View Full Version : Need advice on testing gear. Cant afford a real scope..

04-11-2012, 12:07 AM
So im at the point where Im about to layout my board in KiCad or something similar. Ive been doing alot of research on hi-speed pcb design and decoupling. Im hoping to just hand etch a 2 layer board and use as much extra copper as I can for ground and power planes to create a decent design with some capacitance. Although im not sure its worth it after doing the math and realizing I cant find thinner double layer copper cad. Id really like .8 or thinner with 2 or 3oz copper planes if anyone has a decent source for that.

Anyways I would like to be able to check out my voltage waves going into the propeller and see what difference different pcb designs and decoupling caps make. Im assuming a scope is the only way to do this? I looked into viewport but for some reason the demo couldn't find my prop board even though spin tool and bst can. Anyways I did a little reading on the site and it looks like to me you can only view the data pins, not the actual vcc/vdd pins, and I also remember reading its only good to 80mhz prop clock. So if I can lets say get a prop doing 130mhz would I need a scope that is that fast to monitor the input voltages? Im sorry if this is an idiotic question but Ive never used a scope before. I mean I dont have but loads of cash my top of the line would have to be like 200 dollars or something, and i would like to be able to use the scope with faster arm projects in the future too. Just seems to me like the price of a ghz scope is nuts. Im hoping just a dso or cheap scope storage kit will cut it for my purposes at the moment... or better yet veiw port or some other prop based scope.

average joe
04-11-2012, 04:09 AM
IMO, 160MHZ or better scope should work. Here's something under $200 bucks right now that would do the job. http://www.ebay.com/itm/Tektronix-2465-4-Channel-300-MHz-Oscilloscope-Tested-Fully-Functional-/120890550027?pt=BI_Oscilloscopes&hash=item1c25a3670b

04-11-2012, 04:46 AM
A good scope is invaluable for all sorts of reasons. Remember you need something with at least twice the bandwidth (and prefereably more like 10x) of the signals you want to look at. With Prop2 heading above 200 MHz and ARM going even further you would really want something 1GHz or above

However good design will get you a long way, even if you can't see/measure the waveforms. I don't think its possible to get Propellers to 130MHz but I would love to be proven wrong.

Regarding thin PCBs - The capacitance between two 10x5cm pcb planes (4x2") is about 125pF, or ~250pF for 0.8mm thick PCBs. (ie much less than bypass caps). Do you really need the thinner PCB? They are available in large sheets, I think with the standard 1/2 or 1 ounce copper. You can perhaps electroplate to make it thicker if you really need to.

Duane Degn
04-11-2012, 05:17 AM
Perry made a "Poor Man's Digital Oscilloscope (http://forums.parallax.com/showthread.php?130582-Poor-Man-s-Digital-Oscilloscope-(now-with-VGA-option-and-screen-capture-to-SD))" using a Prop. I'm not sure if it has the resolution you want though.

There are some other links to using Propellers as electronics tools in post #5 (http://forums.parallax.com/showthread.php?135706-Index-Test&p=1049898&viewfull=1#post1049898) of my index.

04-11-2012, 07:54 AM
Capacitance is proportional to the surface area of the conductors (and inversely to the separation distance, as per Tubular's calculations), it's not going to be affected so much by the "weight". The higher weights like 2 and 3 oz. are generally used when more current carrying capacity is needed, as in amps, many amps.

It's not real clear why you want this capacitance anyway. You normally don't want random capacitance all over the place, this leads to noise and crosstalk.

04-11-2012, 10:28 AM
For high frequency signals and decoupling issues the thickness of the copper is (up to a limit) irrelevant - all the current flows at the surface due to the skin-effect. For instance at 10MHz the skin-depth is 18um, whereas 1oz copper is 35um thick. Adding any more copper would only change the effective resistance a few percent.

A ground plane's key properties are low inductance and providing a sink for all the stray capacitance, effectively screening neighbouring traces from each other. Even with a ground plane the stray inductance will dominate the impedances at 100MHz - the resistance of the copper isn't very relevant(*). The ground and supply planes/traces are there to provide a low-inductance path to the decoupling capacitors.

(*) When you have resonant LC circuits the resistance will matter as it sets the Q. Long transmission lines (striplines) also have losses that depend on the resistance. For such situations silver plating is sometimes used as silver has a somewhat lower resistivity than copper. The plating is in the um range.