No RF gear here?

Which ought to have dealt with the "fake news" brigade - but the always just come up with some other nonsense. You can't win!

My own daughter sort of believed that the lunar stuff was faked. I had a discussion with her that involved explaining the technology available in the 60's and early 70's......the clincher, We watched one of the original Star Trek episodes. I explained that this was the best than Hollywood had to offer at the time, and would ANYBODY possibly believe that it was real? Yes Hollywood could create a convincing fake today.....not so easy in 1970.

It took some convincing to tell her that I was talking to a man in space on my ham radio setup too. "How do you know that he is really in space?" UH, a few of us at work built the walkie talkie that he had duck taped to the window in the Space Shuttle.

The first one was a hacked Motorola MX300 on 2 meters. There was an unpublished frequency that the astronauts used to talk to the Motorola Amateur Radio Club. The project was known as SAREX...Shuttle Amateur Radio Experiment. There were unpublished radios in space going back to the Skylab days. I think the club built an HT-220 for one of those, but I wasn't a MARC member at the time.

The first few missions flew under rather low publicity since nobody knew if it would work and if it would interfere with any onboard equipment. Early on the radio's power output was rather low, under 1 watt. In the later days there were so many people calling the shuttle at the same time on the same frequency, that they just heard noise.....like a CB set in the mid 70's.
 
I had the only color TV in the neighborhood. It was a 1957 vintage Emerson (RCA built) that I had rebuilt with a 60's vintage all glass CRT. My bedroom was the old garage with some carpet (bare concrete, bare feet, and hot chassis electronics DON"T mix). There were 30 or so people stuffed into it in the middle of the night to watch that first live color broadcast from the moon.

There was an older ham who had built a large dish antenna with lunar tracking to receive the LEM to Earth transmissions. He had recorded them to analog tape and he would play them to an audience in a forum meeting at the Dayton hamfest in the early 2000's. His completely independent transcription of the events helped dispel some of the "it never happened" conspiracies. Don't know what happened to it all when he passed.


It wasn't before Apollo 14 that we had color TV broadcasts from the moon surface. Apollo 11 had a b/w camera with them, Apollo 12 had a color camera, but Alan Bean immediately damaged it during the installation procedure, Apollo 13 failed to land, and here we have Apollo 14!
Best regards!
 
It wasn't before Apollo 14 that we had color TV

The gathering at my house was in 1971. It must have been from the second EVA since it was at night or early morning. Of course memories from that long ago are somewhat fuzzy.


a few of us at work built the walkie talkie that he had duck taped to the window in the Space Shuttle. The first one was a hacked Motorola MX300 on 2 meters.

I remember that one because the guy who actually took the radio to the Cape Canaveral launch site sat across from me in the lab. He had a large picture of the radio duck taped to the shuttle window in his cube for several years. The shuttle was still on the ground when the picture was taken, so it must have been from early testing.
 
I put the cheap DDS on the RP to look at it, and these are the results.

The first bunch is 5.2 MHz - looks pretty good to me, except for the harmonics, everything else is -70 dBm or better. Fundamental is +7 dBm. It is pretty convenient to have the readout direct in dBm; should make level matching easier. The white background screen is an alternate spectrum analyzer program available on the RP Bazaar.

23.2 MHz does not look as good, a little less output ( 0 dBm ) but other than the harmonic, and whatever is going on at 56 MHz, all the junk is still below -50 dBm. I have not bought any of those Epcos 23.4 MHz bandpass filters, but will probably get some since they are pretty cheap, and some rf amp chips, and take another try at a couple of watt broadband output to drive the linear.
 

Attachments

  • 5 FFT.jpg
    5 FFT.jpg
    127.1 KB · Views: 122
  • 5.2 Oscope.jpg
    5.2 Oscope.jpg
    121.8 KB · Views: 119
  • 5.2 FFT alt (2).jpg
    5.2 FFT alt (2).jpg
    98.7 KB · Views: 122
  • 23.2 FFT.jpg
    23.2 FFT.jpg
    145.9 KB · Views: 120
  • 23.2 Oscope.jpg
    23.2 Oscope.jpg
    119.6 KB · Views: 120
  • 23.2 FFT alt (1).jpg
    23.2 FFT alt (1).jpg
    87.7 KB · Views: 57
I put the cheap DDS on the RP to look at it, and these are the results.

The first bunch is 5.2 MHz - looks pretty good to me, except for the harmonics, everything else is -70 dBm or better.
I have to wonder, (how) does the spectrum change with frequency? What's the DDS input frequency? I can imagine it looking clean if you're generating an exact submultiple of the input (like f/10), but move around a little and there might be grass everywhere.
 
In general, with the DDS, I only checked around 5 and 23 MHz, and the upper limit of the DDS, because those were the only frequencies of interest to me at the moment. I put it at 5.2 and 23.2 to get off exact multiples a little bit. I don't recall seeing any increase in junk as it is tuned around.

It is kind of an inverse of the Si5351, which tends to look better as you go higher in frequency.

The DDS looks worse as you go higher in frequency. It has an RF board to amplify the output, and this is where all those harmonics are coming from.

My measuring capabilities and experience are quite limited, so keep that in mind ...

I've actually considered using the DDS as a 5 MHz tunable, mix it with an 18 MHz crystal oscillator ( built those last summer with the LM375 chip ), band pass the output, and use that for 23 MHz high side injection. It should be pretty easy - the DDS output is +7 dBm for an SBL-1, and since the output will need to be attenuated a lot for the AD831 and NE602 inputs, the SBL-1 output port can be broadband terminated with a big 50 ohm pad, before the bandpass.

Edit: of course I thought this whole thing would be a three or four week project tops, so ....

Hope that helps.
 
Last edited:
If I want to make another single conversion front end, using one of the Mini Circuits SBL-1 DBM's; 5 MHz LO , 9 MHz IF, and 14 MHz operating frequency, do I need to be concerned about all of the harmonics coming out of the DDS?

Should I low pass / bandpass the LO before it goes into the mixer, or just let the crystal filter deal with it at the output?

Also, I think I have some level 13 SMD DBM's; what adverse consequences can I expect, if any, if I only have level 7 drive? Increased conversion loss?
 
Somewhat stalled, no new failures to speak of, but no real successes either.

My wife has discovered that the redneck entertainment mecca of the region - Branson, Mo. - is but two hours from our second house, so we have been taking our six year old up to Silver Dollar City and Branson most weekends this summer. Mowing is taking a lot of my other free time, made worse by having to make frequent repairs to a brand new piece of expensive equipment, and it is a real knuckle busting pig to work on.

I'm stuck for now at a few milliwatts out of the transmitter. I think I could have a narrowband TX strip up and running pretty fast, but I want to learn elementary broadband techniques. Right now all I have to show for that is a pile of toasted transistors.

Both the perf and SMD test boards are here at the office where I have set up a work station, but I haven't done much with it. I've assembled a modest, but reasonably capable, mobile duplicate parts inventory and modest test equipment setup, including the RP, that fits into a couple of toolboxes to use here at the office, or carry to the second house, but I just haven't done much with it, other than carry it around from place to place.

There are some things I like about that cheap AD8950 DDS that has come into the picture lately. I am considering going back and making another single conversion transverter board, using the DDS as a 5 MHz VFO, and taking another try at diode DBM's, this time with some of those SMD TUF-3 DBM's that I purchased a few weeks ago. That could feed either the SG-9 board or the Red Pitaya as an IF. That would free up the Si5351 frequency synthesizer for a 6 meter transverter.

I'm also considering something even simpler, maybe - using the DDS as a DSB direct conversion transceiver. The drive level out of that thing is high enough and clean enough that I think DDS => balanced modulator => one low level amp => 10 watt PA could work. Then maybe play some with image reject modulators / product detectors to cancel the image sideband.

Likely nothing worth photographing will pop up for at least another month. Too many things to try, too little time, not enough knowledge and experience ...
 
Somewhat stalled, ... Likely nothing worth photographing will pop up for at least another month. ...

OK, so my wife decided to switch our primary house to fiber. So with her at the lake, I had to stay at the house all day today.

I didn't have either of the experimental transverter boards here at the house, so I decided to just start the build out of an actual board to put in service.

This will be much smaller, and hopefully somewhat more polished. I got the receive RF amp done. It's just a generic dual gate mosfet RF amplifier, using an NXP dual gate FET that is not obsolete, and available from Mouser.

The circuit is generic. The signal gate is biased with a 100K resistor. The control gate is biased with a 33K/10K voltage divider. The source resistor is 270 ohms. The tuned circuits are low L, hi C, and input and output is through a link on the primary, so it is low impedance in and out. The bandwidth of each tuned circuit is about 250 KHz, so it needs to be stagger tuned.

The FET is SMD, and the other bits are 0805 SMD, except for a temporary flying 33K.

It seems to work. With 0.2 volts p-p in, I get 3.5 volts p-p out at the secondary of the output tuned circuit, at 14 MHz. Lots more on the primary. Current draw is 2.5 ma at 12 volts. The waveform and output FFT look good.
 

Attachments

  • IMG_20190823_160450179.jpg
    IMG_20190823_160450179.jpg
    915.5 KB · Views: 151
  • IMG_20190823_160509044.jpg
    IMG_20190823_160509044.jpg
    883.6 KB · Views: 156
  • IMG_20190823_160526423.jpg
    IMG_20190823_160526423.jpg
    781.9 KB · Views: 140
  • IMG_20190823_160548917.jpg
    IMG_20190823_160548917.jpg
    982.6 KB · Views: 143
  • IMG_20190823_160531925.jpg
    IMG_20190823_160531925.jpg
    741.5 KB · Views: 147
Of the mixers I've played with, I like the AD831 eval board, best. But if I used one of those here, it would take up almost all of the remaining space on the board. I didn't want to do that, because I want to preserve the option for double up conversion in the future.

As a practical matter, the IF port needs to exit to the center of the board to allow board space for double conversion. Of the mixers I've played with, when the pieces are put on the board to see how things will fit, only three conveniently put the IF port to the center: a dual gate mosfet; the TUF-3 DBM; and the HP IAM81008 DBM.

I played with a cascade noise figure calculator, and it seems that as long as the first RF stage is quiet, noise in subsequent stages does not matter much, even going out to a couple of conversions. The NXP BF992 mosfet used in the RF stage should be 1 dB or less NF at HF. So I ignored NF. The mosfet mixer takes a lot of parts, has mediocre port isolation, but has gain and is easy to build. The TUF-3 diode DBM would probably be the best performer, but I haven't played with it at all - the ports would have to be matched, and it has 5 dB conversion loss. To get enough drive from the DDS, I would have to use low side injection and change the BFO crystal. The HP has minimal parts count, good port isolation, and conversion gain. The data sheet and app notes claim port matching is unneeded, and won't improve performance. It need a separate voltage from the RF stage.

Of the three, after thinking about it for the last week, I've used the HP. Even though it claims port matching is not needed, I left a pad for an LC network at the input, assuming I can get a soldering iron back in there - I didn't have any SMD inductors in my travel box. Since I gave up a pad for a possible LC network on the input, I took a pad back by vertically mounting the R part of the RC network used in the low frequency bypasses, that are required below 50 MHz. I put all the parts on the board, and then dropped the chip and soldered it in last.

HF RF comes in at the top. LO comes in at the left. The HP wants to see -10 dBm so there is a dc blocked 10 dB attenuator pad between the LO input and the LO port; either the Si5351 VFO ( 6 meters or up conversion ), or the DDS VFO (HF single conversion) can be used simply by plugging into the board. IF port exits to the right.

The layout and remaining space should work for an SA602 second mixer in the future. That chip lays out so that 45 MHz input would be on the right, and second IF output would be at the left / center. It has an onboard transistor for a crystal oscillator, so that saves a lot of space.

My soldering iron went kaput, so I couldn't finish the DC wiring, therefore it has not been tested. It's a two hour round trip to get into town for a replacement, so not practical. I haven't decided if the low voltage for the HP chip will be generated on the board, or brought in from outside, so not a big deal that the iron died on me.

I thought the toroids might be too fragile for a mobile rig, but I dropped it three times from a height of about four feet unto concrete, and it was unscathed, although the tuning was knocked a bit out of kilter. The locking connectors should keep any of the wiring from coming loose. I can get 175 Khz -1dB, and 400 Khz -2dB bandwidth by stagger tuning the coils. They are low L Hi C; under each trimmer is a 120 pF padder. Removing the padding caps and (possibly) putting in smaller value trimmers will put it on 6 meters.
 

Attachments

  • IMG_20190901_100529904.jpg
    IMG_20190901_100529904.jpg
    836.9 KB · Views: 67
  • IMG_20190901_125035179.jpg
    IMG_20190901_125035179.jpg
    827.3 KB · Views: 65
This batch of HP chips came from a stateside vendor, and, if fake, at least some effort was made to produce a quality packaged fake.

The BF992 seems to have quite a bit more gain than the 40673 in more or less an identical circuit, at an identical Vds. In this screenshot, the RP is putting 100 mv ( as low as it will go ) at 14.2 MHz into the RF amp, and it comes out at -25dBm.

The LO is at 23.2 MHz - 10dBm. LO at the RF port is -40dBm; LO at the IF port is -37dBm. This looks to be right on as per the spec sheet.

RF at the IF port is -75 dBm; this is considerably better than the spec sheet. At 5 volts, current draw is 12 ma which is right on the spec, so these appear to be good chips, as best I can tell.

The output spectrum looks good - the pic is the unfiltered output at the IF port. This gave me a chance to work the RP pretty hard - at one point it had both signal generators going, and both spectrum displays running at once. Very handy. Still very pleased with the RP as a piece of portable test equipment.

The BF992 is now end of life. Mouser is out of stock, but is still taking back orders. There is still stock on BF1105, a 5 volt part. It is an enhancement mode part, and the sheet says it has the bias resistors on the chip. I have some, but haven't used them yet. Looking at the NXP web site, it looks like all of their FET products are end of life, or not recommended for new design.
 

Attachments

  • IMG_20190906_213328847.jpg
    IMG_20190906_213328847.jpg
    110.3 KB · Views: 55
  • IMG_20190906_210943847.jpg
    IMG_20190906_210943847.jpg
    142.8 KB · Views: 70
Member
Joined 2018
Paid Member
[...]
The BF992 is now end of life. Mouser is out of stock, but is still taking back orders. There is still stock on BF1105, a 5 volt part. It is an enhancement mode part, and the sheet says it has the bias resistors on the chip. I have some, but haven't used them yet. Looking at the NXP web site, it looks like all of their FET products are end of life, or not recommended for new design.

The address where I get small signal (mos)fets is Transistors - FET and MOSFET | Buy on-line | rf-microwave.com
A favorite is the BF987, a single gate low-noise mosfet of which I bought ~100, enough for life. Stock still is impressive, I've used the device with success in an FM tuner, and various antenna amps where the BF982 failed (due to lightning)...
 
... The layout and remaining space should work for an SA602 second mixer in the future. That chip lays out so that 45 MHz input would be on the right, and second IF output would be at the left / center. It has an onboard transistor for a crystal oscillator, so that saves a lot of space.
...

I went ahead and dropped an SA602 on the board to see if this would work. Based on how much space the RF amp took, it looks like there is adequate room for a 45 MHz filter, and, if needed, an IF amp between the IAM81008 output and the SA602 input. The space below pins 6 and 7 look adequate for a crystal and the other bits needed to enable the onboard oscillator.

In the meantime, I can use the SA602 as the TX mixer. The screen shot is 9 MHz 100 mv on the RF port simulating the SG-9, 23.2 MHz -10dBm on the LO port. Input and output are single ended - no filtering on the output port. The RP is supplying the RF and LO.

This is just a quick test hookup. The SA602 is of dubious Chinese eBay origin - at some point before it goes in actual service, I'll replace it with one from a reputable supplier.

I'm appalled this has taken two years, but it is sort of starting to come together now.
 

Attachments

  • IMG_20190908_125256468.jpg
    IMG_20190908_125256468.jpg
    130.7 KB · Views: 56
  • IMG_20190908_125400352.jpg
    IMG_20190908_125400352.jpg
    138.9 KB · Views: 59
I played around with it some, and substituted a real ( I think ) SA602 for the eBay sourced part, and honestly could not tell much difference between the two. The eBay part may have had a slightly higher noise floor, but it was right at the bottom of what the RP can see, so not really conclusive of anything.

After switching between parts a few times, I left the genuine part in, and played with LO drive level. On this SA602 chip, the optimum LO drive level is -21 dBm. The RP makes it possible to adjust the drive in very fine increments, and it can output significantly lower voltages than the function generator can. The previous iteration of the SA602 TX mixer likely was not optimized for LO drive, because of this.

Increasing the drive level doesn't add much of anything to the desired output, but markedly increases the visible spurs the RP sees. Decreasing LO below this level decreases desired output significantly. The drive level at the RF port has a profound effect on output, but the SG-9 puts out what it puts out, so not much I can do about that.

The RF input is nulled 29 dB at the output, and the LO is nulled 31 dB. I could increase the LO null with a balance control on the input, but probably won't.

The screenshot is the genuine SA602 at -21 dBm LO and 5 volts. Interestingly, the optimal voltage for this chip looks to be around 3.8; that nulls the LO another 6 dB or so from what is shown, doesn't create any spurious outputs, and only gives up about 1 dB on the desired output.
 

Attachments

  • IMG_20190908_175841322.jpg
    IMG_20190908_175841322.jpg
    137.2 KB · Views: 52