[time-nuts] IRIG-B audio decoder circuits and ICs sought

[Joseph Gwinn]

Responses interspersed below.  Joe


On Sat, 23 May 2015 11:54:53 -0400, time-nuts-request at febo.com wrote:
> Today's Topics:
> 
>    1. Re: IRIG-B audio decoder circuits and ICs sought (Tim Shoppa)
>    9. Re: IRIG-B audio decoder circuits and ICs sought (Esa Heikkinen)
>   15. Re: IRIG-B audio decoder circuits and ICs sought (Hal Murray)
> 
> 
> ----------------------------------------------------------------------
> 
> Message: 1
> Date: Fri, 22 May 2015 06:48:59 -0400
> From: Tim Shoppa <tshoppa at gmail.com>
> To: Discussion of precise time and frequency measurement
> 	<time-nuts at febo.com>
> Subject: Re: [time-nuts] IRIG-B audio decoder circuits and ICs sought
> Message-ID:
> 	<CAJ_qRvZ+1Zep_Wx8k+o-qY=h3dxgdnLcpJ20tuaW_nCffmiRtA at mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
> 
> On Thu, May 21, 2015 at 10:22 PM, Joseph Gwinn <joegwinn at comcast.net> wrote:
> 
>> The definition of "good" here is tenth-microsecond alignment between
>> the 1PPS output of the decoder and the incoming IRIG-B12x signal.
>> 
> 
> 
>>> From: Tim Shoppa <tshoppa at gmail.com>
>>> To: Discussion of precise time and frequency measurement
>>>       <time-nuts at febo.com>
>>> Subject: Re: [time-nuts] IRIG-B audio decoder circuits and ICs sought
>>> 
>>> See for example the Truetime 820 decoder. Discriminators, One-shots, and
>>> Flip-Flops with pots to tweak the levels.
>> 
>> Hmm.  Interesting.  URL?
>> 
> 
> I'm not sure 0.1us is a reasonable expectation to get PPS from the 1kHz AF
> variant of IRIG-B.

In 1950, it was millisecond only.  But people now use it for 
microsecond level timing, and some COTS units do guarantee 50 
nanoseconds one-sigma between IRIG in and 1PPS out.  For instance, the 
TSync-PCIe, according to Spectracom.

The trick is to phase-lock a 10 MHz TCXO to the 1 KHz IRIG-B12x signal, 
and use the decoder to deduce which 10 MHz cycle was the beginning of 
the second using a slow delay-lock loop.


> 1kHz AM IRIG-B was usually distributed across a site/range by
> telephone-type wiring using interspersed audio transformers for isolation
> in the long haul. It was used to drive simple displays that used 60's
> transistors or 70's SSI chips, and also recorded on parallel tracks on
> telemetry recorders. It's pretty cool because when playing back old
> telemetry tapes we would just use the same clock display we did when live,
> but when fed the audio from the recorder it showed us the time of
> recording. (In some cases in the 80's, I got to work with telemetry tapes
> that were 20 years old! Today they'd be 50 years old and you could still
> play it back while watching the display). Sometimes we would slow down the
> tape and on the pen recorders optimistically we could eyeball times with a
> resolution finer than the 10ms bit rate, but not better than 1ms.

 All true, but see above.

> 
> ------------------------------
> 
> Message: 9
> Date: Sat, 23 May 2015 00:31:32 +0300
> From: Esa Heikkinen <tn1ajb at nic.fi>
> To: Discussion of precise time and frequency measurement
> 	<time-nuts at febo.com>
> Subject: Re: [time-nuts] IRIG-B audio decoder circuits and ICs sought
> Message-ID: <555FA034.9010108 at nic.fi>
> Content-Type: text/plain; charset=UTF-8; format=flowed
> 
> Joseph Gwinn kirjoitti:
> 
>>> I prefer the DC level shifted variant of IRIG-B.
>> I like and use IRIG-B00x too, but it only reaches a few meters, versus 
>> the required tens of meters.
> 
> It's differential RS485-alike bus (with TS2100 at least) using 5V 
> signalling level. It works easily more than few meters with twisted pair 
> cable. For RS485 they claim more than kilometer if the speed is less 
> than 100 kbit/s and here it's only 100 bit/s. So only tens of meters 
> there's no problem.

When I last used IRIG-B005, the vendor (Symmetricom?) said it was good 
for a few meters only on shielded twisted pair. I recall that the 
handling of shield grounds was strange.   This was OK, because the 
signal was confined to one cabinet, and it worked just fine.

But there was no mention that the B005 followed RS422 or RS485 or 
anything else.  Nor does IRIG 200-04 say this, so I suspect that each 
company solves it differently.

 
> I think the AM was originally intended for transmitting IRIG-B 
> wirelessly or analog tape/film soundtrack recording...

That is correct.

 
>> Yes, but I must have IRIG-B12x (Amplitude modulated 1 KHz sine wave), 
>> and the analog processing complicates things.  I think that one best 
>> implements the IRIG decoder in a DSP chip.
> 
> Yes and there will be delays. As long decoding delay remains constant, 
> it's easy to compensate. In my implementation the decoding delay is 48 
> cycles (9,6 usec) and remains constant. Actually this is delay from 
> rising edge verification to timer setting point.

Yes.  One designs such that the delays are predictable and constant.

 
> When decoding IRIG-B it's one second behind. Timekeeping functions are 
> needed anyway. Data verification system comes as a side product. It's 
> only needed to compare last received IRIG-B frame with timestamp of 
> passed second. If times differ then there's bit errors OR timekeeping is 
> out of sync. Code has to decide when it should just ignore the received 
> data and when it should synchronize the internal timekeeping.
> 
>> TS2100s are generating a lot of replacement business for GPS vendors.
> 
> Yes, planning to buy at least one spare unit when their prices will drop 
> on Ebay... :)

The problem is that they don't handle GPS week field overflow gracefuly.

 
>> To be modern, one must code in C?  Isn't that true?
> 
> I use assembler only with these 8-bit PIC's. It's little bit special 
> case, there's no point to use assembler with any larger processors. This 
> little baby has only 35 assembly commands. Because most of them run in 
> single cycle it's quite easy to write time critical code like this 
> IRIG-B decoder.

Ahh, well, we still must tease assembly both assembly programmers and 
Python programmers.


> 
>> 
> 
> ------------------------------
> 
> 
> ------------------------------
> 
> Message: 15
> Date: Fri, 22 May 2015 19:01:51 -0700
> From: Hal Murray <hmurray at megapathdsl.net>
> To: time-nuts at febo.com
> Cc: Hal Murray <hmurray at megapathdsl.net>
> Subject: Re: [time-nuts] IRIG-B audio decoder circuits and ICs sought
> Message-ID:
> 	<20150523020151.16A77406057 at ip-64-139-1-69.sjc.megapath.net>
> Content-Type: text/plain; charset=us-ascii
> 
>>> It's util/tg2.c
> 
>> How do I get hold of the code?  Is it in the NTP distribution,
>> versus all by itself somewhere?
> 
> It's in the normal distribution.  Tar file available from:
>   http://support.ntp.org/bin/view/Main/SoftwareDownloads
>

As I suspected.  Thanks.


> 
> ------------------------------
> 
>
> 
> End of time-nuts Digest, Vol 130, Issue 33
> ******************************************