for the adjustment. LE-LEB negative means /LEB is pulled so many tens of mV above G ND. Not much hysteresis to speak of pulling LEB below GND.

an AND. Input abouve lower thresh AND below upper thresh gets a logic "1" o utput, etc.

scheme.

At 0.3K/W that gap pad looks the miracle material. Do you need silicon grea se between the pad and the aluminum?

Reply to
bloggs.fredbloggs.fred

ins

and

d I

s?

o adjust hysteresis Fig. 10 shows LE pulled hard high by connecting to GND. Then the resistor divider voltage is applied to /LEB ==LEB for the adj ustment. LE-LEB negative means /LEB is pulled so many tens of mV above GND. Not much hysteresis to speak of pulling LEB below GND.

AND. Input abouve lower thresh AND below upper thresh gets a logic "1" outp ut, etc.

hough you don't need the complementary output, it's still a good idea to lo ad it to maintain current balance on the power supply distribution in the c hip and eliminate inductive ripple you can't decouple.

It doesn't pull anything either. The NPN is the push part of a push-pull co mplementary output, so I just extended the terminology to this case.

Reply to
bloggs.fredbloggs.fred

That 3G Shielding material is about 5 watts/mK, so a few square inches of thin stuff conducts a lot of heat. The material that I have is sticky/tacky on both sides, so it will stick to the PCB and the aluminum pretty well. The other side of the aluminum plate will press against the inside of the extruded box, which should conduct heat pretty well. Silicone grease would be a mess.

Here's my board so far. I managed to get all the parts on the top side.

formatting link

I like to lay out a board now and then, but I'd rather be skiing. But I made a dynamite batch of turkey broth too.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

I don't see how you're going to slide the board with its sticky/tacky stuff, pressing hard and compressing against the bottom of the extruded box, in and out of the box.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

The sandwich is PCB, gap-pad, aluminum sheet, stuck together.

The idea is to assemble and squeeze the stack just before installation. The gap-pad is foamy and compresses, to half or maybe a third of its relaxed thickness. Now the stack slides into the extruded box. Then, over maybe 10 minutes, the foam expands and pushes the aluminum plate against the bottom of the box.

I guess the alum plate could be skipped, but that would make PCB installation tricky and removal impossible.

That 3G stuff is great. It will compress to a fraction of its native thickness then slowly recover. It needs to be compressed to maybe 0.5 or 0.6 of its basic thickness to meet its theta specs; they don't say so, so I tested some. Of course, squeezing it is a double benefit: specific thermal conductivity increases and thickness decreases. I wouldn't want or need a lot of force on my PCB, so I'll keep it reasonable.

I wish these little boxes had removable lids. We may design our own to replace the stock extrusion.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

niptechnology.com:

formatting link

Reply to
Lasse Langwadt Christensen

how good is its thermal conductivity when not compressed?

--
  When I tried casting out nines I made a hash of it.
Reply to
Jasen Betts

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.