Time track with Rigol DSO

[bitrex]

I have a Rigol DS1102E. I know that some DSO have a "time track" feature
that allow you to demodulate and plot the RMS value of a PWM signal with
respect to time on a cycle-by-cycle basis for e.g. tuning control loops.

I don't immediately see a feature like that available on this scope, is
there a way to coax it into providing similar info that would be
valuable for that task using any of the other math and/or plotting
features it does have?

 

[bitrex]

On 10/29/19 12:41 AM, bitrex wrote:

I have a Rigol DS1102E. I know that some DSO have a "time track" feature
that allow you to demodulate and plot the RMS value of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width with respect to time

 

[Winfield Hill]

bitrex wrote...

On 10/29/19 12:41 AM, bitrex wrote:
... I know that some DSO have a "time track" feature
that allow you to demodulate and plot the RMS value
of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width
with respect to time.

Can you mention a few models with that feature?


--
Thanks,
- Win

 

[bitrex]

On 10/29/19 5:22 AM, Winfield Hill wrote:

bitrex wrote...

On 10/29/19 12:41 AM, bitrex wrote:
... I know that some DSO have a "time track" feature
that allow you to demodulate and plot the RMS value
of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width
with respect to time.

Can you mention a few models with that feature?

There's an article on e.g. EDN about it, helpfully in EDN fashion they
offer no insight into what models/price range of scopes offer this
feature or even what scope the author is using for the examples:

"An oscilloscope's track and trend features add two measurement-based
math functions that you can use to gain insight into measurements."

<https://www.edn.com/design/test-and-measurement/4460232/Trend-and-track--Two-useful-oscilloscope-diagnostic-tools>

So whatever oscilloscope or class of DSO this guy uses that he assumes
everyone has has this feature I guess, sorry I can't be more specific.

I did a little more research and it looks like plotting tracks of the
measurement parameters may be possible on Rigol's lower-end scopes using
their Scopeview software that connects to them which is also reported to
be clunky and a pain to set up unnnnngh

 

On Tue, 29 Oct 2019 00:44:25 -0400, bitrex <user@example.net> wrote:

On 10/29/19 12:41 AM, bitrex wrote:
I have a Rigol DS1102E. I know that some DSO have a "time track" feature
that allow you to demodulate and plot the RMS value of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width with respect to time

Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.

My Rigols will measure most any thing, but with numerical display, not
a trace.



--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[bitrex]

On 10/29/19 11:35 AM, jlarkin@highlandsniptechnology.com wrote:

On Tue, 29 Oct 2019 00:44:25 -0400, bitrex <user@example.net> wrote:

On 10/29/19 12:41 AM, bitrex wrote:
I have a Rigol DS1102E. I know that some DSO have a "time track" feature
that allow you to demodulate and plot the RMS value of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width with respect to time

Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.

My Rigols will measure most any thing, but with numerical display, not
a trace.

I'm having some difficulty setting it up to measure somewhat small
changes in pulse-width, tenths of a percent which the numerical display
measurements seem in theory to have enough precision to do.

If I set it to normal-acquire though looks to be some noise causing the
pulse width readout to jump around randomly in the tenths of % place
swamping the measurement. If I set it to time-average the waveform with
equi-time sampling then it looks like I'm losing precision, the width
measurement parameter seems to only step in increments of 1% or greater
when it's averaging.

Being able to plot the pulse width wrt time on a cycle-by-cycle basis
would help a lot

 

[bitrex]

On 10/29/19 1:34 PM, bitrex wrote:

On 10/29/19 11:35 AM, jlarkin@highlandsniptechnology.com wrote:
On Tue, 29 Oct 2019 00:44:25 -0400, bitrex <user@example.net> wrote:

On 10/29/19 12:41 AM, bitrex wrote:
I have a Rigol DS1102E. I know that some DSO have a "time track"
feature
that allow you to demodulate and plot the RMS value of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width with respect
to time

Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.

My Rigols will measure most any thing, but with numerical display, not
a trace.




I'm having some difficulty setting it up to measure somewhat small
changes in pulse-width, tenths of a percent which the numerical display
measurements seem in theory to have enough precision to do.

If I set it to normal-acquire though looks to be some noise causing the
pulse width readout to jump around randomly in the tenths of % place
swamping the measurement. If I set it to time-average the waveform with
equi-time sampling then it looks like I'm losing precision, the width
measurement parameter seems to only step in increments of 1% or greater
when it's averaging.

Being able to plot the pulse width wrt time on a cycle-by-cycle basis
would help a lot

The PWM frequency is 500kHz

 

[Cursitor Doom]

On Tue, 29 Oct 2019 08:35:30 -0700, jlarkin wrote:

My Rigols will measure most any thing, but with numerical display, not a
trace.

That's no good, then. I think you need to get yourself a Tektronix.



--
This message may be freely reproduced without limit or charge only via
the Usenet protocol. Reproduction in whole or part through other
protocols, whether for profit or not, is conditional upon a charge of
GBP10.00 per reproduction. Publication in this manner via non-Usenet
protocols constitutes acceptance of this condition.

 

[John Larkin]

On Tue, 29 Oct 2019 13:45:42 -0400, bitrex <user@example.net> wrote:

On 10/29/19 1:34 PM, bitrex wrote:
On 10/29/19 11:35 AM, jlarkin@highlandsniptechnology.com wrote:
On Tue, 29 Oct 2019 00:44:25 -0400, bitrex <user@example.net> wrote:

On 10/29/19 12:41 AM, bitrex wrote:
I have a Rigol DS1102E. I know that some DSO have a "time track"
feature
that allow you to demodulate and plot the RMS value of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width with respect
to time

Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.

My Rigols will measure most any thing, but with numerical display, not
a trace.




I'm having some difficulty setting it up to measure somewhat small
changes in pulse-width, tenths of a percent which the numerical display
measurements seem in theory to have enough precision to do.

If I set it to normal-acquire though looks to be some noise causing the
pulse width readout to jump around randomly in the tenths of % place
swamping the measurement. If I set it to time-average the waveform with
equi-time sampling then it looks like I'm losing precision, the width
measurement parameter seems to only step in increments of 1% or greater
when it's averaging.

Being able to plot the pulse width wrt time on a cycle-by-cycle basis
would help a lot

The PWM frequency is 500kHz

Period is 2 us. 0.1% of that is 2 ns.

Are you sure the PWM isn't jittering that much?

One trick with the Rigols (and many other digital scopes) is to always
trigger off a vertical channel, not external trigger. External will
have a full sample-period of jitter, but they do some Shannon thing to
reduce trigger jitter from a vertical input. I've measured 30 ps RMS
on a modest Rigol.

--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[John Larkin]

On Tue, 29 Oct 2019 18:18:17 -0000 (UTC), Cursitor Doom
<curd@notformail.com> wrote:

On Tue, 29 Oct 2019 08:35:30 -0700, jlarkin wrote:

My Rigols will measure most any thing, but with numerical display, not a
trace.

That's no good, then. I think you need to get yourself a Tektronix.

I got rid of my Tek, for a much better Rigol.

Can your Tek scope display duty cycle vs time? To 0.1%?

What does a non-numerical measurement mean anyhow?

--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Phil Allison]

jla...@highlandsniptechnology.com wrote:

-----------------------------------------

I have a Rigol DS1102E. I know that some DSO have a "time track" feature
that allow you to demodulate and plot the RMS value of a PWM signal


Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.

** Oh dear - more wrong advice from he who must not be doubted.

The RMS value of a rectangular pulse wave is proportional to the *SQUARE ROOT* of the duty cycle.

Connects with the I squared R rule.

Power in a resistor varies linearly with duty cycle.



..... Phil

 

[bitrex]

On 10/29/19 2:54 PM, John Larkin wrote:

On Tue, 29 Oct 2019 13:45:42 -0400, bitrex <user@example.net> wrote:

On 10/29/19 1:34 PM, bitrex wrote:
On 10/29/19 11:35 AM, jlarkin@highlandsniptechnology.com wrote:
On Tue, 29 Oct 2019 00:44:25 -0400, bitrex <user@example.net> wrote:

On 10/29/19 12:41 AM, bitrex wrote:
I have a Rigol DS1102E. I know that some DSO have a "time track"
feature
that allow you to demodulate and plot the RMS value of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width with respect
to time

Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.

My Rigols will measure most any thing, but with numerical display, not
a trace.




I'm having some difficulty setting it up to measure somewhat small
changes in pulse-width, tenths of a percent which the numerical display
measurements seem in theory to have enough precision to do.

If I set it to normal-acquire though looks to be some noise causing the
pulse width readout to jump around randomly in the tenths of % place
swamping the measurement. If I set it to time-average the waveform with
equi-time sampling then it looks like I'm losing precision, the width
measurement parameter seems to only step in increments of 1% or greater
when it's averaging.

Being able to plot the pulse width wrt time on a cycle-by-cycle basis
would help a lot

The PWM frequency is 500kHz

Period is 2 us. 0.1% of that is 2 ns.

Are you sure the PWM isn't jittering that much?

I was able to figure out what I needed to know by zooming in on the
pulse and using the infinite persistence so I could see better what the
extremes of the PWM were as the circuit outputting the PWM did its
things, that's really all I needed to know. I can then read off from the
timebase and do the math to get what RMS voltage deviation we're looking at.

Saving screencap .bmps to USB stick is also nice feature, whew 21st
century stuff! however they mounted the USB port on the front
upside-down. :|

One trick with the Rigols (and many other digital scopes) is to always
trigger off a vertical channel, not external trigger. External will
have a full sample-period of jitter, but they do some Shannon thing to
reduce trigger jitter from a vertical input. I've measured 30 ps RMS
on a modest Rigol.

 

On Wed, 30 Oct 2019 01:34:08 -0400, bitrex <user@example.net> wrote:

On 10/29/19 2:54 PM, John Larkin wrote:
On Tue, 29 Oct 2019 13:45:42 -0400, bitrex <user@example.net> wrote:

On 10/29/19 1:34 PM, bitrex wrote:
On 10/29/19 11:35 AM, jlarkin@highlandsniptechnology.com wrote:
On Tue, 29 Oct 2019 00:44:25 -0400, bitrex <user@example.net> wrote:

On 10/29/19 12:41 AM, bitrex wrote:
I have a Rigol DS1102E. I know that some DSO have a "time track"
feature
that allow you to demodulate and plot the RMS value of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width with respect
to time

Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.

My Rigols will measure most any thing, but with numerical display, not
a trace.




I'm having some difficulty setting it up to measure somewhat small
changes in pulse-width, tenths of a percent which the numerical display
measurements seem in theory to have enough precision to do.

If I set it to normal-acquire though looks to be some noise causing the
pulse width readout to jump around randomly in the tenths of % place
swamping the measurement. If I set it to time-average the waveform with
equi-time sampling then it looks like I'm losing precision, the width
measurement parameter seems to only step in increments of 1% or greater
when it's averaging.

Being able to plot the pulse width wrt time on a cycle-by-cycle basis
would help a lot

The PWM frequency is 500kHz

Period is 2 us. 0.1% of that is 2 ns.

Are you sure the PWM isn't jittering that much?

I was able to figure out what I needed to know by zooming in on the
pulse and using the infinite persistence so I could see better what the
extremes of the PWM were as the circuit outputting the PWM did its
things, that's really all I needed to know. I can then read off from the
timebase and do the math to get what RMS voltage deviation we're looking at.

What generates the PWM?

Saving screencap .bmps to USB stick is also nice feature, whew 21st
century stuff! however they mounted the USB port on the front
upside-down. :|

I like to put a post-it on the screen and take a pic with a real
camera. Otherwise I have a bunch of files like TEK0001 and TEK0002,
not unique over time, and I have to keep notes off to the side
somewhere about which is which.

https://www.dropbox.com/s/7gajbmt923oesli/NC7SV74_2.JPG?raw=1

One trick with the Rigols (and many other digital scopes) is to always
trigger off a vertical channel, not external trigger. External will
have a full sample-period of jitter, but they do some Shannon thing to
reduce trigger jitter from a vertical input. I've measured 30 ps RMS
on a modest Rigol.

--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[Winfield Hill]

jlarkin@highlandsniptechnology.com wrote...

https://www.dropbox.com/s/7gajbmt923oesli/NC7SV74_2.JPG?raw=1

Is that the input clock waveform? What'sNC7SV74_1.JPG, etc.?


--
Thanks,
- Win

 

[George Herold]

On Wednesday, October 30, 2019 at 10:55:48 AM UTC-4, jla...@highlandsniptechnology.com wrote:

On Wed, 30 Oct 2019 01:34:08 -0400, bitrex <user@example.net> wrote:

On 10/29/19 2:54 PM, John Larkin wrote:
On Tue, 29 Oct 2019 13:45:42 -0400, bitrex <user@example.net> wrote:

On 10/29/19 1:34 PM, bitrex wrote:
On 10/29/19 11:35 AM, jlarkin@highlandsniptechnology.com wrote:
On Tue, 29 Oct 2019 00:44:25 -0400, bitrex <user@example.net> wrote:

On 10/29/19 12:41 AM, bitrex wrote:
I have a Rigol DS1102E. I know that some DSO have a "time track"
feature
that allow you to demodulate and plot the RMS value of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width with respect
to time

Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.

My Rigols will measure most any thing, but with numerical display, not
a trace.




I'm having some difficulty setting it up to measure somewhat small
changes in pulse-width, tenths of a percent which the numerical display
measurements seem in theory to have enough precision to do.

If I set it to normal-acquire though looks to be some noise causing the
pulse width readout to jump around randomly in the tenths of % place
swamping the measurement. If I set it to time-average the waveform with
equi-time sampling then it looks like I'm losing precision, the width
measurement parameter seems to only step in increments of 1% or greater
when it's averaging.

Being able to plot the pulse width wrt time on a cycle-by-cycle basis
would help a lot

The PWM frequency is 500kHz

Period is 2 us. 0.1% of that is 2 ns.

Are you sure the PWM isn't jittering that much?

I was able to figure out what I needed to know by zooming in on the
pulse and using the infinite persistence so I could see better what the
extremes of the PWM were as the circuit outputting the PWM did its
things, that's really all I needed to know. I can then read off from the
timebase and do the math to get what RMS voltage deviation we're looking at.

What generates the PWM?

One could make a gated current source into cap and then use the
'scope trigger level to only trigger on pulses of a given length.
(time to amplitude converter)

George H.

Saving screencap .bmps to USB stick is also nice feature, whew 21st
century stuff! however they mounted the USB port on the front
upside-down. :|


I like to put a post-it on the screen and take a pic with a real
camera. Otherwise I have a bunch of files like TEK0001 and TEK0002,
not unique over time, and I have to keep notes off to the side
somewhere about which is which.

https://www.dropbox.com/s/7gajbmt923oesli/NC7SV74_2.JPG?raw=1


One trick with the Rigols (and many other digital scopes) is to always
trigger off a vertical channel, not external trigger. External will
have a full sample-period of jitter, but they do some Shannon thing to
reduce trigger jitter from a vertical input. I've measured 30 ps RMS
on a modest Rigol.





--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[John Larkin]

On Tue, 29 Oct 2019 18:00:47 -0700 (PDT), Phil Allison
<pallison49@gmail.com> wrote:

jla...@highlandsniptechnology.com wrote:

-----------------------------------------



I have a Rigol DS1102E. I know that some DSO have a "time track" feature
that allow you to demodulate and plot the RMS value of a PWM signal


Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.




** Oh dear - more wrong advice from he who must not be doubted.

Oh, don't be constantly obnoxious. You'll turn into a sad old prune
like Sloman.

The RMS value of a rectangular pulse wave is proportional to the *SQUARE ROOT* of the duty cycle.

Connects with the I squared R rule.

Power in a resistor varies linearly with duty cycle.

I meant power. PWM is great for things like temperature controllers
with resistive heaters, because it keeps the loop linear.

https://www.dropbox.com/s/k1u77jqlenfuxvm/Chimera_Top.JPG?raw=1

--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[John Larkin]

On 30 Oct 2019 11:26:26 -0700, Winfield Hill <winfieldhill@yahoo.com>
wrote:

jlarkin@highlandsniptechnology.com wrote...

https://www.dropbox.com/s/7gajbmt923oesli/NC7SV74_2.JPG?raw=1

Is that the input clock waveform? What'sNC7SV74_1.JPG, etc.?

That's the Q output of the flop, when the flop is cleared and then
clocked high.

The speed amazed me.

https://www.dropbox.com/sh/lgsbt8qaftj6sdq/AAAA0tfZUDksN1BzKQgtJjaNa?dl=0


The pickoff resistor may have a bit of shunt capacitance, but it's
still really fast, for a 16 cent flipflop.


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Winfield Hill]

John Larkin wrote...

On 30 Oct 2019 11:26:26 -0700, Winfield Hill <winfieldhill@yahoo.com
wrote:

jlarkin@highlandsniptechnology.com wrote...

https://www.dropbox.com/s/7gajbmt923oesli/NC7SV74_2.JPG?raw=1

Is that the input clock waveform? What'sNC7SV74_1.JPG, etc.?

That's the Q output of the flop, when the flop is cleared and then
clocked high.

The speed amazed me.

https://www.dropbox.com/sh/lgsbt8qaftj6sdq/AAAA0tfZUDksN1BzKQgtJjaNa?dl=0


The pickoff resistor may have a bit of shunt capacitance, but it's
still really fast, for a 16 cent flipflop.

Thanks John, I saved the images. Should come in useful!


--
Thanks,
- Win

 

[bitrex]

On 10/30/19 10:55 AM, jlarkin@highlandsniptechnology.com wrote:

On Wed, 30 Oct 2019 01:34:08 -0400, bitrex <user@example.net> wrote:

On 10/29/19 2:54 PM, John Larkin wrote:
On Tue, 29 Oct 2019 13:45:42 -0400, bitrex <user@example.net> wrote:

On 10/29/19 1:34 PM, bitrex wrote:
On 10/29/19 11:35 AM, jlarkin@highlandsniptechnology.com wrote:
On Tue, 29 Oct 2019 00:44:25 -0400, bitrex <user@example.net> wrote:

On 10/29/19 12:41 AM, bitrex wrote:
I have a Rigol DS1102E. I know that some DSO have a "time track"
feature
that allow you to demodulate and plot the RMS value of a PWM signal

Correction - I mean plot the cylce-by-cycle pulse width with respect
to time

Add an outboard lowpass filter to see duty cycle, or some sort of ramp
thing to see pulse width.

Or suck out waveforms and process externally.

RMS is usually linear on duty cycle.

My Rigols will measure most any thing, but with numerical display, not
a trace.




I'm having some difficulty setting it up to measure somewhat small
changes in pulse-width, tenths of a percent which the numerical display
measurements seem in theory to have enough precision to do.

If I set it to normal-acquire though looks to be some noise causing the
pulse width readout to jump around randomly in the tenths of % place
swamping the measurement. If I set it to time-average the waveform with
equi-time sampling then it looks like I'm losing precision, the width
measurement parameter seems to only step in increments of 1% or greater
when it's averaging.

Being able to plot the pulse width wrt time on a cycle-by-cycle basis
would help a lot

The PWM frequency is 500kHz

Period is 2 us. 0.1% of that is 2 ns.

Are you sure the PWM isn't jittering that much?

I was able to figure out what I needed to know by zooming in on the
pulse and using the infinite persistence so I could see better what the
extremes of the PWM were as the circuit outputting the PWM did its
things, that's really all I needed to know. I can then read off from the
timebase and do the math to get what RMS voltage deviation we're looking at.

What generates the PWM?

XOR as phase comparator

Saving screencap .bmps to USB stick is also nice feature, whew 21st
century stuff! however they mounted the USB port on the front
upside-down. :|


I like to put a post-it on the screen and take a pic with a real
camera. Otherwise I have a bunch of files like TEK0001 and TEK0002,
not unique over time, and I have to keep notes off to the side
somewhere about which is which.

https://www.dropbox.com/s/7gajbmt923oesli/NC7SV74_2.JPG?raw=1


One trick with the Rigols (and many other digital scopes) is to always
trigger off a vertical channel, not external trigger. External will
have a full sample-period of jitter, but they do some Shannon thing to
reduce trigger jitter from a vertical input. I've measured 30 ps RMS
on a modest Rigol.

 

[John Larkin]

On 30 Oct 2019 12:43:39 -0700, Winfield Hill <winfieldhill@yahoo.com>
wrote:

John Larkin wrote...

On 30 Oct 2019 11:26:26 -0700, Winfield Hill <winfieldhill@yahoo.com
wrote:

jlarkin@highlandsniptechnology.com wrote...

https://www.dropbox.com/s/7gajbmt923oesli/NC7SV74_2.JPG?raw=1

Is that the input clock waveform? What'sNC7SV74_1.JPG, etc.?

That's the Q output of the flop, when the flop is cleared and then
clocked high.

The speed amazed me.

https://www.dropbox.com/sh/lgsbt8qaftj6sdq/AAAA0tfZUDksN1BzKQgtJjaNa?dl=0


The pickoff resistor may have a bit of shunt capacitance, but it's
still really fast, for a 16 cent flipflop.

Thanks John, I saved the images. Should come in useful!

There is a thing that I call actual or critical setup time, which is
the actual time between D and CLK where the output transition happens,
or gets flakey or metastable. That can be positive or negative for
various parts. It is seldom specified. The data sheet required
setup/hold times presumably straddle that critical setup time.

I've seen parts that had a combination of prop delay and critical
setup time such that making a shift register is unreliable.

--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com