Power supply for time source concerns
On 12/23/19 5:01 AM, Achim Gratz wrote:
Am 22.12.2019 um 20:37 schrieb Taka Kamiya via time-nuts:
Most commercially manufactured time and frequency sources use
switching power supply.
[...]
The suggestion to "just use a linear power supply", especially if it was
designed some decades ago is probably not getting the desired results
these days. There is a lot more conducted noise on the power lines than
there used to be, starting from several kHz right into RF territory.
An excellent point - The plethora of "offline SMPS" such as phone
chargers, and various and sundry USB power supplies, means that even if
they have reasonable conducted emissions, there's a lot of them all
contributing.
Linear regulators generally have bad PSRR at higher frequencies (often
starting to degrade in the low kHz region already) and unless you have a
clean input supply you pass any HF noise almost directly to the load.
SPS are not necessarily worse than linear supplies when it comes to
noise, but it generally shows up in different places frequency-wise (the
detailed characteristics depend a lot on the exact topology chosen, so
that's a whole 'nother dimension of things to consider when chosing). It
is also harder to contain the switching noise as you have several
high-current loops typically whose area must be kept as small as
possible to not radiate noise. Conducted noise can be more easily
filtered, but it gets out both through the input and the output side.
The input side is often neglected a bit, which can come back to bite you
when you have multiple supply rails in the system. Most switching
topologies (or at least the ones that can be "clean" enough for the
purposes under discussion) will produce a triangular ripple at the
output at the switching frequency. You can make it smaller and smaller
at the expense of load regulation, but it usually is easier to just deal
with whichever number of millivolts that you're left with by putting an
LDO post-regulator directly at the load (you put it at the load so you
have better regulation and you can also use the supply line between the
SPS and the LDO as part of a Pi filter). The LDO needs to have good
PSRR at the switching frequency and maybe the first or second harmonic,
the rest of the spectrum should already have been dealt with by filtering.
It is tough to build a "small" input filter that has good rejection at
low frequencies (<100kHz?)
What we wound up doing for a low noise design was cascade the following:
Unregulated 9-15V bus
input filter
Buck converter to 8V (boost converters seem to have more noise)
Buck converter to "usage voltage" + 0.5-1V (i.e. if it's destined for
3.3V, buck down to around 4-4.3V)
Pi filter with 60 dB "design" attenuation (probably less on the board)
Board to board connector
Pi filter with 60 dB "design" attenuation
LT3042 LDO to load.
We did some experimentation to measure the voltage drop across the
filters, and then adjusted the output of the second buck converter so
the LDO would have enough headroom over temperature.
The layout was 4 layer boards, nothing exotic, other than the usual
separation of pre & post regulation power & ground planes.
Late Jim Williams' (RIP) application notes are always a good read:
https://www.analog.com/media/en/technical-documentation/application-notes/an70.pdf
https://www.analog.com/media/en/technical-documentation/application-notes/an101f.pdf
https://www.analog.com/media/en/technical-documentation/application-notes/AN118fb.pdf
On 12/23/19 5:42 AM, Bob kb8tq wrote:
Hi
If you dig into the app notes on the LT304x parts after a lot of extolling the
wonders and virtues, they eventually get into magnetic coupling between the
“upstream” and “downstream” bypass components. Taking care of that with
proper layout is possible, but non-trivial.
And not entirely predictable by modeling, even with the fancy
multiphysics signal integrity modeling.
It is one of those things where you will probably need a few iterations.
Is milivolts of noise “ok”? Maybe it is. There are a lot of devices out there that
run with 10’s of milivolts of noise on the supply line. Do you need <100uV p-p
over 10KHz to 100 MHz? There are devices that do. If that’s what you need,
it’s going to be a struggle. There is no one single “good enough” number.
The power supply world still stops at pretty low frequencies. Very common IC’s
these days are quite happy to produce crud at the “many GHz” level. There are
a few folks who carry around cell phones that put out signals up there as well.
Regulator IC’s will only do part of the job, filtering in one form or the other (likely
several forms) is still needed.
Yes, I think, though, that filtering (even over multiple decades) is
easier at higher frequencies - Assuming you can tolerate the IR voltage
drop through the filters.
Past that system layout begins to get into the act. You can spend a couple of fun
weeks in intro level classes on this sort of stuff. You can be the prof presenting the
class and pretty quickly get to “that would require some research” as the answer
to this or that seemingly basic question. It’s not in any way a trivial topic.
So what to do?
Bottom line is still that test equipment is your friend. Spectrum analyzers that cover
a wide range of frequencies (and have good sensitivity) are part of the “kit” that
allows you to keep things quiet. You can only go just so far ( = not very far) before
you are into the “verify” stage of things. Indeed sound cards and SDR’s can collect
some of this data so there are a range of tools you can use.
Checking cabling / grounding / cross talk is very much part of setting up any RF
bench. It always has been. Tracking down that stupid whatever on the other side
of the room that is putting out a ton of noise is part of the drill. Repeating the
process from time to time also is part of the fun. Looking at the data you collect to
spot “strange” stuff is part of the deal …… There is work involved and you need
to understand what you are doing.
I think understanding how the effect of unwanted interference (by
whatever path) manifests itself in the desired output is part of the
challenge - In many cases, the most sensitive detector of interference
is the unit under test (certainly this is the case for a low noise
receiver).
One thing that can help evaluate a design is it artificially introduce
interference or noise into the power supply - if you can't see a change
in the output of the UUT, then the filtering/isolation scheme must be
working. If the artificial interference is at a level that is >> the
level you'd see in expected use, then you're in great shape.
I think the conversation is going into an area way beyond my initial intent.
My goal is to have a "good enough" source for 24V to run multiple time sources. Yesterday, I measured the output of switcher I've been using for testing purpose only. Holy cow it's dirty! 120mV peak-to-peak wide band noise. Put one ferrite bead on plus side and it had an amazing result. Broadband noise disappearance and periodic pulse (still 120mV p-p)
My goal is to have about 5mV p-p or lower capable of delivering 20Amp in switching arrangement. It's easy to do with linear regulators but it makes the room awfully warm. Why 5mV? Because it is good enough. If more is necessary, it will have to be dealt with locally inside the time source itself. Cabling can pick up that much easily.
(Mr.) Taka Kamiya
KB4EMF / ex JF2DKG
On Monday, December 23, 2019, 1:09:58 PM EST, jimlux <jimlux@earthlink.net> wrote:
On 12/23/19 5:42 AM, Bob kb8tq wrote:
Hi
If you dig into the app notes on the LT304x parts after a lot of extolling the
wonders and virtues, they eventually get into magnetic coupling between the
“upstream” and “downstream” bypass components. Taking care of that with
proper layout is possible, but non-trivial.
And not entirely predictable by modeling, even with the fancy
multiphysics signal integrity modeling.
It is one of those things where you will probably need a few iterations.
Is milivolts of noise “ok”? Maybe it is. There are a lot of devices out there that
run with 10’s of milivolts of noise on the supply line. Do you need <100uV p-p
over 10KHz to 100 MHz? There are devices that do. If that’s what you need,
it’s going to be a struggle. There is no one single “good enough” number.
The power supply world still stops at pretty low frequencies. Very common IC’s
these days are quite happy to produce crud at the “many GHz” level. There are
a few folks who carry around cell phones that put out signals up there as well.
Regulator IC’s will only do part of the job, filtering in one form or the other (likely
several forms) is still needed.
Yes, I think, though, that filtering (even over multiple decades) is
easier at higher frequencies - Assuming you can tolerate the IR voltage
drop through the filters.
Past that system layout begins to get into the act. You can spend a couple of fun
weeks in intro level classes on this sort of stuff. You can be the prof presenting the
class and pretty quickly get to “that would require some research” as the answer
to this or that seemingly basic question. It’s not in any way a trivial topic.
So what to do?
Bottom line is still that test equipment is your friend. Spectrum analyzers that cover
a wide range of frequencies (and have good sensitivity) are part of the “kit” that
allows you to keep things quiet. You can only go just so far ( = not very far) before
you are into the “verify” stage of things. Indeed sound cards and SDR’s can collect
some of this data so there are a range of tools you can use.
Checking cabling / grounding / cross talk is very much part of setting up any RF
bench. It always has been. Tracking down that stupid whatever on the other side
of the room that is putting out a ton of noise is part of the drill. Repeating the
process from time to time also is part of the fun. Looking at the data you collect to
spot “strange” stuff is part of the deal …… There is work involved and you need
to understand what you are doing.
I think understanding how the effect of unwanted interference (by
whatever path) manifests itself in the desired output is part of the
challenge - In many cases, the most sensitive detector of interference
is the unit under test (certainly this is the case for a low noise
receiver).
One thing that can help evaluate a design is it artificially introduce
interference or noise into the power supply - if you can't see a change
in the output of the UUT, then the filtering/isolation scheme must be
working. If the artificial interference is at a level that is >> the
level you'd see in expected use, then you're in great shape.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
On 12/23/19 10:17 AM, Taka Kamiya via time-nuts wrote:
I think the conversation is going into an area way beyond my initial intent.
My goal is to have a "good enough" source for 24V to run multiple time sources. Yesterday, I measured the output of switcher I've been using for testing purpose only. Holy cow it's dirty! 120mV peak-to-peak wide band noise. Put one ferrite bead on plus side and it had an amazing result. Broadband noise disappearance and periodic pulse (still 120mV p-p)
My goal is to have about 5mV p-p or lower capable of delivering 20Amp in switching arrangement. It's easy to do with linear regulators but it makes the room awfully warm. Why 5mV? Because it is good enough. If more is necessary, it will have to be dealt with locally inside the time source itself. Cabling can pick up that much easily.
Check out Jim Brown K9YC's compendia of filters - he's done measurements
on a variety of ferrite mixes in a variety of configurations (number of
turns, number of cores, etc.)
http://audiosystemsgroup.com/RFI-Ham.pdf
http://k9yc.com/2018Cookbook.pdf
He's done a lot of work to take the raw material data sheets and turn it
into practical recipes.
Of course, you could get yourself a copy of Snelling's book on "Soft
ferrites: properties and applications" which is the recognized tome on this.
Am 23.12.2019 um 16:03 schrieb jimlux:
It is tough to build a "small" input filter that has good rejection at
low frequencies (<100kHz?)
Yes if you want a passive filter, but you can view an LDO as an active
low-pass in some sense. Again you need to take care that noise cannot
bypass it, but most of it should be conducted noise in the LF case.
Active filtering is also a good option if the power you need to deliver
is low.
Buck converter to 8V (boost converters seem to have more noise)
That again depends on topology and control type. The canned converters
are almost always optimized to have the lowest number of switches and
work with cheap magnetics (single coil) without easily entering
problematic operation modes, noise is only a secondary concern.
Producing less noise in the first place is possible if you change those
constraints; i.e. allow more switches, more complicated energy storage
(capacitors and magnetics) or employ control algorithms that are less
robust and hence need to be tuned to the load to be stable.
--
Achim.
(on the road :-)
That again depends on topology and control type. The canned converters
are almost always optimized to have the lowest number of switches and
work with cheap magnetics (single coil) without easily entering
problematic operation modes, noise is only a secondary concern.
That depends a LOT on which canned converter you decide to buy,
if you only go after price, or W/mm³ capacity, then certainly yes.
But for a one-off application like this, any money saved on a
cheap model is easily lost many times over in the trouble it will
cause.
But returning to the original post: Has anybody ever characterized
how much difference it makes to use two different PSU's for heater
vs. electronics sides of telecom Rb's ?
I'm sure there is a reason why they make it two different pins ?
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
I, too, am baffled by why some Rb standards use separate pins for oven power
and electronics package power, especially since they don't seem to use
separate
return pins as well.
At Arecibo we had a utility PRS-10 mounted in a box with fan along with 1
each
switching supply (for the oven) and linear supply for the electronics. It
seemed
not to have any problems.
I also would like to know whether or not there is risk of damaging a PRS-10
if
one or the other of separate supplies fails and the unit spends some time in
that situation (like hours or days). I particularly worry about the case
of the oven
supply staying on while the electronics supply had gone off- could the oven
control then put full power to the oven indefinitely, risking destruction
of the
physics package?
2nd question: can anybody tell me what level of switching-frequency
sidebands on
the 10 MHz output of an Rb is generally considered "bad" versus "good"?
Dana
On Tue, Dec 24, 2019 at 6:05 AM Poul-Henning Kamp phk@phk.freebsd.dk
wrote:
That again depends on topology and control type. The canned converters
are almost always optimized to have the lowest number of switches and
work with cheap magnetics (single coil) without easily entering
problematic operation modes, noise is only a secondary concern.
That depends a LOT on which canned converter you decide to buy,
if you only go after price, or W/mm³ capacity, then certainly yes.
But for a one-off application like this, any money saved on a
cheap model is easily lost many times over in the trouble it will
cause.
But returning to the original post: Has anybody ever characterized
how much difference it makes to use two different PSU's for heater
vs. electronics sides of telecom Rb's ?
I'm sure there is a reason why they make it two different pins ?
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Hi
On Dec 24, 2019, at 6:40 AM, Poul-Henning Kamp phk@phk.freebsd.dk wrote:
That again depends on topology and control type. The canned converters
are almost always optimized to have the lowest number of switches and
work with cheap magnetics (single coil) without easily entering
problematic operation modes, noise is only a secondary concern.
That depends a LOT on which canned converter you decide to buy,
if you only go after price, or W/mm³ capacity, then certainly yes.
But for a one-off application like this, any money saved on a
cheap model is easily lost many times over in the trouble it will
cause.
But returning to the original post: Has anybody ever characterized
how much difference it makes to use two different PSU's for heater
vs. electronics sides of telecom Rb's ?
The “old time” answer was that a poorly regulated / poorly filtered supply was
considered “ok” for a heater. For the active electronics you wanted something
nice and stable / clean. To your point, once you get around to measuring
this, stability wise that answer often does not hold up. Noise wise, you are right
back to “what frequency?” ….
The somewhat more complex “old time” answer was that you don’t want the
honking big current of the heater coming off the supply you have tried so hard
to super-regulate. ( = it’s the supply that’s the issue not the Rb it’s self). Obviously
that’s going to depend on how the supply was designed.
Of course next layer to the onion is …. where does the ground current go? …. hmm….
Bob
I'm sure there is a reason why they make it two different pins ?
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
I just found something funny. I have been thinking "clean" power supplies that connects to AC mains. Then I thought, what about lead acid batteries?? So I went to my lab and took some measurement. This is a 12V 7A lead acid sealed battery, the kind commonly found on UPS devices.
The result? Surprise? The battery is oscillating at 5MHz and noise level is 15mV peak-to-peak!!!!
Of course, not...! Battery is pure DC and while voltage might drift, this is not that. For the record, a charger of any kind is not hooked up. It's one battery all by itself. Battery is not oscillating but that's what the measurement actually shows. That brings another point in my quest to "clean" power source. It's not just the power supply but the whole lab eco system has to be considered. Having one master 24V source (my original plan) is not the answer if mV level noise is going to be a problem.
This "discovery" puts whole new layer to having a nice power supply.....
Just as a point of reference, I hooked up a common cheap float charger. The charger itself has 2V p-p noise. Connected to battery, it still have 100mV p-p noise. There goes battery = noise sponge theory....
(Mr.) Taka Kamiya
KB4EMF / ex JF2DKG
On Tuesday, December 24, 2019, 3:00:50 PM EST, Bob kb8tq <kb8tq@n1k.org> wrote:
Hi
On Dec 24, 2019, at 6:40 AM, Poul-Henning Kamp phk@phk.freebsd.dk wrote:
That again depends on topology and control type. The canned converters
are almost always optimized to have the lowest number of switches and
work with cheap magnetics (single coil) without easily entering
problematic operation modes, noise is only a secondary concern.
That depends a LOT on which canned converter you decide to buy,
if you only go after price, or W/mm³ capacity, then certainly yes.
But for a one-off application like this, any money saved on a
cheap model is easily lost many times over in the trouble it will
cause.
But returning to the original post: Has anybody ever characterized
how much difference it makes to use two different PSU's for heater
vs. electronics sides of telecom Rb's ?
The “old time” answer was that a poorly regulated / poorly filtered supply was
considered “ok” for a heater. For the active electronics you wanted something
nice and stable / clean. To your point, once you get around to measuring
this, stability wise that answer often does not hold up. Noise wise, you are right
back to “what frequency?” ….
The somewhat more complex “old time” answer was that you don’t want the
honking big current of the heater coming off the supply you have tried so hard
to super-regulate. ( = it’s the supply that’s the issue not the Rb it’s self). Obviously
that’s going to depend on how the supply was designed.
Of course next layer to the onion is …. where does the ground current go? …. hmm….
Bob
I'm sure there is a reason why they make it two different pins ?
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Hi
Filtering is part of all this. If you are distributing a standard around the lab, cable management
will be part of it as well. Having a volt p-p pop up due to an unterminated cable (of the right length)
is not at all unheard of.
Bob
On Dec 24, 2019, at 3:48 PM, Taka Kamiya via time-nuts time-nuts@lists.febo.com wrote:
I just found something funny. I have been thinking "clean" power supplies that connects to AC mains. Then I thought, what about lead acid batteries?? So I went to my lab and took some measurement. This is a 12V 7A lead acid sealed battery, the kind commonly found on UPS devices.
The result? Surprise? The battery is oscillating at 5MHz and noise level is 15mV peak-to-peak!!!!
Of course, not...! Battery is pure DC and while voltage might drift, this is not that. For the record, a charger of any kind is not hooked up. It's one battery all by itself. Battery is not oscillating but that's what the measurement actually shows. That brings another point in my quest to "clean" power source. It's not just the power supply but the whole lab eco system has to be considered. Having one master 24V source (my original plan) is not the answer if mV level noise is going to be a problem.
This "discovery" puts whole new layer to having a nice power supply.....
Just as a point of reference, I hooked up a common cheap float charger. The charger itself has 2V p-p noise. Connected to battery, it still have 100mV p-p noise. There goes battery = noise sponge theory....
(Mr.) Taka Kamiya
KB4EMF / ex JF2DKG
On Tuesday, December 24, 2019, 3:00:50 PM EST, Bob kb8tq <kb8tq@n1k.org> wrote:
Hi
On Dec 24, 2019, at 6:40 AM, Poul-Henning Kamp phk@phk.freebsd.dk wrote:
That again depends on topology and control type. The canned converters
are almost always optimized to have the lowest number of switches and
work with cheap magnetics (single coil) without easily entering
problematic operation modes, noise is only a secondary concern.
That depends a LOT on which canned converter you decide to buy,
if you only go after price, or W/mm³ capacity, then certainly yes.
But for a one-off application like this, any money saved on a
cheap model is easily lost many times over in the trouble it will
cause.
But returning to the original post: Has anybody ever characterized
how much difference it makes to use two different PSU's for heater
vs. electronics sides of telecom Rb's ?
The “old time” answer was that a poorly regulated / poorly filtered supply was
considered “ok” for a heater. For the active electronics you wanted something
nice and stable / clean. To your point, once you get around to measuring
this, stability wise that answer often does not hold up. Noise wise, you are right
back to “what frequency?” ….
The somewhat more complex “old time” answer was that you don’t want the
honking big current of the heater coming off the supply you have tried so hard
to super-regulate. ( = it’s the supply that’s the issue not the Rb it’s self). Obviously
that’s going to depend on how the supply was designed.
Of course next layer to the onion is …. where does the ground current go? …. hmm….
Bob
I'm sure there is a reason why they make it two different pins ?
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.