Photomicrosensor slower than datasheet specs

[GreyArea]

I'm trying to set up a circuit using the EE-SX1042 photomicrosensor/optical encoder (https://www.mouser.com/ds/2/307/sx1025-1190149.pdf).

Working through the datasheet I came up with the circuit attached.  The problem I have is that the fall time seems particularly slow (~500us) when compared to the datasheet (4us).

After noticing that the R_L value of the Condition/Combination table at the bottom of page 2 was 100R, I replaced R2 with that value.  Now the high (0) signal is only 235mV and isn't usable logic levels.

So, what's wrong with my circuit?

[Kleinstein]

The load resistor at the detector side is likely much larger than in the DS. They usually specify the speed the a low resistance (e.g. 50 Ohms). The main part would be to avoid the transistor to go into saturation (e.g. voltage lower than about 0.2 V).

[danadak]

This might help -


https://www.vishay.com/docs/83590/fastswit.pdf


Regards, Dana.

[GreyArea]

What a fantastic document!  Thanks all

(puts the coffee machine on)

[GreyArea]

OK, so after reading the application note (I must admit the analysis theory went largely over my head) my options seem to be:

a) Reducing the emitters I_F (dimmer IR LED).
b) Tuning R_L to reduce the saturation capacitance at the base of the phototransistor.

I've now tried different values of R_L, from 100K to 100R and it seems to make no difference to the turn off time at all, I've also reduled I_L down to 10mA (R1 = 380R in my circuit).

According to the datasheet, 100R R_L should be down at ~3uS but by my reckoning, I'm like 3 orders of magnitude out with Vcc at 5V.  If I was a 200-500us out, that would be fine for my application, but thousands of us just won't do.

[BrianHG]

I'm curios, as a test, remove any obstruction from the encoder, then pulse the IR led with a square wave from a function generator & examine the speed of the waveform.

[GreyArea]

Ah, that never occurred to me - I haven't even used my signal generator in anger -  until now.  It seems the issue was in fact mechanical, rather than electrical.

Using a 10K R1 I can now trigger with about a 250uS fall time which is perfectly fine for my needs.  Along the way I've learned about phototransistor base saturation and that if the detector doesn't poke into the gearbox as far as it needed to be, it won't work

[BrianHG]

Glad to be of inspirational help.