Interesting GPS module with onboard antenna

A couple are on Ebay for less than $35

formatting link

--
You can't have a sense of humor, if you have no sense.
Reply to
Michael A. Terrell
Loading thread data ...

I did a chip design for SiRF _years_ ago... small world ;-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

I bought some Trimble 'Lassen LP Low Power GPS Receiver' boards on Ebay to build a GPS derived 10 MHz in house frequency standard before I saw those listed on Ebay. The Trimble boards are new, but discontinued so I bought a spare.

--
You can't have a sense of humor, if you have no sense.
Reply to
Michael A. Terrell

On a quick look, the unit referenced by the O/P only seems to provide the 1pps (TM output). Much faster settling time results from using a module with a higher rate e.g. the 10kHz of the Rockwell/Conexant Jupiter modules.

Reply to
who where

The 1 PPS output is a standard. It has been used to sync a TCXO to

10 MHz for well over a decade. You can find several circuits online if you're really interested. BTW, I was the OP in this thread.
--
You can't have a sense of humor, if you have no sense.
Reply to
Michael A. Terrell

Thought you were referring to bees!

Reply to
Nemo

Our gateways have 10Mhz standard and 1pps pulses are the standard as well. The signals are distributed through the racks with a simple, analog, passive distribution device packed with simple splitter/combiners.

All of those signals run on quick and dirty BNC cables, because they are cheap.

The RF stuff goes over ultra low loss SMA cabling.

So one only uses what one needs as overkill has a high attached cost.

So, our GPS and other sync signals, etc. all pass over cheap cables and at relatively low frequencies. So, there is no need to bottle up those signals with expensive gear. The cheap BNC cables are good enough, and the 1pps is absolutely the standard and also good enough.

Reply to
TunnelRat

If you want to build a Brooks-Shera type unit, 1pps may well serve your purpose. No matter how you look at it, the lower the reference frequency the higher the potential jitter, which in the B-S case is only addressed by long conditioning times (hours).

Of course if the 10kHz is itself derived from the 1Hz then the jitter in the 1pps is in turn multiplied. But it is my understanding that the Jupiter's 1Hz and 10kHz are both derived from a higher sat-locked frequency.

I originally had built a TV-derived reference which locked to the line frequency of PAL analog TV transmissions - often in turn locked to a decent source. That used 7812.5 Hz (half the line frequency) to phase lock a local oscillator, from which 10MHz and sub-multiples were obtained. With the availability of a near frequency (10kHz) from the Jupiter modules, and a more reliable signal source I went the Jupiter route, as did James Miller

formatting link
whose article I found after I had built my re-jigged reference.

Pic of proto at

formatting link

The advantage (for me at least) is that this unit does not need to operate 24//7. I use this only occasionally, so I can power it up and get a reference available in minutes.

Yes, I appreciate that Unfortunately "O/P" is an ambiguous term - I was referring to the original post - maybe I should have said referenced IN the O/P.

Reply to
who where

No bees around here, just angry wasps. :(

--
You can't have a sense of humor, if you have no sense.
Reply to
Michael A. Terrell

who where wrote:

I worked with low phase noise synthesizers for about four years. The whole thing will only draw a couple watts, so there is no reason to turn it off. The fact that it is using a 10 MHz crystal gets rid of a lot of noise and jitter. There is more than using a simple low pass filter to reduce phase noise & jitter, including some things that the engineers who designed the thing didn't catch. The fact that the PC board fab added heat relief to the mounting holes didn't help. They raised the phase noise by over 40 dB. Another was that the original design used an uncased disk capacitor that went unobtanium. The MEs decided to just use the same value of SMD caps, standing on edge, next to the 1/8" vias that were intended for soldering the original parts to the boards. The entire test run was so out of spec that I started looking for the causes. I noticed the noise dropped when I stuck a probe in one of the open Vias, so I filled the hole with solder. It helped, so I did the other five, one at a time. Each had about the same effect on the noise level, no matter what order they were filled. The engineers finally aggreed that the additional inductance in the ground plane was the cause, and the holes were eliminated in the next rev of the board. The VCO was 300 to 700 MHz in several overlapping bands, and those caps were where the diodes used for band switching were mounted. The band was changed by forward biasing a diode to ground additional capacitors for the lower bands.

The long settling time is nothing new to me. I worked at Microdyne building telemetry equipment. Some of the AGC circuits averaged at over

100 seconds, so it took a while to test the time period, even with a HP 3325B. The commercial GPS derived standard at Microdyne took a while to acquire the birds, and meet specs.

It could be worse than the other method, if they don't use a crystal in the loop.

Those were also built for NTSC TV. I was a broadcast engineer, and the video proc amp at the transmitter regenerated the sync at the transmitter site. It used 4X colorburst crystal without an oven. The only time you could depend on that unit was on live network feeds, at stations that routed the feed directly to the transmitter.

--
You can't have a sense of humor, if you have no sense.
Reply to
Michael A. Terrell

(snips)

(snip)

"Most" of our local transmitters were locked to the upstream network standard, "most" of the time. That sort of certainty was a distinct drawback, and with digital transmissions looming the move to a local GPS-derived 10MHz reference on the bench was not only wise but inevitable.

Interestingly, I wrote (yes, letters!) to the network engineers at all four national networks asking whether the derivation of a similar (to analog line) frequency reference was feasible in the planned digital transmissions. Two replied, neither seemed to possess any real info on the digital signal structure - to them, it was all black box PnP engineering.

Reply to
who where

Today's TV 'Network Engineers' have 'computer science' degrees. The transmitter manufacturers tell the local engineer what boards to replace when they have problems. They have dumbed down the industry to the mess you see today. When I worked in Broadcast, any decent engineer could design, build & maintain a TV station with little outside help. The exception was tower design, since a PE had to sign off on tower design & wind loading of the antennas.

--
You can't have a sense of humor, if you have no sense.
Reply to
Michael A. Terrell

line

Script kiddies.

?-((

Reply to
josephkk

Well, since these days to much of it is just digital data flying around (at very high speeds), I suspect that even the network engineers who might have the skills to debug something like that simply aren't going to have the equipment to do so with?

I'm told a lot of station owners felt it was too expensive to have to keep a Real Engineer around at each station, so they pushed the FCC to eliminate such requirements and then leaned on the manufacturers to make the products robust enough that anything could be fixed by nothing more than board swapping.

Reply to
Joel Koltner

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.