Opto-Coupler Variations
I only bought 4 pieces of CNY17-3, so they are likely to come from the same production batch.
I have measured them all, the aim being to generate a 2mA current at the photo transistor with ~2.1V Vce.
The LED Vf varies from 1.087 to 1.095V; and the CTR from 0.93 to 1.01.
As they are, when put into the circuit as posted, the output bias will vary between 1.267A and 1.299A.
(Simulation says 1.16A. So a difference of 12%.)
One can say that 32mA difference in bias is negligible, which is true.
But the measurement does not say anything about batch to batch variations.
So other examples bought from different suppliers might differ, by up to 20% according to datasheet.
But at least variations within a batch seem acceptable.
Patrick
I only bought 4 pieces of CNY17-3, so they are likely to come from the same production batch.
I have measured them all, the aim being to generate a 2mA current at the photo transistor with ~2.1V Vce.
The LED Vf varies from 1.087 to 1.095V; and the CTR from 0.93 to 1.01.
As they are, when put into the circuit as posted, the output bias will vary between 1.267A and 1.299A.
(Simulation says 1.16A. So a difference of 12%.)
One can say that 32mA difference in bias is negligible, which is true.
But the measurement does not say anything about batch to batch variations.
So other examples bought from different suppliers might differ, by up to 20% according to datasheet.
But at least variations within a batch seem acceptable.
Patrick
I wonder how much variation there would be with a ZTX450/BC547
I ran my batch of four 4N35 through a test rig similar to its operation conditions: variable bench-top supply through 100R to LED, 24V bench-top supply through 15K to BJT.
I measured the voltage over the 15K at LED voltages of 1.05, 1.10, 1.15 and 1.20.
Sample 1: 1.3, 4.2, 9.6, 17.4
Sample 2: 1.3, 4.2, 9.8, 17.3
Sample 3: 1.4, 4.2, 9.7, 17.0
Sample 4: 1.2, 4.0, 9.1, 16.0
So 3 are virtually indistinguishable on my test rig, but the 4th is about 10% off. At a target of 1.3A bias my sim shows that I'll get 1.29A vs 1.41A (nearly 4 times Patrick's result).
This could be the afore-mentioned batch-to-batch variation (ie: different wafers between the first 3 and the 4th). In any case mine are now matched and there seems little to gain by mixing them up again.
Cheers,
Jeff.
I measured the voltage over the 15K at LED voltages of 1.05, 1.10, 1.15 and 1.20.
Sample 1: 1.3, 4.2, 9.6, 17.4
Sample 2: 1.3, 4.2, 9.8, 17.3
Sample 3: 1.4, 4.2, 9.7, 17.0
Sample 4: 1.2, 4.0, 9.1, 16.0
So 3 are virtually indistinguishable on my test rig, but the 4th is about 10% off. At a target of 1.3A bias my sim shows that I'll get 1.29A vs 1.41A (nearly 4 times Patrick's result).
This could be the afore-mentioned batch-to-batch variation (ie: different wafers between the first 3 and the 4th). In any case mine are now matched and there seems little to gain by mixing them up again.
Cheers,
Jeff.
I've been following this thread for the last few weeks and have some boards in hand, but have been unable to read any of Patrick's attached .asc files. I'm using a Mac, and have tried TextEdit and another web app, but I'm not getting any apparent results, just some meaningless lines of text. Any Mac guys have a tip for me? (Aside, of course, from using Windows.)
Not only for Windows, but also for Mac. You can download the LTspice for mac here: LTspice | Design Center | Analog Devices
