And a small correction to this in case anyone else gets inspired to do anything. Grounding just one of the crystals improved the performance by about 8dB - but did not get as good as the KSYN3 with a shield. It clearly needs the full treatment.
73, Andy, G4PIQ From: [email protected] <[email protected]> Sent: 01 January 2021 20:57 To: '[email protected]' <[email protected]> Subject: RE: K3 Tx Spurious on Specific Frequencies on 160/80/40 I think I have worked it out. Taking the lid off and poking about showed that if I touched the crystals at the DDS end of the filter on the synthesiser board, the spurious levels rose. Upon taking the lid off the other (newer) radio, I realised that the KSYN3 in the new radio (SN 8529) had a shield installed that neither of the KSYN3s in the older radio (SN 4198). Looking at both construction manuals - somewhere between 2010 an 2013 a shield was added to the KSYN3, so this looks like leakage from the DDS pre-filtering, and some aluminium is the pragmatic answer to get the issue down to acceptable levels. As an experiment, I have just grounded the 4 crystal cases (recognising this may change the characteristics of the passband a little) - and it has improved the performance by 15dB - putting the two radios in line with each other. 73, Andy, G4PIQ From: [email protected] <mailto:[email protected]> <[email protected] <mailto:[email protected]> > Sent: 01 January 2021 18:31 To: '[email protected]' <[email protected] <mailto:[email protected]> > Subject: RE: K3 Tx Spurious on Specific Frequencies on 160/80/40 Thanks Rick. To answer those questions. I'm measuring this with the radio into a power attenuator and then with additional attenuation into a Perseus SDR operating as a spectrum analyser - well below its overload threshold and with plenty good enough phase noise and dynamic range to see this. I've also confirmed the signals are there on the second K3. I ripped out all the normal station cabling as part of the testing - so - yes everything is tight. I suspect I haven't noticed this before for two reasons. Firstly - the level of the spurii are very variable by specific frequency in the band - so you have to be a bit unlucky - and unless you're listening for it yourself on site - most times it would be too weak for people to comment / skimmers to pick it up. Also - the times I have run this radio with a second receive in band while transmitting it's normally been with the run radio above 7025 where the spurii are at a much lower / no existent level. Secondly - I wonder if something is degrading in that radio. The fact that the spurious level changes significantly with drive levels suggests this is being generated by a non-linearity in the PA chain, and I wonder if I'm short of gain somewhere in the driver strip so the ALC is forcing the low levels in the strip to be driven harder. It is present on both radios, but is 15dB worse on one radio than the other. Andy, G4PIQ ------------------------- NK7O wrote :- I don't have an answer, I have no test gear that would even allow a look (other than another radio). But I'm curious how you're measuring and observing (not doubting) these results. Except for the reports seen on the air to validate your observations; if the radio change barely affects the results and the condition does not, it's either a failure of both radios or the testing environment (i.e. transmitting into the same antenna). What happens with changes in power output, different (dummy) loads? Then the eternal question; since this hasn't been noted before at your station; what else is different or has changed? Have you tightened your antenna connectors lately? HNY, Rick NK7I On 1/1/2021 9:30 AM, g4piq--- via Elecraft wrote In general I've found the spurious performance of the K3 to be pretty good, but I've noticed something today that I thought it would be worth asking more widely if anyone else has seen and fixed. I was calling CQ on 40m around 7014 kHz and I noted that several skimmers also reported me weakly on c. 600 Hz either side of this frequency. Sometimes skimmers have bad spurious responses, but seeing this on more than one skimmer caused me to look more closely at my transmit spectrum. I discovered that - particularly on 40m - there were specific frequency ranges where a pair of sidebands appeared that move at 150% of the change in carrier frequency. So on 7012 kHz they are at roughly +/- 500 Hz, at 7011 kHz they are at roughly +/- 1 kHz, at 7010 kHz they are at roughly +/- 1.5 kHz. They cross over with the carrier just below 7013 kHz. Similar patterns recur in other parts of 40m, but they are strongest near the bottom of the band. There are dramatic step changes in the level of the spuria at the boundaries between synthesiser segments - So I see spurii from 7000-7004, none from 7004-7009, spurii 7009 -7015, etc. They peak in strength when they are about +/- 500 Hz from the carrier, but drop significantly in level as you get further from the carrier. The level of these spurii is what was concerning - at their worst they were less than 30dB down on the transmit signal. They grow somewhat non-linearly with power. At 5W they are -40dBc, at 12W they are -32dBc, at 13W they are -35dBc and at 100W they are a little less than -30dBc. I can see the same style of spuria in some segments on 80m and 160m - but at lower levels. I assumed this was a fault with that specific radio - and in some ways it is - but the same spurii are present on both K3s that I have here - but on the second radio they are no worse than -45dBc. I've disconnected everything connected externally from the radio, tried different power supplies, and recalibrated the synthesiser and Tx gain with no significant change. Both radios have the original synthesisers - not the KSYN3A. Any leads from anyone on this one? 73, Andy, G4PIQ ______________________________________________________________ Elecraft mailing list Home: http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/mmfaq.htm Post: mailto:[email protected] This list hosted by: http://www.qsl.net Please help support this email list: http://www.qsl.net/donate.html Message delivered to [email protected]
