tag:blogger.com,1999:blog-953397841255562532.post5947020853197516379..comments2023-09-30T14:59:05.385+01:00Comments on G4ILO's Blog: Calibrating the FT-817NDAnonymoushttp://www.blogger.com/profile/11984840704237681015noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-953397841255562532.post-67272846696558290822009-12-02T09:22:40.242+00:002009-12-02T09:22:40.242+00:00Hello Maurice. Thank you for taking the time to co...Hello Maurice. Thank you for taking the time to contribute this comment. It is very illuminating. I think we all tend to think that having taken the trouble to tweak our rigs down to a fraction of a Hz that they will stay there, at least once warmed up. It isn't immediately obvious, until it is pointed out, that a stability of 1ppm means a potential error of 10Hz at 10MHz - quite significant in terms of WSPR operation - and a whopping 140Hz on 2 metres, which is very significant if you're looking for weak beacons using a narrow filter.<br /><br />I found the stability of the FT-817 to be quite good. Once set it tended to stay within +/- 0.15Hz of the frequency. I think the long term stability isn't bad either. In fact when I started it was less than 10Hz off at 5MHz, and it had never been calibrated since it left the factory several years ago.<br /><br />However I now have renewed interest in the availability of the option to lock my K3 to a frequency standard!Anonymoushttps://www.blogger.com/profile/11984840704237681015noreply@blogger.comtag:blogger.com,1999:blog-953397841255562532.post-90797502036438371382009-12-02T01:23:11.677+00:002009-12-02T01:23:11.677+00:00Hi Julian, Yet another interesting item o...Hi Julian, Yet another interesting item on your blog!<br /> <br />My opinion, as (now retired) former Head of Laboratory at British Telecom Electrical Standards Laboratory maintaining standards for test equipment used by BT in the UK.<br /> <br />You were fortunate in being able to set your master oscillator to 0.15Hz at 10MHz, which is an accuracy of 0.015ppm (parts per million). The specification for stability of the FT-817 after warm up at 25C is 1ppm per hour - you will probably find it is better than this but not better than 0.1ppm/hour. Indeed, the more expensive option TCXO-9 is specified at 0.5ppm/hour. You were trying to achieve an accuracy of around 1 part in 10^8 (0.01ppm) which can be achieved with professional test equipment but at a cost over the price of consumer electronics equipment like the FT-817. Then how long will it keep to this accuracy? If you take the spec literally this would be 36 seconds, but even at 0.1ppm/hour this works out to 6 minutes.<br /> <br />The Yaesu spec should ideally include contributions for temperature coefficient, supply voltage and ageing of the crystal oscillator in the FT-817. You can't do much with the first two parameters, but at least with the last one you can compensate for it. Now you have set it to "nominal", repeat you test after a month and see how far it has drifted. Don't be surprised if it has drifted 10Hz, after all that is only 1ppm. Do not adjust it yet, leave it another month and see if the drift is repeatable. Assuming it is, then readjust the oscillator so that it will drift towards nominal during a month (or whatever period you choose) and then, after twice this period, it should have drifted the same amount the other side of nominal. Now is the time to perform the adjustment again!<br /> <br />This is the technique used for calibration of frequency counters used by BT technicians while I still worked there.<br /> <br />This sounds quite involved, but I wouldn't like you to think that once you have achieved the accuracy of 0,15Hz at 10MHz it will actually stay there!<br /> <br />Regards Maurice G4DVM.Anonymousnoreply@blogger.com