No RF gear here?

The DDS chip is the Si5351. I'm only using one output at the moment.

Win W5JAG
Is that actually a DDS? Looking at the datasheet it says "high resolution MultiSynth fractional divider architecture" - is that their buzzword for a DDS, or is this something else? It seems they're vague about exactly what it is.

It sounds like NOT a DDS. I've heard of fractional dividers, but only for setting an "average" clock speed for a microcontroller (Freescale some years ago), and as I recall it inherently introduces jitter, but I can't imagine this chip would do that.

https://www.silabs.com/documents/public/data-sheets/Si5351-B.pdf

Good news for that chip, Adafruit has a breakout board (and Arduino library) for it.
SI5351 | Hackaday
 
Filter the LO?
Win W5JAG

That won't help. The LO voltage in the diode ring is squared up by design.
You can see that with a scope. The more power you try to pump into the
ring mixer, the more square the LO voltage becomes. The ring mixer
simply won't let it grow above a certain value.

A multi-pole low pass is probably what helps best. And shielding.

If you do not have a sweeper, you can use the DDS to tune the filters.

My Red Pitaya now has a 40 mm fan. It is much to strong.
But I have decided to first clean up my low noise amplifiers.
Too many projects in parallel.

Gerhard, DK4XP
 
... Is that actually a DDS? ...

Beats me, Ben. I'll have to defer to the more knowledgeable among us on that point. It's sold as a DDS in the ham community. It looks like it's the frequency element of choice for simple ish homebrew rigs right now, usually monoband, with 9 ish MHz IF's or SDR. Folks seem to love it in that application.

I can't find anything on the web where it has been used as an upconverter for a low VHF IF.

A multi-pole low pass is probably what helps best. And shielding.

Wait, I'm confused. Low pass the LO or the front end? It's not coming in on the antenna.

If the LO can be filtered, I've been looking at both bandpass and low pass filters. I don't know if I can get a sharp enough cutoff with a low pass, since I need to pass 75 Mhz if I want to do 10 meters. It will be in a metal box, eventually, but I would like to resolve this before that point. I may not have suitable coils in stock.

Too many projects in parallel.

I can relate ....

Win W5JAG
 
Last edited:
Is that actually a DDS? .....It's sold as a DDS in the ham community.

It's not a true DDS. A DDS uses a master clock to scan a sine (or other waveform) table into D/A converter.

The "MultiSynth fractional divider architecture" is pretty much what it says. You start with a clock frequency in the hundreds of MHz and use a fractional divider to divide that down to any lower frequency subject to the resolution of the fraction. That comes down to the number if bits in the numerator and denominator.....and a few other tricks like multiple accumulators.
 
Member
Joined 2002
Paid Member
I looked a little more, they use Magic Sauce.

What a useful sauce jar !
https://www.silabs.com/documents/public/white-papers/MultiSynthWhitePaper.pdf

Silicon Labs seems to do a great job. Their Si4734 broadcast receiver chip is performing very well in this portable radio.
Amazon.com: TECSUN PL-310ET FM Stereo/SW/MW/LW World Band PLL DSP Radio Black by allnice: Home Audio & Theater

Now I am waiting for a DDS receiver to arrive (in a week or so)
Excellway(R) Ham Radio Receiver 100KHz-1.7GHz Full Band UV RTL-SDR USB Tuner Receiver Sale - Banggood.com


Gerhard
Thank you for informing us on the Red Pitaya tool. Enjoy it.

George
 
Thanks for the suggestions, fellas - morning update:

Early on I had some problems with the microprocessor on the DDS board, and kludged around them, but never got around to changing it out, even though the vendor sent me a replacement. I was going to try out another IF frequency, and going through the programming, I noticed output 2 was on, around 42.8 MHz, and the programming wasn't responding properly to change it, so I changed out the microprocessor.

Last night I was optimistic that resolved the problem, but a brief check this morning, and it hasn't. It did make some improvement - for example, I can now receive the 1.23 MHz station, which looks like a worst case relationship for images, provided I bypass the hi pass filter so it gets full strength signal.

Filtering off that third harmonic just doesn't look that practical if I want to get 10 meters in the radio. I suppose this issue could resolve itself when it's boxed in the metal chassis - I'm a little reluctant to go that far without it being resolved; just seems dicey.

Plan B, I guess, would be to use a typical IF frequency - 9 or 10.7 MHz. Maybe a second conversion to 455 KHz. I have filters on hand for both these, either single or double conversion, so not a big deal either way. That's just a re hash of something I've already done; likely a much improved re hash, but still a re hash ....

We had a big temp drop yesterday - it was 58 F yesterday morning, and about 16 F this morning, so I probably won't be in the shack much until it warms back up; that gives me some time to ponder this. I did go ahead and order another SBL-1.

Still open to suggestions - I expected some issues to have to work through, but not this one. I've built quite a bit - haven't run into this before. I just don't see much opportunity for FM BCB being picked up on these little boards.

Win W5JAG
 
Between 1 and 30 MHz, I counted OVER 99 instances of serious full quieting, hi fidelity, FM BCB breakthrough, using an Icom IC-R3 as a wideband FM 45 MHz IF, this with NO antenna, and a 51 ohm resistor on the SBL-1 IF to make sure termination was not an issue.

It's just not happening with this device, with a 45 MHz IF.

With high side injection to a 9 or 10.7 MHz IF, it's usable. It's better with low side injection.

I hastily rigged up a quickie 40673 mixer, using high side injection to a 10.7 MHz IF, with a single tuned circuit at about 25 MHz on the LO gate to clean up some of the harmonic snarf, and it's not too bad. Drive level isn't matched. Listening to the Collins net on 14.263 MHz with this setup right now ....

Win W5JAG
 
Last edited:
The mosfet mixer moved to the active board - left a bit of room in the corner for better matching into it and / or an RF amp. Listening to 75 meters with a piece of wire; 10.7 MHz IF; high side injection, same wideband input filter, sidebands flipped.

The "prototype" was wack job on a piece of scrap, worked fine right off the bat, moved it to the main board, tried to make a neat, compact, layout - took three tries in as many hours to make it stable. Go figure .....

When I'm sure it's right enough, I'll change out the transformer to a 45 MHz LC load, and try again at 45 MHz IF.

Went to the lake house Saturday and went by the last "parts" place in the area, hoping they would have an si5351 breakout board for Arduino. They had about everything but that.

Win W5JAG
 

Attachments

  • 40673 mixer top rev 3.jpg
    40673 mixer top rev 3.jpg
    927 KB · Views: 200
Last edited:
I found some other stuff there, including some needed parts for another project that has been swept aside by this train wreck.

I got some copper clad board, and I'm going to try again, deadbugging a single band upconverter, to a 45 MHz IF, on the copper side, and just forget about general coverage. This noise maker will never be suitable for general coverage. At this point, I don't have much confidence it will be suitable for a single band with a VHF IF .

The existing board can be converted over to a conventional single band PLL LO, bandpass front end, with a 10.7 MHz IF. Maybe a ring mixer. So, two paths to the same end, being built in parallel.

Win W5JAG
 
Fellas ... we have up conversion to 45 MHz.

The first iteration of the deadbug was just terrible - it was still one of the hottest, non selective, FM BCB receivers I've ever run across.

When all else has failed, miserably, do arithmetic .....

It appears the fundamental problem was the SBL-1 was being overdriven. After nearly ripping my hair out, I finally decided there was a pretty good chance the output of the Si5351 was about +13 dBm, where the SBL-1 wants to see + 7dBm at it's LO port, so I rigged up a 6 dB attenuator, and, like that, the FM breakthrough was gone. I thought I had drive level issues eliminated with my craphouse pot attenuator, but apparently not.

Still early days, but things are at least looking up. It seems to be about 10 dB more sensitive with a 10.7 MHz IF, but that may be an artifact of my FT-950 "IF" being less sensitive at 45 MHz. Right now I have it parked on 1.23 MHz listening to a ball game, with no interference from the image of that 100 KW FM station, 45 MHz IF, simple 3 pole low pass filter at 30 MHz, non selective antenna.

Interestingly, the 40673 must have been pretty severely overdriven as well, but behaves better in spite of it. I would have expected the opposite. Perhaps that is why I had such a tough time stabilizing it.

I plan to move forward with both mixers - think I'll put about 9-12 dB attenuation on the LO gate of the 40673 and see where that goes.

Win W5JAG
 

Attachments

  • deadbug rev 1.jpg
    deadbug rev 1.jpg
    527.8 KB · Views: 189
In the USA, the FCC allows operation of low power FM transmitters, but anything over 100mW must be licensed.

I'm sure most countries have similar restrictions, whether or not they offer licenses would be up to the individual to check with the appropriate ruling body.

In the US, the legal limit is not measured in TPO, but in radiated field. The maximum field for Part 15 devices within the permitted 200 kHz band shall not exceed 250 microvolts per meter at 3 meters. This limits the range of a part 15 transmitter to about 200 feet radius. Someone once calculated the TPO into a 36" dipole radiator needed to equal that field strength and the power required was said to be 19 MICROwatts.
 
Member
Joined 2002
Paid Member
The maximum field for Part 15 devices within the permitted 200 kHz band shall not exceed 250 microvolts per meter at 3 meters.

This is a frightening low number :confused:

As I read in these
https://transition.fcc.gov/Bureaus/Engineering_Technology/Documents/bulletins/oet63/oet63rev.pdf

47 CFR 15.209 - Radiated emission limits; general requirements. | US Law | LII / Legal Information Institute

The limit for 190kHz-490kHz is 2400/f(kHz) µV/m @ 300 m.
Doing the arithmetic for 200kHz the limit is 12 µV/m @ 300 m.
Applying the inverse law, the limit is 120000 µV/m @ 3 m.
I think the resulting number for the estimated ERP is close to 2mW
Of cource I may be way off :D

George
 
The MOSFET mixer at 45 MHz.

12 dB attenuation on the LO gate; general coverage front end filter; 25 ish foot random wire antenna; 45 MHz L/C single tuned circuit; taking the output with a couple of turns on the main coil. I get a smooth peak at 45 MHz with this coil / cap / link combination.

Sensitivity seems adequate for now - running the FT-950 straight through, the background noise comes up when I clip the untuned wire to the antenna input.

Do I still need to follow the output with a xtal filter, or will the single tuned circuit be sufficient selectivity to go to the next IF? I haven't thought much on the second IF - I have 10.7, 9, and 455 as options. I'd lost confidence in actually getting to this point.

9 could go straight into the Mizhuo board, so that is probably the easiest choice.

Win W5JAG
 

Attachments

  • mosfet mixer rev 4 45 MHz.jpg
    mosfet mixer rev 4 45 MHz.jpg
    630.8 KB · Views: 115
And another variation of the MOSFET mixer, this time with the 45 Mhz xtal filter as the load.

Earlier in the day I had a 45 MHz IF amp running, but the footprint was bigger than I wanted with the filter and all the coils, so I ripped everything back out to try to simplify it a bit, and shrink the size.

I don't think I need the high pass filter, so at some point that will probably come out.

Win W5JAG
 

Attachments

  • mosfet rev 5 45 mhz.jpg
    mosfet rev 5 45 mhz.jpg
    784.5 KB · Views: 103
I think I've learned more about the MOSFET mixer than I planned.

The L/C load has better conversion gain, than the xtal filter load. I've not matched the filter perfectly - I'm using the ECS-45-K20A which calls for 910R/2.5 pF, and the closest I have is 1K and 3 pF. I like the simplicity of just using the filter as a load.

Wow - just heard a VK on 20 with the 40673 as the front end ....

Anyway, I pulled the attenuator back out - the impedance of the LO gate is so high that the Si5351 can't put any appreciable power into it. The conversion gain is greatly influenced by the LO voltage, so to a point, more is better. The scrap mixer I made last weekend worked well, and used a tuned circuit on the LO gate. I was trying to get rid of the harmonic snarf from the Si5351, but in hindsight the beneficial effect of this was probably that the link coupling into the L/C circuit likely boosted the LO drive voltage into the MOSFET.

The LO drive level does not seem to affect noise performance.

I haven't fooled any with the RF input port.

The RF to LO port isolation .... I think is OK enough. I set my SteppIR to bi-directional, and walked around in front of and in back of the antenna with a receiver tuned to the LO, and could hear the LO quite easily, but it nulled when I rotated the portable antenna horizontal, so I don't think the LO is coming back out the antenna.

Alarmingly, I heard a bunch of strong FM broadcast stations between 60 and 85 MHz on my Icom IC-R3, which is not a great receiver, but not a POS, either. There is a big cell site / pubic service transmitter site a couple of hundred yards away ( they like this hill for the same reason I do ) and I'm wondering if they have a transmitter there spewing crap.

The standing current in the MOSFET looks to be somewhat critical. The cookbook value is 470R in the source, with a bypass of 0.01 uF. I found that more current increases conversion gain and noise, but after a point, just noise that can completely swamp weak signals. I wound up putting in a 1K trim pot to try to get the optimal point. I'm using a genuine RCA 40673, and for this transistor, it's about 1.2 ma. Mixers farther from the antenna might benefit from more current, where low noise is not that important.

I got interested in the current in the transistor, because there are still about six spots where I get FM breakthrough with the MOSFET that I can't seem to resolve. Other than that, I'm pretty satisfied with it at this point. I think it's working the way it is supposed to. Now I'm wondering about that xmtr site.

Win W5JAG
 
Not into RF design here, but highly involved in how easy it can be to be on the wrong side of the rules.

I've seen an AM transmitter's harmonics trash GPS reception. That's a long way off!

I've seen smoke detectors ( :eek: ), unloaded wall warts, and tabletop amplified RX TV antennas crush cellular.

-116dBm is expected for the full licensed uplink BW at most 4G cell antennas. That's pretty quiet!

It is easy to go rogue, respect to the designer that can keep it clean. As tech progresses, so does sensitivity.