Those resistor values are for a tester board I'm specing, which if it works will then be tested at the higher 500mA spec. Which will be a minimum of 20-ohm.
It is an interesting part. I think I'll put it into the initial test board, so I can play with it. It is likely we will use some of these lower power rheostats for something around here, and if we can build 2 rheostats out of just one IC and less than a dozen neighboring passives, they could be quite useful (even if the IC is not the cheapest thing in the world). Thanks.
So I finally was able to get around to assembling the circuit with the photovoltaic isolators, and it isn't working.
Here is the diagram:
I've double checked wiring (it's acutally a circuit board, so I checked for shorts, opens, and correct parts). On the shift register side of the isolator, I get the expected 1.4V (1.475V actually) drop across an active LED. On the other side, I get 28mV when on, and an expected 0V when not. 28mV is not quite enough to activate my MOSFET.
The other circuit idea I had (the one where I ran the gates off of MOSFET drivers) worked, but it's on the edge of usability due to the amount of resistance on the MOSFET because the gate is only a few volts above the source. So I was hoping this would work (even though the isolators are annoyingly large).
Try using higher-valued gate resistors (100K or more). The photovoltaic isolators will not provide very much current, and your 10K resistors may be swamping them.
I'll try that. I have to find some 0805 resistors to put in there. Though, the math doesn't really work out. 10K-ohms with 28mV, your looking at only 78nW. So bumping up to 100K would only produce 88.5mV (if my math serves me right). I'm looking for a 1M resistor, that'll give 280mV, a far cry from sufficient voltage. So unless there is more than just a simple voltage/wattage correlation (like at higher voltages, it can produce more wattage)....
The datasheet says it should produce 8V with a 10M-ohm load, which is 6.4µW, 80+times more than what I'm getting, even though the 8V is for a 5mA input, and I'm giving it 18mA (I didn't catch the 10M-ohm load the first time around, else I probably would have started with 1M or 10M).
Once I get different resistors to put in there, I'll have more answers.
Looking at the opto's datasheet, it appears that you can simply omit the gate resistor altogether. In the coupled characteristics section, there's a spec for "off state clamping resistance", which can range from 100R - 3.3K. I presume that to mean that it will discharge the gate through that resistance when not turned on. The fact that the turn-off time is so short compared to the turn-on time adds further credence to this hypothesis.
Comments
It is an interesting part. I think I'll put it into the initial test board, so I can play with it. It is likely we will use some of these lower power rheostats for something around here, and if we can build 2 rheostats out of just one IC and less than a dozen neighboring passives, they could be quite useful (even if the IC is not the cheapest thing in the world). Thanks.
Here is the diagram:
I've double checked wiring (it's acutally a circuit board, so I checked for shorts, opens, and correct parts). On the shift register side of the isolator, I get the expected 1.4V (1.475V actually) drop across an active LED. On the other side, I get 28mV when on, and an expected 0V when not. 28mV is not quite enough to activate my MOSFET.
The other circuit idea I had (the one where I ran the gates off of MOSFET drivers) worked, but it's on the edge of usability due to the amount of resistance on the MOSFET because the gate is only a few volts above the source. So I was hoping this would work (even though the isolators are annoyingly large).
Any ideas?
-Phil
The datasheet says it should produce 8V with a 10M-ohm load, which is 6.4µW, 80+times more than what I'm getting, even though the 8V is for a 5mA input, and I'm giving it 18mA (I didn't catch the 10M-ohm load the first time around, else I probably would have started with 1M or 10M).
Once I get different resistors to put in there, I'll have more answers.
-Phil