7
\$\begingroup\$

Most NPN silicon phototransistors that I see (Everlight, Kingbright, etc.) have fairly poor rise and fall times, on the order of 10us.

The best one I've seen so far is the QT Brightek QSD8T120B with rise and fall times of 7 us. Manufacturers always specify these times given certain load current and resistance and collector voltage, and in this case they are 200uA, 100R and 5V respectively. 7us would allow for a best-case data transfer rate of about 71kbit/s.

What is the limiting factor here? Junction capacitance? Is there a different technology for IR detection that I should look for to get theoretical data rates of at least 500kbit/s?

\$\endgroup\$
  • 2
    \$\begingroup\$ look to using a photodiode for higher speeds. \$\endgroup\$ – Jasen Jan 2 '18 at 3:36
  • 2
    \$\begingroup\$ May be related to electronics.stackexchange.com/q/136928/17429 \$\endgroup\$ – akohlsmith Jan 2 '18 at 3:46
  • 1
    \$\begingroup\$ Additional search words: bootstrapping, cascode. \$\endgroup\$ – Janka Jan 2 '18 at 4:10
  • \$\begingroup\$ Phototransistors have a huge base to catch as much light as possible. \$\endgroup\$ – CL. Jan 2 '18 at 13:21
6
\$\begingroup\$

Photodiode + transimpedance amplifier can give you much faster response. Quite a bit of good information has been written. Anything by Phil Hobbs will be useful.

By keeping a constant bias voltage across the PD you can reduce the effect of the diode capacitance. A constant and large reverse bias will further reduce the capacitance.

More light will allow you to use a smaller (lower capacitance) PD and lower value feedback resistor, so you'll get faster response.


A phototransistor is like a PD between the collector and base of a transistor so you get the PD capacitance multiplied by the Miller effect. If you bias a phototransistor so Vce does not change you can get faster response out of it.

\$\endgroup\$
  • 1
    \$\begingroup\$ I've actually worked with Phil Hobbs for a short consulting time, years ago. Very helpful. Anything he writes is worth the time to read. \$\endgroup\$ – jonk Jan 2 '18 at 5:14
  • \$\begingroup\$ How does one bias so that Vce doesn't change? \$\endgroup\$ – Transistor Jan 2 '18 at 7:25
  • 2
    \$\begingroup\$ @Transistor Cascode (eg with a BJT) or transimpedance amplifier with an op-amp, similar to a PD. \$\endgroup\$ – Spehro Pefhany Jan 2 '18 at 7:38
  • \$\begingroup\$ Thanks. World of a difference. The PD204-6B claims a response time of 6ns! \$\endgroup\$ – Reinderien Jan 2 '18 at 14:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.