12m XTAL Controlled QRSS / WSPR / QRP / SDR

In a previous post, I described the use of CD-crystals for 17m weak signal operations. This post is attempting to do the same but for the 12m.
The crystals I am mentioning here are also used in CB-rigs, although harder to find. Here's would be a possible source.

QRSS: 37.340MHz => 12.446MHz x 2 = 24.892MHz
QRP/SDR: 37.350MHz => 12.450MHz x 2 = 24.900MHz
WSPR/QRSS: 37.395MHz =>  12.465MHz x 2 = 24.930MHz

SW+ 40 possible WSPR/QRSS mod

Regular readers of my blog will know that a PSK-Warbler could easily be modified into a 40m WSPR transceiver. I believe that Dave K1SWL even mentioned this mod during his talk at the FDIM-QRP-Convention.
Very unfortunately, the Warbler has been retired later that year. It seems the days for easy 40m transceivers are over.

However, there is hope, the 40m SW+.

The trx employs a 4MHz crystal filter for reception, since the SW+ is a superhet A1A transceiver, there is no if-filter in the tx-train.
The transmitter consists of the VFO, and up-converter (NE612), a link-filter, linear buffer and driver and a class-C power amplifier.

This is what I see, could be done to the rig... please grab a circuit diagram (can be found on K1SWL's homepage).




Concept


We want to have zero-beat at 7038.6kHz. So the sum of intermediate frequency and local oscillator should end up there. With the 4MHz intermediate frequency the kits comes with, we would need a frequency source of either 3038.6kHz or 11038.6kHz. Both are somewhat hard to reach on standard crystals.
Now, lets have a look what get's us close...
Lets check out the options:

1. 5.000 + 2.048 = 7.048
2. 5.200 + 1.843 = 7.043

Option one will require penning down the 2.048MHz crystal, however would open the option of using a 2.000MHz crystal for the lower band edge QRSS range. Additionally, a 2MHz ceramic resonator oscillator would enable operation in the regular CW portion of the 40m band.
Option two would probably work by just severely pulling of a 1.843MHz crystal.



VFO


This is were the mod begins. The VFO should be skipped; an external xo will be doing this job. Having the local oscillator external will provide some advantages for QRSS operations. First of all, temperature stabilization should be applied. Secondly, one may consider switching between several crystals.
Most importantly, an extra 7dBm output should be added for a transmit DSB modulator.


BFO

In a side-band transceiver, an offset between the BFO and the up-converter is not wanted. In a CW transceiver, one wishes to have an offset of about 800Hz between RX and TX. Since the SW+ is a CW transceiver, we would like to get rid of the off-set. The easiest way of doing this would be to buffer and feed the BFO into the up-converter.


Receiver

The receiver is pretty much were we would like to have it. All crystals should be changed according to the mod you would like to perform.


Transmitter

This is were the choice of option will become important. Buffer and driver look just fine, however, it is definitely required to convert the PA from class C into class AB for linearity. It might be easiest to just design a linear PA, potentially external, and feed the buffer into it. Another option would be to merge designs. Genesisradio.com.au once sold a qrp-transmitter kit, the Q5, which was equipped with a 500mW class A linear PA. The PA design itself genesisradio took from their SDR-TRX.

Superhet crystals for QRP

This post is incomplete and will be updated in a later stage.



160m

• 1.843

80m

• 3.560 = 13.560 - 10.000
• 3.560 = 11.059 - 7.500 = 11.059 - (15.000/2)
• 3.566 = 8.000 - 4.4336
• 3.554 = 6.5536 - 3.000
• 3.690

60m

• 5.360 = 15.360 - 10.000

40m

• 7.03(0/4) = 17.734 - 10.700
• 7.090(0/6) = 4.096 + 3.000

30m

• 10.106
• 10.116

20m

• 14.060 = 11.059 + 3.000
• 14.285 = 9.216 + 5.0688
• 14.28(5/8) = 12.288 + 2.000

17m

• 18.086
• 18.096 = 14.000 + 4.096
• 18.096 = 16.000 + 2.097
• 18.106 = 10.106 + 8.000
• 18.130

15m

• 21.060 = 11.059 + 10.000
• 21.28(5/8) = 12.288 + 9.000

12m 

• 24.906
• 24.950

10m

• 28.060
• 28.365

6m

• 50.060
• 50.285

2m

• 144.060
• 144.285

Frequencies in MHz. Italics denote J3E frequencies.

new blog created

Dear ham radio operator,
Lately I filled this blog, my radio blog, with some computer / IT related posts. This was due to a lack of a dedicated place to publish those posts.
It seems inappropriate to continue with this practice, hence, I decided to create another blog concerned with this sort of technology:
http://homebrew-it.blogspot.com/
Hopefully I will have to report something radio in the near future.
73

ccrrrcrcrc crcrrcccrccc ... SILENCIUM!

My workstation employs a mouse which performance-wise, I really like. It is a super cheap Logitech M100. The only thing that is really annoying is the sound and cheap feel of the scroll wheel.
I know, this is a matter of taste. However, here's one to try for yourself (and this may apply, mutatis mutandis, to other computer mice too!).

The feel and sound is created by a spring mounted against the inner corrugated surface of the wheel (of course, first you need to remove the single screw on the bottom of the device... you know all about this....

overview of the mouse

close-up of the click device

Key point of this very simple mod, compared to other attempts available in the mists of the internet, is to remove the bloody click thing all together.

In order to achieve the goal, the only thing to do is gently slide out the scroll wheel assembly. There is nothing holding down said assembly, just grab it and off it goes. The photograph below shows the assembly, the clicky spring device still in place, whilst the load spring held on the axle just fell off (this spring we actually want to place into its original place when assembling the mouse again!).

the wheel assembly taken removed from the mouse, note the load spring

Now, gently pull the actual wheel from the assembly. The clicky-clacky (ccrrrcrcrc crcrrcccrccc) spring will most likely fall off in the process. Do to the severe complexity of this stage, I forgot to take a picture... sorry for that ;-)

The following step will be to put the wheel back in its original position in the assembly, w/o the crcrcr-spring of course.

As a last step, we slide the wheel assembly back by using its guard rails. Mind the "scroll wheel click load spring"!

now w/o the noise device

Again, I leave it all to you to put the lid on it all and screw it all down by the single mounting means we had to dismantle in the first place.
As a result, I feel personally very positive about the modification. Not only is the bloody noise gone, the now freely spinning scroll wheel provides a real smooth experience.

Adding noise to the server room

The last few days have seen me changing my entire IT, not in terms of the system, rather the location. A land line connection needed to be pulled for the phone (skype - land line DECT combi phone), the RED ethernet had to be rewired and finally an ethernet switch was made redundant.

All in all, I am still running the following system:

• smoothwall express with RED and GREEN connections, running a squid caching proxy
• my old WiFi-router as an wireless access point (WPA2 of course)
• the good ole ACER H340 running NAS4Free with a RAIDZ-1 ZFS pool using four 1TB disks, serving as a media server, a file server and a backup server
• an AMD E-350 system running FreeNAS with a RAID-0 scrap data storage with two 2 TB disks
• my good ole Buffalo NAS (1TB)

All storage devices are running NFS, SMB and AFP, expect for the Buffalo, which has nof NFS.
To me it is somewhat clear, the H340 is the main, safe, data storage. However, it happened today, one of the 4 HDDs stopped its service. We had a nice warm summer day, temps at about 27 centigrade. The HDDs in the H340 were running at about 50º centigrade, which is somewhat warm.
I shutdown the server, let it sit to cool for some while and fired it up again. The HHD was back, however, some data was potentially written when the HDD took a break. Luckily, I am running a ZFS pool, hence, the disk could be fixed, in ZFS this is called "clear". Everything is back to fine again.

The incident made me rethinking about the H340. There is a post on the internet somewhere, in which someone discloses a mod which changes the H340's case fan.
That seems to be exactly what I need to do, I figured. Indeed, the H340's regulated case fan sucks air out of the case. The fan is arrange such, that air is forced along the HDDs, before exiting the case. Well, that makes some sense :-s Drawing a regulated low amount of hot air along some disk drives...
In my scrap box, there still was this unregulated fan which came with the Antec water cooling used in my workstation. This thing is a real blower, however quite noisy, hence, I did not use it in my workstation PC. However, now that all my servers are in a dedicated noisy room, why not using it in the H340.
This noisy Antec fan is now doing service to blow cool air into the case, over the HHDs, with the following impact, on that same hot summer day (eve):

Temperatures before mod:

• CPU: 45º C (essentially independent of the ambient temp)
• HDDs: 47º - 50º C (during idle)

Temperature after mod:

• CPU:  24º C (essentially the ambient temp)
• HDDs: 29º - 36º C (during a ZFS pool scrub)

And here is the downside of it. A very quiet device device is now turned into a very noise but well cooled NAS.
Mind you, the H340 is a perfect machine for running NAS4Free! (see earlier post)

Turn Your Old WiFi-Router Into a Wireless-AP

You may have noticed that I am busy renovating my IT. Some days ago, I posted a short text about the use of smoothwall as a caching web-proxy. Now it seems time to also hand over the routing to smoothwall. At the time of writing the last post, I simply hooked up my WiFi-router to the smoothwall box, telling it to get WAN from it (by DHCP). The rest of the routing was done by this trusty but old router. Downside of this router, it is equipped with 100Mbps only, while the rest of my wired network is 1Gbps.

Went shopping today, and grabbed a simple 5 port Gigabit switch, which is now connected to the smoothwall box.

How to get WiFi? I still could use the old router and tell it to get WAN from said switch. To tidy up address space, it would however be nice to have an Wireless access point in place of a WiFi-router.
A quick search in the internet revealed this page:
http://www.smallnetbuilder.com/wireless/wireless-basics/30338-how-to-convert-a-wireless-router-into-an-access-point
Very cool stuff! Works like a charm!

The next step would be to loose the cheap Gigabit switch again and integrate everything in the "production environment".

WiFi for Hacks

The dear reader may have noticed that a concept called hackintosh caught my attention. Actually, some years ago, when Leopard (OS-X 10.5), the thirst OS-X running on x86 hardware, came out, I already built various systems (Intel Q6600, Intel E4600) capable of running OS-X.
That gets us what a hackintosh actually is. It is a PC, mainly based on Intel processors, which is able to run Apple's OS-X under certain circumstances. To learn more, please search the internet, also about the legal implications/requirements.

The whole trick about hackintoshs is to find the right hardware, being compatible with the original OS-X. Mind you, the name hackintosh refers to the hardware, not the operation system, meaning, that the OS remains absolutely original, i.e. unchanged and non-hacked.

There are to mini PCI-e cards which are fully compatible with OS-X 10.8 that I know of (and actually tested):

1. Broadcom BCM94322HML 
2. Atheros AR5BHB92

My hackintoshs are based on GA-Z77N-WiFi and GA-H77N-WiFi mobos. Both boards come equipped with an Intel mini PCI-e WiFi/BT-combo-card. Whilst BlueTooth is working fine under OS-X, the WiFi part of this card is just dead.
Some additional info here, the Intel cards are equipped with 2 coax connectors, which is reflected by the 2 aerials provided with the mobos. Both cards mentioned above have 2 coax connectors.

From here, there are two options:

1. keep the Intel card for BT and use something else for WiFi
2. exchange the Intel with any of the two mentioned above and use something else for BT 

I personally went for option 2. And here is why:
There is a TPlink WiFi PCI-E card which is 100% compatible and works really well. However, there is only one PCI-E slot with the mobos mentioned above. Using this slot forces the use of the low power on-board graphics. There are other options like compatible USB WiFi devices, however, I believe that the bandwidth would be compromised here.
The second option provides full bandwidth for WiFi, leaving the single PCI-E slot for a GPU. On the downside, one now has to look for an alternative BlueTooth device. It happens that I own a couple of old USB BT devices. All those devices seem to be compatible with OS-X. Due to the low bandwidth of BT, I can easily live with the USB solution.

Modding Stuff for Astrophotography

Dear regular reader, by now you should be aware that I tend to think outside the box. My affiliation to astro-photography should be known, at least to readers of the knightsQRSS list of a couple of years ago, when I proposed stacking spectra to improve the SNR, as it is done in webcam-astro-photography.

So, here's another one for you, now we are crossing computing, the extreme corner of modding that is, with astronomy / astro-photography.

As we all know, human (biological) night vision (scotopic vision) is controlled by a molecule called "rhodopsin" which takes a couple of ten minutes to build up (cf. adaptation). It triggers the rods to go into night mode. In human vision, the rods create a gray-scale visual impression, like a good ole B&W-TV. The problem with scotopic vision is, that is disappears when a certain amount of light within the scopic wavelength range is seen (literally). Luckily, however, the scoptopic range ends at about 620nm. Meaning that red light will not affect scoptopic vision, i.e. red light will preserve the rhodopsin level in the rods. This is actually the reason why the night illumination on yachts, warships and vessels alike is red.
Now, check out the spectrum of red LEDs.

Now, lets turn to computer modding, something I never actually understood. To me, a computer has to perform and that's about it. Overclocking is somewhat understandable, so is watercolling, which are both concerned with performance.
Modding seems to be about visual appearance of a computer, which I personally never cared about. However, look at the recent IBM mainframes, i.e. zEnterprise architecture, these boxes are all odd, prismatic and what not. Did IBM copy the needs of modders to apply it to boxes worth millions of bucks? It seems so!
Modders are also interested in illumination effects of their boxes. Here it starts becoming interesting for the astronomer. Next to mean green and cool blue, there also is hot red illumination available for the modder.
Check this out!

Here comes the link between the two, modding and astronomy: in a modern observatory a computer is mandantory, to check sky-charts, control the dome, control the telescope mount or record data from a digital camera. A computer is found in every observatory I know, be it professional, be it hobbyist, be it amateur.
So, instead of illuminating the inside of your computer, you could equally use the device to illuminate the outside of your computer sitting in the observatory, hence the observatory, with a chain of red LEDs, preserving your scotopic vision during your astro session.

Delidding Photos (i5-3570k, i3-3225)

As promised, here are some photos taken during the process of delidding Ivy Bridge CPUs.

beginning to de-lidd the i5-3570k, knife stuck good, grease still present

different corner now, grease removed - mind you cutting the corners first does the trick

heat spreader and PCB apart - nasty black glue and grey thermal paste all over the place

glue removed (finger nails)

thermal paste removed by means of "akasa TIM clean" (citrus based)

same done to the i3-3225, although this was the second delidding, I cut myself, which did not happen during the first