I need 10MHz - how hard can it be! FE-5680A
HTML-код
- Опубликовано: 23 июн 2013
- gerrysweeney.com/?p=1197
It all started when I wanted to calibrate my HP 53131A universal counter, which as it turns out probably has one of the crappiest and most disappointing standard oscillators ever put into a frequency counter, HP you should bow your head in shame....oh of course I forgot, a half reasonable oscillator is an "optional extra" when you by HP/Agilent - of course it is....anyway on with the job at hand....FE-5680A Rubidium Frequency Standard, OCXO 
Наука
I'm late to the party, here, but with 12 years working in a USAF lab calibrating/repairing counters, I've learned a little, or, enough to be dangerous. Some basic counter advice.... If you're going to calibrate a counter, don't stack it on anything; let it sit by itself, even in use. Keep your room temperature steady if possible. Our lab was kept at 74 degrees F. Allow Room Temp oscillator (RTXO)-based counters 24 hrs warm-up, oven-based (OCXO) at least 72 hrs. Agilent RTXO trimmer cap is an abomination to adjust (I pined for a trimmer cap with a worm drive!). I was happy at +/-2 Hz on those. Use a brass, ceramic, or other non-ferrous screwdriver. Agilent 53132A OPT1 counters were a breeze... one would plug in an appropriate external 10 MHZ standard and "tell" the counter to calibrate itself (all of 5 minutes!)! The Racal-Dana (R-D) counters' OCXO fine adjustment can be very touchy, too; you would adjust it, but then it would take off again! Don't over-power the R-D counter's input 50 Ohm function... you'll burn up the input resistors (easily fixed on the R-D's though, with a pair of 100 Ohm half watt in parallel. Those R-D push-buttons got to be un-obtainable for me in the States... had to condemn a few. Made me furious 'cause they were (are) a good counter. Check/calibrate any counter at least once a year against a rubidium or GPS-slaved oscillator. Get a Rubidium and use it as a "shop" standard as an external input for all your lab 10 MHz equipment.. you shouldn't have to worry for a LONG time. ALL digital displays have an error of +/- one count of the Least Significant Digit (LSD).
Hi Gerry just bought one that looked like the first one you had without 10MHz!. But the rivet marks on the back confirmed I have the correct one. Great video and very reassuring too!
I realise that this video is over 10 years old, but my method may still be of use:
I haven't seen this complete video yet but what I have always done to get an accurate 10Mhz reference is by using the cesium atomic clock driven carrier signal from WWV, by adjusting my oscillator to this carrier frequency using the beat frequency method. This is even more accurate than a rubidium frequency standard.
The method I used was by tuning into WWV 10Mhz signal using a short wave radio then the device that needed its oscillator adjusted was brought near to the short wave radio and its oscillator was adjusted to provide a zero beat signal. I was located in New Zealand , so had to wait for the best time of day/night to get the best signal from WWV of course.
Superb level of detail.
Thanks for the video! I learned quite a lot about frequency counters and standards. I appreciate the explanation of warm up periods. I'm getting ready to purchase my first frequency counter so this will come in handy.
Again, for the videos, lighting and close up shots would be appreciated. A lot of times when you say "you can see" we actually can't. Still better than I could do though. :-)
Thanks!
Hi Gary, thanks for the feedback. The first one is definitely a 1PPS version, there is an 8.3Mhz signal inside and it is tuneable using software and internal RS232 connection but at the time I wanted one that was specifically designed to put out 10Mhz which is why I asked the supplier to send me the correct one as advertised. I will get around to hacking the other two at some point and see what I can do with them. Gerry
Thank you for the feedback, I appreciate it. Gerry
looking forward to your further videos on the subject
thanks
I just stumbled upon your channel excellent detail exactly what I want to do to my HP that I just ordered off eBay. I believe Dave from EEVBlog did the same update upgrade he got the in ternal timer off eBay for the higher accuracy I believe he used a rubidium for his source. In another one of Dave‘s videos or another technical channel someone use the rubidium as a reference timer I believe for all three pieces of task equipment with feeds into the back of their unit. The rubidium timer was plugged into the oscilloscope it was plugged in to the counter and it was plugged into the frequency generator simultaneously so they were all synced off the same rubidium reference.
Thanks for the information, very useful pointers. Gerry
Nice Video Gerry. I have been doing similar things lately. Following Dave Jones video, I bought a Rubidium oscillator several years ago and only used it a little. More recently I've invested in some older but quite high spec equipment and have been curious how accurate they were even though some have included calibration certificates. Just recently I also purchased one of the second hand OCXO's pulled from equipment and I too found them to be very good. I have video's some of my findings but not edited the clips yet. One other comment why not just use one of your oscillators, maybe even the rubidium standard and distribute that to all of your equipment? I would be interested in how you would tackle it. I am thinking of taking the 10MHz input and using some buffers for multiple outputs and distribute that to my various pieces of test equipment, counters, spectrum analyzers etc. One other thought, do you need to convert the sine wave output to square wave for any equipment? and is there a min/max voltage range required for your equipment, I know some must not exceed certain levels.
You are welcome my friend, glad its all working from you. Those CB's are going to be spot on in tune now. I must dig out my old Cobra 148 GTL DX some day, maybe do a teardown on it :)
Thank you, glad you found it useful.
Hi stuckinpants, thanks for the pointer, I found these. I guess my only concern is they are rated at 0.5% which is not that great, The OCXO I bought allows you to calibrate it with 0-10v +/- 0.3 so the tolerance of the TCX0 is the allowable range for the tuning of the OCXO. and the OCXO can be bought for about £20 so I would say they are probably not worth £11 each. Gerry
I remember when i had a Blackstar meteor 100 and a blackstar meteor 600. Always prefered the 100 as it was plenty for what i needed but the warm up time was alot quicker on the 100mhz counter than the 600mhz.
Great Video, Thanks
Do you know, how stable such a OCXO is over a year?
Does it require re-calibration from time to time?
I saw a few on ebay and they need 24 V supply
How critical is the voltage supply?
This is just for historical reference, but back in the days of analog TV. The way I calibrated my FC was I would take a color tv set and hook up my FC to the chroma oscillator and tune the TV to a high powered OTA TV station (very important). Then I would adjust my FC until it read 3579545. ICYW: The FCC required all high powered TV stations (10kW>) to discipline all there frequencies to a rubidium standard.
The good ol days !!!
Yeah thats pretty common approach still but you need to know which transmitters are good and you need a receiver and so on so its not as easy for many who just need 10Mhz :) which is where I came at this from.
Gerry Sweeney will I get oscillations with just a crystal and a battery?
+Owen Chase science channel No.
***** ok thanks
Thanks for the feedback, I appreciate the comment. Gerry
Gerry, I used to modify a lot of sets, upping the power, adding in more channels. My fav amplifier was a 500w beast which I used to use as a seat! Used to have a lot of fun bouncing Tx off the ionosphere around the world.....even with only 4w. Those were the days.
Can you help Me, how to do this so I Design System can be generate signals Up to 10MHz Frequency Of 0MHz using AD9850 DDS system with miccrocontroller ATmega16 ?
please replay it?
A fascinating video and very helpful
Thanks for the reply! Yes. It took me a good bit of time to understand measurement confidence. But I still don't undersand why you need to measure frequency unless it's and RF application. For micro-controllers - and the things you seem to work on, where is the application?? PS - love the videos. Long, but worth the watch. Thanks!
Hi Alexie, thanks for the info. I will probably write some portable code (perhaps in Qt) for Windows, Mac and Linux to make this easy to do, I will test out your math and Mattias and see what works. Thanks for posting. Gerry
Tom, yes it needs to be 15v. In the related blog article I listed the PSU I used for the project, its 15v 2.1A and can be got for £10 if you are in the UK. Gerry
Hi, thanks for the comments. What I found amazing was how good the Racal 1999 is with its own crystal oscillator. I guess its probably a TCXO but even so pretty impressive, only 16Hz out on a 10Mhz verified against the RBS....Gerry
Yes I found that too when I looked again. In fact, I seem to have lost the document that had the unreadable page in it! Who knows....thanks for pointing it out. Gerry
That was almost an hour of my life well spent. And thanks for the D class amplifier explanation in another video. I didn't realise how stable them oven oscillators are , amazed even, and then you can cal them too. Whats the part no for the crystal one? Nice vids keep them coming. I'm sure Dave would like to know how you mod the counter. I'm subbed to Dave, Mike and you, you all entertain and help me in different ways with you great uploads. Wish you well.
Not sure about being talented but thank you Tom, glad you find the vids useful. Gerry
Thats an interesting idea...sounds like it could work. Gerry
You can use modern replacements for those oven-based TCXOs. If you want to make a board anyway, you might consider to chose a modern part. Better accuracy, smaller, way lower power consumption.
Hi George, I would like to know that when I measure a frequency its correct to the best of my capability given the tools that I have. Its not just about accuracy, its also about stability over a longer term. If you are measuring something today and next week its nice to know that your test gear will give you the same result. Accuracy can be important when you are frequency multiplying or building accurate timing applications. Gerry
a Lot of work as gone in to this garry ,,,a star video,,,i feel just the same on my
racal as you garry,,had my racal for 5 or 7 yers ,,,and all my buddy have
HP/Agilent,,,but am poooooooooor
I haven't looked inside the HP 53131A yet. If the capacitor fell apart, then is it possible to replace the cap (I assume it's a variable cap) with a better quality one? Or replace it with one that trims more accurately? I bought two FE-5680A from ebay (from a different seller), and both have an SMA connector on the back. The seller (that I bought from) actually tests their units in advance. Mine work beautifully. They have no trouble with locking from 8 MHz to 12 MHz. Only issue was that one didn't have the 1 pps output. But it wasn't much of an issue, since we can always use a divider chip to get 1 pps.
This is a great video, as it offers nice options in case the HP 53131A stops working due to a wardrobe malfunction with the internal oscillator.
+Kenny yes its possible to change the cap if you can get the right part. That aside though, the internal oscillator, even with a good cap is still a very poor oscillator, definitely should have a high stability oscillator in there, it makes all the difference. Check out these articles if you have not seen them already: gerrysweeney.com/tag/53131a/
+Gerry Sweeney Great! Thanks Gerry. I haven't seen those other articles. Thanks for your time with providing that link. I'll check those out. Much appreciated.
Hello, Gerry. I use those Rubidium standards as stand alone standards for my lab and they perform great. I bought two and both work great.
I have in the lab HP counter like yours with highest possible accuracy which costed a fortune and when compared result to Racal-Dana 1991 + Rubidium standard the result was promising. Readings differ only in 0.6Hz while measuring 40MHZ
Hi George, suppose I was to make a frequency counter using a micro controller or a FPGA, I would want to be able to check how it compares. Or if I were to review a frequency counter I would want to know how accurate it is. I wanted to upgrade my HP counter with an OCXO but I could not be sure of the accuracy without a second point of reference which is where the RBS comes in. I would say that any electronics lab *should* have a reasonable frequency standard. Gerry
If you do happen to need 1 pulse per second, many GPS chips have it. You need a GPS antenna and reception. You can also get very precise frequency references off of the US National Bureau of Standards WWV shortwave stations.
Please add the part number (FE-5680A) to the description so the vid can be found.
0MoTheG thats a good point, have updated the description, thanks for pointing out the omission. Gerry
Do you have two working 10Mhz rubidium clocks? Could you run them simultaneously, like you did with oven controlled one?
Hi Alexei, I decided not to try and re-program and modify the 1PPS version because as you say, each unit has slightly different values. Oddly, now I have a known good 10Mhz reference I could probably tackle that task now. I plan to do a teardown on both broken units to identify all the differences so other people will not have to go through the same pain that I (and you by the sounds of it) did :) I referenced the document done by Mattias on my related blog page at gerrysweeney(dot)com
Thanks for the info. No problem to me with the video length or waffle. Alls good :D
I bought a 27 MHz TCXO module on eBay for about £15 and found that it maintained about +-0.5ppm at room temperature relative to the FE-5680A (spec is like +-2ppm over temperature).
They also use almost no power, I'm going to put a few of those in some radios I have and in my frequency counter some time since the accuracy is so much better than standard crystals for almost no penalty.
I guess the advantage of the fixed one is you don't (should not) need a reference to calibrate/check it. That was what I was originally looking for, but it turned out to be less simple than that!
Hi ....just found your channel great stuff
not sure if this is the case but FEI manufacture a variable output version (between 1Hz and 20 MHz) I think it may be possible that your first one sent in the video could be that one might be worth checking into but requires some additional circuitry to operate properly hope it helps ...............great channel
excellent ! I need a counter for very accurate work,and just can't find a reasonably priced unit with a HS oven. So this is very interesting,since i may just have to "improvise" a standard to bring these out of spec over-priced devices into useful range.Agilent (53181A)
I have a PCB or fully build upgrades for the counter you have if you are interested.
gerrysweeney.com/update-diy-hpagilent-53131a-010-high-stability-timebase-option-pcbs-available/
just placed the order. Great work and much thanks! I'm one step closer to getting a pile of receivers done.
Thank you for the order, I appreciate the support. You will love the counter after you upgrade it, totally transforms it. I will get shipped on Tuesday/Wednesday if thats ok as I like to soak test them for 24hours before I ship them
absolutely. Take the time needed. Glad to support all the hard work.
Congratulations for the accuracy and the detail you seek, it is almost unbelievable to have all this powerful technology at home these days. Locking two "free" oscillators is pretty good.
Carmel Pule' thank you for your comments and for watching. Gerry
Hi Alexei, thats not correct, the working one I show in my video is unopened from new, the seal had not been broken and you get 10MHz out on pin 7, thats a fact. What you describe is true of the 1PPS version (the first one I tried). I don't believe I fried any part in the first one as I only connected power to pin 1 and the rest was just passive probing. I will find out more when I tear it down. Gerry
The OCXO board in the manual i mentioned in your other video is very readable if you zoom in a bit, that is what i can see at least.
Only a guess but most likely the synthesisers mixing strategy means the reference frequency shift up has a negative effect on the synthesised frequency.
Hi Gerry
Could you tell me how accurate that OCXO you use in this video is please
like +/- 10Hz when warmed up
Each OCXO is trimmed to the frequency so needs calibrating but they are generally within 100hz without any trimming. They key though is their stability, which is often 10-10 or 10-11 so pretty good.
I think you are good to go with 14.8V so long as its a regulated supply. The cleaner your supply the better. If you are using a linear regulator (7805) you need a heat sink on the thing, it will get warm. Gerry
I enjoyed that. Couldn't help myself from thinking "stop waving those rubidium references around... two near drops and still not learned your lesson". ;)
I'd still quite like to see the video you sent to the ebay seller too. ;) Thanks for sharing!
Ian, I used to do a *a lot* of CB hacking in the day, all those extra channels, replacing SWR boards with 100W power board and all sorts of interesting stuff. MC145106 and PLL02A's and all those EPROMS, life was so simple them days....:)
Yeah true, but did you see my reaction speed :) I will look at the video again and see if its appropriate to publish, I can't see any reason why not, I just made it for the sole purpose of sending to the seller so I don't think it will be that interesting. Thanks for watching. Gerry
Thanks Craig, I suspect the 10Mhz one is not fixable, probably just reached its end of life - these things are only specified to last 20 years or so, the ones we get on e-bay are 20+ years old already! The programmable 1PPS one should be adaptable though - we will see. I am about to do a follow-up on this with a novel enclosure and reference distribution system for my lab - watch this space....
I admire Electronic People....They seem to have so much energy in Them...
I only have a couple scopes :). The oscilloscope I used in this video cost only £175, I bought it no e-bay and its a great analog scope. The Agilent I often use was very expensive - I had to do a lot of soul searching before I wrote the cheque! Thanks for watching
Thats definitely another way of putting it :)
Indeed, I think thats the problem. Almost anything that caries electrons or (light for that matter) is affected physically by temperature. This is why Rubidium and Caesium etc are used, I don't think the transition of electrons are affected by temperature in Rubidium, at least not to the same degree. Gerry
Hi Stuckinpants, I should probably try a 10MHz TCXO, they look pretty good but are fairly expensive, they are definitely better than the standard XO, which at least in the HP I want to mod is absolutely terrible. I will hunt around for a TCXO. Gerry
Hi John, wow you are right, I cannot imagine what happened there....perhaps they will comment. Thanks for feedback and for watching. Gerry
That sig gen is a nice bit of retro kit, would have killed for one of those 30 years ago when I was CB hacking myself :)
No it will not be too hot, just warm. with 10V drop across the regulator at 80-100mA the regulator will dissipate about 1 watt in heat. Not too much but enough to be a problem if you don;t cool it a little. Let me know when you get it all working. Gerry
If you get the 10Mhz, you need to supply the unit with both 15V and 5V. Pins are as follows...
1 = +15V (about 1.5A)
2 = GND
3 = Lock (active low)
4 = +5V (about 80mA)
6 = 1 PPS Out
7 = 10Mhz Out
The Lock output on pin 4 is not designed for much current, I would use a low current LED (2mA), Cathode to pin 3, Anode to 1K5 to +5V
You can use a simple buck regulator board to stop down 15v to 5v as I did in my video.
Gerry
Thinking about I'm pretty sure I actually read 0.05ppm in that test... the seller I bought from doesn't have 10 MHz ones but I was going to buy from a listing called "2pcs TCXO 10MHz 10.000M Ultra precision Crystal Oscillators DIP 0.5ppm(NT)"
I got same - there are no boxes that have 10MHz output on D connector. You have to add SMA output yourself and then you need to reprogram it which can be easily done over RS232. You most likely fried 74series flip flop chip inside that generates 1 second signal. (I had to re-solder mine :-) ). Also it generates 8.xxx MHz sine wave signal which is not very easy to change to 10MHz because they all UNIQUE, you have to calculate divider value and then reprogram it by setting correct 64 bit divider.
I mean, Yes it is. I think it cost me $60 on e-bay. They are all second hand so the price does vary wildly - hunt around for a good price. No they are not fake-able cheaply, they cost $1000's new.
Hi, great vid, thanks for the shout out...Gerry
I reckon you set the *standard* for this genre of RUclips videos. Pun intended. Cheers mate!
qwertyscope thanks for the feedback and for watching. Gerry
Hi Brian, a GPS and a Rb have different characteristics. A GPS can take a long time to lock so should probably be left switched on, whereas the Rb takes 1-2 minutes to warm up and lock but once locked its ready to use. Also while GPS has excellent long-term stability, the Rb short term stability is far better. All in all, I would say an Rb is a better shop reference. The best would be a Rb synced to a GPS but for me thats definitely overkill. I have a Trimble Thunderbolt I will do a video soon
I just bought a GPSDO that's accuracy is rated at 1 PPB (Billion) The GPSDO locks in well under a minute, has two outputs and it's output frequency is programmable. £150 you can't go wrong.
Gerry,
Did you ever get the one repaired? I had one that wouldn't lock, and it's just a simple trim-cap to adjust the sweep range. I guess as they age, there's some drift due to age, and you just have to make sure the sweep actually crosses 10Mhz. If you hook it to a counter, you'll see the output doesn't cross 10Mhz, so it never locks. Very simple fix...
www.ko4bb.com/dokuwiki/doku.php?id=precision_timing:fe5680a_faq#what_if_my_5680a_output_does_not_lock_up_after_several_minutes
stazeII hi, I never got around to poking around with it, thanks very much for the link it definitely looks worth a try, I may well give that a go. Thanks again. Gerry
Gerry Sweeney No problem. Like I said, worked for me (I bought mine as not working (so it was cheap)), and was very easy. Just had to realize I had to remove the top, and not the bottom plate. Since then I've had no issue at all getting it to lock. Guessing the heat from the unit just causes things to fall out of spec.
Hi Jerry,
First you don't want to let the Rubidium standard get too hot, you really need to heatsink it. I didn't watch the complete video so maybe you did that. The OCXO heats to a point that is its stabilized voltage, once it reaches that it the heater resistor only needs to come on to occasionally bring the case up to working temp.
Rob
+mosfet500 yep fitted a heatsink :)
+mosfet500 in this video ruclips.net/video/chrzrod3tQY/видео.html
Gerry, don't tell me you were into CB's as well..........I was into it big, my old man had a trade business supplying Scotland with CB gear.......as a result I had all the toys! Still got a few Uniden's lying around.
The 7805 *is* a linear regulator, you will need a heat sink on it because it will run warm. Gerry
Hi, thanks for watching. Sorry about the length of the vids, sometimes I waffle a bit and I am working on that :) The part number is OCXO 8663-XS, there is a link to the data sheet on my blog page link just below video) Thanks for the sub, I appreciate it.
... Hi Gary, after 10 years, you may have answered this question... nevertheless, I'd like to know this ... why is canal 1 AC coupled and canal B, DC coupled .?
I love the video - but still don't understand why you need time domain accuracy?!?!? To 10Mhz to a 1 hz resolution?
O well ... I did all that work, should have just asked to send me the right one. Maybe coax gives me cleaner output.
Another thing-you may find a 5 pin header right next to Analog Devices DDS chip. (same one you will have to use to reprogram it over RS232). If you get a dent on metal cover it may touch the header and as result there are no output on D connector at all. Device may still work if you remove cover. There is a pdf doc online describing similar mod made by Matthias Bopp.
Maybe I am missed something in the video, but why not get a shop GPS reference. Can by a GPS 10Mhz reference off ebay at reasonable prices.
Brian, I think yoiu've nailed it. His one reference will handle several counters, etc by itself and all would be "automatically" on frequency. When a USAF"s Type II lab I was working at was re-built several years ago, I , being the "counter specialist", advised management to use a shop reference (GPS-slaved rubidium) a distribution amp, and a BNC outlet at each tech's bench. Worked like a champ and we had 10 to the -12th accuracy.
That’s what I would do
The 1PPS unit does *not* have a sine wave out, you need the 10Mhz unit for calibration. If you get the 1PPS unit and need to convert it to a 10Mhz its possible but its a lot of aggravation. If you are new to electronics and just want to cal you FC, make sure you get the 10Mhz version - trust me on that one. Gerry
You had to need it to buy it.
I know, thats when HP built great kit, well worth the $400 I would say
Your audio is regularly clipping. It would be worth just reviewing the mic levels when recording videos.
+Dr. Lecter Not to mention his camera/renderer combination leads to awful deinterlace artifacts...
You are so rich to have so many oscilloscopes!
Is that a FE-5680A Rubidium Atomic Frequency Standard? How much does it cost you? The price range is a bit weird from 85euro to 500euro for the same "thing". Are there Chinese fakes around?
OMG! So many comments, such an old post... so many things... screw it. Let someone else point it all out.
Thank you for the video, but I know of a much cheaper (N. America)way to calibrate a counter. Get a WWV receiver ($5-$20) and bring it close to your counter. Listen to the WWV receiver, which runs on an atomic standard (10mhZ) and heat the beat. It may sound like wahh wahh wahh etc. Turn your counter until it is steady and presto, your calibrated to 10 mhZ!
Sure that works too, but not here in the UK and not useful when you need a 10Mhz reference :)
Yeah, mobile phones and the internet soon took over...but fun while it lasted
I need 2 kW
Showwww
lol...
Write them and cancel the offer, they should be OK.