LC filter in a design, why not an RC filter?

Status
Not open for further replies.
ericgibbs said:
hi Nigel,

Sounds fun, any particular configuration or parameters would you like to see.
Click to expand...

I don't really know, I suppose a simple rising edge on the input to the choke, and a scope display on the capacitor?.

Is it possible to simulate the characteristics of AVR (or PIC) outputs?.
 
hi,
This is what LTS shows.

Simulating a short pulse on the the paralleled pins of 3V, the Vout [analog] rings at approx 2Khz.

The 47K simulates the loading across the 100nF.

E
 

Attachments

  • AAesp01.gif
    38.8 KB · Views: 160
Rings a bit doesn't it

Can you do the same for the current? (directly from the source of course).
 
Nigel Goodwin said:
Rings a bit doesn't it

Can you do the same for the current? (directly from the source of course).
Click to expand...


Showing current at MCU pins [V1]

Sorry about delay in replying, working on a PCB artwork.

E

EDIT: Added MCU internal clamp diodes, 2nd image
 

Attachments

  • AAesp04.gif
    44.1 KB · Views: 157
  • AAesp05.gif
    36.7 KB · Views: 148
So what about current if you replace the choke with a resistor? - it's good this remote control simulation
 
Nigel Goodwin said:
So what about current if you replace the choke with a resistor? - it's good this remote control simulation
Click to expand...

I think what designer may have in mind is to use the inductor to reduce the inrush current drawn from the MCU pins that are charging the 100nF, when the pins are set low to high.

Also it will minimise the voltage loss across the inductor, if its resistance is low.

I have increased the load current to approx 1mA.

Without the inductor or resistor connected between the MCU and 100nF, the MCU pins are required to try to output a 300mA current spike.

I have added some source resistance to the MCU pins to limit the s/c current to 25mA/pin for the sim.
 

Attachments

  • AAesp06.gif
    36.8 KB · Views: 158
Last edited:
Interestingly this is the MCU pin current spike with the 47uH inductor and the MCU source set to only 1R, its well within the MCU specification.

NOTE: the inductor self resistance is set as zero ohms

E
 

Attachments

  • AAesp07.gif
    35.2 KB · Views: 165

That was Mr RB's thought - and the simulation seems to show that even a 47uH with only a 0.1uF does reduce the surge.

As to whether it makes any practical difference though is a different matter

But it's probably cleared up the reasoning behind it.
 
Hello again,


Well, there is more than one way to set the output drive logic for most uC chips. So far we only assumed it was providing a drive pulse of zero to plus supply voltage. But the pin can easily provide a high impedance to plus supply voltage pulse. That would make the 'off' time look like a high impedance rather than a pull down to zero volts.

With a high impedance during the pulse 'off' time and the plus supply voltage during the pulse 'on' time, we would see a more conventional output that looks more like a PWM modulated DC voltage. That's not to say that is really what they are doing here, but that is what we would see should we drive it that way instead.
This could be simulated to some degree with a diode between the pulse source and the inductor capacitor circuit.
 
I wonder why not to use a mosfet to switch analog on and off. This makes for more robust system with less parts.

Inductor, capacitor, two resistors and 3 pins

or

Mosfet and 1 pin.

It must be something else that made them go the distance.

The pin's clamping diode does make it into a full buck SMPS configuration, so switching the pins on and off is capable of producing less voltage. Is it a 5V MCU? Making 3V for analogs? Or less? It is a battery application and they seem to be paranoid about saving power.
 
samovar said:
I wonder why not to use a mosfet to switch analog on and off. This makes for more robust system with less parts.

Inductor, capacitor, two resistors and 3 pins

or

Mosfet and 1 pin.
Click to expand...

Using a mosfet will still produce an identical surge, so would still require the choke - assuming that's what it's there for, and that it's actually having any effect.

A project I'm involved in does use a mosfet to switch the power, actually to a GSM modem chip, so there's a very substantial surge there - yet no need for a choke.
 
hi Al,

Using a switch thats closed or open to simulate a MCU pin set High and Tristate , this what I see.

If you think about it the Vo on the 100nF cannot go higher than 3V+ diode clamp voltage.

I am assuming like most MCU's the pins are internally clamped.

E.

BTW: I have tried a range of switching frequencies
 

Attachments

  • AAesp01.gif
    33 KB · Views: 146
Nigel Goodwin said:
Using a mosfet will still produce an identical surge, so would still require the choke - assuming that's what it's there for, and that it's actually having any effect.
Click to expand...

Yes, but the mosfet will take power from the rail, so the surge will not affect the pin(s).

Microchip in their data sheets for PIC24H (that's what I use) suggest using a choke for analog power. I tried and I didn't see any difference.
 
ericgibbs said:
Using a switch thats closed or open to simulate a MCU pin set High and Tristate , this what I see.
Click to expand...

I think the diodes should be connected to the point labelled mcu1, rather than Vo.

I would also put the voltmeter to Vo instead of mcu1 because that's what the analog devices will see.
 
samovar said:
Yes, but the mosfet will take power from the rail, so the surge will not affect the pin(s).
Click to expand...

I think the idea was to prevent surges from the battery, which was tiny - no problem for such a tiny surge via the pins, which will self limit it anyway.

Interesting about the 24H - perhaps the AVR datasheet suggested the same thing, which is the sole reason it's there
 
samovar said:
I think the diodes should be connected to the point labelled mcu1, rather than Vo.

I would also put the voltmeter to Vo instead of mcu1 because that's what the analog devices will see.
Click to expand...

The diodes should be the other side of the inductor, but the difference in Vo is not discern able.

The voltmeter Plot already shows Vo not the MCU voltage
 

Hi Eric,

Yes that looks much better now. But im still a little puzzled by the very low value capacitor of 100nf. That tells me that they wanted to be able to change the DC level very quickly. But it's almost too fast, and the values shown would probably require a switching frequency of 100kHz or something. That's a lot of housekeeping for the uC i think unless this analog output is a very important feature. Unless maybe they use another cap as part of the load.

But at least we are now seeing PWM, and the internal diode to ground means it's a real buck circuit.

BTW where did that 3k resistor come from? Is that the load now?
 
Status
Not open for further replies.

Similar threads

P
  • Question
Linear analog filter with zero amplitude in stopband
Replies
14
Views
3K
Tony Stewart
S
  • Question
Help with high pass filter in bass guitar
Replies
5
Views
2K
Scornn
S
  • Question
Fg limit frequency calculation. Active band-pass filter
Replies
18
Views
3K
Tony Stewart
F
  • Question
Need to design and build 250W BLDC inverters for drones
Replies
2
Views
2K
Papabravo
S
  • Question
What is your favorite analog filter design software?
Replies
7
Views
10K
transconductance
T
Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…