RC Low pass filter Q

[thomas]

I have a RCR-433-HP radiotronix receiver which outputs digital hash (random
high-lows) when there is no transmitter in range. This makes using it as a
digital input for a microcontroller troublesome.

I put a RC low-pass filter on the output and was able to smooth the hash to
a nice 1.5-2.5V ripple, yet maintianed a nice 0V and 5V signal when the
xmitter is in range. Unfortunately this ripple is right at the 'trip' level
of the mcu so I still get some noise. How can I reduce the voltage level
(~1V) to avoid the false trips?

other info....I'm a newbie so I totally guessed on the RC filter values
using a scope and a pile of Rs and Cs, ended up with 330 ohms and 94uf. My
goal is to transmit ON and OFF (I guess this is considered DC), not serial
data.
thx

 

[John Popelish]

I have a RCR-433-HP radiotronix receiver which outputs digital hash (random
high-lows) when there is no transmitter in range. This makes using it as a
digital input for a microcontroller troublesome.

I put a RC low-pass filter on the output and was able to smooth the hash to
a nice 1.5-2.5V ripple, yet maintianed a nice 0V and 5V signal when the
xmitter is in range. Unfortunately this ripple is right at the 'trip' level
of the mcu so I still get some noise. How can I reduce the voltage level
(~1V) to avoid the false trips?

other info....I'm a newbie so I totally guessed on the RC filter values
using a scope and a pile of Rs and Cs, ended up with 330 ohms and 94uf. My
goal is to transmit ON and OFF (I guess this is considered DC), not serial
data.
thx

Most microprocessor inputs are very high impedance, so a pretty
significant series resistance will not drop much voltage because of
the input. Try something like a 10k resistor between signal and micro
input, with a .47 uf cap to ground. This should give approximately
the same result as the one you have now. Then you can tie an
additional 20k resistor between the input and ground, so that the
average voltage has to go higher to produce the same input voltage.

If you have a spare output you can tie the 20k resistor to that, and
every time you read the input, copy the state to that output. This
will require the signal to have a larger change to swing from a high
to low or vice versa.

 

[thomas]

Most microprocessor inputs are very high impedance, so a pretty
significant series resistance will not drop much voltage because of
the input. Try something like a 10k resistor between signal and micro
input, with a .47 uf cap to ground. This should give approximately
the same result as the one you have now. Then you can tie an
additional 20k resistor between the input and ground, so that the
average voltage has to go higher to produce the same input voltage.

If you have a spare output you can tie the 20k resistor to that, and
every time you read the input, copy the state to that output. This
will require the signal to have a larger change to swing from a high
to low or vice versa.

That worked very nicely.
I did have to keep the 94uf to maintain a nice smooth signal. The hash is a
square wave with unequal 5V and 0V duties, perhaps that is why.?

10K and 22K brought me down to around 1V and no more false trips! (I had
tried something like this but working against my original 330 ohm I wasn't
able to get it to work). Would there be any hazard in using a smaller "22K"
to bring the voltage down a little more, as a safety factor? OR is it
better to go with your suggetion of using an extra output as 'feedback'.

Thx a bunch

 

[John Popelish]

That worked very nicely.
I did have to keep the 94uf to maintain a nice smooth signal. The hash is a
square wave with unequal 5V and 0V duties, perhaps that is why.?

10K and 22K brought me down to around 1V and no more false trips! (I had
tried something like this but working against my original 330 ohm I wasn't
able to get it to work). Would there be any hazard in using a smaller "22K"
to bring the voltage down a little more, as a safety factor? OR is it
better to go with your suggetion of using an extra output as 'feedback'.

Thx a bunch

You can lower the 22k as much as necessary to prevent false high
states, until you reach the point that a high is not possible. Keep
in mind that the logic input does not have a precisely repeatable
threshold voltage, so if you expect this trick to work with many uP
units, you need a comparator between the filter and the uP or you need
ot use the a comparator input function on the pic to increase the
decision precision. Using the output to reinforce the last measured
state just increases the ripple that can be tolerated without giving
false state changes. This can also be done with a comparator. It is
called hysterisis. You might look up the key word, "schmitt trigger".

 

[Dbowey]

Thomas posted:

<< I have a RCR-433-HP radiotronix receiver which outputs digital hash (random
high-lows) when there is no transmitter in range. This makes using it as a
digital input for a microcontroller troublesome.

I put a RC low-pass filter on the output and was able to smooth the hash to a
nice 1.5-2.5V ripple, yet maintianed a nice 0V and 5V signal when the xmitter
is in range. Unfortunately this ripple is right at the 'trip' level of the mcu
so I still get some noise. How can I reduce the voltage level
(~1V) to avoid the false trips?

other info....I'm a newbie so I totally guessed on the RC filter values
using a scope and a pile of Rs and Cs, ended up with 330 ohms and 94uf. My
goal is to transmit ON and OFF (I guess this is considered DC), not serial
data.
The problem I see is that it uses a super-regenerative receiver, and they
generate lots of broad-spectrum noise in the absence of a signal. The
schematic of the receiver is rather poor - you cannot be certain about the
stage coupling. You don't say what you are using for an antenna. Reducing the
antenna coupling might reduce the regeneration, giving you some improvement in
idle noise, but at the expense of range.

Bottom line is you probably need a better receiver.

Don

 

[thomas]

Thomas posted:

<< I have a RCR-433-HP radiotronix receiver which outputs digital hash (random
high-lows) when there is no transmitter in range. This makes using it as a
digital input for a microcontroller troublesome.

I put a RC low-pass filter on the output and was able to smooth the hash to a
nice 1.5-2.5V ripple, yet maintianed a nice 0V and 5V signal when the xmitter
is in range. Unfortunately this ripple is right at the 'trip' level of the mcu
so I still get some noise. How can I reduce the voltage level
(~1V) to avoid the false trips?

other info....I'm a newbie so I totally guessed on the RC filter values
using a scope and a pile of Rs and Cs, ended up with 330 ohms and 94uf. My
goal is to transmit ON and OFF (I guess this is considered DC), not serial
data.

The problem I see is that it uses a super-regenerative receiver, and they
generate lots of broad-spectrum noise in the absence of a signal. The
schematic of the receiver is rather poor - you cannot be certain about the
stage coupling. You don't say what you are using for an antenna. Reducing the
antenna coupling might reduce the regeneration, giving you some improvement in
idle noise, but at the expense of range.

Bottom line is you probably need a better receiver.

Don

I realize it's a bottom-of-the-barrel receiver but this is more for learning
than anything else. I'm using a 1/4w LINX rubber duckie. Radiotronix has an
app note that suggest using a SAW filter on the TX and RX to narrow the
bandwidth. I think I'll try it. Although after reading your post I don't
think that is exactly the broad spectrum you are referring to.
t

 

[Dbowey]

Thomas posted:
<<

Thomas posted:

<< I have a RCR-433-HP radiotronix receiver which outputs digital hash (random
high-lows) when there is no transmitter in range. This makes using it as a
digital input for a microcontroller troublesome.

I put a RC low-pass filter on the output and was able to smooth the hash to a
nice 1.5-2.5V ripple, yet maintianed a nice 0V and 5V signal when the xmitter
is in range. Unfortunately this ripple is right at the 'trip' level of the mcu
so I still get some noise. How can I reduce the voltage level
(~1V) to avoid the false trips?

other info....I'm a newbie so I totally guessed on the RC filter values
using a scope and a pile of Rs and Cs, ended up with 330 ohms and 94uf. My
goal is to transmit ON and OFF (I guess this is considered DC), not serial
data.

The problem I see is that it uses a super-regenerative receiver, and they
generate lots of broad-spectrum noise in the absence of a signal. The
schematic of the receiver is rather poor - you cannot be certain about the
stage coupling. You don't say what you are using for an antenna. Reducing the
antenna coupling might reduce the regeneration, giving you some improvement in
idle noise, but at the expense of range.

Bottom line is you probably need a better receiver.

Don

I realize it's a bottom-of-the-barrel receiver but this is more for learning
than anything else. I'm using a 1/4w LINX rubber duckie. Radiotronix has an app
note that suggest using a SAW filter on the TX and RX to narrow the bandwidth.
I think I'll try it. Although after reading your post I don't think that is
exactly the broad spectrum you are referring to.

What happens to the noise if you remove the antenna? Super-regen receivers are
quite sensitive, but very broad without some filtering such as the saw filter,
or just an L/C filter. If you remove the antenna and observe the result, you
will know if a saw filter will help.

Don