Kingofkya your account has too many stored PMs so I can't send you one. PM me your email.
Ceer123 wrote:Kingofkya your account has too many stored PMs so I can't send you one. PM me your email.
Try again:P
madsci1016 wrote:My design makes use of 2 diodes and a resistor to bring up the voltage by 1 diode drop, which (should) be just enough to be in tolerance. I've been running it with somewhat slow fades for weeks now but I can go really slow just to be sure.
Did you happen to look at the output with an scope? My circuit is designed around the 2N7000 transistor. However the connections are different than the one in the link you provided. Anyway I was seeing strange things happen when using 10K pullups so I
decided to look at the output through a scope. The rise time was slow. I kept playing with different resistor values and with a 1K pullup the rise time was decent but I was looking for it to be sharp. I wanted it to be just like the output would be if I was measuring straight from the GPIO. I went all the way down to 220 ohm and the edges looked crisp they where sharp. The output voltage was running around 5.2V. May be something to keep in mind.
Also one thing to think about what if someone does something stupid? Like some how connects 12v to the output, that diode drop dosen't protect it. I would recommend the fet(or transistor) based shiftier because most can take up to 30ish volts without issue. And you can use almost any fet and any resistor in the 5-40k range.
kingofkya wrote:Also one thing to think about what if someone does something stupid? Like some how connects 12v to the output, that diode drop dosen't protect it. I would recommend the fet(or transistor) based shiftier because most can take up to 30ish volts without issue. And you can use almost any fet and any resistor in the 5-40k range.![]()
For a 500% difference in cost between the solutions it doesn't protect against every failure mode, if they hook a pixel power supply backwards and connect GND to data, and 12V to ground it will blow up the FET and the MCU connected to it. Granted, there's a very low percentage of people that would hook it up wrong at all, so I'm not designing for that very low minority, just trying to keep cost and complexity to a bare minimum.
madsci1016 wrote:For a 500% difference in cost between the solutions it doesn't protect against every failure mode, if they hook a pixel power supply backwards and connect GND to data, and 12V to ground it will blow up the FET and the MCU connected to it. Granted, there's a very low percentage of people that would hook it up wrong at all, so I'm not designing for that very low minority, just trying to keep cost and complexity to a bare minimum.
NO offense, but you have never had to fix something in the middle of a show. very easy to do.
kingofkya wrote:NO offense, but you have never had to fix something in the middle of a show. very easy to do.
Middle of a show? I routinely get the red and black wires mixed up in the middle of the day in the workshop ... I have a cache of 7805s just for that inevitability.
\dmc
dmcole wrote:Middle of a show? I routinely get the red and black wires mixed up in the middle of the day in the workshop ... I have a cache of 7805s just for that inevitability.\dmc
Yeah exactely, exteramly you always have to wary of future self
kingofkya wrote:Yeah exactely, exteramly you always have to wary of future self
The oddest things happen to electronics when they pass from the comfort of inside to the temp swings we have outside, requiring constant maintenance in harsh conditions. Big heavy gloves and a mind on a hot drink when you get back inside has caused many a fuse to pop (go fuses). IMO, more people swap a few wires when in a hurry (ok I am talking about me) than you might think. I have a few dead AC SSRs from minor accidents that my wife will not let me forget.
kingofkya wrote:Also one thing to think about what if someone does something stupid? Like some how connects 12v to the output, that diode drop dosen't protect it. I would recommend the fet(or transistor) based shiftier because most can take up to 30ish volts without issue. And you can use almost any fet and any resistor in the 5-40k range.![]()
I was curious how responsive the circuit you referenced would be, so I punched it into LTSpice and put a typical 2811 signal through it (1000ns high, 250ns low). Rise time was way too slow w/ the 10k's, but looks decent with a 220. Wasn't sure what rise and fall times were out of the ESP8266 so I used 20ns. Haven't replaced my scope yet, so I'm stuck to playing in LTSpice. Model is attached.
I used thesethese buffers on my PixelSticks and they worked great, SOT-23-5 part.

Logic Shifter - 10k by smerrick, on Flickr

Logic Shifter - 220 by smerrick, on Flickr
