WOW, leave it to Andreas, he will tell you EVERYTHING from the least expensive version to affordable price - doable price! I ran the video 2x, since I missed stuff the first time. I might listen again Wednesday! Thanks Andreas, you are the BEST!
Gruss dich Andreas. When I was young we used a cheap FM stereo receiver. We extracted the 38/19kHz stereo pilot-tone out of a national radiostation. These where about the most stable frequencies you could get in the 80s. Then put it into a PLL to get the frequency we wanted. This video brings back good memories. Thanks. I follow you every now and then. But you got a new subscriber today.
Like you, I am a retired RF engineer. About a year ago I built my own GPSDO. Comment: the 10MHz output of a U-blox receiver has a lot of jitter, it is coming from the 26MHz osc. I had the U-blox put out 1MHz and then divided the 10MHz OCXO by 10 and used a phase and phase frequency detector and used an ESP-32 as the controller and web browser system monitor. 1 MHz provided gave sufficient phase range to account for jitter from the GPS and keep the initial lock time down to a few minutes. In software I built a variable speed loop... very fun project. You can get down to about 5.0 10e-11
Unfortunately, I am just an EE engineer with love to RF, but never sufficient know-how .-) And I have to wait a few years for the pension. Your project is very interesting! I saw the horrible jitter, too. And I have a few "ovens" in the mail. So it might be that I also will try to create such a GPSDO. And I still have ambitions to understand and measure the Allen deviation and compare the different devices.
@@AndreasSpiess I should also warn, that I have been told that some of the U-Blox Neo modules are rejects that are sold under another name, then a clone U-Blox sticker is applied, and sold in China. While I don't know if this is true, I can say that the jitter can be >10X worse. If the price is too good to be true.... beware... or buy direct from U-Blox distributors.
They are Swiss company and we wanted to do a video about clones. Unfortunately they never came back... On the other hand I canot believe that there are so many rejects in a mordern manufacturing process.
@@AndreasSpiess I should not have used the word 'rejects'.... U-Blox sells many different grades/variations of their receivers. You can query the part, and it will tell you exactly what version it is. Someone has removed the sticker on the part, and replaced it with another sticker, which represents it as a more expensive, higher specification part. I have 'altered' parts, but electronic signature reports the truth. So, they are not clone silicon, they are parts where they received a 'label' upgrade. :->
Enjoyed watching, thankyou! For sake of interested folks also reading the comment: Only AU cut crystal can go into making of ovenised oscillators as the mean of S curve (drift vs temp curve) pivots at 70 degrees, while for Regular Crystal (AT cut) the mean pivots at 25 degrees. I hope I have saved a curious mind trying to put a regular crystal into an electronic oven hoping to achieve better accuracy.
@@AndreasSpiess, also, above a particular temperature a crystal will drift in frequency in one direction, while below that particular temperature its frequency will drift in the other direction, (the S curve.) But there is a plateau temperature where a crystal is relatively immune to frequency drift with time. Which depending on the Xtal cut is either at roughly 70C or 25C, but not precisely at those temperatures. I assume the most precise OCXO's each have their exact oven temperature hand tuned to put that crystal on that temperature plateau, but I do not know this for certain, as Xtal-standard companies are fairly quiet about proprietary techniques for their manufacturing precision, stability, and manufacturing cost control. But that was how I set up my home-brew 1 MHz standard that I disciplined with a TRF WWVB receiver at 60 kHz before the GPS constellation was launched into orbit. It was one of the most interesting projects I ever built.
@@AndreasSpiess (gasp) Am stunned that this gentleman with a Swiss accent would stoop so low as to watch that foul-mouthed, talking-hands, Canadee-anne destroy perfektly good, Chinesium tools. Howdaya understandem? AvE: "... you know they say, I am not a gynecologist, but I'll take a look."
Awesome video Andreas. Firstly, it is highly informative and useful. Secondly, I could relate to it as it reminded me of a project I did years ago where I had to synchronize a sheet coater with a oven. Although the motors were digitally geared together with incremental encoders, it relied on phase synchronization to keep things in time. The lacquered sheets had to be fed into the holders of the oven at the right position as they came past on a chain. These sheets are what tin cans in various forms and also beer bottle crown tops are made from. Pulses from both machines were used to create sawtooth wave forms and then an error signal was generated by comparing amplitude. A proportional correction was then calculated and added to the digital gearing ratio. The oven being the master generated a signal with a 50% mark space ratio. It was sufficient to only apply correction whilst the signal was high, and then freeze it whilst it was low. So as you say, using a time standard to discipline an oscillator is highly effective, reliable and simple to implement. I love these types of discussions, right up my alley. Thanks for sharing. And I am glad to see your hand seems to be getting better. Best Regards, Duncan South Africa
Another "Master class" Andreas, thanks. The GPS part is so interesting and affordable for everyone doing any kind of timing... I don't know if it's enough for a separate video... [ES] Otra "Clase magistral" Andreas, gracias. La parte del GPS es tan interesante y asequible para todos los que hacemos cualquier tipo de sincronización... No sé si será suficiente para un video dedicado...
@@AndreasSpiess 9.7 over 10. There is noticeable that there is a speaker from a sudamerican country. We, spanish speakers, use the Spanish language in a slightly different way at both sides of the Atlantic as British or Americans do with the English. But those are not errors! The subtitles are 100% comprensibles and only a few words can be more precises from a technical point of view. For the purpose of divulgation is a really good job. Yours faithfully ;-)
I love this topic. I am obsessed with time and have a Pi running a NTP service in my house based on a GPS antenna. Like you did here, I have also used the signal for my own purposes as well.
Andreas Spiess I will link the guide here. I followed the old version of the post and it looks like they refreshed it in 2018, which is good to see updates. It uses the exact hardware I have installed. I have been running this for three years now and it is great. The main idea is that you still connect to an external time server (Stratum-0) to get the base time, but it uses the local PPS signal from GPS in order to maintain that to perfect accuracy. It has been wonderful to have this in my house and be able to point all devices at a local NTP server. ava.upuaut.net/?p=951
I used to live in Switzerland for graduate school, and when I was there I and a few other "time nuts" (there's a mailing list by that name for enthusiasts of precision time and frequency) did a tour of the time lab of METAS, the Swiss national standards lab. The guys in the time lab there were a bit puzzled why we wanted to see their time lab (they don't get many visitors) but they showed us around and we chatted for a while. Nice people. They have a nice set of HP 5071A cesium reference standards, a research-grade cesium fountain clock, and a hydrogen maser that they use to define UTC(CH), and help define UTC in conjunction with other national time labs. They said they used to get a lot of financial support from Swiss watch companies until the companies realized that people aren't buying their watches for keeping excellent time but rather as a fashion item, so eventually the financial support stopped. Now they do a lot of calibration of commercial and industrial time standards, but even that is not as popular as it was because anyone can get intrinsically calibrated GPS-based time and frequency standards and let the USNO deal with all the hard work.
I visited the time lab at the University of Neuchatel. They did work for the Galileo program. But I do not know how they are financed. At least for me it was very interesting (I was allowed to compare my GPSDO with their maser...
This is a very good video. Thumbs up. And of course Swiss people are renown for timing accuracy, so this is consistent with getting the information from a guy with a Swiss accent. On the downside, Swiss people are not known for having lots of sex. For information on that aspect, we will be waiting for a video by a gal with an Italian accent! ;-)
I am so happy to have acquired the same IKEA clock as you, thanks to your video. So far I have not hacked it to diplay another time. The main issue is that it only synchronise once every hour or so. You cannot get it to move the way you want, or you have to cut the power to make it forget, then be the strongest signal to bring it to the time you want.
Dear Andress, Thanks for your kind reply. You are welcome to my Country. I always share your video in our "Robo Course" App. You may download from Play Store.
Andreas, Thanks for this video. For our microwave projects (Pacific Northwest VHF Society) we use the Leo Bodnar GSPDO with a 5-pole low-pass filter to remove the harmonics. We're on 10, 24, 76 and 122GHz DE WA4OSH
EHF :D Take a look at this eval kit: siliconradar.com/evalkits/#sirad-simple www.bay-net.org/uploads/1/2/2/7/122774721/k6ml-122ghzradio-baycon2018.pdf ghz-europe.com/vk2xax-more-tests-of-122ghz-transverter/ It needs a 10MHz reference
@@AndreasSpiess The resolution of the FMCW radar is much higher than the PIR sensors that you tested in April. Also, check out the NXP CompactRadar www.nxp.com/products/rf/radar-transceivers:MITERWAVEICS#/ I'm trying to get an evaluation kit. Home security that does not get triggered by cats, dogs, birds, deer, etc. but properly identifying humans and cars is my goal.
very good summary of a very technical field! Look for "Time Nuts" -- the news group and web site, to find people that spend a great deal of effort on these things, chasing picoseconds... Old HP OXCOs, like the 18011 dual oven 10 MHz oscillator are still superb performers. eBay sourced rubidiums are also good, but the close-in phase noise of rubidiums is far worse than a crystal. With an inexpensive GPS receiver, you can have very good and very stable time and frequency references for your home lab. (If you know what "Allan deviation" is, you're probably a time nut.)
@@AndreasSpiess Yes, but you can use it with electronics and stay away from the dangerouse hobbies! ha ha ha Glad to hear that your doctor gave you a great news!!! We are all happy for you! Stay safe!
If Ublox modules are with programmable GPSDO - why to buy those expensive boxes with OCXOs and Altera? Small cheap Ublox GPS modules will be still more stable and precise. Addition with AD9834 and some frequency multiplier for it (Triple 25MHz made with UBlox) Should be enough to make any GPSDO based frequency... Very nice video about oscilators!!! I am building now my labs Frequency generator with 2x AD9834 (Sometimes need phase shifted signals) and 2x AD9833 (For square wave signals and PWM modulated signals). Now I am thinking about putting in this box UBlox and GPS antenna to get GPSDO based Function generator. Thank You Andreas for great explaining!
@@AndreasSpiess According to Your video at 18:18 its clearly visible that there is really bad 50/50 ratio and phase is completly out of range. Lets try to set 25MHz and repeat the same test. My plan is to use 25MHz setting and after it I would like to put Band filter to 25MHz (That should shift phase but also stabilise it). Another approach can be like You tell about PLL with Low Pass filter - 25MHz with not phase locked loop, but frequency locked loop where error is compensated with some very low frequency filter (VCO). Anyway - Your video was for me very educational and explains me a lot of unknown before things!
It depends what you want to get out of it. For sure it will have much more phase noise than for example a OCXO. There is anothe comment conccerning this topic, too.
The GPS receiver has a jitter of up to 1us. So, actual accuracy "out of the Blox" is only about 1ppm. Only after averaging out on a very stable oscillator, like OCXO for about 3 hours (1E+4 sec) one can get the actual absolute frequency below the 0.01ppm (1ppm / sqrt(10000))
Neat seeing the Adalm Pluto mod you made also. Didn't know about the aliexpress version. Amazing! Amazing what the latest and greatest NEO-#'s are capable of also. Especially if using for tracking with the RTK though that is a completely different topic/subject. I still think there is a way to hack out the modules even with the Digital Block to make a really low noise communication device receiver. Would be awesome to be able to hack into just the RF Block since seems like would be an awesome receiver/detector for a receiver or spectrum analyzer or I guess VNA too.
@@AndreasSpiess Yeah, was thinking at the least can seal up the GPS module in a shielded case and use the antenna input like for the output of a mixer with a fixed frequency design for either sweeping like for analysis or for just whatever transmitter tuned as a for use as a radio receiver. Though since stuck with the digital block format... would have to design the transmission protocol, if using as a sensitive receiver, to encrypt to the Digital Block format so can be decrypted/translated from the Digital Block format using the transmission protocol format. One of those different ideas I had one day when drooling over the noise floor specs of the latest uBlox modules. Albeit the baud rate will be low... though is improving with each new version.
Thank you for the video, I think this is very educative. As always I appreciated the tone and jokes ! I work in a atomic frequency standard lab in Switzerland. I could measure the stability of your GPSDO using one of our hydrogen masers. That way you would exactly know how stable the unit is. What do you think ?
@@AndreasSpiess I know it's been quite some time, but do you have an update on this? Ever got to check the GPSDO precision and stability with @Etienne Batori ?
@@Halabaloosa Yes, was at the University of Neuchatel, and Etienne tested my devices. They were very stable as long as the GPS reception was good. So the antenna position is more important than the device.
Can you please make a video on how to choose the right microcontroller, I recently got an MSP432 for my college project and finding it hard to work with
@Scott Page USMC Thank you for your interest, My problem is mostly with using code composer studio, I don't see many examples on how to program MSP432 whereas STM32 has got lot resources on web so was wondering if I have switch to that. I am very familiar with using arduino boards, but I was told not to use that.
@Scott Page USMC Thank you for your interest sir, TI is sponsoring the kits for us thus we are supposed to use them, I have searched some books and reading those.
Using the GPS clock signal is easier than I thought! (I never know that PPM is also adjustable too) How do you wire the GPS signal to the basement? Do you have to put the module outside and wire the pulse signal into the room? Or broadcast the clock signal as you do with your Ikea clock video?
GPS antennas are "active". They have an amplifier built in. So you can have quite a long cable from the antenna to the basement. But for sure the antenna must "see" the satellites.
Hi. Great explanation in a short time span! Short question: in the video at about 18':15" we see the square wave of the cheap GPS receiver. Would you know why the square wave has/looks irregular width, with 2 wider and 3 shorter positive period square? If this works that would be a very cheap solution to a very high accuracy clock!
@Andreas Spiess I think I know now the Why part. Just watched an interesting video. It seems that the (programmable) output frequency of the gps generator is provided by dividing the main clock of 24MHz. If you choose a frequency that has fraction, then you get a weird mix. For example: 10MHz would be factor 2.4. It seems you get a mix of 12 and 8MHz. More info on: th-cam.com/video/M6SmdqMA-Dg/w-d-xo.html
11:35 how do you determine the accuracy of the most accurate clock ever built? Just wait 30 million years? ;-) What's the reference there to determine the drift?
Have you seen the SiTime MEMS TCXOs? Seems like they'd be a good combo with GPS/atomic sources, they outperform many OCXO and are small SMD components..
No, I did not see them. They seem to be interesting parts. And, as you write, very small and energy efficient compared to OXCOs. And maybe they are good enough for many applications and can replace a GPSDO. However, precision is not as important for a GPSDO. Stability is more important because the absolute frequency comes from the GPS signal.
@@AndreasSpiess I ordered the POTY last week as I prefer your concept with a 80cm dish. And I'am shure you'll qrv before me as I have a lot of ham projects in parallel, even though I watched video # 306 a dozen times...
Andreas, you should have also included Mems oscillators. You can now get these from SiTime. They are available in TCXO's down to 0.5ppm, and now also have sub ppb accuracy OCXO's. These consume significantly less power and have much better over temperature stability, and have none of the Xtal defects such as activity dips. The are also MUCH smaller...if you want to know more you can connect with me in LinkedIn...otherwise take a look at SiTime.com
This is a "mainstream" channel and the art is to leave most things away ;-) Maybe in a few years when we can buy them on AliExpress like the MEMS microphones...
@@AndreasSpiess Andreas, already available in Ali Express, DiGiKey, Mouser, Arrow, etc.....and programmable to ANY frequency by user if required...They are now the worlds largest Oscillator (not Xtal) manufacturer, ship more that Epson, NDK...etc......obviously this is your channel, and you do a great job and this is why you have so many followers...thanks for everything..
Do you have an Aliexpress Link for a good MEMS oscillator? I plan to play around/compare with the different technologies and you seem to know more than me ;-)
I love the topic and your explanations. However, it wasn’t clear to me if using the PPS output of a suitable GPS receiver would produce precision pulses THAT ARE SYCHRONIZED IN PHASE WITH OTHER GPS RECEIVERS LOCATED ELSEWHERE.? Basically, if a person wanted to do Time Difference Of Arrival (TDOA) location of a transmitter, is it sufficient to use the GPS PPS output at each receiver site with confidence that they are synchronized with each other and with limited skew?
From Wikipedia: PPS signals are used for precise timekeeping and time measurement. One increasingly common use is in computer timekeeping, including NTP. Because GPS is considered a stratum-0 source, a common use for the PPS signal is to connect it to a PC using a low-latency, low-jitter wire connection and allow a program to synchronize to it. This makes the PC a stratum-1 time source. Note that because the PPS signal does not specify the time, but merely the start of a second, one must combine the PPS functionality with another time source that provides the full date and time in order to ascertain the time both accurately and precisely.
@@AndreasSpiess Thanks Andreas. Greetings from the USA. Experimentally, using my DSO, I proved to myself that the default 1pps output of two of my GPS modules is usually within 40 nanoseconds of each other. So they ARE "synchronous". I knew that the period was precise but I wasn't sure that I was guaranteed that they are aligned and it appears that they are. Now it is possible for me to receive radio transmissions at multiple different locations and try to determine the origin using TDOA techniques. I did not want to have to try correlate to a transmitter signal sent from a known location for the timing alignment. It seems much simpler to align all receivers with the GPS pps signal and then work from that!
That was very interesting. I was with you for quite a time (no pun intended) but soon after the Crystal Oven which was used in my formative years in Valve Oscilators for AM Ham Radio you went away from me. A very full episode, thanks.
@@michaels3003 Are you - sorry Mr Spiess - of the opinion that I do not? Generally I tag the episodes to their directory in my Email Client because this contribiter is of value to me and sometimes, not always, record as well. As for what this mysterious comment you have made refers to, I have no idea. Are you referring to the bit were I mentioned my early years with Amateur Radio or something else it is not clear. Smart (supposedly) comments are pointless. Sorry Mr Spiess, the poster irritated me. Reminds me of the daft reply to post of "rtfm" or in the case of early Linux read the Manual files. All of which would have been more understandable to a Human type Person if written in Hieroglyphics.
The clocks run slower on earth than in orbit, because the gravitational potential is larger. The clocks in the GPS satellite advances faster than a clock on the ground by about 38 microseconds per day. they would be 45 microseconds fast, but they move at 14,000 km/hr which leads to a time dilation.
11:51 - "Clumsy and expensive" :) I would definitely knock that all over onto the floor, probably with the back of my chair or a coat sleeve! And then what would we use to recalibrate it?! :)
I see we are currently running on the same clock ;-) Iam am also just looking into the GPSDO stuff for an QO-100 SDR setup. I think I will go with the Leo Bodnar Mini @25MHz for the beginning. Have you seen the Lars GPSDO thread on the EEVBlog Forum? There is a ton of information on a (relatively) simple GPSDO Setup with an NEO-xx Module an OCXO and a small Arduino. Took me quite some time to find it, but looks as the best DIY Setup so far. So it should be possible to get a relatively cheap GPSDO for the QO-100 setup. Besides have you seen the setup where someone reprogrammed the Pluto to run on 25 MHz reference clock speed? So you just need one GPSDO! ;-) (It's just a software modification, no deviders, nothing).
This is one of the cleanest QO-100 setups I found so far (look at the small & clean POTY setup, great!): forum.amsat-dl.org/index.php?thread/2936-pluto-box-solution-w-full-gpsdo-lock/ The Transmitter needs to be replaced, I will go with an SG Labs 20W @2400MHz PA and an github.com/db4um/PA_13cm_2W preamplifier. It took me quite some time and tons of reading threads the last weeks to find the best components, hope it helps.
@@AndreasSpiess Yes, a waterproof outdoor case with all the components included would be the best way I think. And then with a LAN Cable to the Homenetwork. So no losses on the antenna cables. I am just not sure if the temperature changes outdoor wouldn't have a too high impact on the system!? BTW I also stumbled over a nice Bias Tee PCB which injects the LNB DC voltage PLUS the 25 MHz from the GPSDO into the RX Cable. There is also a 50 Ohm to 75 Ohm impedance adaption included. (If you are interested drop me a line than I can look it up in my notes.)
Thank you… I bought a GPS-DO after watching this video a few years ago. I just added a new spectrum analyzer to my bench and connected it to my highly precise timing source. I just wanted to thank you for a great channel and this video in particular. 73 from the USA…. aka King Fauci’s Playground of Evil. 🫣😆
@Andreas Spiess, I'm also trying to get external reference to my Pluto but I'm not sure how to wire the UF.L socket in place of the TCXO.. I found some schematics online but I do not understand them nor I can trust the source 😂 so I would be glad if you could explain how you did it, or how it works! :) Thank you!
Just look at the diagram. There you see the capacitor between the crystal and the chip. The pin of the U.FL has to be connected to this capacitor. And the ground to ground. the connector ends up with about 45 degrees turned. According to Analog Devices, the external reference should be 1.3V p2p max and should be applied before boot
@@AndreasSpiess Thanks for your help! I think I'm pointed in the right direction now! I share a PDF with some info here. If you want to check it I would be glad! Feel free to share it as well drive.google.com/file/d/1SDSqL_ZiND_g2IgaK3W1PT4gsxZjgo61/view?usp=sharing
Dear Andreas, the UBLOX use 24Mhz for the internal crystal so if we want 10Mhz we got a lot of jitter. your oscilloscope screen capture does not shows that because it in single capture mode. ** Do you think we can replace the internal UBLOX crystal for 20Mhz so divide by 2 will give us a jitter free 10Mhz? **
@@AndreasSpiess Dear Andreas i replaced the neo7 onboard TCXO with external osc and did got a visual jitter free 10mhz. you can watch this video and read my comment. from what i read the pps is already disciplined but cannot output 10mhz because the neo7 has a fractional n synthesizer that cant divide to exact 10mhz. however it does 8mhz,12mhz jitter free. the internal tcxo is 26mhz . you can watch this video and read my comment. th-cam.com/video/M6SmdqMA-Dg/w-d-xo.html&lc=UgxHxrVXMsb_1tVLoP14AaABAg.9E6zqlDAGUC9E7I3ry0Mr3
Hello Sir, that means I can program a U-blox chip (can this be a cheap clone version?) to output 27.12 MHz that can be the clock input for a cheap LNB? Thank you
How often does the cheap GPS receiver module update/correct it’s output frequency from satellite data? I’d like to see a comparison to the OCXO GPSDO at varying states of signal availability.
I am a noob hobbyist, not a radio-head, so this sort of precision is over and above. Last year though, while debugging a digital FM tuner (RDA5807M) project, I discovered the frequencies of my watch crystals (2^7hz) differed wildly from each other. Sometimes they were off as much as +/- 30hz, throwing off my tuning. Still do not know any simple way to calibrate these (sigh). Cannot see the end of that rabbit hole. Andreas: Do you know the secret to delivering a good joke? timing
True (the timing of jokes). Concerning watches: You are right. Their timing has to be quite precise. I think a GPS module programmed at the crystal frequency and hooked to an oscilloscope would be en interesting experiment...
How accurate is the ublox 1pps compared to your other GPSDO's? The "timenuts" mailing list (posted online and easily found by searching), has a bunch of information about why the ublox 1pps signal doesn't work as a GPSDO. Basically, the module's internal frequency is not divisible to an integer of 1. The modules are not made for this purpose and this results in a lot of jitter. My grasp of the subject is weak, but from what I've read, the 1pps signal quality is not much better than the module without any satellite signal at all. I would love for someone to really break this down and prove whether this is true or not. I know of a lot of people that have tried to build a ublox based GPSDO, but no one that actually uses the device in practice afterwards. I get the impression they all assume they are the exception and must have made a mistake. I don't think anyone has succeeded though. There are enough examples of people attempting to build this project that more and more people keep trying. Several people that have built one of these ublox GPSDO's, later acquire units like you had here. Thanks for the upload. -Jake
The u-blox really has a horrible jitter. But it is precise in the long run. So it depends what you want. For a stable oscillator you need an OCXO to discipline. If you just want to build a frequency meter I assume it would work for most cases.
Have you considered using the PPS (Pulse per second) output on the $12 GT-U7 Gps receiver boards? Its a tiny board and works really well as a GPS with Arduino or equivalent boards. Here is the Amazon link to the product: TeOhk GT-U7 GPS Receiver Module Navigation Satellite Microcontroller GPS Compatible NEO-6M 51 Microcontroller STM32 with 1x Active Antenna www.amazon.com/dp/B07RPVYSG5/ref=cm_sw_r_cp_tai_Yba.EbN4B5YFQ
All GPSDOs use such modules to "discipline" the crystal. But their short-term stability is not great. Maybe I will make a video about that in the future.
Hello, what is the maximum voltage (or amplitude) of a signal that the BGT7TBL FA-3 (or FA-2) frequency counter can handle? I would appreciate your support a lot!
the uBlox modules intrigue me. About them I have the following questions: if you set a frequency, lets say 12.000.000 Hz, will that frequency also be generated after a power-down/power-up cycle? What accuracy can I expect if the module is free running?
For me it was impossible to get a clean 10MHz signal from an Ublox-8. 12MHz is clean, 24MHz is clean as well, but 10MHz is not. When I look to your oszilloscope screen shot at 18:15 I beleave, you have the same problem. The frequency 10MHz is displayed, but the broadness of the individual peaks changes from peak to peak. So it is 10MHz on average, but it is not a 100ns period. Looks like 24MHz (or a multiple of this) is kind of a sampling rate for the time pulse generation. For lower frequencies, this sampling doesn't matter. But anyway, thanks for your perfect tutorials. 73, DC0UB
That is why we cannot use a cheap GPS receiver for radio applications and must build such a GPSDO where the frequency is produced by a proper oscillator and is only "disciplined" by the GPS.
Scullcom Hobby Electronics channel has made a very nice video series how to build such a frequency standard for cheap at home. I really recommend if you don't want to dish our hundreds of dollars but need a "bit" more than an Arduino can provide.
is it critical for measuring instruments or radios to have a square or sine wave ? and if I understand the neo-6 can have it's 1PPS pin reconfigure to become a 10MHz reference clock ??
@@AndreasSpiess thanks for the reply, it's for a ham radio transceiver that has a 10MHz ref input that required -10dBm (approx 100Hz of drift can be observed with the stock TCXO on 1.2GHz), I thought your u-blox M series gps suggestion was a great idea on top of being cheap !! But since I cannot be sure of the "quality" of the signal and if jitter can cause issue out of the PPS pin to my transceiver, I also cannot measure how much dBm is getting out of this little board as I do not have the proper equipement to measure RF power...
On the GPS unit. Does anyone know what the blinking green LED on the bottom of the board rear between the BLOX module and the board corner labeled D2 means? Mine seems to blink at a 1hz rate, but it may mean something else?
I went through all your QO-100 videos again with more intent on learning some specifics. I am currently planning to replace the crystal LO in my Ukrainian transverter with a GPSDO. I have come across an interesting project that is much more than just a GPSDO but that can do that to. Have you heard of RF Zero? It's a Danish ham's creation. Very interesting. 73 DE KJ7LVZ
@Ash If my memory is right, the crystal has two small capacitors, one either side of it. Normally one of them is variable and can be used to trim the frequency when the temperature is stable.
Many OCXOs have a "VFC" pin that allows their frequency to be calibrated by varying the voltage on the pin. That is how the oscillator in a GPSDO is disciplined.
@Esteban: As a consumer it is easy to ask for everything free. It looks a little different if I have to feed my family... And the most important question: What do you contribute?
@@AndreasSpiess Everybody does need to eat and have basic needs. Every bit of knowledge we share for free allows people in a worst situation than us to get in a better situation. Imagine if you as a kid would have access to your videos as an adult. I firmly believe people will contribute when they are able. Those that take advantage and don't contribute are a minority. I'm extremely grateful for your videos being accessible for free and I try to use your links when buying, if in the future I can do more I will. I'm a lucky one from my country (Argentina) and I managed to learn English, electronics and informatics mostly thanks to people like you that share your knowledge on the internet, otherwise I'd probably be flipping burgers.
since f = 1/t then we can say most universal constant yield f, as many are affiliate with time. example: heat 1ml of water up 1 celcius with 4.184 watt (all values are exact). You got 1 Hertz (exact).
Great your TH-cam show! But I would prefer it very much if your videos would continue to be numbered. It's much easier to reference them and much more practical to say to a colleague: "Have you seen video ### by Andreas (the guy with the Swiss accent) yet?". Or sometimes I just write down: "See Andreas Swiss ##" . :-) Many thanks for your series!
@@AndreasSpiess I have built several GPSDOs using surplus 10 MHz OCXOs, low cost GPS modules, and Arduino processors. The project is based on a design originally posted on EEVBlog by Lars Walenius. This has been a very interesting project to update my 40-year old electronics and programming skills. I would be happy to share the details if you are interested.
Andreas, I think you'll love this article about magnetic amplifiers that could be a good source for another video. www.nutsvolts.com/magazine/article/the_magnetic_amplifier
Thank you for the link. I did not know about it but it seems more focused on motor applications and maybe is also outdated because of the cheap silicon parts today.
Great video, as always. Do you now about a basic 10 MHz distribution box, just 10 MHz in and six or similar 10 MHz out? I don't believe that a OCXO is needed for this, maybe to continue if the 10 MHz ref is down.
We always use our clocks continuously here in Switzerland ;-)
4 ปีที่แล้ว
Hello Andreas. Nice video. Your OCXO is precise, but not accurate. I invite you to read further on the difference between accuracy and precision. When it comes to time and frequency measurement, that distinction matters a lot!: en.wikipedia.org/wiki/Accuracy_and_precision Thanks.
The first sentence in the video immediately had me shouting out: that is the most wrong thing I heard all day and in so many different forms of interpretation. In a physical sense time is relative. And in a standards sense there are so many standards: leap seconds, time zones, ...
WOW, leave it to Andreas, he will tell you EVERYTHING from the least expensive version to affordable price - doable price! I ran the video 2x, since I missed stuff the first time. I might listen again Wednesday! Thanks Andreas, you are the BEST!
This is definitively not a simple topic. I plan to go a little deeper in a future video...
@@AndreasSpiess Can't wait! But don't rush it, I have to get some of my brain back - if I can! ha ha ha
Gruss dich Andreas. When I was young we used a cheap FM stereo receiver. We extracted the 38/19kHz stereo pilot-tone out of a national radiostation. These where about the most stable frequencies you could get in the 80s. Then put it into a PLL to get the frequency we wanted. This video brings back good memories. Thanks. I follow you every now and then. But you got a new subscriber today.
Welcome aboard the channel! And here you see how things changed since the '80s...
This channel and project just keep getting better... love that Swiss clock humor!
Thank you very much!
Like you, I am a retired RF engineer. About a year ago I built my own GPSDO. Comment: the 10MHz output of a U-blox receiver has a lot of jitter, it is coming from the 26MHz osc. I had the U-blox put out 1MHz and then divided the 10MHz OCXO by 10 and used a phase and phase frequency detector and used an ESP-32 as the controller and web browser system monitor. 1 MHz provided gave sufficient phase range to account for jitter from the GPS and keep the initial lock time down to a few minutes. In software I built a variable speed loop... very fun project. You can get down to about 5.0 10e-11
Unfortunately, I am just an EE engineer with love to RF, but never sufficient know-how .-) And I have to wait a few years for the pension. Your project is very interesting! I saw the horrible jitter, too. And I have a few "ovens" in the mail. So it might be that I also will try to create such a GPSDO. And I still have ambitions to understand and measure the Allen deviation and compare the different devices.
@@AndreasSpiess I should also warn, that I have been told that some of the U-Blox Neo modules are rejects that are sold under another name, then a clone U-Blox sticker is applied, and sold in China. While I don't know if this is true, I can say that the jitter can be >10X worse. If the price is too good to be true.... beware... or buy direct from U-Blox distributors.
They are Swiss company and we wanted to do a video about clones. Unfortunately they never came back... On the other hand I canot believe that there are so many rejects in a mordern manufacturing process.
@@AndreasSpiess I should not have used the word 'rejects'.... U-Blox sells many different grades/variations of their receivers. You can query the part, and it will tell you exactly what version it is. Someone has removed the sticker on the part, and replaced it with another sticker, which represents it as a more expensive, higher specification part. I have 'altered' parts, but electronic signature reports the truth. So, they are not clone silicon, they are parts where they received a 'label' upgrade. :->
I am curious, how did you measure the resulting accuracy of the clock? Did you compare it to Rb or Cs standards?
I was aware of TCXO/OCXOs, was not aware of the GPSDO/GPS TCXOs. Learned something new today. Great project. Keep up the good work.
Thanks, will do!
Enjoyed watching, thankyou! For sake of interested folks also reading the comment: Only AU cut crystal can go into making of ovenised oscillators as the mean of S curve (drift vs temp curve) pivots at 70 degrees, while for Regular Crystal (AT cut) the mean pivots at 25 degrees. I hope I have saved a curious mind trying to put a regular crystal into an electronic oven hoping to achieve better accuracy.
Thank you for the additional information. I did not know that fact!
@@AndreasSpiess, also, above a particular temperature a crystal will drift in frequency in one direction, while below that particular temperature its frequency will drift in the other direction, (the S curve.) But there is a plateau temperature where a crystal is relatively immune to frequency drift with time. Which depending on the Xtal cut is either at roughly 70C or 25C, but not precisely at those temperatures. I assume the most precise OCXO's each have their exact oven temperature hand tuned to put that crystal on that temperature plateau, but I do not know this for certain, as Xtal-standard companies are fairly quiet about proprietary techniques for their manufacturing precision, stability, and manufacturing cost control. But that was how I set up my home-brew 1 MHz standard that I disciplined with a TRF WWVB receiver at 60 kHz before the GPS constellation was launched into orbit. It was one of the most interesting projects I ever built.
'10Hz without callibration over a few thousand years, good enough for the girls I go out with' AVE would be very proud ;)
So we all know the source channel ;-)
@@AndreasSpiess (gasp) Am stunned that this gentleman with a Swiss accent would stoop so low as to watch that foul-mouthed, talking-hands, Canadee-anne destroy perfektly good, Chinesium tools. Howdaya understandem?
AvE: "... you know they say, I am not a gynecologist, but I'll take a look."
Nice to see an Uncle Bumblefuck reference, Andreas !!
I once worked for Hilti. Is this sufficient ;-)
Excellent video! I'm really enjoying this satellite series. Looking forward to the next one.
Glad you enjoy it!
7:03 - My ambition was to be filthy, stinking rich; I failed, but two out of three is not too bad.
You are right! 66% is already a good result ;-)
@MichaelKingsfordGray I bet you’re fun at parties. It was a joke.
@@joshuarosen6242 With a name like MichaelKingsfordGray .I think he was joking.
Awesome video Andreas.
Firstly, it is highly informative and useful.
Secondly, I could relate to it as it reminded me of a project I did years ago where I had to synchronize a sheet coater with a oven.
Although the motors were digitally geared together with incremental encoders, it relied on phase synchronization to keep things in time.
The lacquered sheets had to be fed into the holders of the oven at the right position as they came past on a chain.
These sheets are what tin cans in various forms and also beer bottle crown tops are made from.
Pulses from both machines were used to create sawtooth wave forms and then an error signal was generated by comparing amplitude.
A proportional correction was then calculated and added to the digital gearing ratio.
The oven being the master generated a signal with a 50% mark space ratio.
It was sufficient to only apply correction whilst the signal was high, and then freeze it whilst it was low.
So as you say, using a time standard to discipline an oscillator is highly effective, reliable and simple to implement.
I love these types of discussions, right up my alley.
Thanks for sharing.
And I am glad to see your hand seems to be getting better.
Best Regards,
Duncan
South Africa
I also love feedback loops. They are one of the most important inventions of humanity!
Another "Master class" Andreas, thanks. The GPS part is so interesting and affordable for everyone doing any kind of timing... I don't know if it's enough for a separate video...
[ES] Otra "Clase magistral" Andreas, gracias. La parte del GPS es tan interesante y asequible para todos los que hacemos cualquier tipo de sincronización... No sé si será suficiente para un video dedicado...
BTW: How is the quality of the Spanish subtitles from your point of view?
@@AndreasSpiess Let me take a look, I never use them...
@@AndreasSpiess 9.7 over 10. There is noticeable that there is a speaker from a sudamerican country. We, spanish speakers, use the Spanish language in a slightly different way at both sides of the Atlantic as British or Americans do with the English.
But those are not errors!
The subtitles are 100% comprensibles and only a few words can be more precises from a technical point of view. For the purpose of divulgation is a really good job.
Yours faithfully ;-)
Thank you for your feedback. I shared it with Luciana the translator.
@@AndreasSpiess You are welcome!
Very useful overview. I'll keep it bookmarked for when colleagues ask me about the subject.
Glad it was helpful! And it is an honor to become bookmarked...
-Time is probably the only global standard
Americans :
- Yes , yes kiloyards per gallon squared.
American here. We would never use a commie (that's a joke) prefix like kilo. We would just say "thousand".
Hey!
You're not free of our colonial caveman units yet. Sure, y'all dislike our foot length fetish, but the clock still has our hands :-)
If anyone remebers Digital Equipment computers, some of thier operating systems had parameters in micro-fortnights.
You forgot to divide by 8.
@David Parrish - I did my degree project on a PDP8. Never got the timing right...
I appreciate the fantastic job you're doing to investigate and bring attention to these fascinating parts and projects. Way to go!
Thank you for your nice words!
Excellent discussion of a topic, GPS timing, that I had been curious about but knew very little. Thank you for you diligent work.
Glad it was helpful!
I love this topic. I am obsessed with time and have a Pi running a NTP service in my house based on a GPS antenna. Like you did here, I have also used the signal for my own purposes as well.
Do you have a link to the project of your NTP server? Maybe I will build also one...
Andreas Spiess I will link the guide here. I followed the old version of the post and it looks like they refreshed it in 2018, which is good to see updates. It uses the exact hardware I have installed. I have been running this for three years now and it is great. The main idea is that you still connect to an external time server (Stratum-0) to get the base time, but it uses the local PPS signal from GPS in order to maintain that to perfect accuracy. It has been wonderful to have this in my house and be able to point all devices at a local NTP server. ava.upuaut.net/?p=951
Thank you for the link!
I used to live in Switzerland for graduate school, and when I was there I and a few other "time nuts" (there's a mailing list by that name for enthusiasts of precision time and frequency) did a tour of the time lab of METAS, the Swiss national standards lab. The guys in the time lab there were a bit puzzled why we wanted to see their time lab (they don't get many visitors) but they showed us around and we chatted for a while. Nice people. They have a nice set of HP 5071A cesium reference standards, a research-grade cesium fountain clock, and a hydrogen maser that they use to define UTC(CH), and help define UTC in conjunction with other national time labs. They said they used to get a lot of financial support from Swiss watch companies until the companies realized that people aren't buying their watches for keeping excellent time but rather as a fashion item, so eventually the financial support stopped. Now they do a lot of calibration of commercial and industrial time standards, but even that is not as popular as it was because anyone can get intrinsically calibrated GPS-based time and frequency standards and let the USNO deal with all the hard work.
I visited the time lab at the University of Neuchatel. They did work for the Galileo program. But I do not know how they are financed. At least for me it was very interesting (I was allowed to compare my GPSDO with their maser...
This is a very good video. Thumbs up. And of course Swiss people are renown for timing accuracy, so this is consistent with getting the information from a guy with a Swiss accent. On the downside, Swiss people are not known for having lots of sex. For information on that aspect, we will be waiting for a video by a gal with an Italian accent! ;-)
As an Engineer, I must ask you for the source, of course. Otherwise it is only your statement ;-)
Wonderful explanation of T&T basics! Thanks very much.
You're very welcome!
I am so happy to have acquired the same IKEA clock as you, thanks to your video. So far I have not hacked it to diplay another time. The main issue is that it only synchronise once every hour or so. You cannot get it to move the way you want, or you have to cut the power to make it forget, then be the strongest signal to bring it to the time you want.
Did you watch my DCF77 Transmitter video?
@@AndreasSpiess Yes, I have it because of your Videos. Never invested in the coil to control the clock.
Great video. A great refresh of things of the past and also an update on new things. Well done.
Glad you enjoyed it!
Hello Guy with swiss accent, very impressive from Guy with Bangladeshi Accent.
Hi, also from Bangladesh.
Thank you! Never been in your country :-(
Does the Bangladeshi accent differ from that Bengali accent?
@@GnuReligion Those who are not Bengali, their accent probably differs.
Dear Andress, Thanks for your kind reply. You are welcome to my Country. I always share your video in our "Robo Course" App. You may download from Play Store.
Really great episode on precision timing. I'll definitely use your investigation and findings to setup my own precision timing at home for my lab.
Enjoy!
So it's true that swiss people love precise timing
yes we do!
Andreas, Thanks for this video. For our microwave projects (Pacific Northwest VHF Society) we use the Leo Bodnar GSPDO with a 5-pole low-pass filter to remove the harmonics. We're on 10, 24, 76 and 122GHz DE WA4OSH
I also added a 25MHz filter to get a Sinus. BTW: Is 122Ghz still VHF? I thought it is at least UHF ;-)
EHF :D
Take a look at this eval kit:
siliconradar.com/evalkits/#sirad-simple
www.bay-net.org/uploads/1/2/2/7/122774721/k6ml-122ghzradio-baycon2018.pdf
ghz-europe.com/vk2xax-more-tests-of-122ghz-transverter/
It needs a 10MHz reference
Thank you for your links. This is mindboggling!
@@AndreasSpiess The resolution of the FMCW radar is much higher than the PIR sensors that you tested in April.
Also, check out the NXP CompactRadar
www.nxp.com/products/rf/radar-transceivers:MITERWAVEICS#/
I'm trying to get an evaluation kit. Home security that does not get triggered by cats, dogs, birds, deer, etc. but properly identifying humans and cars is my goal.
Maybe we will also get cheap stuff from the cars like we get from GSM stations...
very good summary of a very technical field! Look for "Time Nuts" -- the news group and web site, to find people that spend a great deal of effort on these things, chasing picoseconds... Old HP OXCOs, like the 18011 dual oven 10 MHz oscillator are still superb performers. eBay sourced rubidiums are also good, but the close-in phase noise of rubidiums is far worse than a crystal. With an inexpensive GPS receiver, you can have very good and very stable time and frequency references for your home lab. (If you know what "Allan deviation" is, you're probably a time nut.)
Not yet, unfortunately. I still have to learn a lot!
So geil. I was wondering how to get a better reference than tcxo. Thank you so much Andreas.
You are welcome!
Very good, I liked the demo of outputting frequency from the GPS receiver :-)
I thought this is the easiest way to show it.
Great walkthrough as always 👍
Thanks for sharing 👍😀
My pleasure!
Thanks, I learned quite a lot on OCXO's I never knew. Also surprised on the amount of phase drift your OCXO had.
This is an uncommon part and not well known…
Great topic. excellent, made me get better freq reference . Many thanks Simon
Very welcome
Just another GREAT VIDEO Andreas!!! Thank you for sharing. Have a great week and let us know how is your finger healing.
Thanks! The doctor was happy on Friday. But it will take some time till I can drive my Harley...
@@AndreasSpiess Yes, but you can use it with electronics and stay away from the dangerouse hobbies! ha ha ha Glad to hear that your doctor gave you a great news!!! We are all happy for you! Stay safe!
If Ublox modules are with programmable GPSDO - why to buy those expensive boxes with OCXOs and Altera? Small cheap Ublox GPS modules will be still more stable and precise. Addition with AD9834 and some frequency multiplier for it (Triple 25MHz made with UBlox) Should be enough to make any GPSDO based frequency... Very nice video about oscilators!!! I am building now my labs Frequency generator with 2x AD9834 (Sometimes need phase shifted signals) and 2x AD9833 (For square wave signals and PWM modulated signals). Now I am thinking about putting in this box UBlox and GPS antenna to get GPSDO based Function generator. Thank You Andreas for great explaining!
They are good in the longer term. But from what I saw have a horrible phase noise. But I have to do more investigations
@@AndreasSpiess According to Your video at 18:18 its clearly visible that there is really bad 50/50 ratio and phase is completly out of range. Lets try to set 25MHz and repeat the same test. My plan is to use 25MHz setting and after it I would like to put Band filter to 25MHz (That should shift phase but also stabilise it). Another approach can be like You tell about PLL with Low Pass filter - 25MHz with not phase locked loop, but frequency locked loop where error is compensated with some very low frequency filter (VCO). Anyway - Your video was for me very educational and explains me a lot of unknown before things!
It depends what you want to get out of it. For sure it will have much more phase noise than for example a OCXO. There is anothe comment conccerning this topic, too.
The GPS receiver has a jitter of up to 1us. So, actual accuracy "out of the Blox" is only about 1ppm. Only after averaging out on a very stable oscillator, like OCXO for about 3 hours (1E+4 sec) one can get the actual absolute frequency below the 0.01ppm (1ppm / sqrt(10000))
Brilliant video! Loving this content!
Thanks so much!
Excellent overview and presentation! Thank you sir!
Neat seeing the Adalm Pluto mod you made also. Didn't know about the aliexpress version. Amazing! Amazing what the latest and greatest NEO-#'s are capable of also. Especially if using for tracking with the RTK though that is a completely different topic/subject. I still think there is a way to hack out the modules even with the Digital Block to make a really low noise communication device receiver. Would be awesome to be able to hack into just the RF Block since seems like would be an awesome receiver/detector for a receiver or spectrum analyzer or I guess VNA too.
I do not know what is inside a GPS receiver. But I assume it is pretty specialized.
@@AndreasSpiess Yeah, was thinking at the least can seal up the GPS module in a shielded case and use the antenna input like for the output of a mixer with a fixed frequency design for either sweeping like for analysis or for just whatever transmitter tuned as a for use as a radio receiver. Though since stuck with the digital block format... would have to design the transmission protocol, if using as a sensitive receiver, to encrypt to the Digital Block format so can be decrypted/translated from the Digital Block format using the transmission protocol format. One of those different ideas I had one day when drooling over the noise floor specs of the latest uBlox modules. Albeit the baud rate will be low... though is improving with each new version.
Thank you for the video, I think this is very educative. As always I appreciated the tone and jokes !
I work in a atomic frequency standard lab in Switzerland. I could measure the stability of your GPSDO using one of our hydrogen masers. That way you would exactly know how stable the unit is. What do you think ?
Great idea! Can you contact me? You find me on Twitter, Facebook, or in search.ch
@@AndreasSpiess I know it's been quite some time, but do you have an update on this? Ever got to check the GPSDO precision and stability with @Etienne Batori ?
@@Halabaloosa Yes, was at the University of Neuchatel, and Etienne tested my devices. They were very stable as long as the GPS reception was good. So the antenna position is more important than the device.
Can you please make a video on how to choose the right microcontroller, I recently got an MSP432 for my college project and finding it hard to work with
@Scott Page USMC Thank you for your interest, My problem is mostly with using code composer studio, I don't see many examples on how to program MSP432 whereas STM32 has got lot resources on web so was wondering if I have switch to that. I am very familiar with using arduino boards, but I was told not to use that.
I use Arduinos, ESPs, and Raspberries. Mostly because they have a huge community support.
@Scott Page USMC Thank you for your interest sir, TI is sponsoring the kits for us thus we are supposed to use them, I have searched some books and reading those.
Very fascinating; thanks.
Glad you enjoyed it
Using the GPS clock signal is easier than I thought! (I never know that PPM is also adjustable too)
How do you wire the GPS signal to the basement? Do you have to put the module outside and wire the pulse signal into the room? Or broadcast the clock signal as you do with your Ikea clock video?
GPS antennas are "active". They have an amplifier built in. So you can have quite a long cable from the antenna to the basement. But for sure the antenna must "see" the satellites.
Thank you for the answer. 😃
Fascinating !
You are right!
Hi. Great explanation in a short time span! Short question: in the video at about 18':15" we see the square wave of the cheap GPS receiver. Would you know why the square wave has/looks irregular width, with 2 wider and 3 shorter positive period square? If this works that would be a very cheap solution to a very high accuracy clock!
I do not know the reason for the irregularities :-(
@Andreas Spiess I think I know now the Why part. Just watched an interesting video. It seems that the (programmable) output frequency of the gps generator is provided by dividing the main clock of 24MHz. If you choose a frequency that has fraction, then you get a weird mix. For example: 10MHz would be factor 2.4. It seems you get a mix of 12 and 8MHz. More info on: th-cam.com/video/M6SmdqMA-Dg/w-d-xo.html
11:35 how do you determine the accuracy of the most accurate clock ever built? Just wait 30 million years? ;-)
What's the reference there to determine the drift?
I assume there is a reference. If you are interested in the details, you can go to the "timenuts" page (leapsecond.com)
@@AndreasSpiess Cool, thanks for the link! If you're interested, I can post the answer to my question, if they can help me with that.
@@stefanmayer444 I decided not to care for the details. Because everybody involved told me that it is a dangerous rathole ;-)
@@AndreasSpiess Haha, you started that topic in the first place ;-)
We install synchronization equipment that uses Cesium clocks as backup if the GPS receiver is down or jammed
So it must be an important application ;-)
@@AndreasSpiess Stuff that needs to run properly even if the sh!t hits the planetary fan
Have you seen the SiTime MEMS TCXOs? Seems like they'd be a good combo with GPS/atomic sources, they outperform many OCXO and are small SMD components..
No, I did not see them. They seem to be interesting parts. And, as you write, very small and energy efficient compared to OXCOs. And maybe they are good enough for many applications and can replace a GPSDO.
However, precision is not as important for a GPSDO. Stability is more important because the absolute frequency comes from the GPS signal.
Great video....thank you
Glad you enjoyed it
Very useful! Thanks!
You're welcome!
Another excellent video. Great comparison of strengths and weaknesses. We can now meet on satellite... ;-)) Thanks a lot
Unfortunately not yet. I am not ready. What kind of antenna do you use?
@@AndreasSpiess I ordered the POTY last week as I prefer your concept with a 80cm dish. And I'am shure you'll qrv before me as I have a lot of ham projects in parallel, even though I watched video # 306 a dozen times...
I also have a second HAM project. I got a new keyer ( www.i2rtf.com/signature.html ) from my wife and I started to recover my rusty CW..
How was this comment made 17 hours ago when youtube tells me this video went up 3 min ago...
@Luca Ruggier Not much discipline in their timing..
This video has such good humor
Thank you!
Andreas, you should have also included Mems oscillators. You can now get these from SiTime. They are available in TCXO's down to 0.5ppm, and now also have sub ppb accuracy OCXO's. These consume significantly less power and have much better over temperature stability, and have none of the Xtal defects such as activity dips. The are also MUCH smaller...if you want to know more you can connect with me in LinkedIn...otherwise take a look at SiTime.com
This is a "mainstream" channel and the art is to leave most things away ;-) Maybe in a few years when we can buy them on AliExpress like the MEMS microphones...
@@AndreasSpiess Andreas, already available in Ali Express, DiGiKey, Mouser, Arrow, etc.....and programmable to ANY frequency by user if required...They are now the worlds largest Oscillator (not Xtal) manufacturer, ship more that Epson, NDK...etc......obviously this is your channel, and you do a great job and this is why you have so many followers...thanks for everything..
Do you have an Aliexpress Link for a good MEMS oscillator? I plan to play around/compare with the different technologies and you seem to know more than me ;-)
I love the topic and your explanations. However, it wasn’t clear to me if using the PPS output of a suitable GPS receiver would produce precision pulses THAT ARE SYCHRONIZED IN PHASE WITH OTHER GPS RECEIVERS LOCATED ELSEWHERE.? Basically, if a person wanted to do Time Difference Of Arrival (TDOA) location of a transmitter, is it sufficient to use the GPS PPS output at each receiver site with confidence that they are synchronized with each other and with limited skew?
From Wikipedia: PPS signals are used for precise timekeeping and time measurement. One increasingly common use is in computer timekeeping, including NTP. Because GPS is considered a stratum-0 source, a common use for the PPS signal is to connect it to a PC using a low-latency, low-jitter wire connection and allow a program to synchronize to it. This makes the PC a stratum-1 time source. Note that because the PPS signal does not specify the time, but merely the start of a second, one must combine the PPS functionality with another time source that provides the full date and time in order to ascertain the time both accurately and precisely.
@@AndreasSpiess Thanks Andreas. Greetings from the USA. Experimentally, using my DSO, I proved to myself that the default 1pps output of two of my GPS modules is usually within 40 nanoseconds of each other. So they ARE "synchronous". I knew that the period was precise but I wasn't sure that I was guaranteed that they are aligned and it appears that they are. Now it is possible for me to receive radio transmissions at multiple different locations and try to determine the origin using TDOA techniques. I did not want to have to try correlate to a transmitter signal sent from a known location for the timing alignment. It seems much simpler to align all receivers with the GPS pps signal and then work from that!
@@harrymagooslum5770 Good to know. Thanks! I hope the same will apply when your receivers get signals from very different satellites.
Great stuff.
Thanks!
It's that time of the week again.
Glad you like it!
@@AndreasSpiess Definitely, thank you.
That was very interesting. I was with you for quite a time (no pun intended) but soon after the Crystal Oven which was used in my formative years in Valve Oscilators for AM Ham Radio you went away from me. A very full episode, thanks.
Just watch it more than once.
@@michaels3003 Are you - sorry Mr Spiess - of the opinion that I do not? Generally I tag the episodes to their directory in my Email Client because this contribiter is of value to me and sometimes, not always, record as well. As for what this mysterious comment you have made refers to, I have no idea. Are you referring to the bit were I mentioned my early years with Amateur Radio or something else it is not clear.
Smart (supposedly) comments are pointless. Sorry Mr Spiess, the poster irritated me. Reminds me of the daft reply to post of "rtfm" or in the case of early Linux read the Manual files. All of which would have been more understandable to a Human type Person if written in Hieroglyphics.
@James: Glad you enjoyed it!
Interesting fact: Due to relativity, the precise clocks in satellites run a little slower when in orbit than on Earth.
The clocks run slower on earth than in orbit, because the gravitational potential is larger. The clocks in the GPS satellite advances faster than a clock on the ground by about 38 microseconds per day. they would be 45 microseconds fast, but they move at 14,000 km/hr which leads to a time dilation.
SJ, Are you SURE?
Marcus Metzler, thank you.
@@drmocm Yeah. My bad.
11:51 - "Clumsy and expensive" :) I would definitely knock that all over onto the floor, probably with the back of my chair or a coat sleeve! And then what would we use to recalibrate it?! :)
I am not sure if you have to calibrated these clocks as they use a material constant. (At least that is what I understood.)
@@AndreasSpiess Very true. But don't underestimate my clumsiness to take out atomic physics, too! 😂 Cheers, M1GEO.
I see we are currently running on the same clock ;-) Iam am also just looking into the GPSDO stuff for an QO-100 SDR setup. I think I will go with the Leo Bodnar Mini @25MHz for the beginning. Have you seen the Lars GPSDO thread on the EEVBlog Forum? There is a ton of information on a (relatively) simple GPSDO Setup with an NEO-xx Module an OCXO and a small Arduino. Took me quite some time to find it, but looks as the best DIY Setup so far. So it should be possible to get a relatively cheap GPSDO for the QO-100 setup.
Besides have you seen the setup where someone reprogrammed the Pluto to run on 25 MHz reference clock speed? So you just need one GPSDO! ;-) (It's just a software modification, no deviders, nothing).
Lars GPSDO www.eevblog.com/forum/projects/lars-diy-gpsdo-with-arduino-and-1ns-resolution-tic/
One of my Pluto’s already runs on 25MHz;-) And maybe a OCXO alone would be enough for QO100. They seem to be quite precise.
This is one of the cleanest QO-100 setups I found so far (look at the small & clean POTY setup, great!):
forum.amsat-dl.org/index.php?thread/2936-pluto-box-solution-w-full-gpsdo-lock/
The Transmitter needs to be replaced, I will go with an SG Labs 20W @2400MHz PA and an github.com/db4um/PA_13cm_2W preamplifier.
It took me quite some time and tons of reading threads the last weeks to find the best components, hope it helps.
Thank you for the links. This is also my direction. Maybe waterproof.
@@AndreasSpiess Yes, a waterproof outdoor case with all the components included would be the best way I think. And then with a LAN Cable to the Homenetwork. So no losses on the antenna cables.
I am just not sure if the temperature changes outdoor wouldn't have a too high impact on the system!?
BTW I also stumbled over a nice Bias Tee PCB which injects the LNB DC voltage PLUS the 25 MHz from the GPSDO into the RX Cable.
There is also a 50 Ohm to 75 Ohm impedance adaption included.
(If you are interested drop me a line than I can look it up in my notes.)
Thank you…
I bought a GPS-DO after watching this video a few years ago.
I just added a new spectrum analyzer to my bench and connected it to my highly precise timing source.
I just wanted to thank you for a great channel and this video in particular.
73 from the USA…. aka King Fauci’s Playground of Evil. 🫣😆
You are welcome! 73 de HB9BLA
Great series, Thanks Andreas
73 de 7X2UV
You seem binge watching the series. Cool! 73 de HB9BLA
@@AndreasSpiess I appreciate the effort you're doing to simplify things, You're great (y)
Please continue
@Andreas Spiess, I'm also trying to get external reference to my Pluto but I'm not sure how to wire the UF.L socket in place of the TCXO.. I found some schematics online but I do not understand them nor I can trust the source 😂 so I would be glad if you could explain how you did it, or how it works! :) Thank you!
Just look at the diagram. There you see the capacitor between the crystal and the chip. The pin of the U.FL has to be connected to this capacitor. And the ground to ground. the connector ends up with about 45 degrees turned. According to Analog Devices, the external reference should be 1.3V p2p max and should be applied before boot
@@AndreasSpiess Thanks for your help! I think I'm pointed in the right direction now! I share a PDF with some info here. If you want to check it I would be glad! Feel free to share it as well
drive.google.com/file/d/1SDSqL_ZiND_g2IgaK3W1PT4gsxZjgo61/view?usp=sharing
Dear Andreas, the UBLOX use 24Mhz for the internal crystal so if we want 10Mhz we got a lot of jitter. your oscilloscope screen capture does not shows that because it in single capture mode.
** Do you think we can replace the internal UBLOX crystal for 20Mhz so divide by 2 will give us a jitter free 10Mhz? **
To get precise timing you have to use a OXCO or a TXCO and discipline it with a GPS. The GPS module always will have jitter.
@@AndreasSpiess Dear Andreas i replaced the neo7 onboard TCXO with external osc and did got a visual jitter free 10mhz. you can watch this video and read my comment. from what i read the pps is already disciplined but cannot output 10mhz because the neo7 has a fractional n synthesizer that cant divide to exact 10mhz. however it does 8mhz,12mhz jitter free. the internal tcxo is 26mhz . you can watch this video and read my comment. th-cam.com/video/M6SmdqMA-Dg/w-d-xo.html&lc=UgxHxrVXMsb_1tVLoP14AaABAg.9E6zqlDAGUC9E7I3ry0Mr3
Hello Sir, that means I can program a U-blox chip (can this be a cheap clone version?) to output 27.12 MHz that can be the clock input for a cheap LNB? Thank you
As I said: max 24 MHz
How often does the cheap GPS receiver module update/correct it’s output frequency from satellite data? I’d like to see a comparison to the OCXO GPSDO at varying states of signal availability.
I do not know. Maybe I will once compare the different sources I have, but I have to learn a lot before.
I am a noob hobbyist, not a radio-head, so this sort of precision is over and above. Last year though, while debugging a digital FM tuner (RDA5807M) project, I discovered the frequencies of my watch crystals (2^7hz) differed wildly from each other. Sometimes they were off as much as +/- 30hz, throwing off my tuning. Still do not know any simple way to calibrate these (sigh). Cannot see the end of that rabbit hole.
Andreas: Do you know the secret to delivering a good joke?
timing
True (the timing of jokes). Concerning watches: You are right. Their timing has to be quite precise. I think a GPS module programmed at the crystal frequency and hooked to an oscilloscope would be en interesting experiment...
The great thing about standards, is that there are so many different of them...
Not for time, AFAIK. Also the Americans use seconds. And this says a lot, I think.
How accurate is the ublox 1pps compared to your other GPSDO's? The "timenuts" mailing list (posted online and easily found by searching), has a bunch of information about why the ublox 1pps signal doesn't work as a GPSDO. Basically, the module's internal frequency is not divisible to an integer of 1. The modules are not made for this purpose and this results in a lot of jitter. My grasp of the subject is weak, but from what I've read, the 1pps signal quality is not much better than the module without any satellite signal at all. I would love for someone to really break this down and prove whether this is true or not. I know of a lot of people that have tried to build a ublox based GPSDO, but no one that actually uses the device in practice afterwards. I get the impression they all assume they are the exception and must have made a mistake. I don't think anyone has succeeded though. There are enough examples of people attempting to build this project that more and more people keep trying. Several people that have built one of these ublox GPSDO's, later acquire units like you had here.
Thanks for the upload.
-Jake
The u-blox really has a horrible jitter. But it is precise in the long run. So it depends what you want. For a stable oscillator you need an OCXO to discipline. If you just want to build a frequency meter I assume it would work for most cases.
very nice thankyou, ah yes phase lock loop PLL I remember the old cbs also lol
You are welcome!
Have you considered using the PPS (Pulse per second) output on the $12 GT-U7 Gps receiver boards? Its a tiny board and works really well as a GPS with Arduino or equivalent boards.
Here is the Amazon link to the product: TeOhk GT-U7 GPS Receiver Module Navigation Satellite Microcontroller GPS Compatible NEO-6M 51 Microcontroller STM32 with 1x Active Antenna www.amazon.com/dp/B07RPVYSG5/ref=cm_sw_r_cp_tai_Yba.EbN4B5YFQ
All GPSDOs use such modules to "discipline" the crystal. But their short-term stability is not great. Maybe I will make a video about that in the future.
Hello, what is the maximum voltage (or amplitude) of a signal that the BGT7TBL FA-3 (or FA-2) frequency counter can handle? I would appreciate your support a lot!
Its datasheet says: CH1 Input sensitivity：50 OHM: 25mVrms-2Vrms or 1MOHM: 25mVrms-10Vrms
the uBlox modules intrigue me. About them I have the following questions: if you set a frequency, lets say 12.000.000 Hz, will that frequency also be generated after a power-down/power-up cycle? What accuracy can I expect if the module is free running?
I mentioned in the video: You have two frequencies to select. One comes up without GPS fix, the other with GPS fix.
So many types of oscillators now a a days!
XO
TCXO
OCXO
VCXO
SAW
MEMS (SiTime)
Etc etc
I only wanted to cover the "usual suspects". It is already a heavy stuff.
4:47 👈 Hand-ception!
:-))
Next Video: Building a atomic clock cause im swiss and a shift of 2 seconds in 10 billion years is not acceptable 🇨🇭 😂
Exactly! And I learned a lot...
For me it was impossible to get a clean 10MHz signal from an Ublox-8. 12MHz is clean, 24MHz is clean as well, but 10MHz is not. When I look to your oszilloscope screen shot at 18:15 I beleave, you have the same problem. The frequency 10MHz is displayed, but the broadness of the individual peaks changes from peak to peak. So it is 10MHz on average, but it is not a 100ns period. Looks like 24MHz (or a multiple of this) is kind of a sampling rate for the time pulse generation. For lower frequencies, this sampling doesn't matter. But anyway, thanks for your perfect tutorials. 73, DC0UB
That is why we cannot use a cheap GPS receiver for radio applications and must build such a GPSDO where the frequency is produced by a proper oscillator and is only "disciplined" by the GPS.
Scullcom Hobby Electronics channel has made a very nice video series how to build such a frequency standard for cheap at home. I really recommend if you don't want to dish our hundreds of dollars but need a "bit" more than an Arduino can provide.
I saw his videos. Unfortunately he is no more active for the moment. I am not sure if you get it much cheaper (case and PCB included)
is it critical for measuring instruments or radios to have a square or sine wave ? and if I understand the neo-6 can have it's 1PPS pin reconfigure to become a 10MHz reference clock ??
It depends on the instrument.
You can create higher frequencies with the GPS modules. But they create jitter.
@@AndreasSpiess thanks for the reply, it's for a ham radio transceiver that has a 10MHz ref input that required -10dBm (approx 100Hz of drift can be observed with the stock TCXO on 1.2GHz), I thought your u-blox M series gps suggestion was a great idea on top of being cheap !! But since I cannot be sure of the "quality" of the signal and if jitter can cause issue out of the PPS pin to my transceiver, I also cannot measure how much dBm is getting out of this little board as I do not have the proper equipement to measure RF power...
I would not use a plain GPS chip for this purpose. Usually a GPSDO is the better solution.
@@AndreasSpiess Thank you for all your replies and tips !
👀Mindblowing Good 👍👍
Thank you!
On the GPS unit. Does anyone know what the blinking green LED on the bottom of the board rear between the BLOX module and the board corner labeled D2 means? Mine seems to blink at a 1hz rate, but it may mean something else?
It is well possible that it is a 1Hz signal because many GPS receivers are programmed to deliver a 1Hz signal.
I went through all your QO-100 videos again with more intent on learning some specifics. I am currently planning to replace the crystal LO in my Ukrainian transverter with a GPSDO. I have come across an interesting project that is much more than just a GPSDO but that can do that to. Have you heard of RF Zero? It's a Danish ham's creation. Very interesting.
73 DE KJ7LVZ
I just had a look at this project. Very interesting!
Just out of curiosity, how to calibrate the frequency of the OCXO?. What are the tools and methods used?.
A better reference like a GPSDO is compared and the OCXO is adjusted.
@@AndreasSpiess Thanks for you reply, put how does the OCXO get adjusted?
@Ash If my memory is right, the crystal has two small capacitors, one either side of it. Normally one of them is variable and can be used to trim the frequency when the temperature is stable.
Many OCXOs have a "VFC" pin that allows their frequency to be calibrated by varying the voltage on the pin. That is how the oscillator in a GPSDO is disciplined.
Andreas, I know you don't have time, but you would be a great teacher/professor at some institution of "higher learning."
I sometimes do lectures. But I only reach a few people. Here I cna reach thousands. More efficient ;-)
Institutions are a scam. Knowledge should be free as free beer and free as in freedom.
@Esteban: As a consumer it is easy to ask for everything free. It looks a little different if I have to feed my family... And the most important question: What do you contribute?
@@AndreasSpiess Everybody does need to eat and have basic needs. Every bit of knowledge we share for free allows people in a worst situation than us to get in a better situation. Imagine if you as a kid would have access to your videos as an adult. I firmly believe people will contribute when they are able. Those that take advantage and don't contribute are a minority. I'm extremely grateful for your videos being accessible for free and I try to use your links when buying, if in the future I can do more I will. I'm a lucky one from my country (Argentina) and I managed to learn English, electronics and informatics mostly thanks to people like you that share your knowledge on the internet, otherwise I'd probably be flipping burgers.
Other than a counter, what’s the next best measurement for frequency
I do not know.
since f = 1/t
then we can say most universal constant yield f, as many are affiliate with time.
example: heat 1ml of water up 1 celcius with 4.184 watt (all values are exact).
You got 1 Hertz (exact).
You made me laugh. Life does get more difficult :-) Thank you
You are welcome!
@@AndreasSpiess Cheers mate, and keep up the great work.
CB dayz, freaking and peaking 4 to 100w exciter to 1 kw, into neighbors stereo speakers when not even turned on.
I only came to 150 watts with my FT277. But it was illegal enough for my parents ;-)
can you use this to calibrate a bench voltmeter?
I do not think so. You get voltage references for that purpose
2:42 LOL. Sometimes I think you don't take this seriously xD
True! If I stop laughing I am probably dead ;-)
What antenna would be adviseable to use on these ublox modules?
They usually come with an antenna. Because I am in the basement I need an active external antenna and an SMA connector
Great your TH-cam show!
But I would prefer it very much if your videos would continue to be numbered.
It's much easier to reference them and much more practical to say to a colleague: "Have you seen video ### by Andreas (the guy with the Swiss accent) yet?".
Or sometimes I just write down: "See Andreas Swiss ##" . :-)
Many thanks for your series!
They stay numbered, just not the last 10 or so...
yay!
:-)
Make video on GPSDO homebrew
This is probably not a project for this channel. Too specialized...
@@AndreasSpiess your work is great thx for the reply
@@AndreasSpiess I have built several GPSDOs using surplus 10 MHz OCXOs, low cost GPS modules, and Arduino processors. The project is based on a design originally posted on EEVBlog by Lars Walenius. This has been a very interesting project to update my 40-year old electronics and programming skills. I would be happy to share the details if you are interested.
You can always buy a Hewlett-Packard caesium clock for your lab
I am not sure I could convince my wife ;-)
@@AndreasSpiess But, Andreas a clock is a clock... and I hear some about clocks in your country.
But a Swiss Franc is also a Swiss Franc ;-)
@@AndreasSpiess just say it's the king of all clocks
Is it possible to use this with usrp b200 or B210? any help please.
I do not know if those devices have a 10MHz reference input.
@@AndreasSpiess yes it's have
@@rakotondramananasamuel3970 So you have to check if the voltage levels match.
Andreas, I think you'll love this article about magnetic amplifiers that could be a good source for another video. www.nutsvolts.com/magazine/article/the_magnetic_amplifier
Thank you for the link. I did not know about it but it seems more focused on motor applications and maybe is also outdated because of the cheap silicon parts today.
👍👍👍
Thank you!
Great video, as always. Do you now about a basic 10 MHz distribution box, just 10 MHz in and six or similar 10 MHz out?
I don't believe that a OCXO is needed for this, maybe to continue if the 10 MHz ref is down.
I read they sell it also without oven, but I do not know where. I assume the oven does not add a lot of cost...
Look at the distribution amps from tapr.org -- they still have the TADD-1 distribution amp available, which is a low noise, isolated amp for 10 MHz.
❤️
:-)
7:06 My Asian parents: Did he just utter that word in public!?
:-) I have no idea why he was expecting it I find it very hard to come by these days.
@James: I assume you talk about the money ;-)
@@AndreasSpiess :-D
a regular quartz clock would easily end up slowdowns after 1~2 years of continuous use. 🤔
We always use our clocks continuously here in Switzerland ;-)
Hello Andreas. Nice video. Your OCXO is precise, but not accurate. I invite you to read further on the difference between accuracy and precision. When it comes to time and frequency measurement, that distinction matters a lot!:
en.wikipedia.org/wiki/Accuracy_and_precision
Thanks.
Thank you for the link!
So many fun jokes in this video!
Glad you liked it!
The first sentence in the video immediately had me shouting out: that is the most wrong thing I heard all day and in so many different forms of interpretation. In a physical sense time is relative. And in a standards sense there are so many standards: leap seconds, time zones, ...
You are of course right!