1
\$\begingroup\$

I would like the ESP-WROOM-32 to supply more current than a single pin provides. For example, about 100mA. I have seen in its datasheet that the current a pin supplies when it is HIGH is 40mA. If I parallel two pins HIGH and supply 3.3V, could I get the desired current (40+40+40=120mA > 100mA)?

\$\endgroup\$
2
  • 11
    \$\begingroup\$ Not advisable. Use a transistor (either BJT or MOSFET). \$\endgroup\$
    – StarCat
    May 1, 2022 at 19:49
  • \$\begingroup\$ what are you trying to do? \$\endgroup\$
    – jsotola
    May 1, 2022 at 21:05

2 Answers 2

2
\$\begingroup\$

SHort Answer: YES

But don't load it with a large cap as this demands more switched current.

In theory the output ports ought to be well matched and ganging ports lowers the output resistance. If mismatched, CMOS and PCh in this case for high side tend to have a PTC characteristic so they tend towards sharing equally when used as switches.

Fundamentally most FETs are ganged junctions so this yields a similar result.

How is this output impedance measured?

(VDD1=3.3V,VOH>=2.64V, IOL= 40mA) Thus the incremental resistance or RdsOn is (3.3-2.64)/40 mA(min) = 360 mV/40 mA = 9 Ohms .Thus ganging 3 output ports in theory reduces the RdsOn to 3 Ohms.

So drawing 100 mA drops the voltage by 0.3V to 3.0V for VOH if shared equally.

THe first concern is thermal temp rise in the chip and fusing the wirebond. A 3V source supplying 100 mA implies the load is 30 Ohms or ten (10x ) times with 0.3 W dissipated by the load and 30 mW shared by 30 mW outputs. With the tempco for the chip. This does not pose a risk.

The output currents are usually defined by a logic level that gives sufficient margin in the next input , which may have a wide tolerance or hysteresis and tolerance. They are not rated for some arbitrary load. Yet the absolute maximum outputs must be respected with some margin.

Recommendation

  • If transient current is higher and supply load regulated noise matters THEN Use a FET switch.

  • if just driving a 3V LED then be sure of your tolerance errors and assumptions.

\$\endgroup\$
4
  • \$\begingroup\$ Sorry for replying late. I find your comment quite interesting, but I only understand a few things. For example, could you explain to me a little bit about PCh and the PTC feature? On the other hand, I don't know if I have understood correctly. The input/output pins of the ESP32 are made up of FET transistors whose source voltage represents a low or high value on the pin. Would that be it? \$\endgroup\$
    – Sharik_97
    May 9, 2022 at 17:38
  • \$\begingroup\$ The outputs are CMOS which are FETs with Nch & Pch which are voltage controlled switches with an RdsOn. The positive temperature coefficient is PTC occurs above a certain load according the device specs where as switches they can share current, \$\endgroup\$ May 9, 2022 at 22:16
  • \$\begingroup\$ So if you apply a lot of voltage to the gate of the FET you get a very low RdsOn (it conducts) and if the voltage is low RdsOn is very large (it does not conduct)? \$\endgroup\$
    – Sharik_97
    May 10, 2022 at 13:51
  • \$\begingroup\$ At Vgs(th) aka Vt the threshold it is specified to conduct a tiny current. Then the resistance drops towards the rated minimum in a square law curve. Vgs must be at least 2 to 3x the Vgs(th) to function well enough to not get too hot. There are more mathematical ways to describe this that I omitted. Also Vgs(th) has a wide tolerance range you might consider as well. often 2:1 range. \$\endgroup\$ May 10, 2022 at 17:49
1
\$\begingroup\$

Apparently, this chip (ESP-WROOM-32) is being phased out and should not be used for new designs. That being said, it's best to use a driver at the output of your chip.

I don't know what you're driving at 100mA at the output, but there are plenty of driver chips that will give you a 100mA output.

An LYT0006P chip costs about 89 cents and will do this.

Alternatively, I've used MOSFETs and BJTs to do it as well, however, unless you use two transistors in series, it will invert the output. If you invert the output in the software, then a MOSFET or BJT will invert it again.

A 2N7000-D26Z will handle up to 200mA. This is a voltage-driven device, so it must be biased to a turn-on at a little over 3V.

A PN2222ATA BJT will handle up to 1A. At 150mA, the current gain is 100-300 (hfe), so I tiny bit of current will turn it on. This might be a simpler and better solution.

\$\endgroup\$
7
  • 2
    \$\begingroup\$ This is more a comment than an answer. For this to pass as an answer please supply at least one part theat could be used. \$\endgroup\$
    – kruemi
    May 2, 2022 at 4:43
  • \$\begingroup\$ Wouldn't the chips that you have mentioned serve to obtain a larger and more precise current? A power supply connected to them would also be necessary, right? \$\endgroup\$
    – Sharik_97
    May 9, 2022 at 17:46
  • \$\begingroup\$ There are ~9 versions of this chip and some V3's are being phased out, not all of them \$\endgroup\$ May 9, 2022 at 22:21
  • \$\begingroup\$ the 2222's cannot exceed 500 mA for DC and that's at max temperature Tj ~ 100'C temp rise approx or worse. \$\endgroup\$ May 9, 2022 at 22:28
  • \$\begingroup\$ Okay so, apparently, "I have comments." Being new here, I'm not sure what that means, but I suppose I'll respond. \$\endgroup\$ May 11, 2022 at 18:43

Your Answer

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

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