GPS Module for Oscilloscope

[W8LV]

Is there a simple and cheap no brainer plug in GPS Module that will work with all oscilloscopes in general? I have a Siglent 100 mHz scope, with a time base BNC connector plug in on the back. It sits on my bench in the Basement Lab. Above it is a window...And above THAT is a sky full of GPS Sats that I already pay taxes for. Seems like a shame not to take advantage of them.

73 and All the Best!
DE W8LV Bill


Sent from my SM-N920V using Tapatalk

[DaJMasta]

Are there Siglent scopes with a 10MHz external reference input?  Last I checked, their higher end 2000X series doesn't have the input required to use an external time reference.


If I'm wrong and yours does, you'd want to be looking for a GPSDO.  A GPS itself is good for getting time locked to an 'absolute' reference, but for it to be usable by a scope or other device you generally need it to be in the format of a 10MHz signal.  The most effective way to get that easily is to use the GPS output to discipline an OCXO, which gives you the excellent long term stability of the GPS pulse with the short term stability offered by the ovenized oscillator.  If you're looking for one, I'd start with searching around for used units from telecom installation pulls, there are quite a few available for reasonable prices now.

[Ghislain]

Hi Bill,
Regurarly check for one of these on e-bay: Trimble Thunderbolt GPS 10Mhz Disciplined Oscillator GPSDO Frequency Standard.
Very reliable and easy to setup.
You may need to buy a small external antenna to make the package complete (some are offered with)

[nctnico]

There isn't much use to have a scope on an external reference. For starters: the frequency measurement isn't precise enough to show a difference.

[cdev]

The TruePosition telecom surplus GPSDO (I bought mine for around $40 on ebay) is a really great deal if you are willing to put it in your own box. You end up with an extremely precise 10 MHz reference (precise to as many as 11-12 decimal places) and a 1 PPS signal with much less jitter than any GPS by itself. They are overbuilt in a way consumer gear is not.

You will need an amplified GPS antenna and a good sky view free of obstacles for best performance. (The time figure of merit is influenced by the clear sky view, obstacles cause slight delays.. so has nothing to do with signal strength)

[hamdi.tn]

Ublox-M8F have a 30.7MHz deciplined VCXO. with 100ppb stability if you lose GPS signal for whatever reason.

[tautech]

Is there a simple and cheap no brainer plug in GPS Module that will work with all oscilloscopes in general? I have a Siglent 100 MHz scope, with a time base BNC connector plug in on the back.

That's not what the rear BNC is for.
Depending on the model/series you have there maybe two, Trig Out and/or Ext Trig input.

In any case, Siglent scopes (current models) don't provide a frequency Ref input, only their AWG's and spectrum analysers offer this feature.
At anytime you might be concerned about internal frequency accuracy be sure to run the Self Cal but only after having the unit warm up for a while, details of such are in the user manual.

[mikeselectricstuff]

There is no point giving a scope an external reference - it's not a precision tool.

[timb]

The only scope I’ve ever seen with an external reference input is the old Tek 2465 series with Counter Timer Trigger option. Adding an external 10MHz reference will net you an extra digit of resolution on the counter.

[DaJMasta]

A lot of high bandwidth scopes will have a reference input, and it does make a difference if you're zoomed in and want to sync to the signal you're looking at or if you need the extra stability it can offer for long persistence build ups or whatnot.  With a 100MHz scope, I doubt it will be much of a benefit in almost any application, but I also don't think Siglent's current offerings include a scope with an external reference.

[picitup]

I bought the BG7TBL GPSDO from eBay which came with an external antenna.  It works great.  The 1pps drives my £1.00 atomic clock which after 3 months is bang on to the second.

https://www.eevblog.com/forum/projects/t81709/msg1117592/#msg1117592

The 10Mhz drives my tti TF930 frequency counter and sdg1025 awg direct without amplification or buffering.

If you're looking for a scope with an external 10Mhz reference input, the Rigol DS/MSO/4000 series all have it although they are quite pricey.

Cheers

Steve

[Keysight DanielBogdanoff]

It's very common for higher end scopes to have an external 10 MHz reference input BNC, for us it's the 4000 X-Series and up. It's most useful if you want to sync multiple instruments to the same clock, but also gives you the timebase accuracy of your reference if that's a concern. Typically scopes have a timebase accuracy in the 1-30 ppm range, with super high end scopes having a .1 ppm accuracy.

[NivagSwerdna]

Is there a simple and cheap no brainer plug in GPS Module that will work with all oscilloscopes in general?

Yes and No. 
In my very limited experience pure GPS frequency references often have a certain amount of dither... i.e. they will jump occasionally.  What you are probably referring to is a GPSDO.  Ex-mobile GPSDO modules exist on ebay and there are threads on the forum for the Trimbles and Symmetricom etc...  they do need some love and attention as they are just modules but if you get one that works (many have been removed with a sledge hammer from decommissioned mobile equipment) then it as easy as pointing one end at the sky and plugging the other end into the external reference BNC on your scope.
I have a Symmetricom.... but... their accuracy is based on a convergence so they ideally need to be powered on all the time.
An alternative is to use another frequency reference... e.g. a Rb standard... these turn up on EBAY too but are a bit more expen$ive... these units offer quite good accuracy after a few 10s of minutes.
BUT....
You probably don't need one anyway (although it is fun to get a Rb or GPSDO working for the sake of it).  I use my GPSDO occasionally to tell me that my frequency counter is still in calibration... which is pretty much always is! 

[exe]

Yeah, some scopes with horrible jitter could definitely benefit from this 

[Zucca]

for us it's the 4000 X-Series and up.

Why not the 3000 X-Series? Oh well my hack to do list is getting bigger and bigger... 

[W8LV]

Is there a simple and cheap no brainer plug in GPS Module that will work with all oscilloscopes in general? I have a Siglent 100 MHz scope, with a time base BNC connector plug in on the back.

That's not what the rear BNC is for.
Depending on the model/series you have there maybe two, Trig Out and/or Ext Trig input.

In any case, Siglent scopes (current models) don't provide a frequency Ref input, only their AWG's and spectrum analysers offer this feature.
At anytime you might be concerned about internal frequency accuracy be sure to run the Self Cal but only after having the unit warm up for a while, details of such are in the user manual.

Thanks!

73 and All the Best! DE W8LV Bill

[W8LV]

Is there a simple and cheap no brainer plug in GPS Module that will work with all oscilloscopes in general? I have a Siglent 100 MHz scope, with a time base BNC connector plug in on the back.

That's not what the rear BNC is for.
Depending on the model/series you have there maybe two, Trig Out and/or Ext Trig input.

In any case, Siglent scopes (current models) don't provide a frequency Ref input, only their AWG's and spectrum analysers offer this feature.
At anytime you might be concerned about internal frequency accuracy be sure to run the Self Cal but only after having the unit warm up for a while, details of such are in the user manual.

I actually have the Siglant digital AWG... Do you know if a reasonably priced GPS time reference? This would also be valuable to me to sync transmitters for Weak Signal Work...

73 and All the Best!
DE W8LV Bill


73 and All the Best! DE W8LV Bill

[tautech]

Is there a simple and cheap no brainer plug in GPS Module that will work with all oscilloscopes in general? I have a Siglent 100 MHz scope, with a time base BNC connector plug in on the back.

That's not what the rear BNC is for.
Depending on the model/series you have there maybe two, Trig Out and/or Ext Trig input.

In any case, Siglent scopes (current models) don't provide a frequency Ref input, only their AWG's and spectrum analysers offer this feature.
At anytime you might be concerned about internal frequency accuracy be sure to run the Self Cal but only after having the unit warm up for a while, details of such are in the user manual.

I actually have the Siglant digital AWG... Do you know if a reasonably priced GPS time reference? This would also be valuable to me to sync transmitters for Weak Signal Work...

No sorry I don't but there's some forum interest on this topic.
Do a search here for GPSDO and you should find something you can build of at least point you in the right direction.

[nctnico]

Do you know if a reasonably priced GPS time reference? This would also be valuable to me to sync transmitters for Weak Signal Work...

This one is OK:
https://www.ebay.com/itm/GPS-Receiver-GPSDO-10MHz-1PPS-GPS-Disciplined-Clock-with-Antenna-Power-Supply/251802969291

[cdev]

If you have a 10 MHz frequency standard you can use very simple PIC hardware to time-stamp events to microsecond precision  as well as perform multiplication and division on waveforms without any error.

That approach is also really cheap, and from what I have read you don't have to worry about quantization error because the super simple PIC doesn't work that way.

Similarly cheap now is using GPIOs on an SBC like a RaspberryPi along with a GPS (or GPSDO would be better) 1 PPS and the timestamp/timing tools in PrecisionTimeProtocol (PTP) in Linux to measure time intervals.

You could get really high precision either way, considering the cost. A few years ago measuring anything time related that precisely would have required much more expensive ($1000s) equipment.

[nctnico]

PtP only gives reasonable resolution if the network interfaces offer hardware timestamping.

[cdev]

If you can connect a few short headers between it and a Raspberry Pi you can get a Navspark Mini that works great for 1PPS plus a USB-UART adapter for free, plus the $10 shipping cost.

The GPS in the Navspark Mini is Venus 8 based and really good, (especially considering its extremely low price) but it is not a "timing GPS". However, it is good enough to be an good stratum 1 NTP server, however because the network card is USB and does not support hardware timestamping the performance has several areas where it could be improved. That is not the GPSs fault though, at all. Even switching to a timing GPS I suspect would make little difference compared to other optimizations. They should be done if your needs involve instruments or measurements on the raspberry pi directly using its GPIOs, SPI, i2c bus, etc.  For everything over the network the network latency would reduce the precision.

I would be happy to help you get it set up your GPS (any kind) on a Raspberry Pi. Its pretty easy.

You will need a GPS antenna, either active or passive would work fine.

If you get the Navspark mini, its antenna input uses the tiny u.fl plug so you likely need an adapter, which costs around a dollar on ebay.

Virtually any decent GPS antenna, active or pssive will do for the navspark/skytraq GPSs. I use a passive taoglas antenna and my GPS is indoors and quite some distance away from any windows, still as far as I can recall it has not lost its GPS lock ever, without my causing it. It just works. Performance of its PPS pulse as seen by NTP seems to be top notch.

Check out David Taylor's site at satsignal.eu for info on using NTP with a Raspberry Pi.

As others say, the input on your scope likely isn't for a 10 MHz frequency standard.

What could be super useful though is using a PIC or RPI to timestamp events in conjunction with your scope. Then you will get precise numbers you can attach to the pictures your scope gives you.

Is there a simple and cheap no brainer plug in GPS Module that will work with all oscilloscopes in general? I have a Siglent 100 mHz scope, with a time base BNC connector plug in on the back. It sits on my bench in the Basement Lab. Above it is a window...And above THAT is a sky full of GPS Sats that I already pay taxes for. Seems like a shame not to take advantage of them.

It is..

You can take advantage of GPS's in dozens of ways, and I'm sure your scope will be extremely useful.
Lack of a direct 10MHz timebase input is not important.

Even if you are just using a GPS 1PPS to tell your computer's clock exactly when the seconds change right there you are getting a very useful improvement.

Does your computer have a real, (non-USB) serial port? If so, you should get a level changer so it can work with 3.3 volt hardware via its serial port.

An interrupt with low latency that the timekeeping software can see (often the DCD line of a serial port is used)  is important. Thats where desktops are better than laptops because an add on card that adds a real serial port can be used...using the DCD as well as he Txd/Rxd

USB timing is not predictable.

If you are using a laptop that doesn't have a real serial port (or parallel port or anything similar) there may not be any way to get really precise timing data into it unless its timestamped externally somehow, (and in that case its not useful for NTP)

[cdev]

Yes, that is true, so whatever nanosecond timing you need to do needs to be done using the RPIs own GPIOs and ideally if you're doing stuff like that you should recompile your kernel to add support for kernel PPS, which I havent done because as you say, once you're using the network you lose it.

RPI supports it in software on receive, not transmit but- they to shave a few bucks off the cost used a USB/ethernet ethernet card which makes its latency measurably worse than if they hadn't.

This is an area I am not super knowledgeable about, and wish I knew more.

PtP only gives reasonable resolution if the network interfaces offer hardware timestamping.

[exe]

I'm not sure what accuracy you need, but I thought if you want to have stable predictable performance, you cannot rely on Linux. These GPSDOs provide very low jitter... I have doubts that it's possible to improve an oscilloscope's internal oscillator with a Linux-powered clock reference (unless all done in hardware and Linux is only for configuration).

[nctnico]

If you can connect a few short headers between it and a Raspberry Pi you can get a Navspark Mini that works great for 1PPS plus a USB-UART adapter for free, plus the $10 shipping cost.

The GPS in the Navspark Mini is Venus 8 based and really good, (especially considering its extremely low price) but it is not a "timing GPS". However, it is good enough to be an good stratum 1 NTP server

For use as a reference for test equipment NTP is utterly useless.

Always remember the following list:
Best NTP accuracy: 1ms
Best PtP accuracy: 1us
There is a PtP extension in the making which has sub-ns accuracy.