As cheap as PWM duty sweep

[Lord_grezington]

Hello Everyone...

I have this project when I need to try and get a PWM duty sweep starting at around 30% up to 100%. The issue I have is that we done have anything left in the design (BOM Cost). I would normally use a PIC but there are a few issues, firstly I need a operating voltage range of preferably 2-15V (with a PIC would also need a LDO).

Been looking at 555 timers, the ones with the voltage range I need are a bit more expensive but can create the PWM, but any one know how I can create the sweep from 30% to 100%? this can be done anywhere between 500ms to 1s if needed.

Thanks

 

[Alec_t]

Welcome to EP!
Is this a school/college project?
What pulse frequency?
What will control the PWM? Potentiometer? Voltage? ...... ?

 

[WHONOES]

Can be done but more info as above required.

 

[Lord_grezington]

Hello

Not a school project, its being used to control a BLDC fan using MCP8063. I have some PCB's made up but the fan only works if I feed the MCP8063 with a PWM. If I leave the input floating (100% PWM) the fan judders.

Frequency range can be anywhere between 0.02KHz up to 100KHz however we should avoid anything audible (Below 20Khz).

I cant have anything to control the duty, it needs to be done automatically as the supply voltage is switched on (hence the problems with using 555's). I thought about using a capacitor charging to supply the voltage but to get the time constant long enough I will need too large of a capacitor (cost & size constraints).

If you need more information please let me know.

Thanks

 

[Harald Kapp]

I cant have anything to control the duty,

how does that fit in with

I need to try and get a PWM duty sweep starting at around 30% up to 100%.

?

Something hat to control the sweep from 30 % to 100 %.

it needs to be done automatically as the supply voltage is switched on

This automatic circuit you envision is the control you "can't have". How do you expect this automatic function to operate? Sweep from 30 % to 100 % and back? in a linear (triangular) fashion or more smoothly in a sinusoidal fashion? at which rate? Why do you require the sweep at all?
What happens when you feed the MCP8063 with a pwm of fixed duty cycle? A two transistor astable multivibrator could be used in that case. Anything else will in essence require two generators: one for the pwm, one for the varying control signal, e.g. using two 555s.

Frequency range can be anywhere between 0.02KHz up to 100KHz however we should avoid anything audible (Below 20Khz).

The pwm frequency is irrelevant.According to the datasheet chapter 4.1 the pwm freqeuncy has no influence on the control signal for the motor and thus will not lead to audible artifacts from the motor.

 

[Lord_grezington]

Hello Harold

I know something needs to control the sweep, I was referring to no manual control to perform the sweep such as a voltage or pot.

I do not need to go back, just a sweep up to 100%.

The sweep can take any form (either triange or sinusoidal) as it wont really affect the fan much. The sweep is needed as the fan wont start at 100% (it just judders). Tests have shown that it starts ok if I feed it with a PWM acceleration. I mentioned the rate on the first post (full sweep up to 100% to take around 0.5-1s).

If I need two 555's that is what I may need to do, not sure if this is cheaper than the micro controller & regulator though.

Yes, you are correct. As the PWM is not fed into the fan there will not be any audible effects from the fan due to the frequency.

Thanks

 

[Harald Kapp]

I do not need to go back, just a sweep up to 100%.

So you want a soft start at 30 %, then ramp up to 100 %. No sweep (a sweep in my understanding would move alternatingly between the two extremes).
YOu could use a 555 in astable mode with a voltage ramp on pin 5 (CTRL). The ramp can easily be created by an RC element.
Or use this dedicated pwm chip: LTC6992 which is very small.
All these possible solutions however, will not work in the wide voltage range (2 V ... 15 V) you require.

You could try a capacitor from the MCP8063s pwm input to ground. This capacitor will hold the pwm input low when power is first applied, thus the motor will be off. When the capacitor charges through the internal pull-up resistor on the pwm pin, it will eventually reach the threshold to logic high and the motor is powered. Although the datasheet gives no explicit information on the startup behavior, there's a chance that the control algorithm within the controller can deal better with this condition than with a hard turn-on by the power supply alone. No guarantees given.

 

[Alec_t]

If you want cheap .......

Only 2 of the 6 gates in the IC are used. 0% to 100% duty cycle sweep. R1,C1 set the sweep period. R3,R4 set the sweep amount. R5,C2 set the basic PWM frequency. Note: the frequency varies by a 2:1 factor over the sweep period, but this is immaterial for the fan control.

 

[Lord_grezington]

Hello Alec

I have loaded up the simulation and yes, it seems to do exactly what I need. I am buying in some parts to knock it up on a bread board tomorrow.

Thanks

 

[Lord_grezington]

Hello Alec

Just to let you know, hardware has arrived and it works almost identically as the spice simulation (however sweep periods were lower on hardware). I had ti fiddle the values to get C1 as low as possible (due to physical size constraints) but so far it looks like its a winner...

Thanks

 

[Alec_t]

Thanks for the feedback. It's always good to know how close reality is to a simulation.