Timer Circuit Improvements

Hello all,

I have created a prototype board for my dual timer. It has a 3 position switch so that users can choose between two different times. This is my first time creating a PCB circuit and there seems to be many flaws and inefficiencies. I have attached my version 1 schematics. There are two changes I have already made for version 2 (switch from crystal resonator to oscillator and increase resistance on led resistor).

Even though this circuit works as is, I have concerns about expanding on it because the power management seems to be performing at the bare minimum. With 2 fresh aaa batteries (all I can fit) the converter is putting out 3.3v with the led on and 3.8v with it off (I was intending to run on 5v). I choose these parts on my own without a great ability to understand the datasheets. Looking back the D3 diode probably isn't necessary for reverse current protection so I will likely remove it. After the switch is turned on it goes through either D1 or D2(SM4007PL). I measured the input to be 2.3v and the output to be 1.5v of the active side diode. For the voltage boost, I choose the HT7750a because it is available on JLCPCB's SMD service. I copied the test circuit on the datasheet and added the 100uH inductor, 47uF and 22uF capacitors and diode (I used B5819W).

Any suggestions for improving this circuit would be much appreciated.

Hi,
Does your project work?

C5 in the datasheet is 10uF, you have 100uF, have you tried changing that.
Did you build the SMPS first and try it on your PCB before adding the other components.
Have you looked at the datasheet and graphs to see if the 7750 will do the job?

Can you post an exported jpg image of your PCB please.
Have you kept all the SMPS components as close a possible together?

Did you build a prototype before going to PCB?

Thanks.. Tom... :slight_smile:

Why are you even using a voltage converter? Run the thing at 8 MHz (you can use the internal oscillator if 1-2% off is OK with you). The converter is also wasting your battery power due to its quiescent current.

You power supply issues do go to show how hard it is to build a boost (or buck) converter from scratch. It's easier (and far more reliable - as in, it will just work) to mount an existing module on your PCB.

This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.