Jon and Charles: If you have EE degrees and I forgot, my apologies! But
Kevin is one of the guys on this list who could write at least parts of
"The Art of Electronics" while many of us, like me, can only read,
hoping some day to master the equations hiding under the
practical/intuitive presentation of that book. I spent a summer a few
years back getting my math back to where i was when I graduated from
high school. I still haven't gotten my act together to get the math to
where it was when I bombed out of Auburn. The inconvenient detail of
math exams is that they're frequently like programming settings:
"close" is a fail. :-)
Regards,
Pete
On 6/30/20 7:35 AM, Pete Soper wrote:
I was hoping an EE would join in. :-)
Your point is excellent. By the time you have the usable bits on the
ADC side you've either got PWM frequencies that don't happen to be
available w many hobby MPUs or enough complexity that only in an
industrial setting can you say you're saving X pennies by avoiding a
DAC chip. But the extra headache is that most (all?) DAC chips require
at least two pins for control. Anybody know of an exception?
Pete
I was hoping an EE would join in. :-)Your point is excellent. By the time you
have the usable bits on the ADC side you've either got PWM frequencies that
don't happen to be available w many hobby MPUs or enough complexity that only
in an industrial setting can you say you're saving X pennies by avoiding a DAC
chip. But the extra headache is that most (all?) DAC chips require at least two
pins for control. Anybody know of an exception?Pete
-------- Original message --------
From: Kevin Schilf via TriEmbed <[email protected]>
Date: 6/30/20 3:14 AM (GMT-05:00)
To: Jon Wolfe <[email protected]>, Charles A <[email protected]>
Cc: [email protected]
Subject: Re: [TriEmbed] PWM to Analog 0 to 5 VDC?
Hi Chuck,
A few thoughts sight unseen. :-)
Increasing switching speed pushes the switching artifacts out to the
right on a frequency plot giving you greater frequency "space" between
the DC value you want and the switching noise you don't. This gives
you flexibility on the pole frequency/ses and room for the low pass
filter to roll off without having to use a higher-order filter to
shorten the tail. Basically, you want the highest pole frequency with
sufficient roll off using the simplest acceptable filter topology
(single pole RC (simplest), double pole LRC, active filter). You can
use a Spice simulator (LTSpice, etc.) to do trade-offs.
Faster switching implies more power. No free lunch.
Playing devil's advocate, if you control the head end another option
is a simple D2A converter instead of PWM? Many micros now provide
D2A's. You may also be able to digitally filter the A2D readings in
the receiver.
Good luck,
Kevin Schilf
On Tuesday, June 30, 2020, 12:08:32 AM EDT, Charles A via TriEmbed
<[email protected]> wrote:
Well I have the low pass RLC filter into the OpAmp. I get less
ripple at the higher PWM frequencies however at those frequencies I
have much less granularity on the duty cycle. Trying to deal with the
ripple. Guess I should look at different RLC filter values next.
The ADC input is on another board and I have no control over it. I
need to supply a stable DC voltage to it. The current device (obsolete
now) only deviates by 10 mV or less according to its raw ADC reads. I
need to match that or the SW reading the ADC is not satisfied. My low
pass filter attempt so far deviates at best by 50 mV.
On Mon, Jun 29, 2020 at 11:57 PM Jon Wolfe <[email protected]
<mailto:[email protected]>> wrote:
Yeah an rc low pass filter is the way to go. I've done it that
way many times. The unity gain op amp could give you a buffer on
the filter output, but I think ADCs are usually high impedance
inputs. Depending on how fast you need the signal to change could
impact what op amp you would need to use. You could play around in
LT spice with different frequencies and component values to find
something that works. Pete would know better than me, but I think
the downside to higher frequency might be higher power draw, since
more of the signal is going through the cap in the filter.
On June 29, 2020 10:32:55 PM Pete Soper via TriEmbed
<[email protected] <mailto:[email protected]>> wrote:
The classic way to do this is with a low pass filter. If you
google "PWM DAC" you'll find what you need. But the performance
is going to be a function of the PWM frequency and how precisely
you can change the duty cycle.
Pete
-------- Original message --------
From: Charles A via TriEmbed <[email protected]
<mailto:[email protected]>>
Date: 6/29/20 10:23 PM (GMT-05:00)
To: [email protected] <mailto:[email protected]>
Subject: [TriEmbed] PWM to Analog 0 to 5 VDC?
Anyone have a favorite circuit or chip to convert a PWM signal to
a 0 to 5 VDC signal? The resulting voltage needs to be very
stable. It feeds an ADC input. I've looked at an RL circuit into
an OpAmp that also has a cap to ground at the OpAmp input. The
DVM says it's stable but the ADC reading the voltage says it's
not. I'm measuring 100 mV deviations. Would like to get to a 10
mV deviation. I've tried changing cap values on the input as well
as adding caps on the output side of the OpAmp. Made
improvements but still not good enough. So looking for
suggestions please.
Thanks,
Chuck
_______________________________________________
Triangle, NC Embedded Computing mailing list
To post message: [email protected]
<mailto:TriEmbed%40triembed.org>
List info:
http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
TriEmbed web site: http://TriEmbed.org
To unsubscribe, click link and send a blank message:
mailto:[email protected]
<mailto:unsubscribe-TriEmbed%40bitser.net>?subject=unsubscribe
_______________________________________________
Triangle, NC Embedded Computing mailing list
To post message: [email protected] <mailto:[email protected]>
List info: http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
TriEmbed web site: http://TriEmbed.org
To unsubscribe, click link and send a blank message:
mailto:[email protected]
<mailto:[email protected]>?subject=unsubscribe
-------- Original message --------From: Kevin Schilf via TriEmbed <[email protected]>
Date: 6/30/20 3:14 AM (GMT-05:00) To: Jon Wolfe <[email protected]>, Charles A
<[email protected]> Cc: [email protected] Subject: Re: [TriEmbed] PWM to Analog 0 to
5 VDC?
Hi Chuck,A few thoughts sight unseen. :-)Increasing switching speed pushes the
switching artifacts out to the right on a frequency plot giving you greater frequency
"space" between the DC value you want and the switching noise you don't. This
gives you flexibility on the pole frequency/ses and room for the low pass filter to roll
off without having to use a higher-order filter to shorten the tail. Basically, you want
the highest pole frequency with sufficient roll off using the simplest acceptable filter
topology (single pole RC (simplest), double pole LRC, active filter). You can use a
Spice simulator (LTSpice, etc.) to do trade-offs.Faster switching implies more power. No
free lunch.Playing devil's advocate, if you control the head end another option is a
simple D2A converter instead of PWM? Many micros now provide D2A's. You may also be
able to digitally filter the A2D readings in the receiver.Good luck,Kevin Schilf
On Tuesday, June 30, 2020, 12:08:32 AM EDT, Charles A via TriEmbed <[email protected]> wrote:
Well I have the low pass RLC filter into the OpAmp. I get less ripple at the higher PWM frequencies however at those frequencies I have much less granularity on the duty cycle. Trying to deal with the ripple. Guess I should look at different RLC filter values next. The ADC input is on another board and I have no control over it. I need to supply a stable DC voltage to it. The current device (obsolete now) only deviates by 10 mV or less according to its raw ADC reads. I need to match that or the SW reading the ADC is not satisfied. My low pass filter attempt so far deviates at best by 50 mV. On Mon, Jun 29, 2020 at 11:57 PM Jon Wolfe <[email protected]> wrote:
Yeah an rc low pass filter is the way to go. I've done it that way many times.
The unity gain op amp could give you a buffer on the filter output, but I think
ADCs are usually high impedance inputs. Depending on how fast you need the
signal to change could impact what op amp you would need to use. You could play
around in LT spice with different frequencies and component values to find
something that works. Pete would know better than me, but I think the downside
to higher frequency might be higher power draw, since more of the signal is
going through the cap in the filter.
On June 29, 2020 10:32:55 PM Pete Soper via TriEmbed <[email protected]>
wrote:
The classic way to do this is with a low pass filter. If you google "PWM DAC" you'll
find what you need. But the performance is going to be a function of the PWM frequency and how
precisely you can change the duty cycle.Pete-------- Original message --------From: Charles A
via TriEmbed <[email protected]> Date: 6/29/20 10:23 PM (GMT-05:00) To:
[email protected] Subject: [TriEmbed] PWM to Analog 0 to 5 VDC? Anyone have a favorite
circuit or chip to convert a PWM signal to a 0 to 5 VDC signal? The resulting voltage needs to
be very stable. It feeds an ADC input. I've looked at an RL circuit into an OpAmp that also
has a cap to ground at the OpAmp input. The DVM says it's stable but the ADC reading the
voltage says it's not. I'm measuring 100 mV deviations. Would like to get to a 10 mV
deviation. I've tried changing cap values on the input as well as adding caps on the output
side of the OpAmp. Made improvements but still not good enough. So looking for suggestions
please.Thanks,Chuck
_______________________________________________
Triangle, NC Embedded Computing mailing list
To post message: [email protected]
List info: http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
TriEmbed web site: http://TriEmbed.org
To unsubscribe, click link and send a blank message:
mailto:[email protected]?subject=unsubscribe
_______________________________________________Triangle, NC Embedded Computing
mailing listTo post message: [email protected] info:
http://mail.triembed.org/mailman/listinfo/triembed_triembed.orgTriEmbed web
site: http://TriEmbed.orgTo unsubscribe, click link and send a blank message:
mailto:[email protected]?subject=unsubscribe
_______________________________________________
Triangle, NC Embedded Computing mailing list
To post message: [email protected]
List info: http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
TriEmbed web site: http://TriEmbed.org
To unsubscribe, click link and send a blank message:
mailto:[email protected]?subject=unsubscribe
