Electronics > Beginners
DC/DC converter with uC interface
Simon:
As you say a well designed circuit. I think the chip you are using is not the most efficient. It's a BJT for starters so has more voltage drop than a MOSFET might.
aix:
--- Quote from: Simon on May 05, 2019, 06:31:45 pm ---For a DAC you can low pass filter a PWM signal.
--- End quote ---
I would definitely like to learn how to do that, and will read up on it.
In this particular project I happen to be out of uC I/O pins, which is not something I mentioned before. Using I2C doesn't require any additional uC pins since I already have an I2C bus. Using an extra £1 part (MCP4725) is no biggie in my case.
Simon:
You produce a PWM from you µC, you put that through an RC filter that has a cutoff frequency 10-100 times lower than your PWM frequency and voila' you have the equivalent of the duty as a voltage. It's more or less what your buck converter does in a SMPS but obviously it's a high impedance source suitable as a signal only. The higher the cutoff frequency the more ripple you get but the faster it changes when the PWM duty changes, the lower the cut off frequency the smoother the signal but it will take longer to follow any changes in duty, like all things it is a trade off.
free_electron:
you can't speed control a bldc properly by changing its voltage.
get a 4 wire fan and use a pwm on the control line to regulate speed. the way that works is that they are chopping the power rail to the bridge. the rail there has a capacitor to smoothen the voltage. the control ic reads the pwm signal and changes the drive pattern of the mosfets to adjust the speed.
the voltage on a bldc only controls torque. ( indirectly through current)
the rotational speed is purely controlled by speed of commutiation on the phases.
mariush:
--- Quote from: aix on May 05, 2019, 05:53:49 pm ---I've looked at the efficiency of my experimental circuit, and it's 71% at V(out)=5V going down to 58% at V(out)=11V.
I honestly have no idea what I should be expecting, but that does seem a bit low. Would you agree? If you do, how might I go about finding & fixing the root cause?
(The project does have to run from a battery after all, and this part is by far the biggest power drain.)
--- End quote ---
The MC34063 is a shitty cheap ancient switching regulator IC with poor efficiency from the start, and then it actually matters to pick the right inductor and the right timing capacitors and various resistors. IF you put the circuit on a prototyping board, the capacitance of the board itself can actually affect some capacitors.
See my post above where I showed you some more modern regulators like that MP1540, which claim to be as efficient as over 90%, see picture below :
BUT, the inductor MUST be chosen properly to handle the peak currents, and you MUST follow the instructions in the datasheet and have the components as close as possible to the chip, with as tight loops as possible
Another good example would be LT1961 : https://www.analog.com/en/products/lt1961.html#product-overview
Datasheet : https://www.analog.com/media/en/technical-documentation/data-sheets/1961fa.pdf
See page 11 which shows you example circuit layout, and you can see how tight you're supposed to place the components to get decent results.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version