really slow PLL...

[Phil Hobbs]

Joe Gwinn wrote:

On Fri, 22 Jul 2022 21:38:39 -0000 (UTC), \"Don\" <g@crcomp.net> wrote:

Joe Gwinn wrote:

snip

Also, I\'d lose the BNC connectors. Threaded connectors like SMA, TNC,
and Type N are far better.

Or use shielded twisted pair to carry the 1PPS pulses. This would
work better over a backplane.

This is good advice. Even though the lazy guy within me never truly
gives up his fight to take the easy way out with BNC.
Twisted pair (TP) sounds even easier than BNC. So, what\'s the
\"catch\" with TP? Where\'s the \"gotcha\" to make TP harder than BNC?

Depends on what you are trying to do.

For nanosecond edges, coax is pretty useful, but short range and often
mechanically awkward.

For microsecond edges at 1000 meters, RS422 over shielded twisted pair
is pretty good.

For bus length links, LVDS or the like.

And so on. And there is always optical links.

Joe Gwinn

BNCs are the bomb, as long as you aren\'t putting 500 of them in series,
as with the old 10base2 coax Ethernet.

TNCs are a very small niche, and N connectors belong only on spectrum
analyzers.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Don Y]

On 7/22/2022 5:24 PM, Lasse Langwadt Christensen wrote:

lørdag den 23. juli 2022 kl. 02.10.41 UTC+2 skrev Don Y:
On 7/22/2022 4:52 PM, Lasse Langwadt Christensen wrote:
fredag den 22. juli 2022 kl. 23.44.07 UTC+2 skrev Don Y:
On 7/22/2022 1:01 PM, Joe Gwinn wrote:
On Fri, 22 Jul 2022 10:54:30 -0700 (PDT), Lasse Langwadt Christensen
lang...@fonz.dk> wrote:

If your device\'s *timing* was off by 0.05%, would that be consequential?

https://youtu.be/AFaRIW-wZlw?t=54
Note context of post...
My recollection is the for a chorus to be in unison, all the singers
must be within twenty milliseconds of one another.

This is discussed a lot in the computer music literature.
Things get \"painful\" at about 50ms. I suspect your brain tries to
consider them as different events instead of indistinguishable.

afaiu the limit for a phone system is >25ms round trip before you need echo canceling
Yes, but that addresses quality. A phone is still usable with
noticeable echo, esp if the echo is many dB down.

sure, but it seems that below ~25ms it is not noticable

It seems like the brain has some (temporal) threshold beyond which it
can no longer treat things as being concurrent and starts a new \"recognition
process\" (so to speak) to deal with the \"other\" event.

at big concerts with speaker towers for the people far away from the stage
those speakers gets delayed so they are 10-15ms behind the sound from the stage
giving the illusion that all the sound is from the main Pa at the stage

I was near the close-in towers off stage left:

<http://3.bp.blogspot.com/-syHkjhoSSr8/VeV32eq-8EI/AAAAAAABw1w/ippl8bx8kKE/s1600/Grateful%2BDead%2Blive%2Bin%2BRaceway%2BPark%252C%2B1977%2B%252816%2529.jpg>

 

[Don Y]

On 7/22/2022 6:16 PM, Don Y wrote:

On 7/22/2022 5:24 PM, Lasse Langwadt Christensen wrote:
at big concerts with speaker towers for the people far away from the stage
those speakers gets delayed so they are 10-15ms behind the sound from the stage
giving the illusion that all the sound is from the main Pa at the stage

I was near the close-in towers off stage left:

http://3.bp.blogspot.com/-syHkjhoSSr8/VeV32eq-8EI/AAAAAAABw1w/ippl8bx8kKE/s1600/Grateful%2BDead%2Blive%2Bin%2BRaceway%2BPark%252C%2B1977%2B%252816%2529.jpg

Shot from a different viewpoint:
<http://media.gettyimages.com/photos/more-than-125000-rock-fans-from-down-the-road-and-across-the-nation-picture-id127330716>

The fence is made from \"semi\" trailers to give an idea of scale.

 

[Les Cargill]

John Larkin wrote:

Suppose I have several rackmount boxes and each has a BNC connector on
the back. Each of them has an open-drain mosfet, a weak pullup, and a
lowpass filtered schmitt gate back into our FPGA.

I can daisy-chain several boxes with BNC cables and tees.

Each box has a 40 MHz VCXO and I want to phase-lock them, or at least
time-align them to always be the same within a few microseconds,
longterm.

I could call one the leader (not \"master\") and make the others
followers (not \"slaves\") and have the leader make an active low pulse
maybe once a second. A follower would use her (not \"his\") clock to
measure the incoming period and tweak its local VCXO in the right
direction. That should work.

\"Her\" clock is not necessarily as stable as the 1PPS. How quickly does
it need to converge? Is this just a \"make it a wee bit better\" thing or
do you have a specific jitter spectrum in mind?

Don\'t GPS receivers lock their 10 MHz oscillators to a 1 PPS pulse
from the satellites?

\"... and the other is a
time-locked loop in the 1023 bit code space domain with the goal of
tracking both the code and carrier phases for that signal.\"

http://www.gpsinformation.net/main/gpslock.htm

That sounds suspiciously like what I did ( in software ) for a
terrestrial radio system once. The FPGA gave me a few bits of tuning
data and it was a state machine. I *think* the output was a
counter-delta value back to the FPGA. That was in the long ago.

I put trimpots in the management plane , put those in a GUI
and let the FPGA guy tune it. He was ecstatic and it beat
him running back and forth.

He tested it over coax; never did understand why this was needed at all
unless you were over air/free space. The analog PLLs should have handled
it. I think the RF amps we were using were pretty awful and not being
run in a very linear range. System leaned heavily on some bandlimiting
( SAW ) filters.

My system should work from a 1 PPS GPS pulse too, all boxes as
followers.

A second is a long time. GPS is a couple of nanoseconds per day
to a few msec. Yer average computer will go off to see the wizard and
back in a day.

The PLL algorithm might be interesting.

It felt more like a \"goalie\" algorithm than a PLL. But it did show
a measurable improvement. It\'s not really even an algorithm. It\'s a
\"bang bang\" controller. Nine pound hammer in software.

--
Les Cargill

 

[Les Cargill]

Phil Hobbs wrote:

jlarkin@highlandsniptechnology.com wrote:
snip

It dithers around the setpoint but nobody notices.

That\'s what lowpass filters are for.

This is immune to classic control theory so the concept annoys some
people but it works great.

A real old time control guy like Tim Wescott would probably be a fan
too--the great virtue of a bang-bang controller is that (as you say)
it\'s highly resistant to variations in the _plant_.

Well, yeah - it\'s naturally constrained. When I jack the temp target on
the A/C here, it take 30-45 seconds to turn everything off.


Tim used to be a lot of fun and put up with much. FWIW rbj showed up
on Reddit and lasted a couple days.

Your furnace doesn\'t go nuts when you have a Christmas party, even
though all those people generate a lot of heat, and there\'s lots of
opening and closing of doors and ovens.

You\'re just doing trust falls with slew rate limiting. There\'s
probably a PhD paper somewhere with a madman low-pass filtering the
output of a bangbang with a lowpass.

Kinda like... the .047 uf cap in the tone circuit on a Telecaster
It\'s there to limit damage.

Cheers

Phil Hobbs

--
Les Cargill

 

[Les Cargill]

bitrex wrote:

On 7/20/2022 8:22 PM, John Larkin wrote:
On Wed, 20 Jul 2022 19:32:20 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

John Larkin wrote:


Suppose I have several rackmount boxes and each has a BNC connector on
the back. Each of them has an open-drain mosfet, a weak pullup, and a
lowpass filtered schmitt gate back into our FPGA.

I can daisy-chain several boxes with BNC cables and tees.

Each box has a 40 MHz VCXO and I want to phase-lock them, or at least
time-align them to always be the same within a few microseconds,
longterm.

I could call one the leader (not \"master\") and make the others
followers (not \"slaves\") and have the leader make an active low pulse
maybe once a second. A follower would use her (not \"his\") clock to
measure the incoming period and tweak its local VCXO in the right
direction. That should work.

Don\'t GPS receivers lock their 10 MHz oscillators to a 1 PPS pulse
from the satellites?

My system should work from a 1 PPS GPS pulse too, all boxes as
followers.

The PLL algorithm might be interesting.


It\'s certainly possible.  However, within whatever tiny loop bandwidth
you wound up with, the lockers would still have

20 log(40e6) = 152 dB

higher phase noise than the lockee.

GPS has that problem too.


It would be interesting to do the math to see whether it\'s possible to
generate a concensus lock for the group: if you get everybody close
enough, just sum their sine wave outputs and lock each one of them to
that, with some bit of AC coupling or something so that they don\'t all
wander together off to the edge of the tuning range.

Maybe have one doing the locking with a phase shifter and the others
with VCOs, or something like that.

Definitely a partly-baked idea, but surely one could do better than
152 dB!

Cheers

Phil Hobbs

Each box is basically a multichannel power supply, but channels can be
programmed to do stuff in timed sequences. I want different box
outputs to time align within, say, one millisecond longterm once
programs are kicked off together. So, many microseconds of equivalent
RMS phase noise is OK as long as we stay time aligned longterm.

It sounds like you\'re looking for a protocol like DMX if what you want
is to trigger sequences of events across boxes to within a millisecond,
I don\'t understand what this lock-the-40 MHz across boxes is about.

https://en.wikipedia.org/wiki/DMX512

DMX for this is like hunting deer with an artillery piece. DMX is for
the big-ass risk scenarios in distributed topologies; this is a lot
less profound.

--
Les Cargill

 

[Lasse Langwadt Christensen]

lørdag den 23. juli 2022 kl. 03.57.33 UTC+2 skrev Les Cargill:

bitrex wrote:
On 7/20/2022 8:22 PM, John Larkin wrote:
On Wed, 20 Jul 2022 19:32:20 -0400, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

John Larkin wrote:


Suppose I have several rackmount boxes and each has a BNC connector on
the back. Each of them has an open-drain mosfet, a weak pullup, and a
lowpass filtered schmitt gate back into our FPGA.

I can daisy-chain several boxes with BNC cables and tees.

Each box has a 40 MHz VCXO and I want to phase-lock them, or at least
time-align them to always be the same within a few microseconds,
longterm.

I could call one the leader (not \"master\") and make the others
followers (not \"slaves\") and have the leader make an active low pulse
maybe once a second. A follower would use her (not \"his\") clock to
measure the incoming period and tweak its local VCXO in the right
direction. That should work.

Don\'t GPS receivers lock their 10 MHz oscillators to a 1 PPS pulse
from the satellites?

My system should work from a 1 PPS GPS pulse too, all boxes as
followers.

The PLL algorithm might be interesting.


It\'s certainly possible. However, within whatever tiny loop bandwidth
you wound up with, the lockers would still have

20 log(40e6) = 152 dB

higher phase noise than the lockee.

GPS has that problem too.


It would be interesting to do the math to see whether it\'s possible to
generate a concensus lock for the group: if you get everybody close
enough, just sum their sine wave outputs and lock each one of them to
that, with some bit of AC coupling or something so that they don\'t all
wander together off to the edge of the tuning range.

Maybe have one doing the locking with a phase shifter and the others
with VCOs, or something like that.

Definitely a partly-baked idea, but surely one could do better than
152 dB!

Cheers

Phil Hobbs

Each box is basically a multichannel power supply, but channels can be
programmed to do stuff in timed sequences. I want different box
outputs to time align within, say, one millisecond longterm once
programs are kicked off together. So, many microseconds of equivalent
RMS phase noise is OK as long as we stay time aligned longterm.

It sounds like you\'re looking for a protocol like DMX if what you want
is to trigger sequences of events across boxes to within a millisecond,
I don\'t understand what this lock-the-40 MHz across boxes is about.

https://en.wikipedia.org/wiki/DMX512




DMX for this is like hunting deer with an artillery piece. DMX is for
the big-ass risk scenarios in distributed topologies; this is a lot
less profound.

? it\'s a 250kbit uart on RS485, hardly rocket surgery

 

[Les Cargill]

jlarkin@highlandsniptechnology.com wrote:

On Thu, 21 Jul 2022 11:42:28 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:
snip
Phil Hobbs

Mathematicians often like music. In my experience, music fandom is
negatively correlated to engineering design skill. Different brain
structure or something.

Engineering is composition. Composition is the thin edge of the musical
wedge. Musicianship is different; it\'s pattern identification. As is
composition but in a different way. But it is all the same thing.

It all depends on which wall you prefer to have your back against.

One other thing I see a lot is undue respect for standards. As in \"you
can\'t do that because it violates SCPI standards.\" Where are the SCPI
Police when you need them?

Over where they MATLAB.

--
Les Cargill

 

[bitrex]

On 7/22/2022 10:09 PM, Lasse Langwadt Christensen wrote:

lørdag den 23. juli 2022 kl. 03.57.33 UTC+2 skrev Les Cargill:
bitrex wrote:
On 7/20/2022 8:22 PM, John Larkin wrote:
On Wed, 20 Jul 2022 19:32:20 -0400, Phil Hobbs
pcdhSpamM...@electrooptical.net> wrote:

John Larkin wrote:


Suppose I have several rackmount boxes and each has a BNC connector on
the back. Each of them has an open-drain mosfet, a weak pullup, and a
lowpass filtered schmitt gate back into our FPGA.

I can daisy-chain several boxes with BNC cables and tees.

Each box has a 40 MHz VCXO and I want to phase-lock them, or at least
time-align them to always be the same within a few microseconds,
longterm.

I could call one the leader (not \"master\") and make the others
followers (not \"slaves\") and have the leader make an active low pulse
maybe once a second. A follower would use her (not \"his\") clock to
measure the incoming period and tweak its local VCXO in the right
direction. That should work.

Don\'t GPS receivers lock their 10 MHz oscillators to a 1 PPS pulse
from the satellites?

My system should work from a 1 PPS GPS pulse too, all boxes as
followers.

The PLL algorithm might be interesting.


It\'s certainly possible. However, within whatever tiny loop bandwidth
you wound up with, the lockers would still have

20 log(40e6) = 152 dB

higher phase noise than the lockee.

GPS has that problem too.


It would be interesting to do the math to see whether it\'s possible to
generate a concensus lock for the group: if you get everybody close
enough, just sum their sine wave outputs and lock each one of them to
that, with some bit of AC coupling or something so that they don\'t all
wander together off to the edge of the tuning range.

Maybe have one doing the locking with a phase shifter and the others
with VCOs, or something like that.

Definitely a partly-baked idea, but surely one could do better than
152 dB!

Cheers

Phil Hobbs

Each box is basically a multichannel power supply, but channels can be
programmed to do stuff in timed sequences. I want different box
outputs to time align within, say, one millisecond longterm once
programs are kicked off together. So, many microseconds of equivalent
RMS phase noise is OK as long as we stay time aligned longterm.

It sounds like you\'re looking for a protocol like DMX if what you want
is to trigger sequences of events across boxes to within a millisecond,
I don\'t understand what this lock-the-40 MHz across boxes is about.

https://en.wikipedia.org/wiki/DMX512




DMX for this is like hunting deer with an artillery piece. DMX is for
the big-ass risk scenarios in distributed topologies; this is a lot
less profound.

? it\'s a 250kbit uart on RS485, hardly rocket surgery

Right, it\'s glorified MIDI.

 

[Don]

Phil Hobbs wrote:

Joe Gwinn wrote:
Don wrote:
Joe Gwinn wrote:

snip

Also, I\'d lose the BNC connectors. Threaded connectors like SMA, TNC,
and Type N are far better.

Or use shielded twisted pair to carry the 1PPS pulses. This would
work better over a backplane.

This is good advice. Even though the lazy guy within me never truly
gives up his fight to take the easy way out with BNC.
Twisted pair (TP) sounds even easier than BNC. So, what\'s the
\"catch\" with TP? Where\'s the \"gotcha\" to make TP harder than BNC?

Depends on what you are trying to do.

For nanosecond edges, coax is pretty useful, but short range and often
mechanically awkward.

For microsecond edges at 1000 meters, RS422 over shielded twisted pair
is pretty good.

For bus length links, LVDS or the like.

And so on. And there is always optical links.

Joe Gwinn


BNCs are the bomb, as long as you aren\'t putting 500 of them in series,
as with the old 10base2 coax Ethernet.

TNCs are a very small niche, and N connectors belong only on spectrum
analyzers.

The lazy guy within me always tries to use an N connector to BNC
adapter on my boat anchor spectrum analyzer. He convinces himself he\'s
only interested in frequencies less than 2 GHz, so, what\'s the harm?

High performance WiFi antennas also use N connectors to squeeze out
every last iota of performance. You need a DIY N connector to reverse-
polarity SMA to connect such antennas to consumer WiFi devices.

Danke,

--
Don.......My cat\'s )\\._.,--....,\'``. https://crcomp.net/reviews.php
telltale tall tail /, _.. \\ _\\ (`._ ,.
tells tall tales.. `._.-(,_..\'--(,_..\'`-.;.\'

 

[whit3rd]

On Friday, July 22, 2022 at 2:38:45 PM UTC-7, Don wrote:

Joe Gwinn wrote:

snip
Also, I\'d lose the BNC connectors. Threaded connectors like SMA, TNC,
and Type N are far better.

Or use shielded twisted pair to carry the 1PPS pulses.

Twisted pair (TP) sounds even easier than BNC. So, what\'s the
\"catch\" with TP? Where\'s the \"gotcha\" to make TP harder than BNC?

Biggest \'gotcha\' is the lack of good shielded TP connectors. I had only
UHF-style twisted pair shielded connectors last time I wanted some, and
that\'s a polarity-insensitive connector. We applied paint markings
to get it straight.

MiniDIN 3 (don\'t trust the shield connector) was what Apple used for their
LocalTalk/Appletalk hardware, and of course there are microphone connectors (big \'uns))
but for cheap \'uns, RJ--11 (6P4C and use the second pair) wasn\'t good (I needed to
pass significant current) and Lemo offerings were... neither inexpensive nor locally stocked.

 

[Clifford Heath]

On 23/7/22 02:02, Don wrote:

For some unknown reason, the only thing to truly satisfy me is to
tune instruments by ear.

Many instruments (including all stringed instruments) exhibit
inharmonicity. In the case of strings the harmonics are progressively
sharper than the numerical frequency multiples. So to produce an
even-tempered tuning requires balancing the matching of fundamentals
with the harmonics. It\'s not an exact science.

Piano\'s are \"stretch tuned\" by as much as two semitones between the
bottom octave and the top one, and professional pianists have personal
preferences for more or less stretching.

You could program an electronic tuner to do this of course, but you
can\'t easily tell it how much stretching you\'d like.

Clifford Heath

 

[Clifford Heath]

On 23/7/22 03:03, John Walliker wrote:

There is cross-correlation between left and right ears in the brainstem which
is involved in determining the direction of sound sources. Detectable time
differences are remarkably small. (I would have to look it up to give a number.)

It seems like a chip-based cross-correlator would be pretty trivial.
A pair of bucket-brigade lines running parallel but in opposite
directions, and a correlator at each node. Frequency response of each
correlator would be based on the delta-wavelength of one bucket in the
chains.

You could read out the relative directions of all sound sources by
looking at the left-right correlation histogram

CH

 

[Gerhard Hoffmann]

Am 23.07.22 um 08:35 schrieb Clifford Heath:

On 23/7/22 02:02, Don wrote:
     For some unknown reason, the only thing to truly satisfy me is to
tune instruments by ear.
Many instruments (including all stringed instruments) exhibit
inharmonicity. In the case of strings the harmonics are progressively
sharper than the numerical frequency multiples. So to produce an
even-tempered tuning requires balancing the matching of fundamentals
with the harmonics. It\'s not an exact science.

Piano\'s are \"stretch tuned\" by as much as two semitones between the
bottom octave and the top one, and professional pianists have personal
preferences for more or less stretching.

You could program an electronic tuner to do this of course, but you
can\'t easily tell it how much stretching you\'d like.

Most seem to like a lot. Remember WhiteSnake\'s text (in Kittens got claws):

\"with her g string tuned to A\" !


Gerhard

 

[Martin Brown]

On 22/07/2022 19:21, John Larkin wrote:

On Thu, 21 Jul 2022 18:21:51 +0100, Martin Brown
\'\'\'newspam\'\'\'@nonad.co.uk> wrote:

On 21/07/2022 16:31, John Walliker wrote:

The Group Execute Trigger command does allow quite tight synchronisation
between different GPIB devices.

GPIB flat out on a good day could manage 1Mbyte/s but in real world
situations with interconnect cabling you would be lucky to get 500kb/s.
It\'s best feature was that it ran at the maximum speed the receiving
device could handle (assuming that the controller was fast enough).

Synchronisation to a GET command would be probably be better than 1us
but would depend on the decoding time in each individual box. Some GPIB
devices were rather pedestrian at accepting commands.

IEEE488 was good in its day but a bit long in the tooth now. Still on
some test equipment in service today and was provided as standard on NEC
9801 PC\'s in Japan although hardly ever used by their customers.

Ever read the actual 488 spec? There is a state diagram that could
wreck your sleep for a week.

I\'ve implemented it on several chipsets including Motorola 68488, Intel
8292 and TI 9914. We even implemented full pass control between peer
controllers and our own bus analyser to capture bus transactions (which
after a redesign could not be blinded by HP\'s IFC tricks).

Later we reworked things for IEEE488.2 in the early 1990\'s but I dropped
out of that game not long afterwards to go and work in Japan. Ethernet
had made significant inroads into the instrument market by then.

488 has a hardware \"accepted\" line, but for some reason SCPI in other
contexts is send-and-pray.

And the accepted line makes sure the slowest thing on the bus gets the
message which is why it tended to slow down as longer cables and more
kit was added. It tolerated much longer cables than the official
standard permitted with only modest loss of speed.

488 is rare on new instruments, which are ethernet and USB. A Rigol
scope makes a great USB power supply for fans and charging phones.

It surprises me that it is still present on *any* modern instruments.
I expected it to survive on useful legacy kit until about now though.

Plenty of labs will have good working kit that still uses that interface
and it is plenty fast enough for a lot of ordinary lab use.

--
Regards,
Martin Brown

 

[server]

On Fri, 22 Jul 2022 21:10:35 -0500, Les Cargill <lcargil99@gmail.com>
wrote:

jlarkin@highlandsniptechnology.com wrote:
On Thu, 21 Jul 2022 11:42:28 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:
snip
Phil Hobbs

Mathematicians often like music. In my experience, music fandom is
negatively correlated to engineering design skill. Different brain
structure or something.


Engineering is composition. Composition is the thin edge of the musical
wedge. Musicianship is different; it\'s pattern identification. As is
composition but in a different way. But it is all the same thing.

It all depends on which wall you prefer to have your back against.

I\'ve always wondered about musicians. They have to play a piece
hundreds of times to get it right. Some have surely performed
something thousands of times. Don\'t they get bored? Apparently not.

I design something, finish, and then want to design something entirely
different.

It depends on boredom thresholds.

One other thing I see a lot is undue respect for standards. As in \"you
can\'t do that because it violates SCPI standards.\" Where are the SCPI
Police when you need them?

Over where they MATLAB.

SCPI is send-and-forget. There is some query you can send to ask if
the last command worked. And you can have an error queue that you can
interrogate now and then for historical forensics.

I told the customer that damn the specs, every command is going to
reply with data, an error message, or \"OK\". They agree.

 

[Phil Hobbs]

Martin Brown wrote:

On 22/07/2022 19:21, John Larkin wrote:
On Thu, 21 Jul 2022 18:21:51 +0100, Martin Brown
\'\'\'newspam\'\'\'@nonad.co.uk> wrote:

On 21/07/2022 16:31, John Walliker wrote:

The Group Execute Trigger command does allow quite tight
synchronisation
between different GPIB devices.

GPIB flat out on a good day could manage 1Mbyte/s but in real world
situations with interconnect cabling you would be lucky to get 500kb/s.
It\'s best feature was that it ran at the maximum speed the receiving
device could handle (assuming that the controller was fast enough).

Synchronisation to a GET command would be probably be better than 1us
but would depend on the decoding time in each individual box. Some GPIB
devices were rather pedestrian at accepting commands.

IEEE488 was good in its day but a bit long in the tooth now. Still on
some test equipment in service today and was provided as standard on NEC
9801 PC\'s in Japan although hardly ever used by their customers.

Ever read the actual 488 spec? There is a state diagram that could
wreck your sleep for a week.

I\'ve implemented it on several chipsets including Motorola 68488, Intel
8292 and TI 9914. We even implemented full pass control between peer
controllers and our own bus analyser to capture bus transactions (which
after a redesign could not be blinded by HP\'s IFC tricks).

Later we reworked things for IEEE488.2 in the early 1990\'s but I dropped
out of that game not long afterwards to go and work in Japan. Ethernet
had made significant inroads into the instrument market by then.

488 has a hardware \"accepted\" line, but for some reason SCPI in other
contexts is send-and-pray.

And the accepted line makes sure the slowest thing on the bus gets the
message which is why it tended to slow down as longer cables and more
kit was added. It tolerated much longer cables than the official
standard permitted with only modest loss of speed.

488 is rare on new instruments, which are ethernet and USB. A Rigol
scope makes a great USB power supply for fans and charging phones.

It surprises me that it is still present on *any* modern instruments.
I expected it to survive on useful legacy kit until about now though.

Plenty of labs will have good working kit that still uses that interface
and it is plenty fast enough for a lot of ordinary lab use.

Plus you can get GPIB-Ethernet adapters for fairly cheap. We use one
made by the estimable Abdul at Prologix, which works pretty well--we\'ve
used as many as three instruments with it at a time--two scopes and a
spectrum analyzer. Did it just yesterday, in fact.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Phil Hobbs]

Don wrote:

Phil Hobbs wrote:
Joe Gwinn wrote:
Don wrote:
Joe Gwinn wrote:

snip

Also, I\'d lose the BNC connectors. Threaded connectors like SMA, TNC,
and Type N are far better.

Or use shielded twisted pair to carry the 1PPS pulses. This would
work better over a backplane.

This is good advice. Even though the lazy guy within me never truly
gives up his fight to take the easy way out with BNC.
Twisted pair (TP) sounds even easier than BNC. So, what\'s the
\"catch\" with TP? Where\'s the \"gotcha\" to make TP harder than BNC?

Depends on what you are trying to do.

For nanosecond edges, coax is pretty useful, but short range and often
mechanically awkward.

For microsecond edges at 1000 meters, RS422 over shielded twisted pair
is pretty good.

For bus length links, LVDS or the like.

And so on. And there is always optical links.

Joe Gwinn


BNCs are the bomb, as long as you aren\'t putting 500 of them in series,
as with the old 10base2 coax Ethernet.

TNCs are a very small niche, and N connectors belong only on spectrum
analyzers.

The lazy guy within me always tries to use an N connector to BNC
adapter on my boat anchor spectrum analyzer. He convinces himself he\'s
only interested in frequencies less than 2 GHz, so, what\'s the harm?

High performance WiFi antennas also use N connectors to squeeze out
every last iota of performance. You need a DIY N connector to reverse-
polarity SMA to connect such antennas to consumer WiFi devices.

Danke,

N connectors are almost identical to BNCs internally--in fact you can
mate an N male to a BNC female very nicely.

I use SMA for everything above about 2 GHz anyway. (Some of my test
gear uses expensive K (2.8 mm) or very expensive V (2.4 mm) connectors,
but those all have SMA adapters on the front.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

 

[Joe Gwinn]

On Fri, 22 Jul 2022 17:16:00 -0700 (PDT), Lasse Langwadt Christensen
<langwadt@fonz.dk> wrote:

fredag den 22. juli 2022 kl. 22.37.08 UTC+2 skrev Joe Gwinn:
On Thu, 21 Jul 2022 04:17:14 -0700, jla...@highlandsniptechnology.com
wrote:

On Wed, 20 Jul 2022 23:49:40 -0700 (PDT), whit3rd <whi...@gmail.com
wrote:

On Wednesday, July 20, 2022 at 4:21:08 PM UTC-7, John Larkin wrote:
Suppose I have several rackmount boxes and each has a BNC connector on
the back. Each of them has an open-drain mosfet, a weak pullup, and a
lowpass filtered schmitt gate back into our FPGA.

I can daisy-chain several boxes with BNC cables and tees.

Each box has a 40 MHz VCXO and I want to phase-lock them, or at least
time-align them to always be the same within a few microseconds,
longterm.

If you can tolerate \'a few microseconds\' on a 40 MHz signal, that\'s not a phase-lock
problem, it\'s a frequency-lock problem. Why not just run an up/down counter
to generate a correction voltage for each non-leading VCO?

It\'s actually a time lock problem. If a follower box starts up and
sees its first 1 PPS input, it can thereafter declare 1 PPS internal
events, based on its local VCO, and then do successive early/late
comparisons to the external pulses. And trim its VCXO accordingly.

Yes, exactly. And the drift between two reasonably good clocks is
slow, so the correction need not and should not be all that fast.

What I\'ve done in real applications is to periodically measure the
offset between when the external 1PPS is predicted to happen and when
it actually does, and adjust the VCO frequency such that in say 50
seconds of roughly linear convergence they will coincide (and keep on
going). The process is repeated every few seconds (exact interval not
important as it is measured).

This is roughly the algorithm a helmsman uses while steering a
sailboat, where effect is very much delayed from action.

In many computer systems it is quite difficult to do anything on a
strict time mark, but easy to measure that actual elapsed time, using
the actual clock that is being steered - it all still converges, so
long as one doesn\'t try too hard.

So, in your example, the local clock would come from a 40 MHz VCO of
good manufacture (probably needs to be a TCXO of some sort). The 40
MHz output would be fed to a divider that puts out a 1PPS pulse train.

During initialization, when the first external 1PPS leading edge is
received, reset the divider, and start counting 40 MHz cycles. Maybe
wait for things to stabilize.

Thereafter, measure signed offset between external and internal 1PPS
leading edges, and compute how much change (plus or minus) in current
VCO frequency is required for zero offset to occur in say 50 seconds
(make this time-to-zero an adjustable parameter), and change the VCO
control voltage accordingly.

A few seconds later (also an adjustable parameter), repeat the above
adjustment process, again looking 50 seconds into the future from now.
Repeat forever.


isn\'t that kind how NTP works?, speeding up or slowing down over some period
to sync time with the server without big jumps and always increasing (except possibly at start up)

Yes. This is exactly how part of NTP works, in particular the FLL
(Frequency Lock Loop). Battle-tested. That\'s where I got the idea.
NTP was developed in the early 1980s, shortly after Ethernet made such
a thing useful.

This kind of thing is extensively discussed on Time Nuts.

come to think of it, some closed loop servo systems (and step generators) for things like CNC machine work similarly
at a fixed interval, say a few kHz, current target and actual position is compared, from that (usually with a PI loop)
the speed of the motor (or frequency of steps) is set so the position will hit the next target at the next timer tick.

Yep.

I forgot to mention one thing, a way to speed initialization up:

The external 1PPS pulse-train is taken as gospel. If one counts local
40 MHz oscillator cycles between any adjacent pair of 1PPS events, one
will get a very accurate measurement of the local oscillator signal
frequency. Knowing that it is supposed to be 40 MHz, one can compute
how far off correct (as a ratio) that local oscillator is from truth.
This can be used to jump far closer starting frequency to correct
without waiting for convergence to get there.

Joe Gwinn

 

[Lasse Langwadt Christensen]

lørdag den 23. juli 2022 kl. 21.00.37 UTC+2 skrev Joe Gwinn:

On Fri, 22 Jul 2022 17:16:00 -0700 (PDT), Lasse Langwadt Christensen
lang...@fonz.dk> wrote:

fredag den 22. juli 2022 kl. 22.37.08 UTC+2 skrev Joe Gwinn:
On Thu, 21 Jul 2022 04:17:14 -0700, jla...@highlandsniptechnology.com
wrote:

On Wed, 20 Jul 2022 23:49:40 -0700 (PDT), whit3rd <whi...@gmail.com
wrote:

On Wednesday, July 20, 2022 at 4:21:08 PM UTC-7, John Larkin wrote:
Suppose I have several rackmount boxes and each has a BNC connector on
the back. Each of them has an open-drain mosfet, a weak pullup, and a
lowpass filtered schmitt gate back into our FPGA.

I can daisy-chain several boxes with BNC cables and tees.

Each box has a 40 MHz VCXO and I want to phase-lock them, or at least
time-align them to always be the same within a few microseconds,
longterm.

If you can tolerate \'a few microseconds\' on a 40 MHz signal, that\'s not a phase-lock
problem, it\'s a frequency-lock problem. Why not just run an up/down counter
to generate a correction voltage for each non-leading VCO?

It\'s actually a time lock problem. If a follower box starts up and
sees its first 1 PPS input, it can thereafter declare 1 PPS internal
events, based on its local VCO, and then do successive early/late
comparisons to the external pulses. And trim its VCXO accordingly.

Yes, exactly. And the drift between two reasonably good clocks is
slow, so the correction need not and should not be all that fast.

What I\'ve done in real applications is to periodically measure the
offset between when the external 1PPS is predicted to happen and when
it actually does, and adjust the VCO frequency such that in say 50
seconds of roughly linear convergence they will coincide (and keep on
going). The process is repeated every few seconds (exact interval not
important as it is measured).

This is roughly the algorithm a helmsman uses while steering a
sailboat, where effect is very much delayed from action.

In many computer systems it is quite difficult to do anything on a
strict time mark, but easy to measure that actual elapsed time, using
the actual clock that is being steered - it all still converges, so
long as one doesn\'t try too hard.

So, in your example, the local clock would come from a 40 MHz VCO of
good manufacture (probably needs to be a TCXO of some sort). The 40
MHz output would be fed to a divider that puts out a 1PPS pulse train.

During initialization, when the first external 1PPS leading edge is
received, reset the divider, and start counting 40 MHz cycles. Maybe
wait for things to stabilize.

Thereafter, measure signed offset between external and internal 1PPS
leading edges, and compute how much change (plus or minus) in current
VCO frequency is required for zero offset to occur in say 50 seconds
(make this time-to-zero an adjustable parameter), and change the VCO
control voltage accordingly.

A few seconds later (also an adjustable parameter), repeat the above
adjustment process, again looking 50 seconds into the future from now.
Repeat forever.


isn\'t that kind how NTP works?, speeding up or slowing down over some period
to sync time with the server without big jumps and always increasing (except possibly at start up)

Yes. This is exactly how part of NTP works, in particular the FLL
(Frequency Lock Loop). Battle-tested. That\'s where I got the idea.
NTP was developed in the early 1980s, shortly after Ethernet made such
a thing useful.

This kind of thing is extensively discussed on Time Nuts.


come to think of it, some closed loop servo systems (and step generators) for things like CNC machine work similarly
at a fixed interval, say a few kHz, current target and actual position is compared, from that (usually with a PI loop)
the speed of the motor (or frequency of steps) is set so the position will hit the next target at the next timer tick.

Yep.

I forgot to mention one thing, a way to speed initialization up:

The external 1PPS pulse-train is taken as gospel. If one counts local
40 MHz oscillator cycles between any adjacent pair of 1PPS events, one
will get a very accurate measurement of the local oscillator signal
frequency. Knowing that it is supposed to be 40 MHz, one can compute
how far off correct (as a ratio) that local oscillator is from truth.
This can be used to jump far closer starting frequency to correct
without waiting for convergence to get there.

yeh, get an initial guess and then switch to a slow correction so any jitter on the 1pps
gets \"filtered\"