Levels measured in a Softrock TX, Diode Probe, 1dB Point, Faultfinding.

Some checks if the Softrock is not working.

This also covers all other Softrock TX stages, just different component numbers.


I decided to record these measurements from a working Softrock. I actually used a v 6.1 but they all have similar circuits.

I used IQout http://www.m0kgk.co.uk/sdr/download.php to produce an IQ signal.
any SDR TX program may be used but this makes it easy to set.
These are my settings. I chose a fairly low frequency as many digital meters on the AC range will work at audio frequencies. 
Select your soundcard and drivers in use.
Connect the Softrock TX input labelled "Line Out" to the computer Line Out.
Connect a power meter and dummy load to the Softrock output.
Power up the Softrock and USB.  GFGSR http://home.ict.nl/~fredkrom/pe0fko/CFGSR/ may be used to set the frequency to the lowest range covered. The Tune tab has buttons to make band changes quick. AND on the "Test" Tab it has a PTT button.

The Ensemble does not have COM port PTT so  you can use a cliplead to ground the top left-hand lead of U4 (Pin 4), activating PTT. Thanks for the pictures Robby! BUT see CFGSR above! http://www.wb5rvz.com/sdr/ensemble/index.htm

TOOLS:- The AC range of many modern digital meters is specified to work at audio frequencies. If you have one of these then it will probably tell you the approximate audio  RMS voltage present.

For the RF measurements this can not be used. An oscilloscope or just a simple diode probe will give a good idea of the levels present. The oscilloscope and it's probes will need to be specified to work on the test frequency. A diode probe should cover any HF frequency.
Diode probe example http://www.n5ese.com/rfprobe2.htm  (Lots of things on this site!) This will read RMS with a 10M Ohm digital meter. Multiply by 4 if you have a 1M Ohm meter.
Below, my build of one of these.

A piece of copper clad board cut to fit inside a pen shell. Lands made with small saw blade.

Components soldered. A stainless pin or needle may be difficult to tin with solder. A suitable flux may be needed.

Assembled. Not seen is the ground wire and clip that goes to ground on the item under test.
This probe actually has slightly different components and seemed to work better.
It is a single diode probe with a low-value resistor, not critical, 15K used here. It will read PEAK voltage.
A complete unit. This one is an integrating probe, a series 47K resistor followed by a 0.1uF to ground. The red clip is the ground connector, I'd run out of black sleeves! This may be used for measuring DC voltages where RF is present, like on the Softrock divider and mixer.


Set IQout "Volume" to a low setting, start it and enable PTT. Gradually increase this level until 500mW (1/2 watt) output is achieved. 500mW is 5V0 RMS, 7V1 peak measured at the antenna terminal with the dummy load in place. If this is not reached I hope the measurements will give a clue where to look! Note:- Do not exceed an audio input level of 0.7 volts RMS.
Set the audio to 0.5 volts RMS and see where they vary greatly from mine, below.
I have taken the schematic from Robby's SDR site. http://www.wb5rvz.com/sdr/ The red dots show where the exposed "hairpin" of the resistors are.

Audio voltages may be measured at the hairpins of R26, 25, 28, and 27.
For 500mW out I measured about 1V2 peak-peak on an oscilloscope and about 0V5 RMS with a digital meter.
Voltages at pin 2 and at the audio inputs will be similar.
If the input voltage is lower than this then check the soundcard mixer settings.
Some cards will go much higher,
March 2015:- Note tests done by Warren, 9V1TD indicate that to conform with FCC spurious emission requirements this level should never be increased above 0.7 volts RMS. If 1/2 to 1 watt output is not obtained then it is likely there is something wrong. For the 15/12/10 metre versions more drive may be used, the spurious signal is in fact lower at higher frequencies.


With an oscilloscope:- Voltages  PEAK TO PEAK. Make sure the oscilloscope and probes are specified for the frequency used. Frequencies higher than this may be used but sensitivity will be less and readings lower.
Normally a x10 probe will be used to obtain the specified frequency response. Check your measurements using the scopes "Calibrate" signal or put it to DC input and check the 5V line or a battery to be sure you are reading correctly.

Transformers will show a reading on three of the six connections, at one end of each winding..

One of my diode probes measures approximate PEAK voltage the other RMS. [RMS probe 0.7 times lower voltage. These RMS voltages were confirmed using a Bradley RF voltmeter with an unterminated probe] I found my RMS diode probe gave unreliable results at T2 where voltages are low.
Tests at 3.5MHz.
 Test Point  Oscilloscope Peak to Peak  Probe PEAK Reading  Probe RMS Reading
 Each End of T2 Primary. (R30,31)  0V4  0V3 0V2 
 T2 Secondary (C20/L1)  0V6 0V3 0V3 
 Junction of L1 and C21  1V0 0V5  0V4 
 Q6 Emitter and Base  0V8 0V5 0V4 
 Q6 Collector  4V0 2V6  1V9 
 T3 Primary (R44 hairpin, same as Q6 collector)  4V0 2V4 1V9 
 Each end of T3 Secondary. (Q7,8 Gate)  2V0 0V6  0V7 
 Each End of T4 Primary (Q7,8 Drain)  6V5  4V1  3V0 
 T4 Secondary   8V0 5V1 
 Antenna (With 50 Ohm Load)   6V5 3V6 

These points are easily identified underneath the board, not so easy from the top but at least some of them are accessible. You do not need ALL these measurements. 
All measurements need a fine probe and care must be taken not to slip and short two points together.
NOTE Some of these readings do not seem consistent. Take them as approximate. The important thing is that they should increase as you progress through the transmitter.

I did some tests on my Softrock 6.3, which is really the same as an Ensemble but with plug-in coils. This seems to show that the Softrock TX is limited on the high bands.
These are derived from the amount of drive required using WSPR. This has a drive control calibrated in dB which seems reasonably close. (I measured -29db over it's range of -30dB, most likely the 1dB difference is my error.)
Using a fixed audio level I recorded the following WSPR attenuator values.

 1 Watt out on 80/40m  -11dB
 1 Watt out on 15m  -8dB This was the 1dB compression point
 1 Watt out on 12m  -5dB
 1 Watt out on 10m  -1dB The 1dB compression point was at 500mW so 1 watt required an extra 8dB.

I measured at the equivalent points below. Again the measurements from the probe, probably a different one than used above, do not match the scope readings at low voltages. 
Test at 28MHz.
 Test Point  Oscilloscope Peak to Peak  Probe PEAK Reading
 Each End of T2 Primary. (R30,31)  0V7 0V6
 T2 Secondary (C20/L1)  1V4 1V6
 Junction of L1 and C21  

 Q6 Emitter and Base  2V0 1V6
 Q6 Collector  5V0 2V8
 T3 Primary (R44 hairpin, same as Q6 collector)  

 Each end of T3 Secondary. (Q7,8 Gate)  3V0 1V4 
 Each End of T4 Primary (Q7,8 Drain)  8V0 4V9
 T4 Secondary  11V0 5V4
 Antenna (With 50 Ohm Load)  12V0 6V3

Driven to about 500mW on 28MHz. The audio drive was 1V8 peak to peak on the scope. This is only just below the 0.7 volts RMS maximum allowable figure.

You will realise that a diode probe is ideal for setting low TX powers. One may be made with the addition of a 50 Ohm load with a coax lead for the Softrock output. The dummy load may just be a 1/2 watt 50 Ohm resistor or a parallel combination. Even 4 off 220 Ohms and a 560 Ohm in parallel would suffice.

My method of checking the 1dB compression point.
This is a test to determine the maximum linear  output of an amplifier. An oscilloscope will not show small levels of distortion.
I measure the RF output across a 50 Ohm load reasonably accurately. I use an old Bradley RF meter, convert to dB.
A diode detector, like the probe above should measure PEAK RF voltage with a 10M Ohm meter. But check this. Calibration may still be necessary. 
There are lots of conversions two here http://www.tecmag.com/pdf/dbm_v.pdf Peak V = dB = mW
http://www.eeweb.com/toolbox/rf-unit-converter  Enter a figure, the rest is calculated. 
Using something like IQGen (see above) I increase the drive dB by dB. As power increases there will come a time when the output does not rise a whole dB. When the RF output lags the input by 1 dB then that's the 1dB point. 
This occurs at something like 1 watt on some low bands on some of my RXTXs but not that much in the mid-high bands.
Then a scope may be used to check the level. Do not exceed this level for linear modes like SSB, PSK.

Warren's way of checking linearity with a second Softrock http://groups.yahoo.com/group/softrock40/message/70260

And Ed's way using a scopehttp://groups.yahoo.com/group/softrock40/message/70268

To get a rough check of whether a soundcard SDR receiver is working connect phones to the output of it. Tuning through a busy band should produce some high pitched noises that may well sound something like radio signals. Stereo phones should reproduce sound in both channels. If your ears are old this needs care! The high frequencies may not be heard, or maybe just in one ear. Use a good antenna. Ensure the band is "open". 

If you have now found the SDR is not functioning properly then fault-finding will be required.
If it is a Softrock see if Robby lists it here http://www.wb5rvz.com/sdr/ go through the various checks. Commonly faults are due to bad solder joints, bridges or solder splashes. Check with an eyeglass.  Sometimes component values are confused. Check ICs are the correct way round. Resistors should be measured before fitting. A fault causing no reception on RX/TX Softrocks can be due to faulty BS170s, used as switches in the antenna changeover. Bob's drawing of the switching in the PDF.   http://groups.yahoo.com/group/softrock40/files/G8VOI/    Ensemble RXTX PTT fault finding.
One way of testing can be to lift the end of L4 that goes to Q10 Q11. Connect the antenna to the free end.

Transmit output. To avoid any software an initial very basic check could be to feed audio to the Softrock's socket labelled "Line Output". Using the PTT method described at the top of this page two RF signals should be produced. Two signals because the audio will not be IQ, 90° separated.

If you still have problems join the Softrock grouhttp://groups.yahoo.com/group/softrock40/ and ask.
Other SDR kits will have their own group.

These notes refer, in particular, to Softrocks.

A minimum of a general purpose TEST METER for voltage, current and resistance is required. Most will use a digital meter. Be aware that some readings will vary between different Softrocks. Readings made where square waves are present, on the dividers and mixer, may vary a lot. 
METER PROBES. Use great care. Do not short out anything when testing a live circuit. Use suitable small tipped probes. Ensure the meter is on the correct range. Current ranges in particular should not be used for testing when voltage measurements are required. 
Also, a GENERAL COVERAGE RECEIVER may help to diagnose and confirm operation.
A SIGNAL SOURCE such as a transceiver running low power into a dummy load can be useful for testing and diagnosis. If not available a simple signal generator or oscillator will do. I use a DDS generator like this http://www.midnightdesignsolutions.com/dds60/index.html I also use one like this http://www.qsl.net/k5bcq/Kits/Kits.html This uses a Si570 so only works above 3.5MHZ unless followed by a divider. A 7474 as in the Softrocks may be used.
A pin held in your fingers can act as an audio generator. Test the Ensemble RX op-amp:- just touch pins 2 and 6 of the audio amp U11 with a pin. The SDR display will show the effect and you will hear a hum if you tune to the centre. If nothing is found check by touching the pins of the "Line In" jack.
Also, try U10. Pins 2,4,5,7,9,10,13,14 also produce a hum. This is a rough test of U10 and the components after it.

OSCILLOSCOPE Robby's construction tests shows the use of an oscilloscope. I do not consider this essential. Indeed unless you are experienced in interpreting what you see they can obstruct progress. Recently someone showed a superb oscilloscope picture of a faulty transmit signal. It was three weeks before someone interpreted it correctly. In the meantime a few simple checks would quickly have pointed to the faulty audio cable.
Very occasionally a scope may help.

A METER. Note the "PA Standing Current" test can give larger variations than quoted in Robby's pages. This is not considered significant. The specification for these FETs gives a large tolerance for this value. 
A RECEIVER IS ESSENTIAL in order to setup the image rejection. With SDR transmitters this is always a manual adjustment.
A signal source will be useful, as above.
Oscilloscope, same comments as above.