Twiddling Thumbs

Sent emails to all your past customers letting them know that if they need you, you're available?

Maybe it's time to start your own semiconductor company -- figure out what's missing from the TinyLogic lineup (a decent 3-state phase comparator would be my first choice), make it, and sell it.

I've toyed with that idea. ...Jim Thompson

During High School years, I had a night time job delivering Pizzas. During really slow periods, we got creative. A look at the previous month's orders gave phone numbers and addresses.

We'd call the numbers to confirm a non existent order. These calls triggered an impulse to buy in about forty percent of the people called. Apologies were given to all for the 'mix up' in the receipt boxes.

It's not something I'm particularly proud of having done....the need for disposable cash made it happen. Am I going to Hell with all the Monsanto executives?

mike

I want a low-skew true/complement buffer or gate.

I have the opposite problem, so much work I can't see straight. Mostly new projects from old customers, but some new ones too. It's astonishing how much business you can get from a few important people who like you.

I just hired two kids, one male MSEE from around here, one female BSEE who recently graduated from a college in Mexico. Both are smart and nice and seem to have really good analog instincts, but training them of course takes more time than doing things myself; it's an investment.

would this work?

-Lasse

Well-designed XOR's do that... particularly ECL/PECL. ...Jim Thompson

Training today's crop of engineering graduates analog is quite a challenge. ...Jim Thompson

Can you design an ASIC array that hashes the Scrypt algorithm at better than 3 megahashes per second and consumes less than 40 watts, and comes in at under a 100 dollar price point?

That's not analog. ...Jim Thompson

Oh...for some reason I thought you did digital too. SORRY!

I ordered one of the recently released Scrypt cryptocurrency miners from Shenzen, just out of curiosity. Who knows if anything will actually arrive in my lifetime:

I think the difficulty has already increased from when the first batch shipped. At 1.5 MH/s one could be earning a whole $1.50 a day in cold hard doge:

Job for an FPGA?

John Larkin Wrote in message:

I guess even FPGAs are too inefficient to make using them worthwhile. To get the kind of hash rates they advertise at the power usage they do it has to be a custom, hand optimized chip. I don't really know enough about digital ASIC design to know what I'm talking about though.

I can do mundane clocked combinational stuff, but I don't do any of the digital descriptive languages. I have an associate, Todd James, in Columbus, OH, who cranks that stuff out for me.

...Jim Thompson

Time to start writing that book we were talking about a few years back.

No time like the present.

Cheers

Phil Hobbs

I thought I'd wait until 85 >:-} ...Jim Thompson

Lazy slob. I started in 1994, and sent it off to the publishers in 1998.

Cheers

Phil Hobbs

Probably so. The gate-gate skew is probably low, and rise/fall asymmetry would usually be OK, with such a fast part.

...snip...

I've toyed with that idea. ...Jim Thompson

There is a REAL demand for FAST turn around miniaturization to create a first-time=operating ASIC, in small volume.

How to do.

Take a look at David Lam's compnany, Multibeam, a direct write - no mask opeartion using e-beam. Say hello from me when you're there.

What I'm saying is combine the ability to create ANYTHING you want per wafer without investing in a mask, with your ability to design the ASIC so it works the first time thru, and seems there's a niche market.

If you cn't do the ASIC design, because of high demand, at least create the tools [sequence] to enable others to do such.