1
\$\begingroup\$

Required is a timer circuit to switch on an LED with 2 time constants, when a switch is depressed as follows:

  • when switch is closed, light an LED 10 ~ 20mA
  • when switch is opened again, keep the LED on for approximately 5s more
  • then fade LED off for approximately 10s.
  • circuit voltage = 3v
  • assume a standard red LED , Vf =~ 1.8v
  • prefer to have maximize brightness
  • small SMD low leakage caps <~22uF
  • no ICs that consume continuous quiescent current; but accept minimum idle current (< 0.25uA @ 25'C)

I'd like to use a mosfet to minimize voltage drop and maximize brightness.

I think that it would involve some chain of transistors or mosfets that are held on by capacitors, but I can't figure out how to keep it on full brightness for the first 5s, as they would start to fade immediately as the voltage drains out of the capacitor.

enter image description here

\$\endgroup\$
  • \$\begingroup\$ What does "virtually no current" mean? I don't think we can design to that requirement. \$\endgroup\$ – Elliot Alderson Jun 17 at 1:15
  • \$\begingroup\$ Design Rule # 1. Never specify component types unless absolutely necessary. ONLY give all a) Input b) function and Output specs... plus other major restrictions. Rule #2 when you specify the wrong parts read Rule #1 again. Ceramic has a leakage time constant less than 20 seconds. But that's OK. There's another way. REWRITE your question using a datasheet point form approach with ALL SPECS including environment and tolerances for time and voltage, cost and mechanical construction. Now! s'il vous plais! \$\endgroup\$ – Sunnyskyguy EE75 Jun 17 at 1:26
  • \$\begingroup\$ If you outline in simple point form all parameters and tolerances in next 1hr I will award 1 pt otherwise -1 pt. Others may agree. Thankyou \$\endgroup\$ – Sunnyskyguy EE75 Jun 17 at 1:32
  • \$\begingroup\$ How complex are you willing to go? Discrete, I see three NPN, one PNP, and one NFET which is in keeping with your capacitor limitations. \$\endgroup\$ – jonk Jun 17 at 5:16
  • \$\begingroup\$ Are you not aware of the leakage time constants in capacitors especially ceramic? Show me one capacitor with a time constant in the range you want. ( I know what to expect, but I want you to look) again you specified ceramic when it is size, and function that is important. Film caps are the only ones with long Δt=C*ΔV/I_leak \$\endgroup\$ – Sunnyskyguy EE75 Jun 17 at 17:49
1
\$\begingroup\$

It looks like I found a solution highly dependent on the device parameters.

enter image description here Low leakage parts are the critical factor. Take careful selection of Ids @ Vt and Vgs

Two inversions are done: The 1st must have a gate threshold of << 1/3 to 1/2 of Vbat and not 2 to 4V. This is what causes decay on the 1st time constant.

  • hi-side mechanical switch,
  • low-side Nch (inv)
  • hi-side Pch (inv)

The critical factors in the selection of both FETS are:

  • the IDSS leakage current must be extremely low at Vgs near 0V.
    • Carefully select with Drain Resistor to result in a voltage drop less the next gate threshold.
  • the threshold Vt or Vgs(th) @ Ids is critical because this acts as the voltage comparator for the 1st stage and the gradient declining LED intensity for the 2nd.
  • Both have long time constants so capacitor leakage resistance at Vcap needs to be estimated for the ambient temp. Film caps are best but are low Dk large parts.
\$\endgroup\$

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.