Yup, another diode question

[Psycogeek]

ok

i understand more now.

i got that "cds" pack from radioshack, it had ONE primadonna big one in it. and a bunch that would work good with supporting silicoln switching stuff.

BTW i am using 1-3W EACh leds, and arrays of smaller ones that ammount to 1 and 3 watts, various types and sizes aquired to test, tossed into the Home lighting stuff after testing.

right now there is almost 40W of leds lighting up the whole kitchen.

a FEW would have been sufficient, but the homeowner was used to 160W of florescent, so trying to get this TEST room converted to solar and low power, it still has to meet the Critics opinion of lighting :-)

[ghurd]

You could probably do 50W with the IRF510.

Searching the data sheets shows specs all over the board.

From a max of 4.5A to 8A, so depends on what Radio Shack got a deal on that day.

The CDS has less than 1ma through it, so even the little ones should work fine.

And at a certain point, far below 20W, a fluorescent will make more light with less power than LEDs.  Changing the 160W to thin-lite type balasts would be more efficient.

My house, I split the 4 at 40W tubes fixture to 2 circuits on seprate switches.

Very rarely need all 4 tubes on at once.

G-

[ghurd]

'Power' is power, like amps.

'Signal' is for very small tiny power, like what comes through you car radio antenna. And they may turn on and off faster.

Ebay often names stuff wrong anyway.

A TO-220 8A 600V sure sounds like a power device to me!

Sounds like either would be fine.

Most surplus places sell the right diodes cheaper, after shipping, than Ebay.

And you know for sure what they are sending.

G-

[Nando]

The diodes should be rated to be at least 3 times the working current to minimize the voltage drop and power dissipation.

Interesting converter that requires High Voltage DC for the grid connection -- make sure that the proper isolation is used or you may short circuit the system to the grid.

Nando

[ghurd]

Most UL grid tie inverters are high voltage input.

[Psycogeek]

they DO latch, if you got clean dc, and the power doesnt get removed from the source, and the drain load doesnt change.

i am getting further along here, thanks to Scotts, and Ghurds info. and my test shows that after you trigger the gate on this thing i have, it stays triggered.

a 500K resister to ground pulls it back down.

i am SOOO happy, this is an amazing little piece of silicoln.

I purchaced this MOSTFET, its a NPN type IRF540N (cheap cheap)

the data sheet is here

http://www.irf.com/product-info/datasheets/data/irf540n.pdf

and it is triggering with lower voltage than i thought it would, and has VERY low resistance.

i am very close to using it to replace the relay, and mabey other relays in my system, depending on how low the reistance is.

THANKS for the help. i ALMOST feel like i know what i am doing :-)

[scottsAI]

LOL LOL man you had me going.

The MF gate is not latched.

The input impedance is so high, once you put a charge on it, the gate will float and stay floating until something pulls it low. Using a low impedance driver, when it goes hi and low the MF will turn on and off. No latching.

I expect you will fry a couple before you get it.

I wonder how skinney this will get?

Have fun.

[Psycogeek]

impedance?  driver?

oh come now dont confuse me, with the facts. i just stick these wires on it and thats what i get

tests results i put <--- over in the main comments.

[Psycogeek]

ok so here are my totally unscientific tests of 3 diode type of things, that i applied to my (very) low-voltage system, to see what has the least Volt drop, or losses in MY situation.

normal cheapo rectifryer diode to start

the shotkey had (about) 1/3rd the voltage drop, pushing the same current through it, with the same current specs.

its package size was about 1/2 as big, with less Loss there is less Heat for it to dissapate.

the Mosfet (burn my fingers will ya) had (about) 1/3rd the drop of the shotkey BUT only after i triggered the gate High enough to get it to fully switch, .   umm right BEFORE that, there was flesh burning smells :-)  , but i quickly sent more voltage to the gate.  the mosfet was capable of handling 10TIMES the current, of the other diodes, and its a good thing too, because untill i hit it with high enough gate voltage it was very high in resistance (dissapating much power in heat)

So basically what i learned was, dont stick your finger . . . oops, not that,  i learned that shotkeys blow a normal diode away for this, and mosfets blow the shotkey away, as long as you have reasonable voltage.  (in my particular application)

[ghurd]

You need 2 resistors.

One from pos to gate, to supply V and I.

One from gate to negative, to get rid of the V and I,

to stop that 'latching' or floating.

The resistors can be huge values, just have to decide a ratio between the 2 to make it work and stop when you want.  Like 100K and 500K.

Also, 'latched' or floating will have a higher Vdrop than 'On'.

G-

[ghurd]

Revised Resistor Values,

for anyone going to try the auto on circuit.

Made the 'auto on at night' LED setup last night.

The resistor is better at 150K to 220K.

The 150K comes on a little early, off a little late.

A 99 cent 5 pack of 82K resistors would be good for experimenting.

Also, it will depend somewhat on the CDS cell.

G-

[scottsAI]

Very good.

One fry done, two to go.

Heat sinks are your friend.

Have fun.

[Psycogeek]

ahh thanks  gotta keep the power on the gate, so all the cows can come home, or it starts getting muddy ,and they slow down

i discovered even a finger can start sending the gate down, high humidity, bumping it, about anything.  certannly not a good way to keep the resistance all the way down.

[Psycogeek]

oops, that would be the VOLTAGE on the gate. as not much current is required.