Amazing Discovery

Sorry but thats waaay too high! Let someone else do the soldering. For hand soldering 330 C is the ideal temperature. At higher temperatures the flux will burn up too quick.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

John, Let us know if you find out anything interesting about the 2412. I keep them around as band-aids for lab groups who can't seem to get a TTL signal to travel reliably across the lab (throgh 100 feet of cable, taking a few turns around a pump, etc.) The 1414-fiber-2412 solution is expensive and slow, but it usually works. Until the idiot physicists tie a knot in the fiber...

-Jim MacA.

[snip] [snip]

You'd think Larkin would have a stronger wrist ;-) ...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

| 1962 | I love to cook with wine. Sometimes I even put it in the food.

location?

Maybe it's the batch of copperclad that I have, but I can't get that to work. The copper is stuck too hard and tears off in bits. I can peel it if I heat it with a soldering iron, but this score-and-rub thing takes less hands.

I used to have a sheet of copperclad teflon, microwave stuff. It would cut easily, and then peel beautifully. 50 ohm traces were wide, too. It was soft, more like copperclad leather than fiberglass.

John

Excellent! Nice to hear you got a good deal. And, btw ... double drool!

Ed

I sell a little e/o converter gadget, the "J720" tube in the pic

ftp://jjlarkin.lmi.net/BB_HFBR.JPG

whose typical output is about one milliwatt, maybe two, depending on how hard you drive it. I have a customer at LLNL who is using these with 2412s as the receivers. He called me and is reporting very long prop delays, like microseconds sometimes, and was asking for ideas. We conjectured that the HFBRs get weird at high optical input levels, so I got some and set it up.

I can't see anything radical going on. A low-level (say, 2 volts, fraction of a mW optical) 100 ns positive electrical pulse into the e/o makes about a 150 ns negative output pulse from the HFBR. As the drive increases to 4 volts, probably about 2 mW optical, the early/falling edge gets sooner by maybe 20 ns, and the rising/later edge gets later by about 30. No surprises.

We make o/e receivers that are much faster than the 2412s, but they are a lot more expensive. We use a silicon or GaAs PIN diode, a tia, and a fast comparator. We get around the need to design a low-noise high-gain front end by blasting them with a milliwatt of light, and get prop delays of a few ns and link jitters like 10 ps RMS. The 2412 has a lot more gain, so can work at -30 dBm or so.

He's driving our e/o with an SRS DG535 delay generator, and I'm inclined to blame that. If he would just buy our DDG and our optical receivers, he wouldn't be having problems like this, and we could go to lunch at a higher class of restaurants.

John

location?

I found that the thinner the copper, the harder it is to peel off without tearing into little bits. But, it also removes easier with heat than the thicker copper.

_Microseconds_? Maybe if there's a cracked cap in a bias network someplace, but I expect that would look more like a dropout than a delay.

There's no place to put microseconds worth of signal in there, so the only way that could happen is if something suddenly gets much much slower.n

At my PPOE I had a DG535. It was okay for what I wanted it for, but not all the output settings worked, and the jitter was fairly horrible, like

200 ps.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

Nice! One addition to the "click trail" above: MECH-14 > CLASS LESSONS > PCB Fabrication

Ed

I'm a scalpel fan, myself; a number 15 blade does a good job of scoring copper and can be wielded chisel-like to lift a bit of foil. The best tool, though, is a 2mm square-shank chisel, like the old Vector slit-n-wrap kits came with; this was intended to cut wire in tight quarters, and is rigid enough to get better tactile control than with an X-acto or scalpel.

Carbon-steel scalpel blades work better than stainless, but are more brittle than X-acto blades, so use eye protection.

I would have gone with too much optical power, too, but failing that, are you sure he's pulling up the output of the 2412 nice and hard, with something like 560 Ohms to +5V? I've accidentally installed pullup resistors in the KOhms and seen a very sluggish response. And is he powering from +5V? I know, these are a bit obvious, bit it does sound like an implementation problem with the 2412.

-Jim MacA.

It's OK, but the user interface is a PITA. Having a digital delay generator around one's lab great, even a DG535. They have a new/improved model, but I haven't played with one. Their technique has a lot of inherent insertion delay and rep-rate limits and jitter, and I appreciate their not copying ours.

The DG535 has a mode where the output is a current source, not a 50 ohm source. Maybe they're running it in that mode, and the cable capacitance is slowing things down. The numbers don't quite add up to microseconds, but it's the only idea I have.

John

I sometimes use a Dremel to make big, linear copper cuts. I hold the Dremel horizontal on the bench and push it down against some spacer, like a pad of paper, and adjust the cutter height to be a little less than the board thicknes. Then I slide the board under it to cut the copper. That works like a horizontal milling machine. Depth control is good, but you can't do fine work that way.

I use x-acto #16 blades, and I've never broken one. They are short and fairly thick, very strong.

Dremel makes little routing attachments. That, and a straightedge, might be interesting, if there is a really small diameter cutter somewhere.

John

--- If you've got a drill press, you can buy a 1/4" or so end mill, (per Speff's suggestion) chuck it up, set the quill stop so that the end mill doesn't go much past the copper and just spot-face off the copper wherever it is you want to glue down the dead bug.

For mini end mills with a 1/8" shank, which will fit into the large Dremel collet,:

-- JF

Yay! Thanks for the link, John.

--
My pleasure. :-)

Thanks. Now I need a small, affordable lathe for some projects.

--
It's easy to think outside the box, when you have a cutting torch.

Haven't checked to see if the price is better, but McMaster has 'em pretty cheap if you ever need larger quantities:

Yup, just heard from LLNL: it was the DG535 being weird.

John