Sunday, 22 January 2017

TS590 climb down, and WSJT-X 1.7.0 outrunning my PC.

"Oh Lord, it is hard to be humble
when you are perfect in every way"

So sang Mac Davis many years ago. I have to agree with him of course. I am not quite so sure about the next bit of that song.

"I can't wait to look in the mirror
'cause I get better looking each day."

Anyway, it is so hard to admit to imperfections (especially as I am so close to perfection almost all the time).

There are a few things in my life which, on reflection, I might not have done the way I did at the time. That is, knowing what I know now.

I suppose, to mention a few, I could have done without getting robbed at gunpoint. Or maybe I could have handled it better. I was careless enough not to take notes about the gun, as the police asked me to describe it later. And then there was getting my upper jaw broken. I might have been better taking the ambulance to the right hospital, rather than getting taken to the nearest one by car. I did not want to travel in the same ambulance as the woman who caused the broken jaw, so I took a lift instead. At the nearest hospital they called out the duty dentist who tried to stitch together the two bones by trying to hammer a needle and thread through the upper jaw bone. Having a needle hammered in the roof of my mouth was painful. He bent several needles before he gave up. I think all the broken teeth were getting in the way, and the bone wasn't fixed of course. You cannot stitch bones together, nor hammer through solid bone, or not easily. I should have had the composure to point that out to him.

They had to break the jaw again later and re-set both parts. So while my upper jaw was wired to my lower jaw as a splint for 6 weeks, I could not open my mouth. CW only for 6 weeks! What a joy.

Surely then I have to admit that everything in my life has not been perfect, and there are a few things I might have done differently. Like not breaking the jaw in the first place, or not getting knocked unconscious on a different occasion.

There are other things, but they are far more gruesome and therefore not to be mentioned here.

Yes, despite my near perfection of judgement, not everything in life has gone spiffingly.

So I can now admit that selling the Kenwood TS-590 3 years ago was not my greatest judgement call. I was young and foolish. So, given a bit of prompting lately, I have bought another one.

I will report on the TS-590 later. This one is second hand, but very low mileage. It is a TS-590SG, which has a few tweaks compared to the 590S which I had before. The differences are slight, but they all seem to be useful for me. Maybe that is why I changed the 590 for a newer one.

It was quite a struggle to get the CAT control to work, but finally after a full reset of the rig and changing the port settings over to "Hardware" handshake everything started to work.

First observations are that it works beautifully. The receiver is a joy, as you might expect looking at the specifications. For the type of operations I do it needs the optional TCXO oscillator as it seems to drift quite a lot on 6m after switch on. For normal SSB use this would not be such a problem, as we are talking about 100hz or so. We shall see whether the TCXO does the trick.

I have been warned off cheap TCXOs being sold on eBay. It seems that they generate a lot of phase noise. Going onto YouTube I found an interesting post showing someone carving up a generic eBay TCXO. It looked like a crystal oscillator with a frequency divider. I could not see how it was compensated, and it performed badly compared with OCXOs. We will see how I get along with the official Kenwood one when it arrives. I am not against adding an oven as well, if I need to, but let us wait and see.

Will I regret selling my FT-450? Maybe because there was nothing wrong with it. I have not actually sold it yet but it is on eBay and I have a bid, so it will sell in due course. What it lacked were the transverter sockets, plus it is generally a much lower-specified rig than the 590. Very good value all the same and it came fitted with a tcxo.

So there you are I have admitted to my mistake and corrected it.

Hey, now that I am not claiming to be perfect, I do not have to follow the song's theme and it should be easy to be humble now.

No, I think I will give humility a miss. Where is the mirror so I can admire myself?
WSJT-X 1.7.0 is a great thing, but it is running rings round my computer.

I have mentioned this before, but I have now examined things more carefully.

Here are two plots showing the received signals, both running simultaneously, receiving noise on MSK144 mode. The top one using MSHV, and the lower one is using WSJT-X...
They are both set, of course, to 15 second receive segments. But if you look a the lower of the two traces on each, MSHV manages to fill the 15 seconds time with received signal, whereas the WSJT-X one stops short at just over 12 seconds. This is strange, as the received audio was timed by me at exactly 15 seconds in both cases.

It is the same with the both rx segment traces. I could not quite get to the full 15 seconds on the upper ones as that would have cancelled the trace, but on MSHV I was getting almost all the way to 15 seconds, whereas on WSJT-X I could only get to just over 12 seconds again.

So the WSJT-X trace only gets to around 12 seconds along, even though this represents 15 seconds of recording. It seems to me that either the software is missing the packets or it is not displaying then. The timing cannot be wrong when the tx/rx switching is happening at exactly 15 seconds every time, whatever the display shows.

The situation gets worse if I increase the load on the PC by, say, widening the FTol to 200, and it gets a bit better if I reduce it, such as by reducing the decode depth. The load as shown on the processor varies between 40% and 80%. Setting FTol at 200 pushes it to 80% most of the time.

It is pretty clear to me that WSJT-X 1.7.0 puts too much load on my processor, whereas MSHV, which does not load the processor so much, does not.

This effect is present when I only have one instance of WSJT-X running on MSK144. It does not get worse if I run several instances running WSPR at the same time, but it does get worse running two instances of WSJT on MSK144. Leaving the multiple instances issue to one side, and only running one instance, WSJT running MSK on its own has this problem. If I close down any processes and lighten the load on the processor, the trace goes further along the graph, if I increase load the problem gets worse.

I doubt very much if this is actually affecting the processing of WSJT or the decoding of signals. I guess it is just the display which is getting behind. Nevertheless, it is not a good sign, and the PC has crashed a couple of times for no good reason other than it was using a lot of processor capacity.

The actual PC display goes through a separate IDE display board intended for gaming with its own processing on board. I doubt if that can be at fault as the MSHV display is fine.

This computer is no slouch. Well, not by the standards of a man who once owned a Sinclair  ZX81. It is an ASUS board with Windows 10 Pro, AMD FX 4 core processor running at 4.2GHz, fully loaded with 8Gb of RAM. I have added 4Gb more space using Ready Boost and a USB flash card. I think it has gone as far as it can go, and the basic problem is that the machine was never configured for speed. It was always intended as a general purpose computer.

So I am working on transferring in a machine used for video editing, which has more RAM. Having been built by me originally using gaming standards it was generally built for speed. I think its weakness is in the disc transfer speeds but let us see how I get on with higher processor capacity.

Buying the TS-590 and the TCXO has emptied my wallet. I need the TCXO as the TS-590 is not stable enough to turn down the FTol on WSJT, and that in turn is overstressing the PC. The two issues are linked.

I do not think that WSJT-X was specified to require a gaming computer to drive it. I fancy that this is exposing some overall weaknesses in my general "office" computer. But nevertheless the arrival of the TS-590 and WSJT-X, and my ever-increasing processing demands, are all pushing me into investing more money. So paradoxically, I cannot now afford the 2 metre transverter, which is why I bought the TS-590 in the first place.

I might just go back to an FT-101E. Never had issues like this then.




  1. Ha! Nothing wrong with admitting a mistake, "The most valuable thing you can make is a mistake - you can't learn anything from being perfect" - that's what I tell the wife.

    Nice radio though, I am thinking of trading in my IC756pro2 for a 590sg.

  2. I bought a used TS590s and i love it so far. Not able to get FT8 to work on it yet though. N1SFE

    1. Paul. Sorry to hear you are having problems with FT8 and the TS590. If you wish to drop me an email via the address on I'll try to help. 73. Jim