Help duplicating a smd pcb

I'm surprised, I got know one's opinion. The probe is good up to 1.2 GHz, I don't expect mine to be that good, but I'm trying to do well. I ordered thick film SMD's resistors, because of price and availability. Is there a better choice for resistors at a GHz?

Mikek

Strange, something on that page is crashing my Chrome.

Oh hell... looking at the source, it contains NO CLOSING TAGS. I wonder if Chrome might be stack-overflowing or something?

It's also HTML 3.2? Was that *valid* back then?! (Checking, I guess it was!) Holy hell, how long has this website been up for?!

OTOH, at least the source is readable!

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com

Doubt that is the cause. html5 doesn't require an end tag on .

Yes, it was probably more valid then.

Is there any reason it wouldn't be? "

Yes. Why create all that capacitance ? Most likely a simple single sided board. Actually using SMDs there is less inductance floating all over the place and no holes to drill.

I say it would be a waste of copper and process.

Thanks for the input. Mikek

Most likely no full ground plane (see jurb's post for reasons).

Other similar probes may have some partial ground planes or shielding traces. The RF front end (where the diodes are) should be free from ground plane capacitances however.

Here's a picture of the inside of another such probe, rated 100kHz to

1.2GHz. It's not from Bontoon, but Tonghui and belongs to a TH2268 type millivoltmeter. The construction is a little simpler and the output is asymmetric (goes to a BNC), but the working principle is more or less the same. It's also housed in an aluminum tube, only that the front end has a smaller diameter. It does work at 1GHz, but I don't have a reference meter to reliably check the calibration up there.

As you see, there's some rather myopic board edge cutting, but it seems to have just enough ground trace left to still work. They decided to put a shield trace right under the first resistors, then again under the second one, probably to keep the RF out of the DC path. That's easily done with through-hole parts, but could be done with 1206 size SMD resistors too if needed.

Regards Dimitrij

I ended up removing the ground plane under the diodes and the first resistors and left it for everything else. Waiting for my parts to arrive. Thanks, Mikek

On a sunny day (Fri, 28 Feb 2014 00:59:47 +0100) it happened "Dimitrij Klingbeil" wrote in :

Mine is even simpler, tested to several GHz:

Somehow I deleted Jorge's message, so I can't respond to it. Here's mine, unbuilt, untested, waiting for parts. I have since redesigned the pcb to get rid of the 1/2in of wire One more sleeve of heatshrink, over BNC, but installed so tip can be

removed. Note: I artificially added part layout and cable using Paint.

I think the input to my Boonton is differential? So it takes two plus ground at the probe output.

I expect to drill a small hole on each side of copper tube so I can solder ground on PCB to copper pipe.

Then I need to find someone to test and calibrate. I'll see what my local state college has to help. Want do it on the cheap.

btw. I layed out a pcb and tried the printed page ironed on to copper and I couldn't get the toner to stay on the copper. I have special paper, but I think my toner cartridge is on empty or the paper is just to old. I've done it many times, tried twice just didn't stick. I ended up exacto knifing the whole thing.

Mikek

Mikek

Several things to note:

- You need to get the copper really clean... and micro-roughening it helps. Scrubbing with a green abrasive pad helps. Scouring powder helps, if rinsed very thoroughly. Dipping briefly in etchant and then immediately rinsing helps.

- Reportedly, Brother printers don't work well for this... something about the toner formulation. Classic HP printers work pretty well.

I've done a few this way. Next one I do, I'm going to laserprint on a transparency and use presensitized copperclad and a fluorescent light, and see how that process works.

I did use the green abrasive pad, didn't do an etchant dip.

Well, I do have a Brothers printer, but I have made many boards on it. My toner cartridge is a refill, and it has been long enough since I made boards that I don't remember if I always used refills or new cartridges for boards.

I used Expresspcb as my cad program. Anyone know how to copy and past a board? I'd like to put 10 circuits on one board. Mikek

Never mind special paper. I've found that the kind they print supermarket fliers on works well. Just make sure it's glossy or semi-glossy. Print right over the ink, it won't come off. Soak in hot soapy water for a while and peel it off. Some experimentation will be needed.

- YD.

Here's mine, unbuilt, untested, waiting for parts.

I have since redesigned the pcb to get rid of the 1/2 in of wire One more sleeve of heatshrink, over BNC, but installed so tip can be removed. Note: I artificially added part layout and cable using Paint.

I think the input to my Boonton is differential? So it takes two plus ground at the probe output.

I expect to drill a small hole on each side of copper tube so I can solder ground on PCB to copper pipe.

Mikek

Maybe you didn't apply enough heat? I use a hotted-up laminator, and it still needs 3 passes.

A friend imports his ExpressPCB designs into Robot Room Copper Connection to make Gerbers. If you take that path, there's a variety of Gerber tools that could help.

Clifford Heath.

There are some suspiciously stripline-looking things in there...

--
"Design is the reverse of analysis" 
                   (R.D. Middlebrook)