Call Search

New to Ham Radio?
My Profile

Friends Remembered
Survey Question

DX Cluster Spots

Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement

donate to eham
   Home   Help Search  
Pages: [1]   Go Down
Author Topic: High stability Oscillator question  (Read 13766 times)

Posts: 165

« on: December 08, 2004, 05:42:39 PM »

I am thinking about adding a TXCO-9 oscillator to my FT-857D for digital modes such as JT65.
I dont know very much about high stability oscillators.Would the general performance of the radio improved ? I have heard somewere of a lower
noise floor but i cannot found a technical article about it.
Any comments ?
Dave VA2DV

Posts: 240

« Reply #1 on: December 08, 2004, 07:14:39 PM »

The High Stability Reference Crystal is possibly useful if the radio is used in the field where temperatures change constantly.  If you are using your 857 as a base spend your money on a more useful option.

Here is a quote from the FT-897 Forum

 Do I really need the TCXO-9 for normal use of the FT-897?

 NO. The TCXO-9 provides even more stable operation of the FT-897; however, it is not needed for PSK31 or other typical modes with the F-897. For use in non-amateur applications such as CAP and MARS, it may allow the FT-897 to be utilized.


Posts: 21818

« Reply #2 on: December 09, 2004, 08:33:53 AM »

I'd echo those remarks.

I don't know how a high-stability reference oscillator can reduce noise levels, or noise floor, unless the original reference oscillator was very noisy.  Most of the modern programmable reference oscillators, which cost only $2-$3, are very low noise (usually referenced as "jitter") to begin with.  

The difference with a high-stability oscillator is its stability over temperature; in a climate-controlled house (hamshack), it matters little.  You can check this for yourself, though, as I often to do decide that I don't want or need a HS reference oscillator in my rigs:

Connect an audio frequency counter (after a good long warm-up, like 24 hours) across the speaker or headphone line of the receiver/transceiver in question.

Turn on rig from a cold start (no warm-up), dial up WWV at 10.000 MHz (or wherever you can hear it, could be 5.000, 15.000, 20.000, whatever propagation allows), tune it in to a low pitch like 100 Hz, with your rig set to an SSB or CW mode.  Watch the audio frequency counter, and set for some exact frequency (like 100.0 Hz).

Come back 10-20-30-40-50-60 minutes later and log the frequency indicated by the counter.

Is it drifting?  If so, how much?

After 60 minutes, unless your shack temperature changes a lot, almost any modern rig will stop drifting and you'll see the frequency remains very constant.

How stable is stable enough?  I guess that's up to you, but for the duration of a QSO, even using a very quirky digital mode, most all modern rigs are stable enough.

Accuracy is another matter entirely, and rigs can be adjusted for that as they age, if necessary.



Posts: 2832

« Reply #3 on: December 09, 2004, 09:40:18 AM »

Good phase noise performance and high stability, as hams know these terms, often results in design tradeoffs against each other.  Phase noise is limited by the thermal noise floor and how hard you are willing to run the crystal.  Typically, ham sets run at room temperature--adding a crystal oven for stability makes the thermal noise floor even worse, by raising the operating temperature (noise=kTB, where k is Boltzman's constant, T is the temperature, and B is the bandwidth.) The more output you can get from your crystal oscillator, the greater the signal to noise ratio of your signal.  However, a lot of power output also means a lot of power dissipated--higher power means more thermal heating.  The increased thermal heating often results in decreased stability, or drift, as hams know the terms.  A clever technique used with simple QRP radios is to use a very low power oscillator in order to minimize the thermal heating problem.  Phase noise isn't a problem--a 1.5 watt transmitter can afford 30 dB more noise before becoming as troublesome as a 1.5kW transmitter.  Similarly, simple receivers rarely have the mixer dynamic range that would allow the phase noise to become noticeable.

Phase noise of the simple programmable oscillators can be a problem--the PLL designs used are often inadequate for amateur receiving and transmitting applications. This is especially true of VHF or microwave local oscillator chains, where the phase noise may be multiplied by 6dB/octave or 20dB/decade.

Posts: 40

« Reply #4 on: June 25, 2005, 07:46:30 AM »

the TCXO will minimalise the freq drift of youre tcvr down to only a few hertz maximum .

using WSJT modes and especially JT65 EME working requires good freq stability in order to obtain optimum results .

most modern tcvrs can run JT65 perfectly ok without the TCXO installations but at high rf  output levels some modern tcvrs drift a fair amount in freq between transmit and receive periods .

 the IC746 is noted to have inherent freq stability problems at high rf output levels on 2m and probably all other freqs it will run on .

the drift of modern eqpt may be only 40hz or so but it is nonetheless noticeable on EME using JT modes .

i use an ic706mk2g on 2m eme with JT65B and without the TCXO installed it drifted by as much as 50hz between RX >TX periods even when i only used 2.5w output to drive  my amplifier , with the tcxo installed i am more or less on the same freq all the time when running skeds with other stations who have extremely stable systems .

you would benefit from installing a TCXO and it will make youre freq stability improve .

as for noise floor improvements i doubt you will see any from it .

 if you want to improve "noise figure" of system  then use a low noise external pre amplifier and switch off the tcvr internal pre amp .

some TCXO unit are not easy to install , IC706MK2G is tricky due to no space around shielding can to view position of legs which pass thru pcb and also you may have to drill out holes in pcb where you need to solder due to existing holes being full of solder .

i fitted TCXO into an ic756 recently, again tricky to do due to smallness in size of legs to be soldered in place .

if youre yaesu only requires a plug in TCXO unit then id advise you to buy one and fit it .

if you want a cheaper $20 "halfway'route  between the two kinds of options then look at KUHNE ELECTRONIC in germany and see theyre QH40A temp controlled crystal heater kit [look under SPECIAL section ] .

i fitted one of these to my JRC JST-245  tcvr easily and it improvedfreq  stability somewhat .

hope this helps you .
73 ian



Posts: 0

« Reply #5 on: September 01, 2009, 10:10:34 AM »

Out of the box, my FT-817 (which also can be upgraded with the TCXO-9) was within 5 Hz accuracy when checked against WWV at 15 MHz. (That's better than 0.5 PPM initial accuracy!) When it warms up, it moves 4 Hz to within 1 Hz of the 15 MHz WWV signal (after two hours) and from there on it stays there. An excellent job Yaesu did on that one. My conclusion: you do not need to buy the TCXO-9 high-stability oscillator at all, the master oscillator that is in the rig in stock condition is perfect.
The method I use to check a rig agains WWV is simple but I have never seen described in any Ham literature, weird enough. It works as follows.
Tune the radio to 15000.30 MHz LSB (thus 300 Hz above 15 MHz) and store this mode & frequency in a free memory location. Then, tune the radio to 14999.70 MHz USB and store this in a second, adjacent memory location.
Then, compare the two tones you hear when switching between the two memory locations. If the radio would be exact on frequency, you would hear exactly the same 300 Hz tone differential when switching between the two. If you hear two different tones when toggling between the memories, then tune one of the frequencies such that the WWV differential tones sound equal again after retuning one frequency. (so for instance, let's say you had to tune the 15000.30 LSB frequency to 15000.45 LSB to make it sound equal with the 14999.70 USB tone. The rig's error frequency (at 15 MHz) is then the difference between 15000.30 and 1500045 divided by two. So in our example this would be 15 Hz divided by two is 7.5 Hz)
As you can see, this is also a perfect method to adjust the master oscillator. Program the two frequencies; let the rig warm up for two hours, and then tweak the TCXO capacitor such that the two tones sound exactly the same. You have then adjusted the rig to within half a hertz (at 15 MHz), as the human ear is able to easily discern between a hertz when listening to a tone frequency of 300 Hz.
Now you will also understand why you need to do this with the highest WWV frequency you can receive. Because (as an example) if you would calibrate the radio to within a 1 Hz error at 10 MHz, this error (in absolute terms) will be 3x more at 30 MHz, being 3 Hz. So make sure you pick the highest WWV signal you can receive when performing this calibration.
A note of caution (no pun intended): Some people are completely tone deaf and unable to discern between two different tones. It's rare but it exists. If this is the case with you, ask a friend to help you.
Another word of caution: The above USB/LSB method works ONLY on modern rigs that have all internal frequencies derived from ONE master oscillator, including the SSB product-detector injection frequency. This is the case with most modern rigs, where the product-detector frequency is derived from the master oscillator using a DDS. For some older rigs, this is not the case. For instance, my Icom 765 has a master oscillator which is used for the DDS in the main VFO, but the rig's product detector uses fixed crystals for the USB and LSB injection frequencies instead. In such a case, this method does not work, because the crystals easily introduce an error of 10 Hz or more. Check your transceiver block diagram if you want to be sure.
Needless to say this method absolutely does NOT work for older rigs like the KWM2A and the like (not even for the KWM-380, it uses a separate crystal for the product detector). Good luck!

Posts: 1897

« Reply #6 on: September 01, 2009, 11:04:17 AM »

adding a crystal oven for stability makes the thermal noise floor even worse, by raising the operating temperature (noise=kTB, where k is Boltzman's constant, T is the temperature, and B is the bandwidth.)

Note that's ABSOLUTE temperature, (Kelvin or degrees Rankine).

So, for a hypothetical application (likely not ham radio though) where both stability and low noise floor were critical could one refrigerate the crystal, perhaps w/ a Peltier cooler controlled by a well tuned PID controller (say to 250K [-23°C/-9°F] vs. heating to -- what? -- perhaps 320K [+47°C/117°F])?

Would be expensive though!

Posts: 6834

« Reply #7 on: September 01, 2009, 11:54:01 AM »

OK, raising or lowering the xtal temp is OK for frequency stability, but we are looking at HF/VHF/UHF here! Atmosphereic noise controls the noise floor. The biggest improvement would probably be some good filters for the FT897/857.  They would make the most sense even from a noise floor point of view.  Decide what passband you need and go from there.


Posts: 728

« Reply #8 on: September 02, 2009, 08:21:27 AM »

The far out phase noise of crystal oscillators is generally not as good as L-C oscillators unless an extra crystal is added as a filter. But unless the stock oscillator is very poor, it is unlikely to be the dominant factor in the receiver or transmitter phase noise performance - that is usually frequency dividers, phase detectors and active loop filters, as well as the VCO. Even with a DDS, it's the digital circuitry that tends to dominate. There is a tendency to use high value resistors and low value capacitors in the PLL loop filter when the designer wants to get power consumption down, but this is the wrong thing for phase noise. As a result, it's always a compromise.

Posts: 712


« Reply #9 on: September 03, 2009, 01:50:38 AM »


The TCXO option is only worth considering if you intend to use data modes. It will not make any difference to your TX or RX performance in terms of noise performance, but it does help improve frequency stability.

I have an FT-857D which I bought second hand to use for WSPR and other data modes. I already have an FT-897D, so I assumed I was getting the same basic rig when I decided to get the 857. What I didn't realise was that unlike the 897. which had the TCXO option fitted as standard when they introduced the D version. The 857D didn't have it fitted.

Now the 897D with the TXCO is just about adequate for WSPR use on 50MHz, marginal on 2m and inadequate on 70cm. This is because the TCXO temperature is not really high enough to lift the Xtal above the ambient temperature inside the case, when it's been on transmit for a short time. The very short term stability is OK but the drift over a few minutes becomes noticeable.

The FT-857 in basic form is marginal for PSK and unusable for WSPR and other very narrow band data modes on the higher HF bands). This is because the reference Xtal oscillator is mounted at the rear of the rig, directly adjacent to the air exhaust vent over the rear heatsink. As a result as soon as you go on TX the hot air flowing over the Xtal causes it to drift off frequency.

Adding the TXCO option improves this considerably, to the point where it is usable for WSPR, but I think Yaesu could have done a better job for the price of the option (about $130 at current UK prices).

One cheap fix I considered before I bought the TCXO was to fit a 12V 25 ohm PTC thermistor (Mouser 527-3005-12V25) near the Xtal on the existing board. However The cost of obtaining this part in the UK made it uneconomical just to try, so I choose the safe (but much more expensive) option.

Incidentally if you do fit a TCXO-9 you will need to set it on frequency using something like WWV. I used a GPS locked spectrum analyser whilst TXing on 70cm where the frequency error was greatest. I could set the TX frequency within about 20HZ of the desired frequency, but the trimmer cap on the TCXO board is C**P and adjustment not smooth, so it's very hard to get it set spot on frequency.

Hope this helps you to make a considered choice.


Martin - G8JNJ

Posts: 680

« Reply #10 on: September 03, 2009, 04:24:17 AM »

"So, for a hypothetical application (likely not ham radio though) where both stability and low noise floor were critical could one refrigerate the crystal, perhaps w/ a Peltier cooler controlled by a well tuned PID controller (say to 250K [-23°C/-9°F] vs. heating to -- what? -- perhaps 320K [+47°C/117°F])?"

You mean like a cryogenic sapphire oscillator (CSO)?

There are lots of solutions to temperature effects and phase noise, and the trades between them. For instance, with a modern rig, the software could measure the temperature and use that to adjust the DDS frequency to compensate for the freq vs temp curve of the oscillator, allowing the use of a drifting, but quiet oscillator.

Or, one can get yourself a GPS disciplined oscillator or a Rubidium source (<$2K brand new, much less used), both of which will be more accurate and cleaner, in general.

Or, if ham rigs had external reference inputs, there's lots of things one could do.

But ham gear is designed to a cost point, and part of the trade is oscillator performance

Posts: 1

« Reply #11 on: October 07, 2009, 02:04:30 PM »

The Old FT-897 can use a TCXO-9, but the FT-897D has it installed.
"The new FT-897D version includes coverage of the U.S. 60-meter (5 MHz) band, along with the 0.5 ppm TCXO Unit, at no additional charge!"   From

Posts: 712


« Reply #12 on: October 08, 2009, 09:32:15 AM »

Hi All,

Beware although the FT-897D does have the TCXO-9 option fitted as standard, the FT-857D does not.

It caught me out !


Martin - G8JNJ

Posts: 1

« Reply #13 on: January 24, 2010, 03:00:48 PM »

And to answer to all concerned by the hypothetical increase of noise caused by thermal considerations, the TCXO-9 is NOT an oven oscillator... it is a temperature-compensated oscillator... there are NO heating elements in it... it runs at the same temperature of the standard oscillator that comes with the rig....

Bottom line : absolutely not necessary for SSB and CW work. Must have for WSPR and other narrow band digital modes.

73  Alberto  I2PHD
Pages: [1]   Go Up
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!