delay line logic

something like this?

-Lasse

Not necessarily cheaper but better. I have designed out numerous silicon delay lines because their phase noise spectrum rivaled that of Niagara Falls.

Never heard that one but back then most delay lines were made by Belfuse or by Kappa.

In the middle ages I did a PLL (to decode Manchester) that used gates and two delay lines (the digital-in/digital-out type) to create two VCO's that could be "jam" locked, giving direct synchronous clock creation.

...Jim Thompson

Cool. Thanks. Google done let me down somehow.

John

Yup. I used one of the Maxim parts once and jitter was ghastly.

Seems the one I was asking about is a Maxim silicon part.

John

this?

might not be the same part I played little around with the part number to find it

-Lasse

I've had a case where phase noise was so bad that their newly develoepd system was even failing self-cal regularly. Traced it to a bunch of silicon delay lines of the ritzy kind, at about $65 a pop. Replaced the whole enchilada with some servoed discrete phase shifters. Phase noise gone, system always calibrated within milliseconds and as a bonbon this area of the board (now widely vacant) went from $200 to $20. Their CFO really liked me :-)

on

I never liked those chips and did what little I could do at the design review to stop the product. Jitter was on my list of complaints. But marketing people know better. ;-) The Dallas parts have the same issues. Think of comparing a relaxation oscillator to a tuned circuit.

As I've mentioned here before, the 232 chips with onboard charge pumps have a ton of jitter on them. I laugh at the designs I see on the net that can just cheat and not use a 232 chip, but go out of their way to add a 232 with onboard charge pump because they think the design will be better. Now if you have to drive a long line, yeah, get a 232 chip. But for a few feet, that is nuts.

I like SN75155 as an RS232 transceiver. No charge pump, cheap, clean, tough.

John

That would be the "Cheapskate Financial Officer"? ;-)

I've never used any of those. Always ye olde 1488/89 plus rolled my own voltage converters. Why is it that younger engineers seem to shy away from inductors so much? Do they teach at universities that evil spirits live in there?

That client was quite generous. After another one of those "Oops, we accidentally dropped the cost by another hundred bucks" they sprung for a kegger at the park on a Saturday. The good stuff, microbrew.

If all you've got is 3.3 or 5 volts, an integrated charge-pump RS232 chip can be handy. It would take a fair amount of parts to generate

I've done the RS232 output thing with an opamp section, and the input with just a resistor.

John

Yep, I always liked "ye olde 1488/89" ;-)

Seems there's some rule against ultra-reliable... "cheap" is king :-(

...Jim Thompson

On a sunny day (Thu, 16 Apr 2009 08:36:41 -0700) it happened John Larkin wrote in :

If you are lucky and have hardware and software access to DTR and RTS, then this will work fine:

Picture of the 'board':

By the time universities are teaching circuit design (rather than just "generic circuit analysis" using KVL/KCL), these days most of them are concentrating on IC design, where -- barring some RF designs where the inductors are in the single-digit nH range or less -- you're very strongly discouraged from using inductors, since they'll have to be "outboard." :-(

Many EE students probably never even see a buck converter unless they specifically opt to take some of the "power" courses.

I doubt any university course even mentions things like magnetic amplifiers anymore.

Still, on the bright side, the resources for self-education today are greater than ever before, so anyone who really wants to learn all about magnetics has a better shot than they might have 20+ years ago.

---Joel

I saw that coming in the early 90's. This was when I got the first design from a new client that "didn't work for production". It had a cleverly designed IC on there. Problem was, this IC assumed near ideal outside conditions, no noises other than Gaussian, no cell phone rat-tat-tat, and so on. The only option I saw was rolling a discrete design and the client had to write off the (really expensive) mask costs and such. Still in production ...

Coincidentally it was this project where a Fairchild sales engineer tried to talk me out of using CD4000 logic, and that all the new single gate-style stuff would be "the" future. Yeah, it costs more and I'd need a regulated VCC but it's the future. Other engineers would be converting in droves. And yada, yada, yada. Then I knew that I'd have lots of work for years to come :-)

True. As a student I really had to save hard in order to be able to buy something like a radar book. $100 out of a $600 monthly budget that had to cover rent, utilities, food, everything, was tough. Now a few clever mouse clicks gets you all that for free.

I can remember when we did this with discretes. Half a dozen parts, but essentially bullet-proof. ;-)

Cheers! Rich

s.

aper

licon

ara

But that part needs dual supply. My point is if you are just "talking"

To use a logic gate to receive 232, you need to pad it with a voltage divider and snub negative voltage from the logic gate input. Many of the 232 CMOS chips do just that.