3
\$\begingroup\$

Ideally I would use a low dropout regulator but I can't. I also don't have access to a suitable buck boost converter. The best alternative I have (and also my plan B) is to boost the Li-ion battery voltage to some 4.3V and then buck it down to 3V and power the ESP8266. I'd rather not do that.

schematic with ESP8266 and battery

This is what I want to do instead. I want to put a series diode (1n4007) between the battery and the device. I couldn't find information online that documents dropping voltage with diodes for dynamic loads. That's why I wanted to consult more experienced people before committing to it.

  • The worst case of fully charged battery and low forward voltage (4.2 - 0.6 = 3.6) is within specs of the esp module.
  • The average diode power drop (0.7V * 80mA = 0.05W ) is well below the maximum of the diode(3W).
  • The current draw of ESP8266 varies from a few uA in sleep mode to 500mA. Using the inbuilt ADC and some software, I can ensure that the ESP will not sleep when battery is fully charged to avoid less than 0.6V drop in the diode. For higher transient current draws, the diode might increase the forward voltage to 0.9V. I have kept the parallel capacitor hoping to smooth that transient voltage drop.

detail of 1n4007 datasheet regarding forward voltage drop in relation to current draw

  • Even at a junction temperature of 100°C the forward voltage might drop by 100mV (Not in the datasheet, I'm using the rule of thumb mentioned in some article). It is still not difficult to make sure the voltage is still within specs.

So my question is, can I expect this setup to work reliably? I'm hoping for answers that correct my assumptions if they are false and point out flaws in this design.

ESP8266 current consumption

This is the ESP8266 current consumption profile taken from this page posted for reference.

\$\endgroup\$
9
  • \$\begingroup\$ How are you charging the battery? Is it possible to just set the termination voltage to 4.1V, then the diode solution will be fully within spec, even when sleeping. \$\endgroup\$
    – anrieff
    Oct 25, 2018 at 6:40
  • \$\begingroup\$ @anrieff It seems that it is possible with the charger IC I'm planning to use albeit with a fairly convoluted process. I might have to do it in the end, but I am hoping that I don't. Thanks for letting me know. \$\endgroup\$
    – Nirav
    Oct 25, 2018 at 7:02
  • 1
    \$\begingroup\$ "I would use a low dropout regulator but I can't.": You need a low-dropout voltage regulator. Use one. If you can use an extremely complex piece of hardware like an ESP8266, then there's no logical reason to not use a voltage regulator. I'm really curious what technical restriction says you can't use an LDO? \$\endgroup\$
    – Marcus Müller
    Oct 25, 2018 at 8:10
  • 1
    \$\begingroup\$ Your problem is what happens at currents off the bottom end of the diagram in your datasheet. The curve shown goes down to 10mA, but the ESP8266 can go as low as 20uA in sleep mode. What's the voltage drop on your diode in that case? \$\endgroup\$
    – Jules
    Oct 25, 2018 at 10:19
  • 2
    \$\begingroup\$ Seriously, the cheapest voltage regulators that the major international electronics distributors carry would do here; there's hundreds of clones of TLV1×××. Even if you're in one of the very few countries that these don't ship to, these should be relatively easy to source. Again, your problem says you need an LDO. The cheapest, easiest, easiest to source solution is buying an LDO. Period. \$\endgroup\$
    – Marcus Müller
    Oct 25, 2018 at 13:56

2 Answers 2

Reset to default
4
\$\begingroup\$

Using a diode in series with PSU to "adjust" its level is always a bad idea, especially when the load varies significantly. But you can use a Lithium iron phosphate battery instead, it has 3.2V nominal level, and a fairly flat voltage discharge profile.

enter image description here

\$\endgroup\$
3
  • \$\begingroup\$ You also get the other advantages of LiFePO4 batteries: longer cycle life, and safer charging due to less volume change in the chemistry. \$\endgroup\$
    – Jules
    Oct 25, 2018 at 10:23
  • \$\begingroup\$ Could you please elaborate on the first sentence? Using LiFePO4 battery is an elegant solution to this problem but I'm really just trying to design around the limitations of the diode if I can. I suppose it's a silly thing to do but I'd like to do it anyways. \$\endgroup\$
    – Nirav
    Oct 25, 2018 at 12:37
  • \$\begingroup\$ @KittyHawk, "to design around the limitations of the diode" -- diodes were around for several decades, and so far no simple design workarounds were found. The best workaround is called "ultra-low LDO". \$\endgroup\$
    – Ale..chenski
    Oct 25, 2018 at 19:00
1
\$\begingroup\$

A diode drop here almost works...it is tempting to pursue.
The variation of voltage drop on that 1N400x diode is one detractor. It will vary with temperature, and it will vary due to its I vs. V curve.

One thought might help these variations a bit by adding a shunting transistor to take more of the current. Here's a few variations on the basic idea:

schematic

simulate this circuit – Schematic created using CircuitLab
D1 shares a portion of current with Q1. Both should be at the same temperature: choose your own parts that can handle the maximum current. It would be far better if D1 & Q1 were integrated on the same die. Or just use a power transistor having a low saturation voltage. Voltage regulation can be a bit better than a power diode, but temperature variations still exist. There may be better choices than TIP31.

A simulation in LTSPICE with 2SCR573D improved the voltage drop (compared to a raw diode) by 186mV at a fixed 27C temperature. Load was 10K in one case, 10 ohm in the other case. Voltage regulation is not wonderful. And as @jules has suggested, when ESP8266 is put into low-current mode (sleep?) ESP8266 may see an over-volt on its DC supply pin.

\$\endgroup\$

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.