Paul.Rako

's profile
image
Writer

Paul Rako is an engineer that writes and a writer that engineers at Rako Studios.


Paul.Rako

's contributions
  • 10.08.2013
  • True or false: High-power LEDs don’t generate IR heat in the forward direction like a filament lamp
  • Fantastic work. I agree, an ounce of trial is worth a pound of theory. To the fellow that says there is no IR from a blue-white LED, well maybe he read the spectrum output of the phosphor, not the LED itself. I did a CO2 sensor proposal for the space shuttle, and the filaments that made the IR we shined through the sapphire windows did not get even close to incandescence. You could not see them glow with the room lights out. Anything hot makes some IR. I do agree that an aluminum sheet will also turn the entire visible light spectrum to heat. But you can't argue when the author points out a clear 100C plastic melts when held over a 60C LED. And kudos to the author, for teaching us this, warning us, and even giving up what could be trade secrets about organic binders. He definitely supports my biased opinion: "Engineers, highest form of life on the planet." To figure out how much IR is pouring off the hot emitter, I would propose to reverse-bias the LED, until it avalanched, maybe 20 or 30 volts. Then modulate the current so the heat sink reaches the same heat in the same conditions. The heat is coming out of the junction, although I admit the junction is wider since the reverse voltage pulls the "plates" of the P and the N material apart. Still, you could then measure the IR going forward, and understand it is not a product of the light made in the junction.
  • 11.12.2016
  • The PCB is the most important component of your design
  • "I'm pretty sure that was promulgated so that schematics could be printed on daisy-wheel printers." Oh gosh I needed a good laugh this morning. While I have some friends that like to use Boolean expressions for PALs and such, I always like old-fashioned gates, since you can picture a box called a D-flip-flop holding a bit, and you can imagine a bit moving down a shift register with the clock. You folks have me so jazzed I will do another article about schematic symbols. I have the key figure in my head already. On the left, a plain box with pins labeled 74S34. It will be subtitled "useless". Next to that the ANSI symbol with the caption "worse than useless". Then the gates drawn out in one part, then the gates drawn individually, and then this "power gate" idea, assuming I can get OrCAD 9.2 to play nicely. And yeah GSKrasle, I too draw a little package pinout on the schematic, if only so I assigned the right footprint-- TO92-CEB TO92-BEC TO92-ECB...... I also did it so I knew what I was probing. When backside silkscreen became pretty much free, I would put the part numbers and collector base emitter initials on the backside near the pins. That was based on the great Manny Realyvasquez teaching me we don't have to always probe from the top-side, it can be easier to do it from underneath.
  • 11.12.2016
  • The PCB is the most important component of your design
  • Agree salbayeng, GSKrasle's "5-gate" idea is great. I can see how it would make gate or amp swapping easy. My worry is I will forget to place the "power gate". Still, I learned to put all the unused gates on the schematic, and I agree, sticking the power gate with the decoupling caps. And yeah, every engineer I know have been burned by invisible power pins that get hooked to a net based on name. That is why I always put the pins in the symbol explicitly.
  • 11.12.2016
  • The PCB is the most important component of your design
  • Very good point MWagner_MA. I have an ME pal that went over to his service bureau. He wanted the young fellow to re-assign the origin of the layout, I forget why. The kid looks at him with a straight face and says, "You can't do that." I can't remember if it was PADS or OrCAD, but my buddy walked the kid thought the menus, even though my pal had never seen the program before that moment. So sometimes the service bureau is not even skilled at running the tool. The kid said "Wow, that is really useful, a lot of customers ask for that." So tell the boss he has to pay you to go sit with the guy while he does placements, and maybe you can teach him enough routing discipline while you are there, or at least to have the bureau send you a drawing with placements, and maybe the rat's nest, so you can make sure they are on the right track. Doing that has much lest schedule hit and money than a re-spin.
  • 11.18.2016
  • Use plastic optical fiber for isolation
  • Jacob Beningo is a good man, I met him when he visited Silicon Valley. He lives in Michigan near the GM test track in Milford, which I have been to when I worked at GM. I love his little system. My pal at an aircraft navigation module had best luck with Silicon Labs, for its common-mode interference resistance. He said they had to worry about lighting strikes on the wings. If this is the same chip-- the one my friend used was just a simple RF swith-- a "1" on one side was made into an RF signal inside the chip, and received a few mm away and turned back into a "1". Your article is equally cool. I never knew about the Siemens IL300 analog isolator. I did know about the Linear Technology LTC1531 Self-Powered Isolated Comparator, but they obsoleted it.
  • 11.12.2016
  • The PCB is the most important component of your design
  • My schematic symbol conventions are nothing too radical. For digital chips, I always draw the gate inside the symbol, as opposed to that IEEE monstrosity convention that TI used. I also break all multi-part packages into individual gates, so the wires in the schematic don't have to go to one place despite the gates being used all over the place. In general, I put power pins on the top and ground/negative supply pins on the bottom. Input pins on the left side and outputs on the right side. I put clocks and enable pins on the left towards the bottom. For chips with buses, well then it gets interesting. If the bus has more data going out of the chip then in, I will put the pins on the right. The pins go on the left if I think the chip is mostly reading data off the bus. I have been known to put buses on the top and bottom of a bunch of chips so I can string chips along a horizontal bus, both above and below it. I have used the DeMorgan equivalent function in OrCAD 9.0 to make alternative op-amp symbols where the minus input pin is above the plus pin. If you just flip the symbol in the schematic it puts positive power at the bottom and I hate that. This gets to a good argument-- sure, you need to break a quad op-amp into the four parts, but do you put the power pins on all 4 parts? I say yes, since some engineers are not diligent, and don't explicitly place all the spare parts, so if you only put power pins on one of the 4 parts, they might get forgotten about and left off. A similar case for no invisible power pins letting the names connect to a net. That is just asking for trouble.
  • 11.18.2016
  • Use plastic optical fiber for isolation
  • Yes, you are right about MIDI data rates. The MIDI stuff is here because the article is about cheap ways to achieve isolation, and MIDI is even cheaper than POF. When I talked about "pushing it" I was not implying you could "push" the MIDI standard, which is, after all, a standard. Sorry to be unclear, I meant an engineer could buy cheap MIDI cables and cheap MIDI receivers and transmitters, and make a proprietary isolated interface that runs much higher than 31.25kbps. Its not like the hardware would break at 31.126kbps. You also right, and I apologize, that there are no MIDI transceiver chips-- but there are things called MIDI transceivers as a box with two connectors that transmits and receives MIDI. I guess I baked in a duplex requirement when I thought about how I would design a link. And hardly irrelevant, a multi-drop isolated interconnect bus is absolutely relevant and fascinating. Did you vamp off an RS-485 interface? A contributed article would be wonderful, you are a great writer.
  • 11.12.2016
  • The PCB is the most important component of your design
  • Thinking_J; I also re-do almost every library schematic signal. I am a big fan of "inputs on the top and left, outputs on the bottom and right". I also like a little more detail inside the symbol to show maybe an open-collector output, or a simple block diagram of a switching power controller IC. And just like you said about routing, its easy once you have good placements, drawing the schematic is easy once you have all the symbols built and placed.