Thru hole vs surface mount R's at ~100MHz.

o 100

?=20

That would be a good resistor to few hundred MHz. At 2 GHz it is more=20 then half way inductive. Surprisingly there is quite a lot of ringing.

Do they switch the power off afterhours in SF ?

VLV

Reply to
Vladimir Vassilevsky
Loading thread data ...

They just take in the sidewalks ;-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

What, shut down the tens of thousands of cocktail blenders? I don't think so.

John

Reply to
John Larkin

Phil Hobbs Inscribed thus:

I recall that information from an article published quite some time ago in MW&RF. But hey I'm still learning. :-)

Highest I've ever played with was 24Ghz. Yes I could get a signal with modulation to travel about 80ft and detect the trees moving in the wind. But apart from that, there ain't no one to talk to up there...

--
Best Regards:
                Baron.
Reply to
baron

On a sunny day (Mon, 15 Nov 2010 15:48:11 -0800) it happened John Larkin wrote in :

That is a nice test, very informative, but your soldering sucks :-) I am using a couple of those type of resistors as dummy load at 27 MHz, soldered to a PL259.

Reply to
Jan Panteltje

Written by a guy who can not draw a legible schematic !!!

h

Reply to
hamilton

On a sunny day (Tue, 16 Nov 2010 06:55:11 -0700) it happened hamilton wrote in :

I hope you realize that if the schematic is great, but the soldering sucks, that then it will fall apart. If the schematic is not readable by you, but the hardware works, then there is no problem :-)

BTW I *can* draw clearer schematics, you just did not pay for it.

Reply to
Jan Panteltje

Dummy load using puny 0.25W resistor attached to a connector with RF parameters not even stated for it. The 27 MHz is essentially DC, try doing that at 800 MHz.

VLV

Reply to
Vladimir Vassilevsky

Indeed. Consider the clearness/cost ratio as a figure of merit.

Reply to
Spehro Pefhany

On a sunny day (Tue, 16 Nov 2010 08:20:24 -0600) it happened Vladimir Vassilevsky wrote in :

Hey, the topic was about ~ 100MHz. May even work at 100.

I am sure it does not work right at 10 THz either. hehe

Reply to
Jan Panteltje

On a sunny day (Tue, 16 Nov 2010 09:39:24 -0500) it happened Spehro Pefhany wrote in :

I guess Hamilton is just pissed because Vettel won F1 ;-)

Reply to
Jan Panteltje

I learned to solder when I was three years old. It's my photography that sucks.

John

Reply to
John Larkin

On a sunny day (Tue, 16 Nov 2010 10:26:18 -0800) it happened John Larkin wrote in :

I see nothing wrong with da picture.

Reply to
Jan Panteltje

OK, OK, enough sarcasm from you lot. Here's the 0805:

ftp://jjlarkin.lmi.net/47R_0805.JPG

ftp://jjlarkin.lmi.net/47R_0805_TDR1.JPG

ftp://jjlarkin.lmi.net/47R_0805_TDR2.JPG

On that last one, you can see the hardline, the SMA stuff, the capacitive bump where it hits the PCB, the short trace, then the inductive bump of the resistor. The resistor bump is about 30 ps FWHM, not bad for a 1 cent resistor.

There's a method somewhere to compute L given the area under the bump. Some scopes do it for you.

I've been doing customer support all day and just *had* to solder something.

Hmmm, maybe I should flip the resistor over...

John

Reply to
John Larkin

Yikes!

ftp://jjlarkin.lmi.net/0805_normal.JPG

ftp://jjlarkin.lmi.net/0805_flipped.JPG

John

Reply to
John Larkin

Say John,

Ever taken a look at resistive splitter networks? My boss was mentioning this morning that it had occurred to him that, rather than using the traditional "Y"-style splitter (...but extended to 3 or 4 outputs, since that's what we normally need...), a "Delta"-style splitter could be better because the resistor values go up as you increase the number of outputs (to Z0*(n-1), where n is the number of outputs ) vs. dropping and then coming backward towards Z0 with the Y-style (to Z0*(n-1)/(n+1) ). Hence, if you assume your dominant parasitics are a bit of stay inductance, the resistor looks more ideal Delta-style (e.g., for a 3-way splitter, Y-style you might have resistors that are, oh, say, 25+j10 ohms, but Delta-style they'd be 100+j10 ohms -- closer to "ideal").

He made a test board and said it was looking quite promising in the frequency domain, although the layout was very "uninspired": Not symmetrical and parts only at 0 and 90 degree angles. (...this is why I prefer to lay out my own boards when we're talking really simple circuits like this!... ) He's planning to make a few more with differently-sized resistors... and I've been suggesting using a more symmetrical layout with angled resistors, but he claims that the manufacturing guys hate that because pick and place machines need extra hand-holding if you angle your components at, e.g., 45 degrees or whatever. (Anyone know how much of a problem that is these days?)

The Microwaves 101 guys claim that Delta-style splitters "become a nightmare for more than a 2-way split," but that's really not the case at all. I'm half-tempted to write up a page and submit it to them, but having some more test results to go along with it would be good.

---Joel

Reply to
Joel Koltner

I've done boards with goofy angles. Production didn't mind.

ftp://jjlarkin.lmi.net/AmpTop.jpg

Is there a sneaky geometry for a multiport delta splitter that does all the output port resistor routing without a tangle of vias and traces?

Is there a ring splitter? That would be one input port, N output ports. There would be a resistor from IN to each OUT, and bridging resistors between adjacent outs only.

Rum and math don't get along so well. The rum wins. Next course is cracked crab and beer.

John

Reply to
John Larkin

Are you cleaning ALL the flux out from under the chip?

If no cleaning after the solder attachment takes place, the results will be skewed by leakage and other artifacts that unclean test scenarios can present.

At the bench, heated IPA at 99% or a good brominated solvent (ensolve) with some warmth.

The heat facilitates not only better cleaning, but quicker and more complete evaporation.

This knowledge comes from years of debugging SMD HV multiplier circuits.

Another way to make it easier to perform tests is to place "solder bumps" on the pads where the part goes, and attach the part spaced away from the board a half mm or so. This allows quick, complete clean ups between test scenarios.

Reply to
TheGlimmerMan

We had one where the layout guy got the pins on a SOT-23 footprint 120 degrees out. The SMT stuffer had no problems with the rotation and even though the spacing wasn't quite right, it worked. It got fixed before the product made it to the customer, though. EMI compliance made sure of that. :-(

Reply to
krw

For 3-ports, almost: The end result is a Y inside a delta (6 resistors total, all 100ohms for a 50ohm system impedance). The edges of the delta are the outputs and the center point inside of the Y is the input, so you just need one via down to your microstrip or whatever. (...or if you use a big enough resistor, you might just route the input straight through the gap between one resistor's leads!...)

For 4-ports if you follow a similar approach to the topology, you get four resistors in a cross around the center input point, and then the tips of the cross also connect (draw them at 45 degree angles) to each other... but you also then need resistors from the top tip to the bottom and from the left tip to the right. That's 10 resistors total (150ohms each), and it's still relatively clean; for a symmetrical layout you'd likely put a via at each tip so that the top/bottom and left/right tip resistors could be on the bottom side of the board.

The number of resistors is the sum of n, n-1, ... 1 where n is the number of ports.

So we figure 4 ports is about the practical limit -- which is often the case with a Y-style splitter too: when you consider 12dB power hit, even if the performance was really good for, say, an 8- or 16-port splitter, you start needing pretty high-powered amplifiers behind the splitter, which gets expensive fast compared to just using, e.g., amp->4 port splitter->4 amps->4 more 4 port splitters where the cracked crab and beer.

I'm told there's key lime pie for dessert here tonight!

---Joel

Reply to
Joel Koltner

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.