I dont want to be unkind, but there is so much wrong with your circuit that I dont know where to start...
Are you a devotee of "electron current" by any chance?
First, the LEDs are all backwards. The +48V should feed the anode end of the LEDs, not the Cathode. As connected, you would indeed blow up some expensive LEDs.
The LM317 constant-current sinks are backwards, and miswired. The terminal marked "in" should be connected to the fuse, and the terminal marked "out" should just connect to one end of the 1.3Ω, the other end of the 1.3Ω ties to "adj" and to the drain of the big NFET. btw- are the 317s on a heatsink?
The LT1157 and NFET looks ok, however you overkilled the current spec on NFET (100A), super low RdsON, but shortchanged the Voltage rating (only 55V for a 48V supply).
What is the forward drop on your leds at 1A of current?
Unkind? I consider it helping save me money so I do not fry the expensive LEDs.
First, the LEDs are all backwards. The +48V should feed the anode end of the LEDs, not the Cathode. As connected, you would indeed blow up some expensive LEDs.
The LM317 constant-current sinks are backwards, and miswired. The terminal marked "in" should be connected to the fuse, and the terminal marked "out" should just connect to one end of the 1.3Ω, the other end of the 1.3Ω ties to "adj" and to the drain of the big NFET.
I have not built that part yet, but judging on your comment, I am guessing they should be.
The LT1157 and NFET looks ok, however you overkilled the current spec on NFET (100A), super low RdsON, but shortchanged the Voltage rating (only 55V for a 48V supply).
I had a hard time looking for the correct NFET. I did not know which to use and then I was not sure if I needed a LT1157 or not. Can you suggest a different FET?
What is the forward drop on your leds at 1A of current?
Ok, at 3.7V per LED, times 12 LEDs is 44.4V, so starting with 48V, you only have 3.6V left over. That is just enough for the LM317 to work as a current limiter if you get rid of the 2Ω resistor.
The dropout voltage of the LM317 is 2V at 1A, plus the drop across the 1.3Ω resistor is 1.3V, for a total min drop of 3.3V, so you will have only 0.3V of headroom for the LM317 to do its constant-current magic.
At 1A of LED current, the drop across the LM317 will 2.3V, so it will be dissipating 2.3W. Which package are you using? The TO220 package with no heatsink is spec'ed at 50degC/W, meaning that it go into thermal limiting. You will need to get a heatsink on it which will lower the thermal resistance to under 30degC/W...
Anyone have advice on testing or selction of a MOSFET. I have tried searching online, but can not find a clear site which explains how to select the corrrect MOSFET.
If I were doing what you are trying to do, I'd be selecting a NFET with a Vds>100V, Rds < 0.2Ω, Vg to turn on ~5V, in a T0220 package, in stock, cheap shipping.
I am also working on a project that requires a switch while using a basic circuit with a 96V battery, based on what you have said i found this MOSFET on mouser wanted to ask if this would work what what he needs as i am trying to learn how these data sheets read.
Its Part # 512-MTP3055VL it reads (0.18 Ohm @ 5 V) how ever it says breakdown is 60V does this need to be closer to 100v? if so i found a 512-IRL640A with (0.18 Ohm @ 5 V) with a breakdown of 200V.
Would it be a power nFET or does the Vds>100V imply that?
Another question about MOSFETS:
When the LEDs are on the forward voltage to the FET would be very small. Does the drain need to be a certain amount of voltage higher than the gate for the FET to work corectly? How can I tell this in the specs?
Vds must always be higher than the supply voltage. How much higher? I like 2X.
When the LEDS are on, and there is 2A flowing from drain to source, Vds will be RdsON*Id ≈ 0.2V. The Vgs when the FET is On will be ~8V, per your driver chip. When the FET is OFF, depending on the leakage through the LEDs, Vds could go all the way to 48V.
It's not going to help, the LM317 will limit the current to 750mA and the LEDs will still be toast if it fails, even with the fuse, because the surge current will be high enough to destroy them.
Look at the Rds_on or Rds(on) spec. There will be some parameters listed next to that spec. One of them will be Vgs=?, and will list the voltage that was used to get the resultant Rds_on figure. A logic level MOSFET will have the Vgs= parameter at 5V or lower. You'll notice most MOSFETs are 10V.
So I have changed the power supply from one power supply running at 2A and 48V to two power supplies running at 48V and 750mA. The question I have is do I now need to double the number of channels producing the PWM signal or can I connect both to the same NFET?