4
\$\begingroup\$

I use a lot of "wired or" circuits where rather than using a logic gate in an IC package I simply use diodes (eg. 1n4148) into a destination gate that has a pull-up or pull-down resistor on its input (depending on the logical function I need). I have found this quite reliable and have used as many as 5 diodes (so far) to make such a logical function.

With CMOS circuits I have been delighted that I can do this with very high valued resistors. Where I might use something in the 1K to 5K range in a TTL circuit, I am using up to 100K with CMOS circuits. This of course helps keep current draw down.

My question is "what problems might such an approach expose my circuits to?" Using very high value pull-up/down resistors is something I want to do, but I read one article (from TI) that suggests the 1K to 10K range even for CMOS logic. That seemed counter-intuitive to me. Are there any nasty suprises that may lay in wait for me by using high-valued resistors?

\$\endgroup\$
4
  • 6
    \$\begingroup\$ 100k is usually good. 1M and sometimes 10M may work but specific circumstances may catch you out. || High R values give RC time constants that may start to matter in some cases. | !N4148 is lowish leakage. Be ware of Schottky - low Vf but can have terrible Ir. | Loading may vary results but usually not a major issue. || I'm oldish -= I played such games about 50 years ago :-). || With diode logic play with 74x14 CMOS hex schmitt triggers. You will love what can be achieved with them. Delays, oscillators, monostables, flipflops, stores, PWM mod/demod ! :-), Class D audio amplifier ! :-), ... . \$\endgroup\$
    – Russell McMahon ♦
    Dec 27, 2020 at 10:35
  • 4
    \$\begingroup\$ High impedance circuits are also more susceptible to interference. Very high resistor values might cause inputs to trigger spuriously. \$\endgroup\$
    – Unimportant
    Dec 27, 2020 at 10:51
  • 1
    \$\begingroup\$ As several have said - coupling from other nodes can cause issues as R sizes rise. Both capacitive, inductive and leakage are starters. Leakage currents also have greater effect. I'd expect 100k is usually reasonably safe, 1m usually safe and 10m riskier. Very much cct and context dependant. 10k is barn door safe except perhaps during nuclear exchanges. \$\endgroup\$
    – Russell McMahon ♦
    Dec 28, 2020 at 1:22
  • \$\begingroup\$ This discussion and finally the answer I accepted has been very useful to me. Thanks all. \$\endgroup\$
    – DarylK
    Dec 28, 2020 at 5:45

1 Answer 1

Reset to default
9
\$\begingroup\$

high resistance make the circuit slower and more succeptable to electromagnetic interfereance.

If you circuits are slow and stateless it probably doesn't matter

some stateful parts (like flip-flops) need a certain edge speed and could misbehave if the the edges are too slow.

\$\endgroup\$
2
  • 3
    \$\begingroup\$ Many CMOS parts (just about all modern parts) state the minimum rate of change for the inputs. \$\endgroup\$
    – Peter Smith
    Dec 27, 2020 at 13:51
  • \$\begingroup\$ Yeah "stateful parts" logic parts (like and gates) typically do not. \$\endgroup\$
    – Jasen Слава Україні
    Dec 27, 2020 at 19:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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