With description and schematic of a simple QRP transverter with just 6 active parts
Click on any image to enlarge it
This page is more about my progression on 137kHz rather than an expert's guide. If you are like me and want to experiment and find out how to make improvements then keep checking back to see any progress I make. My ethos has always been "keep it simple", so if I can make something simple work on 137kHz you'll find it here. My real challenge will be getting an effective 137kHz antenna in my small garden without upsetting the neighbours and XYL. I succeeded with a 5m long antenna on 500kHz, so some results are certainly possible, especially with modes like WSPR and QRSS.
Jan 25 2013
For the last few days I have been running my 8W (RF out, not ERP) QRSS3 beacon on 137.766kHz into the 20m spaced earth-electrode "antenna" and waiting for reports. Yesterday I got comprehensive reports from G4FEV (56.7km) and G3XDV (61km) who both copied my signal very well.
I also did some field measurements to establish the ERP of the beacon in the best directions of the "loop in the ground" by comparing the field strength against that of DCF39, the German commercial station just outside the band. Using a well known formula, it is possible to work back and get an ERP figure. Here is how:
Answer? My measured ERP on 137.766kHz is 6.6uW
Jan 22 2013
Today saw my return to 137kHz with the testing of an 8W CQ/QRSS beacon transmitter used in conjunction with my 20m spaced earth-electrode antenna in the garden. Also tested was a new E-field probe attached to a mag-mount with 15cm whip mounted on the car roof. This EFP has proved to be excellent and allows me to do local (up to around 15km) drive-around tests to check coverage when coupled into my FT817 and a laptop PC running Spectran.
Details of the E-field probe is on my blog at http://g3xbm-qrp.blogspot.co.uk/2013/01/137khz-e-field-probe-mag-mount.html .
Sept 18th 2012
It is over a year since I last transmitted on the band. I am working on a new transverter for 472kHz and hope to adapt a version of this for 137kHz by changing the crystal and output network only. As part of the work in preparation for 472kHz I am either going to re-erect my loop (with thicker wire diameter) or a Marconi vertical. Not decided yet.
Watch this space......
Sept 8th 2011
Recently I have been updating my kit for 137kHz. I have now rebuilt the QRSS3/30 beacon to produce around 23W and the transverter used for WSPR has been improved to give 24W from the PA. So far only the beacon has been tested on air but reports from G4AYT (Whitstable) and M0BMU (Tring) today confirm the signal ERP is better. I estimate the ERP as being 0.2mW currently using my 80sq m loop in the garden. The photo below shows the latest version of the beacon TX. Some while back the original beacon was copied in Holland, so with the increased ERP further DX is possible.
The image below is of the latest version of the transverter which produces 24W from the (very efficient) PA. The heatsink hardly gets warm at all yet the 50 ohm load is VERY hot!
Feb 1st 2011
Bert PA1B has done some interesting "reverse analysis" of the WSPR reports I've given people. Working back from these you can estimate the lowest levels (in mW) that stations could have used and still been copyable with me. This also gives a good guide to the likelihood of these stations receiving my weak 50uW ERP signal. It is clear that M0BMU should be able to copy me signal frequently, which he does, but that I'd need a lot more ERP to be able to get reports from the PA3DGO, PA0A, DF6NM or SM6BHZ.
Yet again, this sort of analysis really shows the value of WSPR when making station improvements and comparing results. It is not perfect as there are other variables than just the TX station's ERP and the quietness of the LF receiving station (occasionally large signals nearby can screw up WSPR decodes for example), but overall it gives a clear indication of the work that remains to be done to get consistent results and reports from further afield.
My thanks to Bert for running this analysis. See Bert's QRPp website at http://a29.veron.nl/pa1b.htm.
Jan 30th 2011
Yesterday I combined the RX and TX parts of my 137kHz QRP transverter for the first time and all is working although a few dB in RX sensitivity is being lost in the ultra-simple RX-TX switching method. Despite this, I'm copying DL, SM (951km with 50mW ERP) , PA and G stations on WSPR. My best DX on transmit is still 148km.
Next I need to improve the antenna efficiency as power from the IRF510 FET PA is 8W, but the ERP from the vertical wire loop is just 50uW. Even so, WSPR reports suggest that 100km + should be reliably possible even with the present set-up. With another 6dB or so ERP I could be getting WSPR and QRSS3 reports from continental stations out to around 300km.
My loop ATU uses a capacitance decade box to select a value of C to resonate the loop: with high-Q loops, the voltages (even with QRP) can be very large in the loop and one really needs 500-1000V capacitors that can take the voltage and heat. At the moment the ATU tuning is not always constant as the caps in the decade box are getting a bit stressed. So, the loop ATU needs improvement or I'll try a vertical Marconi again with much more wire in the air and a large loading coil of around 4-8mH.
The latest 137kHz QRP transverter schematic is shown on the right. Click on the image to get it full size.
Jan 14th 2011
Today I re-erected my vertical loop for 137/500kHz and was rewarded with a string of reports from Jim M0BMU who is 69km away near Tring when I was using around 50uW ERP. Reports of up to -22dB S/N suggest that >100km should be consistently possible given some WSPR activity in range from stations using decent receiving systems in quiet locations.
Jan 12th 2011
My conclusions are that an ultra-simple QRP approach to 136kHz TX is just not possible if one is to achieve useful results consistently.
Jan 11th 2011
Today I did a recalculation of my 137kHz ERP using the loop antenna with my WSPR transverter. I have a small RF ammeter (ferrite toroid with 50:1 transformer) in series with the loop wire and I measured the current in the loop wire as 1.3A.
For 137kHz the ERP of a loop is = (I^2 * A^2) / (205 * 10^6)
where I = loop current and A = loop area (see "LF Today" RSGB book p101)
Putting in the numbers ERP = 1.3^2 * 80^2 (205 * 10^6)
So my ERP = 52uW on 136kHz.
A further 3dB could be obtained by thickening up the wire used in the loop taking the ERP up to around 100uW. Other improvements could come from increasing the TX power output from the present 5W.
Later today I'll work on a small ATU to resonate my 5m feeder to the 28MHz halo as a Marconi vertical tuned against all my grounds. First I will work out the antenna current needed to beat the ERP of the loop. At least a Marconi will not be directional, which is a drawback of the loop.
October 25th 2010
A fairly easy way to increase my ERP on 137.5kHz WSPR is to up the supply voltage to the IRF510 PA. Using a 20-24V supply I am able to get another 3-4dB ERP judging by the reports from M0BMU and G3WCD this morning. This evening I'll give it a go with the higher supply to see if I can get any reports from further afield.
Sept 26th 2010 wet and miserable!
137kHz transverter (provisional schematic
(click for full sized image)
Sept 21st 2010 dry and sunny
Real WSPR DX at last!!!
Just had a report from G3YXM who gave me -24dB S/N on 137.5kHz when I was running no more than 20uW ERP. He is now my 5th WSPR report on the band and the best DX by far at 148km. This is so thrilling! The transverter is just sat on my desk and the PA isn't even warm. My expectations are exceeded now by a LONG way.
Since starting on 137kHz WSPR just over a week ago I have now been copied by 5 different stations so far, but Chris G3YXM is definitely the best DX thus far.
Sept 19th 2010
Lots of activity on 137kHz WSPR today, but sadly I only got only two unique reports today, again from G3XIZ and G3XVL. With the ERP I'm running there's just not enough signal level for more distant stations to copy me. Already I've learned how far 10-20uW is copyable and this is not as far as I'd like! 100km is, I think, just possible on WSPR but only under favourable QRN/QRM conditions. Although I can see relatively simple ways of gaining a few more dBs of ERP I have to decide how seriously I want to take 136kHz now: to get more reports at greater distance needs at least 10dB more ERP I think to make breakthroughs in range. This implies running at least 20W RF (6dB) and another 3-4dB in antenna improvements i.e. some significant changes will be needed. Do I want to do this?
Sept 18th 2010
Chris G3XIZ has successfully received my QRPp WSPR beacon this morning on 137.5kHz. ERP is around 10-20uW from the class-E PA producing around 4-5W. The heatsink runs quite cool at this power, barely a few degrees above ambient thanks to the class-E configuration. Chris's most recent reports around lunchtime have been better with up to -20dB S/N. At this sort of level my signal should be copyable in the better directions (E-W loop orientation) out to considerably further, even before raising the ERP. Later in the day my first report on 137.5kHz WSPR from G3XVL near Ipswich, which is well off the best direction for my small TX loop antenna.
This brings my total number of WSPR reports on the band to 3 with the best DX on this mode now 61km.
Sept 17th 2010
This is the TX transverter breadboard currently. Note the T106-2 toroids and the TO-220 heatsink for the IRF510 FET
Sept 15th 2010
Well, I think I've discovered why my FETs have been dying during assembly: my trusty soldering iron earth was not earthed at all and I suspect that every (recent) attempt at soldering devices has been killing them in the process, probably by applying a large AC voltage between gate and source. I've now fixed the grounding problem and I'm happy to report that recent insertions have been fine. So, tomorrow I hope to get the QRSS3 beacon and WSPR beacon TX working and on-air.
Sept 8th 2010 dry
Well, I'm really on a roll now - blowing up FET PAs that is! Out of the blue my IRF510 (my last working one here) decided to die on me in mid beaconing. I then replaced it with an IRF530 but this took 3.5A and appeared dead on insertion, despite careful handling. 2 others followed and appeared dead. I think I have a bad batch as I cannot see anything wrong with the circuit there were fitted in. I'm told that the IRF510 is rugged and not prone to static damage because of the high input and output capacitances. My next batch of 10 is due tomorrow along with heatsinks. Better luck next time?
Sept 7th 2010 dry
Today I made my first proper attempt at a QRP class E PA design for 136kHz. Although the values in the output circuit seemed to tie up well with theory I wasn't able to realise the expected high efficiencies, possibly because of the gate drive design. In fact I managed to blow up my last IRF510 because it got very hot and this meant aborting a WSPR test (at a sensible power level). One resource recommended to me is http://www.alan.melia.btinternet.co.uk/classepa.htm which has a E class design spreadsheet. I also discovered today that 16mm diameter 3C90 toroids do NOT make good LF PA output bifilar transformers: they get very hot. The larger diameter 3C90s appear to be fine. Overall, a frustrating day but I have learnt a bit by making mistakes. I've ordered some heat sinks (to replace the croc clips!) and 10 more IRF510s to play with.
Sept 6th 2010 rainy
Also had this interesting analysis regarding the difference in WSPR reports (over a 3km distance) between 136 and 500kHz from VK2XV:
Many thanks for this Steve.
Sept 5th 2010 mild and dry
Just started up the WSPR beacon into the "in the air" loop antenna and running what must be between 10-20uW ERP. So far the only reports have been from my local friend G6ALB who is 3km away and he is giving me only -20 to -22dB S/N, which is not very good at this distance. ATU tuning seems less sharp that with my lash-up, so I need to see if the step-down transformer needs a different turns ratio. Otherwise, apart from rebuilding the ATU, nothing has changed. Adjusting the turns ratio a tiny bit has now got me -16dB S/N from G6ALB.
I've just discovered that the 3C90 cores (about 16mm diameter) I use in my ATU step-down transformer (top left in the picture below) is getting very hot, suggesting it is lossy. I have one much larger 3C90 core so will wind a step-down transformer on that instead in the morning and see if that is better. I'm surprised that at 5-6W into the transformer this gets hot.
This evening I've been running about 5W PA output (not ERP) WSPR from an IRF510 PA on both 136 and 500kHz WSPR and getting reports from G6ALB who is 3km away. In both cases I'm using the same ATU (just different capacitor settings) into the same loop antenna. On 136kHz the best report was about -20dB S/N whereas on 500kHz it was +10dB. This is a difference of 30dB!! For some reason, 136kHz seems a heck of a lot harder than 500kHz. The radiation resistance is considerably lower on 136kHz, so losses become more significant I guess, but the difference still seems to big.
Sept 4th 2010 mild and dry
The piece of paper at the bottom shows the values needed to resonate the loop on both bands (15.7nF on 137kHz and 1.2nF on 500kHz). It is simple, but works fine at the QRP power levels I'm using (5W max from PAs on both bands). My report from M0BMU at 69km on 500kHz WSPR was -8/-9dB S/N which is pretty good. I'll be trying the loop and ATU with my QRSS3 and WSPR beacons on 137kHz on Sunday and during the following week.
Sept 2nd 2010 warm and dry
Today I converted my earth electrode antenna with elevated feed wires into a full "in the air" loop for both 500kHz and 136kHz. As might be expected, tuning is now very sharp as there is no longer a 50 ohm earth resistance is circuit dampening the Q. At first I found matching and tuning impossible until I decided to match the loop with a step-down transformer (3.5:1 on a 3C90 core) and separately do the tuning. This works a treat and it is now a simple matter to tune the loop either for 500kHz or 136/137kHz. The tuning is done with a capacitance decade box with a 365pF variable capacitor to tune it "on the nose".
First reports on 136kHz with the
QRSS3 beacon suggest the signal is about 6dB stronger, which is a great
result. The loop is wound with quite thin wire (1-2mm diameter), so a
higher efficiency is possible with thicker wire, but with even sharper
Sept 1st 2010 warm and dry
Today I managed, I believe, for the first time to accurately measure the ERP of my QRPp system on 137kHz.This is the method used:
The test site is about 45 degrees off the main line of fire of the antenna, so in the best direction it could be 10dB (?) stronger, i.e. 20uW ERP giving an antenna efficiency of -53dB in the best directions. Frankly I'm amazed that anyone can copy this signal at any distance, so full marks to G3XIZ (48km) and G3XDV (62km).
G3XDV sent me some more screenshots of my signal during this morning's test and on one image you can, looking carefully, make out the "G3X" part of my callsign. Far from a strong signal, but there nonetheless.
Aug 31st 2010
Today I changed the earth electrode antenna by raising the height of the connecting wires to form a (part) loop in the air as well as in the ground. The loop bid in the air has an area of >100 sq metres. This should increase the effective loop area by a factor of at least 2, thereby improving the ERP. Later I'll repeat the tests with just the loop in the air only, so I have 3 different sets of results to compare. On 500kHz WSPR my signal with M0BMU (69km) was 6dB stronger than before with the earth electrode antenna. At the moment I am QRSS3 beaconing on 137.675kHz and seeking reports.
August 29/30th 2010
August 28th 2010 Dry
Having got the TX part of the 136kHz transverter going today I am, for the first time, putting out a WSPR signal on 137.5kHz. Power from the PA is around 4W and the ERP from the earth electrode antenna is around 20uW. I am hopeful that Chris G3XIZ in Biggleswade may be able to decode the signal - he was able to receive the QRSS3 beacon earlier today at a just not readable level (would have been OK in QRSS10) - and possibly others not too far away. In the picture the transverting bit (from the 10MHz band) is on the right and the PA in the top left with low pass filter visible using the T106-2 toroid. Note the heatsink for the IRF510 - a couple of crock clips as this was all I could find in the junk box!
On Sunday I will leave the WSPR beacon on during the morning. Next week I shall try to get a more efficient antenna erected, either the loop or a Marconi properly loaded and matched.
August 24th 2010 (again)
Wow - DX reception!!
Mike estimates another 6-10dB should allow him to read the message properly. Signals were weaker later, possibly because of rain here altering the effective loop size within the ground.
This is the first time my QRPp signals have been received at any distance apart from my own local testing sessions. It has certainly encouraged me to get a bit more ERP and the transverter build is now progressing steadily. I hope to erect the "loop in the air", as opposed to the "loop in the ground" within days.
August 24th 2010 Dry after heavy overnight rain
This test was after a night of very heavy rain and although this would improve the earth probe's contact with the soil it should, theoretically, reduce the loop area formed within the ground using an earth electrode antenna. If anything the loop current was slightly higher (expected) but the field strength was also good suggesting the loop area within the ground had not been significantly altered.
This is the Spectran display when using the 80cm loop to receive the signal.
August 22nd 2010 Dry after overnight rain
Today I attempted to calibrate my 136kHz 80cm receiving loop antenna by measuring the signal strength of DCF39 in Germany on 138.83kHz. I'm told that the field strength of this signal in daytime in this area is around 1mV/m. The level of DCF39 was -4.5dB on the Spectran screen with the settings I used. So, by extrapolation, 0dB on the screen should be close to 1.7mV/m.
I then measured the signal level of my QRSS3 beacon on 137.675kHz to see how much lower the level was at the same point about 1.5km away from the home QTH. My QRSS3 beacon was measuring -42dB approximately, corresponding to a field strength level of 12uV/m at 1.5km. To work back to my ERP from the earth electrode antenna I used the formula ERP = (E * d)^2 /49 and this gives a result of 6.3uW and an antenna "loss" factor of 56dB (power in to RF out).
There are several sources of error:
August 16th 2010
The 20m spaced earth electrodes "antenna" has about 30m of wire coming back at low height to the feedpoint upstairs. I wondered if I could increase the current flowing into this structure by tuning out the small inductance presented by the wire. Using a small capacitance switching box I selected the value that gave most current with the 2W beacon TX. It was 100nF although, because of the resistance of the ground return path the Q is low, the peak is small. This suggests the inductance of the feed wire is only 13.5uH. I also tried matching the loop using a 3C90 transformer with various secondary taps but this made little difference. So, conclusion is that resonating the loop and matching it makes very little difference to the current (and ERP) because of the low Q. With an elevated TX loop with low AC resistance in the wire, then matching the loop will be critical.
August 15th 2010
Having simply tuned and matched my 5m vertical coax feed to the 28MHz halo (the halo is used as a top capacity hat) I'm now putting out the grand level of 8uW (+/-3dB) on 136.93kHz in QRSS3 from my 2W out beacon TX using this somewhat small Marconi antenna. Tuning was with 1.2mH wound on a small piece of ferrite in parallel with some fixed and variable capacitance. Matching was via a 3C90 toroid transformer. Although the power is still very low, I'm hopeful someone may now copy this signal. I'd appreciate any screen shots and reports. I'll compare field strength levels in the local area (1-15km) against those obtained using the earth electrode antenna which got 14.9km a few weeks ago).
LATER: Field results were disappointing with the earth electrode antenna stronger. Back to the drawing board!
The next stage is a loop antenna that can be used on both 500kHz and 136kHz.
August 2nd 2010 - warm with light drizzle
Results were as follows:
Remember, this is ultra simple stuff:
Interesting to speculate what would happen with such a system at 73kHz. Pity there is no allocation there today.
July 30th 2010 - dry after a few rainy days
Today I increased the power from my QRSS3 136.93kHz beacon to 2W out of the PA into the 20m spaced earth electrodes. Another "ride about" test locally with the deaf FT817 and E-field probe RX antenna gave the following results:
At 2.4km good copy on the Spectran display
At 4.6km clear copy on the Spectran display.
At 8.6km the signal was detectable, but weak (see picture attached from Spectran screen).
Further tests will then be on hold now for about 10 days as my young grandson is coming to stay again. After he's gone back home I'll start on the Marconi vertical or a loop antenna, and of course on the full transverter.
July 24th 2010
A repeat of the test on July 19th gave less encouraging results today. Reasonable copy on QRSS3 at 1.5km but zero copy at the same spot 4.6km away from the QTH (within 20m). Not sure what has changed as the earth electrode "antenna" is the same. Maybe the soil is slightly more damp? My next stage (through August) is to make the beacon TX at least 5W out (IRF510) and repeat these tests.
After that I'll make a start on the full transverter to 10MHz so I can run WSPR both on RX and TX. The Marconi vertical I plan (in my head at least) will have a spiral top inductance, 5m vertical section, a 2mH coil (large diameter) a further run of wire to the last vertical down wire (2m long) with a 2-3mH variable tapped coil at the base. I'll have a couple of ground rods and some radials from the ground point This is about the best I can manage in the space I have.
July 19th 2010
In the meantime, if anyone near Burwell, Cambs wants to look for the QRSS3 signal near 136.93kHz I'll leave it running until about 8pm tonight when I have to switch off as my grandson is sleeping in the shack bedroom. Any reports would be amazing.
July 18th 2010
As I had no suitable toroids for the low pass filter on the output I used a small slab ferrite antenna which had some 50uH inductance wound on it.
Using my 20m spaced earth electrode "antenna" on TX and with a PA0RDT designed active antenna with a 19 inch whip antenna feeding into the (deaf) FT817 I could copy the 10wpm CW at 0.4km across the fields.