[ntp:questions] almost decided which new PPS GPS to buy

Ron Frazier (NTP) timekeepingntplist at c3energy.com
Sun Mar 4 23:04:14 UTC 2012


On 3/4/2012 3:18 PM, unruh wrote:
> On 2012-03-04, Ron Frazier (NTP)<timekeepingntplist at c3energy.com>  wrote:
>    
>> Hi all,
>>
>> Some of you know from my prior threads that I've been debating what GPS
>> to buy to upgrade from my USB based BU-353.  I think I've just about
>> nailed it down, based on a number of conflicting factors.
>>
>> I'm getting + / - 6 ms accuracy now with the BU-353.  My goals for the
>> new device are: 1 ms if using USB-serial interface, and 100 us if using
>> a real serial interface.  One problem with the BU-353, is that it's
>> average offset from the internet servers seems to wander around up to 60
>> ms or so.  When that happens, NTPD sometimes marks my GPS as a
>> falseticker.  ESR, the guy mentioned in another thread who wants to
>> connect GPS to 100 routers, called this "wobble", and said it's related
>> to the SIRF chipset.  I haven't done any other research to confirm that.
>>
>> I'm pretty sure I'm going with the Sure board (pun intended) for the
>> following reasons:
>>
>> A) Low cost for a NEW product.
>>
>> B) Antenna and USB cable included.
>>
>> C) No need to add connectors to the board, or a power supply if powered
>> by USB.
>>      
> There is the need to add a connecting wire to the board. You can go with
> David's procedure which involves soldering two wires between legs of
> surface mount devices. Altgernatively you can solder a wire rom the hole
> marked PPS on the board to the dcd pin of the serial port. Both of these
> are no-brainer soldering jobs-- almost impossible to get it wrong.
> (soldering to the surface mount chips is delicate and requires a good
> iron and steady hands). Most serial chips will be fine with the TTL
> level of the PPS signal (Again, does anyone know of some that are not OK
> with it?)
>    

I may try pulling the signal directly from the PPS directly to DCD as 
you describe.  If it works, the procedure might be a bit simpler.  By 
the way, are the data lines on the serial port of the Sure board running 
at standard RS-232 levels?

>> D) Good timing performance.
>>
>> E) High sensitivity for potential indoor operation.
>>
>> F) David Taylor's excellent tutorial on setup:
>> http://www.satsignal.eu/ntp/Sure-GPS.htm
>>
>>   From my reading, it appears that I need to add the following to the
>> Sure Electronics package of parts:
>>
>> 1) DB-9 Male - DB-9 Female serial cable
>>
>> 2) Motherboard header - DB-9 port adapter
>>      
> Sorry, what is this for?
>
>    

Only one of my PC's has a serial port.  There is no external connector 
available on the rear panel.  There's only a row of header pins on the 
motherboard.  This adapter will plug into the motherboard header pins on 
one end and provide a traditional DB-9 male connector on the other end.  
It has a metal plate which screws on to one of the card slots, so the 
DB-9 connector is exposed at the rear of the computer.

>> 3) Plastic box for mounting.  Misc hardware like screws, standoffs, etc.
>>
>> (David T. or others, I don't suppose you can suggest any specific part
>> numbers for those hardware items from Radio Shack [do you have those in
>> the UK], Mouser, or DigiKey.  Just thought I'd ask.)
>>      
> A serial line you can buy at your local dollar/euro/pound store.
>
>    
>> My plan of attack is as follows, although the sequence may vary:
>>
>> A) Get the unit working "at all", using USB NMEA data only.  I would
>> hope to see performance equal to my BU-353.
>>
>> B) Get the unit working with NMEA data from the RS-232 port, going
>> through a Trendnet TU-S9 USB - Serial converter, based on the Prolific
>> chipset.  Performance should, again, be similar to the BU-353 (which
>> also uses the Prolific chipset internally).
>>      
> ?? not sure what you mean by this. Note that you are going to get of the
> order of 10ms ( after correcting for the 500 or so ms offset of the nmea
> output). Part of the problem is that the nmea sentences are of variable
> length. Each character takes about 1ms (at 9600 bd) to come out.
>
> You need the pps to get good timing (sub ms)
>
>    

Internally, the USB GPS I have, the BU-353, is using a Prolific serial - 
USB chipset.  So, when I plug it in, it creates a virtual com port.  
However, it doesn't pass any handshaking signals as far as I know.  Or, 
perhaps they're just not connected.  The TU-S9 USB - serial adapter I 
have uses the same chipset, but it DOES pass all the handshake signals 
coming from the DB-9 port.  So, I want to connect a serial cable to the 
new GPS, and the other end of that to the TU-S9, and the other end of 
that to a USB port on the PC.  If I have PPS active on DCD, this SHOULD 
bring the PPS into the PC through the TU-S9 adapter.  Yes, it will be a 
USB message, not a true hardware interrupt.  But it should at least work.

I'm already getting better than 10 ms just using NMEA data only on a USB 
virtual com port.  I've got the GPS outputting data at 57,600 baud to 
minimize latency and jitter.  I'm also outputting the GPZDA sentence (as 
recommended in one of the Trimble manuals).  I'm pretty sure this 
sentence is fixed length.  It only has time in it, and a couple of other 
minor things.  I'm getting + / - 6 ms accuracy polling the device, on a 
USB port, every 8 - 64 seconds.  Jitter is averaging 2.4 ms or so.  So, 
I'm already less than 10 ms, at least just a bit.  I'm hoping to get + / 
- 1 ms performance once I get USB going through an adapter with 
handshaking output.

>> C) Get the unit working with NMEA data from the RS-232 port, going to a
>> real serial port.  I have to tinker with the motherboard on another
>> computer to do that.
>>
>> D) Get the unit working with NMEA + PPS, with data from the RS-232 port,
>> going through the TU-S9, which claims to pass all the handshaking
>> signals.  I'll have to do the mods that David recommends on the board
>>      
> See above for this.
>
>    
>> for this.  Hopefully, the same procedure will still work on the recent
>> boards.  This is where I'm hoping to get + / - 1 ms performance.
>>
>> E) Get the unit working with NMEA + PPS, with data from the RS-232 port,
>> going to a real serial port on the other computer.
>>
>> This should be an interesting adventure, hopefully ending with a time
>> server running at 100 us accuracy or better.
>>      
> You can get 2-3us if you use a real serial port with dcd interrupt.
>
>
>    

2 - 3 us would be great.  Even 100 us would be great, since my needs are 
not critical.  On a USB port, I will consider 1 ms to be great.

>> Sincerely,
>>
>> Ron
>>
>>      


-- 

(PS - If you email me and don't get a quick response, don't be concerned.
I get about 300 emails per day from alternate energy mailing lists and
such.  I don't always see new messages very quickly.  If you need a
reply and have not heard from me in 1 - 2 weeks, send your message again.)

Ron Frazier
timekeepingdude AT c3energy.com



More information about the questions mailing list