“Those who hear not the music, think the dancer mad”
FireFly SDR
FireFly Kit

FireFly 20M CW SDR

The world of SMD and SDR

FireFly 20M SDR Getting used to handling the components, but despite my best efforts the first C0805 cap went “ping” across the desk. I started with the Weller TCP iron just as an experiment, but as the board is tinned with silver solder it just was not hot enough and it all got messy.

Switched to my new Union iron and selected a smaller bit with one flat side. Tinned one pad then positioned and soldered one end of the cap, then the other and then re did the first end. Better, but it often seemed to take at an angle to the board, the side that was soldered first was always higher than the other end.

It must have been getting very late as I had not drunk any beer and thought the part would lie flat so had soldered the other end before properly checking it.

Only four SMD cap parts have been soldered so far, they appear to be electrically OK although not neat! I won’t attempt any more reworking of them at this point. But I will seek to improve the next ones, holding the cap “down” with tweezers was the problem, as the cap could not be held flat to the board.

PIC and DC checks

Ready for PIC and DC checks

Just getting ready for the first “smoke” test – The 6 and 9 volt regulated supply.

Passed the 6 and 9 volt DC checks.

Used the Atlas component analyzer to check the SMD values.

First CW Side Tone Monitor test

Current drain should be 13.5 to 14.5ma measured as 10.3ma

Side Tone circuit test

Completed the side tone circuit, connected an earpiece and powered up, the keyer sends FB - It works!

Local oscillator test.

Just about to break out my TS530S as it is the only receiver to hand that can RX 20M (My K2 is out on loan).

Connected up a short length of hookup wire to the TS530's antenna socket then powered up the Firefly, after a little tuning the local oscillator appeared. It still amazes me when this stuff actually works.

FireFly S9 Local oscillator
S9 local oscillator

90 Degree Phasing Section and LO Buffer Tests

This test must have got missed..

Detector Circuit Tests

Construction has gone well up to this point, all the previous tests were correctish - FB on power up, the local oscillator etc were all fine.

But now the FireFly is drawing 200ma+ when it should be 27ma. I carefully checked the connections / parts placement, and checked if anything that should not be is grounded. But it is unresolved.

Voltage at either side of R4 2.37 - Correct
Voltage at C24, C27 2.54, 2.51 - Correct
Voltage at C25, C3 .32 - Should be 2.5

Posted to the Firefly Yahoo group

Checking using the circuit diagram

A slightly different approach - Checked the IC pins against the circuit diagram, looking for shorts, checking interconnections etc. But nothing found.

Voltage at C26, C7 .15 - Should be 2.5

“Fueled by Red Label Tea..”

Pin 15 (2OE) Output Enable - 2.51 or .20!

Pin 1 (1OE) Output Enable - .15

Voltage on R4 2.37 on either side - Correct
Voltage on IC9 pin 7 - 2.35 - Correct
Voltage on IC9 pin 9 - 2.34 - Correct

Checked the chip interconnections against the circuit diagram. Noted the numbers of the chips already on the board and downloaded their data sheets, then used the circuit diagram for the remaining chips - All the chips should be different - But 74AC00D was on the board as IC6 and on the circuit diagram as IC7..

Checked the diagram again - IC6 should be a 784AHC04D and has a number of grounded pins.
Time to practice the SMD removal techniques.

Removing IC6

Various methods could be employed but this worked for me.

1. Thread a fine length of enamel wire under the legs on one side, fasten the wire in a loop.
2. Flow a continuous bead of solder along all the legs on that side.
3. While heating the solder bead very gentle pull the wire up to just lift the legs clear of the solder bead.
4. Remove excess solder from the pads with solder wick.
5. Repeat for the other side.
6. Inspect.

Soldered in the correct chip - Tested OK.

Receiver Audio Pre amp Tests

Current drain is 27.4 ma, should be 33 to 35ma. The C34 trimmer is moved once and breaks - No phase adjustment for the moment then.

C94 3.75 - 4.2 Volts, C95 5.1 - 5.3 Volts

The C95 voltage is to high. Should be 1.5 to 3.5

Re solder pin 4 of IC6

Pulled IC9 (TX/RX switch) pin 1 high (TX position) C94 and C95 now at 0 volts, so the TX/RX switch is working correctly.

Local oscillator buffer outputs I and Q on IC6 Pins 2 and 4 - Scope fails to resolve the signal, but there is something there. Switch inputs IC9 pins 2 and 14 - The same.

Now C94 4.2 Volts, C95 4.9 Volts - Still high. Requires further checking. Found the missing .1uf cap.

Further Testing

Attempted more testing but the scope was not working properly. Took it to G4BYE and we traced through it on his scope; all looked fine.

Receiver Front End and T/R Tests

Testing the transmitters voltage controlled crystal oscillator
Testing the transmitters voltage controlled crystal oscillator

L9 Toroid - Managed to get 4 inches and 4 turns messed up in my mind – But of course did not discover until the toroid transformer was soldered in and the wires trimmed. It was easy enough to fix though, as it was one turn too many, not one too few and the solder wick worked magically to clear the hole.

Tests OK.

Transmitter VXO

M3UYC gave me some thinner solder to try, which was alot easier to use.
L1 Toroid - Ok, so it is 40 turns of fine wire on a FT 37 core, but it really is not that bad. Replaced C37 - 68PF as it seems to have got lost! Local oscillator and TX VXO spot can be heard on the TS530S. Remember not to mount the crystals flush on to the board as they may short against the PCB pads.

Transmitter Buffer T/R Switch Tests

IC7 - The chip removed from IC6 is installed, various other caps and resistors are installed. Current drain is 50ma previous stage was 31ma. Current drain should not have risen, so something may be wrong - perhaps IC7 got fried during it's previous outing. T/R Switching works correctly and spot still works.

Firefly PA on the scope showing 7 dahs
7 Dahs - Output of the PA buffer chip

Checked the various inputs and outputs of the buffer chip on the scope, they look OK. Still unsure why the current consumption has gone up.

Transmitter PA Circuit

Sue made a good job of winding the toroids for the bandpass filter, the remaining PA components were then fitted.

Finished the board using .4mm solder (Thanks Dave) which was alot easier.

There was one 33pf - C79 and one 47pf - C32 cap left over.. C79 had a 68pf fitted - The missing 68pf had been found and as C32 was correct - 47pf. That just left the 33pf, and as no caps appeared to be missing it must be surplus.

Boxing the Firefly
Boxing the Firefly

Boxing Up

Tom (G4BYE) assisted with laying out the controls and drilling the box. The Firefly now draws 50ma, which is correct for the completed kit.
Solder wick made short work of clearing holes used for temporary connections.

First Test

Completed the basic wiring on Tuesday and took it down the club, connected it up to a T23 laptop running Rocky, but it did not receive anything at all. Checked things later and there was a problem with the audio cable. Now it will work…

Further Testing

Reconnected to the T23 and PowerSDR receiver software, then checked the software configuration but still nothing; even the net does not show up on the PowerSDR display.

“Current draw was steadily rising!”

Current draw was steadily rising! Removed from the case, current draw was then steady at 50ma, net correctly adds 1ma. The board was "very" close to the case, added additional nut spacers, reinstalled and rechecked current draw, back down to 50ma. Net correctly adds 1ma.

Both the local oscillator and net can be heard on the TS530.

Reconnected, switched on net, this should be a big signal on PowerSDR, but nothing showed.

Wired up the net pot, switched it back on and very carefully tweaked the sharp C87 and C75 preselectors, the net signal now appeared on PowerSDR, with several images on both sides of the centre frequency. At least the basics of the receiver were working.

Image Suppression

Rechecked the spec of the T23 line input to ensure it was stereo. The phase adjustment pot makes no difference to the opposite images; eventually it just squashes the whole signal. PowerSDR calibrate also makes no difference to the opposite images. Disconnecting one of the inputs just reduces the signal strength. The phase adjustment just does not seem to be working correctly.

Transmitter VXO Testing

Following discussions with Bryan (G8DKK) I checked the output from spot using the TS530S, there is a response approximately 8khz above and below the spot signal. This would indicate the Firefly RX is not the cause of these responses.

Spot Off - PowerSDR screen shot

Spot On - PowerSDR screen shot

Perhaps there is just too much coupling between the spot and the Firefly RX. If the coupling to the TS530S is reduced or the RF gain reduced the responses disappear. Keying the Firefly PA into a dummy load gives much the same results, the images are "blurred" and harder to detect but if the coupling is reduced they disappear.

Spot Harmonics

Dan N7VE - "Yes, this happens, especially when the spot signal is very close to the receiver crystal frequency. This is because the detector readily picks up very high order harmonics of the signal. Take for example the receiver crystal frequency of 14.065 MHz. If the the spot signal is 1 KHz away at 14.064 MHz, the second harmonic of the spot signal will appear to be 2 KHz, the third harmonic will be 3 KHz away, etc.

For signals arriving from the antenna, these harmonics are removed by a combination of the transmitter low pass filter and the receiver bandpass filter. The spot signal however, is not received via the antenna. Rather, the
signal is radiated and conducted across the surface of the PC board, bypassing the filtering that over-the-air signals receive. The largest, closest in signal is the spot signal you want. Like I said, this is not normally an issue until the spot signal gets close to the receiver oscillator frequency.

The above example is a bit simplified in that the detector naturally rejects even order harmonics by roughly 40 db. I expect that the frequencies that you see will be the 1x, 3x, 5x, 7x, etc harmonics of the spot frequency."