Response from NOAA + Boston WeFax

After my last blog – Weather Fax from Honolulu – I decided I’d drop the Weather HF Operations in Honolulu an email to see if they could confirm whether it was them that I received or not.

After a few hours I got back the following response from one of their Ops staff:

Aloha Tony!
In reviewing the schedules looking for that particular chart (Honolulu’s 24 hr Wind/Wave), I believe that is only transmitted from Hawaii. While there are Wind/Wave charts sent on all of the stations, the 24 hour Wind/Wave that is produced by WFO Honolulu is only broadcast on the Hawaii frequency.

From this, I’m happy to say it was Honolulu that I received and have noted it as such.

At the time they replied I was actually listening out on one of the NOAA Boston frequencies to see if I could get anything from them.

It was static on 9108.1 kHz (9110 kHz) when I started and was about to give up, when all of a sudden a transmission started part-way through a chart – just like they’d flicked a switch on that frequency transmitter. As you can see below, the chart just started.

Part-way through chart at 2222z on 12th May 2020.

I received a further two charts before the end of that days charts and the service stopped.

Quality not too bad, 4880 km away from here.

Again, all received on my Icom IC-R8500 and Wellbrook Loop, using JWX 3.0 software.

Weather Fax from Honolulu

I’ve been tinkering with Marine weather radiofacsimiles recently. I do this from time to time, especially when the weather is good here and I want to know when it is likely to end – more often than not, very soon, though I’ll have to admit the weather has been amazing in recent weeks.

I’ve used a number of decoders over the years and they have all produced some pretty good results, and each have their different features. Some work with SDRs whilst some don’t – or rather my set up doesn’t let them work with SDRs as I run everything through a M-Track 8 mixer.

Some of the decoders i’ve tried include – Sorcerer, MultiPSK and fldigi. fldigi has been my favourite for some time until recently.

The only receiver I don’t run through the mixer is my trusty Icom R-8500 receiver connected to the PC via the microphone-in input using a mono cable from the record-out on the Icom. I have to say this gives the best results with all the software tried. This is probably because it is less susceptible to PC jitter – when the processor skips due to lag – and the image is put out of alignment. I find, probably because my PC is getting past it, that the processor requirements to run the SDR software is enough to make this happen.

Any misalignment requires some Photoshop manipulation, often quite a few just to rebuild the image. Sometimes this is required anyway to recreate a split image – if the phasing was not automatically detected, the modem could not deduce the beginning of an image causing an an image which is horizontally shifted. This is very easy to fix with Photoshop if this happens.

Some of the software available does the rebuilding for you, but some don’t. Each can control slant error and suchlike, but each is different in doing this. And each Fax sender – Northwood (GYA), DWD and NOAA for example – generally works with a different slant requirement.

The latest software I’ve used – and I think the best to date – is JWX 3.0. I still haven’t managed to get this to work with an SDR but I have tried. But I do know it does work with the AirSpy HF+ Discovery as one of my contacts on Twitter – Gerhard Schweizer – has managed it after I pointed him to JWX.

Anyway. This is all leading to a search I was making for GYA’s fax’s on the 9th May. I wanted to get the 24, 48, 72 and 120 hour surface prognosis charts and had set everything up on 4608.1 kHz USB – the true frequency is 4610 kHz but you have to step down 1.9 kHz in USB mode for a proper decode.

Unusually I didn’t get anything, so after a while I decided to try the other frequencies used by GYA. This included 11084.4 kHz (11086.5 kHz) which I’ve never found anything on before, and even has a question mark next to it in the Wordwide Marine Radiofacsimile Broadcast Schedule produced by NOAA.

On going to 11084.4 kHz, to my surprise, I could hear a very faint signal though I realised quickly that the frequency was off a little. I wondered then if the GYA frequency noted in the NOAA document was incorrect. Turning the dial on 8500 it settled on 11088.1 kHz where a clear-ish image was coming through on JWX.

But the chart didn’t quite look right for a GYA one, it looked more NOAA like, and as it was at the end of the chart, the senders information and chart title box was starting to come through – along with a NOAA badge.

As well as the badge, the ID started to become visible and I was amazed to see NOAA/National Weather services Honolulu – 24 HR Wind/Wave Forecast – pictured below.

First chart received, in progress at 0756z

The transmitter site for Honolulu (KVM70) is located at 21°25’34.82″N 158° 9’12.36″W – next to an impressive weapons storage bunker area – on O’ahu. From my shack this equates to a range of 11,132 km! The schedule states that KVM70 uses a 4 kW transmitter.

In the NOAA schedule it does say many of these charts also broadcast from Pt. Reyes, CA and Kodiak, AK but even so this would be 8118 km from Point Reyes; or 7133 km from Kodiak. However, the Honolulu frequency isn’t listed under their entries in the schedule so I’m not certain whether they are sent in parallel from these transmitter sites at the same time or not. Certainly, the charts aren’t listed under either Point Reyes or Kodiak.

In all honesty, I expect that it is one of these stations that I received, but I’d like to say it was KVM70 🙂

VOACAP grey line at the time of the first reception.

Below are the charts received. As you can see, from the 0856z chart, the signal starts to disappear and they become barely readable – eventually just noise.

0856z 96HR WIND/WAVE FORECAST and the last readable chart.

I did set up the R8500 and JWX the next night to see if I received anything and it drew a blank, but another try today, on the 11th, did produce some weak results. The chart below is the best out of the bunch, received at 0715z. It also shows a good example of phasing error and cpu jitter – the vertical black line should be on an edge so the far left of chart should be on the right; the lower portion has jumped a little due to cpu usage.

The frequency is now stored in CSVUserlist so I’ll revisit it every now and again to see whether I’m able to get some more charts.