[time-nuts] Ground loops in measurements?

Magnus Danielson magnus at rubidium.dyndns.org
Tue May 21 16:28:37 EDT 2013


Hi,

One should be careful with transformers, since they can bite you in 
interesting ways. While they give galvanic isolation, they also act like 
a capacitive divider. Just a primary and secondary coil will connect to 
each other, and it is not uncommon for the connection to be asymmetric. 
A single shield grounded to the other side will still cause asymmetric 
coupling to the other side. A double shield correctly to each side will 
contain the asymmetric to it's own ground, such that the only coupling 
between the sides becomes between the grounds.

So, just going transformer does not necessarily help. A good way to 
assist is to use an RF choke, and it works better if it has a good 
common mode termination, such as that seen from a balanced winding with 
the mid pin connected to ground.

Another approach is mesh-grounding rather than isolating ground. Mesh 
ground strategy is to ground everywhere and with everything such that 
the total wire area becomes large and hence resistance low.

Cheers,
Magnus

On 05/20/2013 05:52 PM, johncroos at aol.com wrote:
>
> Attila Kinali
>
> One effective and sure fire way to deal with ground loops is to use
> a minature 1:1  50 or 200 ohm (not critical as most instruments are not matched
> inputs anyway ) in the line from your distribution amplifier to the driven device. Coilcraft has these
> for both PC and SMT Just drive one side with the source coax and connect the load to the other.
>
> They are about a 1/2 inch cube and the PC mount types have leads to which the coax can be simply
> soldered, or you can get fancy and put in a box with ground isolated connectors.
>
> The band width of these devices is typically 1 to 100 MHz or more. Eliminates all 60 Hz transmission
> that is common mode on the coax. These are not expensive. www.coilcraft.com. They did sell on the web for me about a year ago.
>
> -73 john k6iql
>
>
>
>
>
>
>
> -----Original Message-----
> From: time-nuts-request<time-nuts-request at febo.com>
> To: time-nuts<time-nuts at febo.com>
> Sent: Mon, May 20, 2013 10:28 am
> Subject: time-nuts Digest, Vol 106, Issue 97
>
>
> Send time-nuts mailing list submissions to
> 	time-nuts at febo.com
>
> To subscribe or unsubscribe via the World Wide Web, visit
> 	https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> or, via email, send a message with subject or body 'help' to
> 	time-nuts-request at febo.com
>
> You can reach the person managing the list at
> 	time-nuts-owner at febo.com
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of time-nuts digest..."
>
>
> Today's Topics:
>
>     1. Re: Time Nut Pickens at the MIT Flea ? (J. Forster)
>     2. Re: time transfer over USB (Jim Lux)
>     3. Re: Time Nut Pickens at the MIT Flea ? (Tom Van Baak)
>     4. Re: time transfer over USB (lists at lazygranch.com)
>     5. Re: Ground loops in measurements? (Volker Esper)
>     6. Re: OCXO shock protection (Tom Van Baak)
>     7. Re: time transfer over USB (Attila Kinali)
>     8. Re: OCXO shock protection (Attila Kinali)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Mon, 20 May 2013 05:50:07 -0700 (PDT)
> From: "J. Forster"<jfor at quikus.com>
> To: "Andy Bardagjy"<andy at bardagjy.com>
> Cc: Discussion of precise time and frequency measurement
> 	<time-nuts at febo.com>
> Subject: Re: [time-nuts] Time Nut Pickens at the MIT Flea ?
> Message-ID:
> 	<56273.12.226.214.5.1369054207.squirrel at popaccts.quikus.com>
> Content-Type: text/plain;charset=iso-8859-1
>
> IMO, it would be good to have any gathering place w/in easy walking
> distance of the flea. Parking in Cambridge is not easy in the best of
> times.
>
> YMMV,
>
> -John
>
> =============
>
>
>> Plenty of good stuff in the area; Flour, Area Four, Catalyst, Friendly
>> Toast, Blue Room, CBC, and so on.
>>
>> I live in the area - recent MIT graduate - and would be happy to set
>> something up (though I think my gf's birthday is that day, so maybe not)
>>
>> Andy Bardagjy
>> bardagjy.com
>>
>>
>> On Mon, May 20, 2013 at 8:34 AM, J. Forster<jfor at quikus.com>  wrote:
>>
>>> There are cafe/pizza options in the Strattion Student Center, although
>>> seating is somewhat limited. But there are other places in the same
>>> building.
>>>
>>> It's a short walk from the flea site (Mass Ave at Vasser Street) Bldg
>>> W-20
>>>
>>> -John
>>>
>>> ================
>>>
>>>
>>>> I was there as well, but did not see much of interest. That might be
>>>> because Paul Swed scooped up the good stuff before I could finish the
>>>> first pass :) Nice find Paul, hope they settle down for you. Would
>>> love
>>> to
>>>> see a post-flea get together and put some faces with names, but I
>>> don't
>>>> know of any places nearby.
>>>>
>>>> Paul - K9MR
>>>>
>>>> On May 19, 2013, at 10:22 PM, paul swed wrote:
>>>>
>>>> Good news at least the first EGG RB fired up and after adjusting the
>>>> synthesizer is locked very nicely. Letting it bake and figuring out
>>> the
>>>> dip
>>>> switches. They were wrong or purposely offset. Hard to say. Lamp
>>> voltage
>>>> is
>>>> 12 VDC I suspect thats quite a good number if compared to the old FRS
>>> and
>>>> FRCs I have down in the 5-6V range. The FRS/C run 12 new.
>>>> Can't really find any documentation, but expected that.
>>>> Regards
>>>> Paul
>>>> WB8TSL
>>>>
>>>>
>>>> On Sun, May 19, 2013 at 10:09 PM, Bob Bownes<bownes at gmail.com>  wrote:
>>>>
>>>>> We should plan a Time-Nuts BOF lunch after the next flea.
>>>>>
>>>>> Saw little TN gear @ Dayton save one Efretom RBI time base for $1800
>>> and
>>>>> a
>>>>> few 10811s of dubious quality for $50 ea. the dents put me off
>>> gambling
>>>>> on
>>>>> one since I was given or the afternoon before! ;)
>>>>>
>>>>> On May 19, 2013, at 21:50, paul swed<paulswedb at gmail.com>  wrote:
>>>> _______________________________________________
>>>> time-nuts mailing list -- time-nuts at febo.com
>>>> To unsubscribe, go to
>>>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>>> and follow the instructions there.
>>>>
>>>>
>>>
>>> _______________________________________________
>>> time-nuts mailing list -- time-nuts at febo.com
>>> To unsubscribe, go to
>>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>> and follow the instructions there.
>>>
>
>
>
> ------------------------------
>
> Message: 2
> Date: Mon, 20 May 2013 06:14:12 -0700
> From: Jim Lux<jimlux at earthlink.net>
> To: time-nuts at febo.com
> Subject: Re: [time-nuts] time transfer over USB
> Message-ID:<519A21A4.8020709 at earthlink.net>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> On 5/20/13 2:43 AM, Attila Kinali wrote:
>
>> Oh.. and connect the whole thing to a port on the PC that does _not_
>> have an internal USB hub.
>>
> That's a bit of challenge, I suspect.. A casual look at the PCs  I have
> around here running windows all seem to have on-mobo hubs when you
> check Device Manager.  I suspect they are integrated into one of the
> peripheral chips.
>
>
>
>
>
> ------------------------------
>
> Message: 3
> Date: Mon, 20 May 2013 06:17:23 -0700
> From: "Tom Van Baak"<tvb at LeapSecond.com>
> To: "Discussion of precise time and frequency measurement"
> 	<time-nuts at febo.com>
> Subject: Re: [time-nuts] Time Nut Pickens at the MIT Flea ?
> Message-ID:<A4575A20315944FA80EEB0B6FFAA0FC6 at pc52>
> Content-Type: text/plain;	charset="iso-8859-1"
>
>
> A gentle reminder to keep postings to time-nuts as technical as possible, and
> on-topic. The occasional mention of upcoming local events or conferences is
> welcome (it's a great way to build community), but I suggest subsequent details,
> follow-up, meeting arrangements, and where to get pizza should be done by
> private email. The list now has 1300 members world-wide.
>
> Thanks,
> /tvb
> http://leapsecond.com/time-nuts.htm
>
>
>
>
>
> ------------------------------
>
> Message: 4
> Date: Mon, 20 May 2013 14:06:08 +0000
> From: lists at lazygranch.com
> To: "Discussion of precise time and frequency measurement"
> 	<time-nuts at febo.com>
> Subject: Re: [time-nuts] time transfer over USB
> Message-ID:
> 	<89993712-1369058767-cardhu_decombobulator_blackberry.rim.net-179398076- at b26.c1.bise6.blackberry>
> 	
> Content-Type: text/plain
>
> I suspect the idea is to use a port where no other devices, that is internal,
> are on the hub.
>
> Like you, I never saw a usb port not on a hub.
>
> -----Original Message-----
> From: Jim Lux<jimlux at earthlink.net>
> Sender: time-nuts-bounces at febo.com
> Date: Mon, 20 May 2013 06:14:12
> To:<time-nuts at febo.com>
> Reply-To: Discussion of precise time and frequency measurement
> 	<time-nuts at febo.com>
> Subject: Re: [time-nuts] time transfer over USB
>
> On 5/20/13 2:43 AM, Attila Kinali wrote:
>
>> Oh.. and connect the whole thing to a port on the PC that does _not_
>> have an internal USB hub.
>>
> That's a bit of challenge, I suspect.. A casual look at the PCs  I have
> around here running windows all seem to have on-mobo hubs when you
> check Device Manager.  I suspect they are integrated into one of the
> peripheral chips.
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
>
> ------------------------------
>
> Message: 5
> Date: Mon, 20 May 2013 16:15:00 +0200
> From: Volker Esper<ailer2 at t-online.de>
> To: Discussion of precise time and frequency measurement
> 	<time-nuts at febo.com>
> Subject: Re: [time-nuts] Ground loops in measurements?
> Message-ID:<519A2FE4.8090201 at t-online.de>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
>
> Moin Attila,
>
> yes ground loops can cause serious measurement problems. And solving
> those could fill a hole book. Here's what I do in practical:
> 1.) avoid the loop
> 2.) if you can't, try harder to avoid it
> - depending on the problem: break up the dc loop by using capacitors
> (most often you only need to kill 50/60 Hz so you can possibly insert a
> C in the shield)
> - if you need dc current or extremly low frequencies flowing in the
> shield, use inductors in the shield to get rid of 50/60 Hz
> - if you have to transfer low frequency rectangular pulses, you have to
> decide or even to try, what will be the better choice
> - but that induces new problems if you have to be synchronous to within
> some ns...
> - on the lab bench - if you can't avoid loops - make the area of the
> loop as small as you can to reduce the inducing field ->  keep shields
> together
> - use a well grounded!! metal plate (use iron, if you can) under your
> experiment and lay the coax cables flat down on it
> - as far as you can connect all case grounds at one point only
> - if you are experimenting with low frequency on your bench you can try
> to not connect the shield on one side of the cable - be aware, that the
> current now takes another way, so that is practicable in only few
> situations (and if you fumble around it will change measurement conditions)
> - use floating power supplies - but remember, they can be coupled to
> earth or the power line over the stray capacitance of the transformer
> (rather a problem for higher frequencies than 50/60 Hz)
>
>
> Volker
>
>
> Am 20.05.2013 14:08, schrieb Attila Kinali:
>> Moin,
>>
>> A couple of weeks ago, there was a short discussion on "bad" connectors
>> and cables and the coupled in noise of those. Summarized it said that
>> measurements in the time-nuts scale are very sensitive to even the lowest
>> noise levels and coupled in signals.
>>
>> But, all the measurements we do are done using some sort of coax which
>> have their shield connected to the case of the devices. As the invovled
>> devices in a measurement are also grounded over their power supply
>> this will lead to ground loops and thus a 50/60Hz noise. Also, because
>> loops are good magnetic antennas, a lot of other noise floating around
>> in the ether is coupled in (eg a nearby radio station).
>>
>> How do you handle this kind of problems?
>>
>> 			Attila Kinali
>>
>>
>
>
>
> ------------------------------
>
> Message: 6
> Date: Mon, 20 May 2013 07:35:18 -0700
> From: "Tom Van Baak"<tvb at LeapSecond.com>
> To: "Discussion of precise time and frequency measurement"
> 	<time-nuts at febo.com>
> Subject: Re: [time-nuts] OCXO shock protection
> Message-ID:<66BC12BF125945FFAD05CC5C91DEA924 at pc52>
> Content-Type: text/plain;	charset="utf-8"
>
>> We will be moving to southern California in the near future.
>>
>> My question if for people who live in that area is do you need to do
>> additional shock mounting for OCXO?s because of the ongoing,
>> usually minor, earth tremors that take place?
> Hi Perry,
>
> Your OCXO will be fine, unless it falls on the floor.
>
> "Shock" mounting won't help because earthquakes are more slow rocking, rolling,
> or shaking than shock. The bandwidth of the acceleration is quite low, around 1
> Hz. Your OCXO, and the shelf holding it, not to mention your house and
> neighborhood will all enjoy the same ride. I suppose a helium balloon suspended
> OCXO would be immune.
>
> Here's a back-of-the-envelope calculation:
>
> The typical OCXO has an acceleration sensitivity on the order of 1e-9 per g. You
> can easily measure your OCXO by placing it on each of six sides, computing the
> frequency difference.
>
> If the quake is strong, the temporal acceleration is on the order of 0.1 g. See:
> http://en.wikipedia.org/wiki/Peak_ground_acceleration
>
> This means the short-term frequency *modulation* is down at the 1e-10 level, or
> below (if you orient the OCXO along a less sensitive axis). Most OCXO already
> have this much short-term noise, so you'd never notice the quake. Even with a
> low noise OCXO, it takes a very high-resolution phase meter to detect rapid
> frequency changes at the 1e-11 level. Note there is little or no effect on
> timing, since the sum of the +/- acceleration over the duration of the quake is
> essentially zero. The same goes for net change in frequency, unless the OCXO is
> gets moved or tilted as a result of the quake.
>
> Pendulum clocks are much more fun to measure during a quake:
> http://www.bmumford.com/mset/tech/quake/
>
> My home lab survived a 6.8 quake. If you do sub-cm GPS you may want to double
> check your coordinates after a strong quake, as your house (and neighborhood)
> may have moved:
> http://leapsecond.com/pages/quake/
>
> /tvb
>
>
>
> ------------------------------
>
> Message: 7
> Date: Mon, 20 May 2013 16:56:10 +0200
> From: Attila Kinali<attila at kinali.ch>
> To: Discussion of precise time and frequency measurement
> 	<time-nuts at febo.com>
> Subject: Re: [time-nuts] time transfer over USB
> Message-ID:<20130520165610.ccd34f902f20a6eee55b82f1 at kinali.ch>
> Content-Type: text/plain; charset=US-ASCII
>
> On Mon, 20 May 2013 06:14:12 -0700
> Jim Lux<jimlux at earthlink.net>  wrote:
>
>> On 5/20/13 2:43 AM, Attila Kinali wrote:
>>
>>> Oh.. and connect the whole thing to a port on the PC that does _not_
>>> have an internal USB hub.
>>>
>> That's a bit of challenge, I suspect.. A casual look at the PCs  I have
>> around here running windows all seem to have on-mobo hubs when you
>> check Device Manager.  I suspect they are integrated into one of the
>> peripheral chips.
> In my experience, it's usually two or 4 ports connected to an internal
> hub, where at least one is not on a hub at all. You have to try to find
> out which they are. I dont know any similar tool with windows, but
> on linux (and i guess *BSD) you can use lsusb to give you a hierarchy
> of the USB system on your mainboard. There you can see which ports have
> a hub behind them and which dont.
>
> 			Attila Kinali
>



More information about the time-nuts mailing list