I am trying to select an IR receiver for a remote controlled unit, from this table. Here is one example, but all of the parts exhibit the same seemingly contradictory specification.

Having a look inside TSOP75438's datasheet, we immediately notice the promising:

IR Receiver Modules for Remote Control Systems
Recommended for NEC code: yes

Cool. However, moments later, there is:

For bursts greater than: 70 cycles
a minimum gap time in the data stream is needed of: 4 x burst length

Hmmm. Here is a NEC example message: enter image description here

With carrier frequency 38kHz, the initial 9ms pulse consists of 342 impulses, much more than 70, and the pause of 4.5 ms is much less than 4*9 ms.

Here is some more data, from an application note:

enter image description here

Note that 70 cycles @ 38kHz equals approximately 1.8 ms.

I understand that the above restriction is posed by the Automatic Gain Control unit, in order to distinguish between data and long monotonous sequences of IR noise form e.g. fluorescent lamp or LED backlight. However, my understanding MUST be flawed, as 1) the part is actually recommended for this protocol and 2) I just tested it with a scope and it works perfectly.

So what am I missing?

  • \$\begingroup\$ Just a guess but I wonder if it's based on a 100% duty cycle. There's figure 8 on page 3 that seems to imply for some duty cycles it doesn't matter. \$\endgroup\$
    – PeterJ
    Commented May 29, 2013 at 8:09
  • \$\begingroup\$ @PeterJ, Yes, I asked a similar question yesterday about that figure 8 ;) But shill, the recommended duty cycle for NEC is 1/4 or 1/3, much higher than the permissible ~1/10 in fig8! \$\endgroup\$
    – Vorac
    Commented May 29, 2013 at 8:19

1 Answer 1


From the quotes you show, I would assume that "gap time" is not the time from the AGC pulse to the data start, but instead the gap from end of data stream to next AGC pulse. Think about it: The AGC pulse sets the sensitivity level, and the data should come right after that. If the gap is too big, then the AGC circuitry will have time to recover, and thus the data will not be decoded in context of the initial burst.

The gap time between codes is there both so that the AGC can recover if you move the sender, and so that the decoder can easily sense the difference between successive codes versus a very long pulse train.

  • \$\begingroup\$ This is the only explanation that makes sense to me. Consider perhaps that the AGC's job is to progressively lower the gain to eliminate spurious events. So while ever the AGC sees occasional short pulses it ratchets down the sensitivity to eliminate them. Then it gets the big NEC "leader code" and stops adjusting for the "gap time" (called "idle time" in Vishay doc 80069). If you think in this backwards way (off for signals, on for noise), the explanations start to make sense. \$\endgroup\$ Commented Dec 2, 2023 at 0:04

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