- Sun Jan 31, 2016 11:55 am
#40082
joostn wrote:http://bbs.espressif.com/viewtopic.php?t=49
Thanks! This got me going...
I's more complicated though. This is not your typical garden variety PWM.
This thing does not generate just one on/off cycle each period.
Instead you get a train of short pulses. More or fewer pulses, depending on the data value.
Fine if you want to turn it into a voltage via R/C filter.
So I went at it scientifically... errrm... Well I tinkered until it worked...
The prescaler has to go all the way up for the circuit I was using.
I did not want to use a smaller inductor actually, I liked the frequency nice and slow.
Here are two scope shots. One at 0x1FF30 gives -12.7V.
The other at 0x1FF70 that gives 4.7V. The relay still holds. Nice!
It's nicely visible how all the pulses are the same width, but with gaps.
(CH1=PWM, CH2=switch node, CH3=output)
That's now fine by me. If someone want's to calculate the buck-boost through all the way, be my guest.
For me, it just has to work...
Settings:
Write these values into the sigma delta control register.
0x1FF70 -> 4.6V
0x1FF60 -> 5.4V
0x1FF50 -> 7.0V
0x1FF40 -> 9.3V
0x1FF30 -> 12.6V
0x1FF20 -> 17.6V
Power:
Relay measures at 390R
At 12.7V output, the converter takes 135mA at 3.3V = 545mW
Relay takes 414mW (12.7V^2 / 390R), so 130mW loss.
This is active for maybe a second, then the power is reduced.
To hold the relay, I chose the 9.4V setting.
Converter takes 80mA at 3.3V -> 264mW
This is now less than the relay's nominal power use, so who cares if there are converter losses?
The canonical way would have needed an extra 12V. And typical circuits do not reduce relay power after the relay has engaged.
So I can now go into mass-producing the five units I need....
You do not have the required permissions to view the files attached to this post.
