Don't get me wrong -- I'm a huge fan of Doug DeMaw. His books and articles are a treasure trove for ham radio homebrewers. Also, Doug was an honest guy who admitted in the preface to his QRP book that at times he did not fully understand the circuits he was building; that kind of honesty is rare, and is very helpful to amateurs who struggle to understand the circuits we work on.
But everyone makes mistakes, and Doug made one in his "W1FB Design Notebook." I present it here not as a "gotcha" effort to nitpick or sharpshoot a giant of homebrew radio, but because this error illustrates well the depth of the 75 LSB/20 USB myth, where it comes from, and how important it is to really understand sideband inversion. Here is the mistake:
That's just wrong. A receiver built like this will not allow you to listen to 75 LSB and 20 USB "without changing the BFO frequency." (Am I the first one to spot this error? Didn't anyone build this thing, only to discover that it, uh, doesn't work?)
Here's a little drawing that I think illustrates why the mythical scheme will not work:
All confusion about sideband inversion could be avoided with the simple application of what I think we should call "The Hallas Rule":
"Sideband reversal occurs in mixing only if the signal with the modulation is subtracted from the signal that isn't modulated."
Be careful here: I think some arithmetic carelessness is responsible for much of the myth. Taking the difference frequency is not enough to produce sideband inversion. Read the Hallas Rule carefully: For sideband inversion to occur, the signal with the modulation must be subtracted FROM the signal without the modulation.
---------------------------------------------
About the Swan 240's SSB generation scheme:
I first stumbled on this problem when building my first SSB transmitters in the Azores. I was using a VXO, and a filter pulled out of a Swan 240 (5.173 MHz). I started with VXO crystals at around 12.94 MHz. The rig worked, but I couldn't pull the VXO crystals very far. So I switched to crystals at around 23.3 MHz (you can pull higher frequency crystals farther). But look what happened: My Carrier Oscillator frequency had been set up to receive USB signals on 17 Meters. With the 12.94 MHz rocks, that worked fine: 18.150-12.977 = NO INVERSION. But it all changed when I went to the 23 MHz VXO rocks: 23.323-18.150 = INVERSION! This had me scratching my head a while. I had to draw myself little spectrum pictures (like the one above) before I realized what had happened. To get it to work -- to get it to produce USB on 17 meters -- I had to move the Carrier Oscillator to the other side of the passband. Good thing that Swan 240 came with TWO BFO crystals (5.1768 MHz and 5.1735 MHz). I just had to change the crystal.
For 75 and 20 meters, the Swan 240 uses the correct 5.173 MHz filter with a 9 MHz VFO to get the happy situation of 75 meter LSB and 20 Meter USB WITHOUT changing the BFO/Carrier Oscillator frequency. This is the Mythbuster scheme. Unlike Doug's receiver, it works. The scheme also works in the Swan 240 on 40 meters because for 40 the Swan rig has the VFO running from 12.073 MHz to 12.513 MHz. Here too, no change in the BFO/Carrier Oscillator frequency is needed. But the Swan recommended a modification that would allow operation on 20 LSB and 75/40 USB! It used a BFO/Carrier Oscillator crystal of 5.1765 MHz and a switch mounted on the front panel. Luckily, my junker Swan (acquired from HI8P in the Dominican Republic) had the second crystal -- mine was 5.1768 MHz. It was that crystal that allowed me to get my Azorean SSB transmitter to work using the 23.9 MHz VXO rocks.
If DeMaw did not understand some of the circuits in his book, well then there's hope for me.
ReplyDeleteExactly.
ReplyDeleteAnother way to show this: there’s nothing special about the two inputs to a mixer. The RF and LO ports are the same. So we can treat them the same.
ReplyDeleteIf you’re tuned to an LSB signal, consider what happens when a) the signal goes down in frequency, vs b) the LO goes down in frequency.
A) when the signal goes down in frequency, the audio goes up. Frequency inversion.
B) when the LO goes down in frequency, the audio goes down too. NO frequency inversion.
The same thing happens on a USB signal, but reverse obviously.
So I think this is equivalent to saying that the LOWER of the two input frequencies to a mixer gets inverted, and the HIGHER of the two frequencies does not.
Dunno, I haven’t had my tea yet this morning, I might be thinking about this wrong.
This comment has been removed by a blog administrator.
ReplyDeleteSmitty's right, this is def a post caffiene thing. :)
ReplyDeleteThis spectral inversion thing apparently continues to rear its ugly head and amaze/confuse the DSP crowd in 2008 as well, so we can't be too hard on Doug - see https://www.dsprelated.com/showarticle/51.php, but it will take more than a cup of tea to get through it :)
Is it an oversimplification to say that if the frequency of the unmodulated signal is greater than the modulated signal, and you're selecting the difference, the signal will invert?
Or not :) Scott ka9p
This comment has been removed by a blog administrator.
ReplyDeleteThis comment has been removed by a blog administrator.
ReplyDeleteIt’s odd that he also talks (on page140) about adjusting the BFO to minimize the unwanted side band - he had a pullable vfo crystal and that probably had enough range to flip usb/lsb - maybe it did “work” ? Albeit not quite like he said….
ReplyDeleteThat is odd. I've used that scheme to cover the two bands and it's the dial scale that inverts, not the sidebands. But if anyone deserves a Mulligan, it's DeMaw!
ReplyDeleteThere is a Multiband Sideband exciter in the ARRL Handbook of 1968, pp 272-275, probably also in nearby years, that illustrates this nicely. A modulated USB signal is generated from 2.900 to 3.150 Mc. (recall we are in 1968, I am a Junior in HS!) This is mixed with a crystal oscillator with crystals(in Mc.) 80 (6.9), 40 (10.2), 20 (11.2), 15 (9.15). Clearly adding for 20,15 & subtracting for 80, 40. BTW the editor is Doug DeMaw, W1CER. Even the great Homer nods!!
ReplyDeleteYes, that scheme would work. You'd be getting sideband inversion on 80 and 40 not just because you are taking the difference product, but because you are subtracting the modulated frequency FROM the unmodulated frequency. On 20 and 15, you are just taking the sum freq -- the signal remains as USB. 73 Bill
DeleteExactly! Love your SS book and blog. 73, Dan
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteThe main thing to consider here is the exact pass band of the IF filter. For a BFO frequency of 9.000 MHz, to receive LSB with no inversion, the filter pass band has to be below 9 MHz If it is exactly 3 KHz wide, its lower edge is around 8.9967 MHz and upper edge around 8.9997. Note that the BFO needs to be around 300 Hz from the edge of the pass band (at around -20dB). If you want to use the same filter for an Upper sideband signal, then modulation of a USB signal needs to be in the pass band of the filter. Its is a simple as that.
ReplyDeleteNow to get the full story, one needs to do some maths here. Assume the BFO is fixed at 9 Mhz. I am going to use an unrealistic modulation width to simplify the maths. And assume an LSB filter from 8.9 MHz to 9.0 MHz - ok!
Lets assume the modulation is 100 kHz wide, with the low tones at near zero Hz, and the high tones at 100 kHz.
So a USB signal with its original carrier on 14.000 MHz, will have a sideband going from 14.000 MHz to 14.100 MHz .
Now mix that signal (subtraction) with a 5.000 MHz VFO, and the result is a new signal going from 9.000 MHz (the low tone) to 9.1 MHz the high tone).
But the signal wont pass through the LSB filter from 8.900 MHz to 9.000 MHz - it is outside the pass band - right!.
So in order too get that through the IF strip, the incoming 14 MHz signal has to pass through the LSB filter.
So it has to come out of the mixer between 8.9 MHz and 9.0 MHz .
If you now move the VFO to 5.100 MHz, you now get 14.000 - 5.100 = 8.900 MHz and 14.100 - 5.100 = 9.000 MHz .
Now the signal is inside the filter - but the low tone is now at 8.9 MHz and the high tone at 9.0 MHz .
There has not been a sideband inversion - just a shift down.
Now in order to recover the audio the right way round, the BFO has to be where the original carrier was suppressed
and that is near the low tone at 8.9 MHz not 9.0 MHz . So the BFO must move down.
And furthermore, the original signal was generated at 14.000 MHz so your dial has to read 14.000 not 100 kHz higher.
Your VFO has moved to 5.1 MHz . So the dial is wrong by 100 KHz!
How do you overcome that? Well in the old days, rigs like the FT101B, with only one filter, had 3 red marks on the dial knob
which were the USB, LSB, and centre dial positions against the frequency scale.
But the UNIDEN 2020 had 3 filters - LSB, USB, AM. (CW inside one of the sideband filters) so the dial
did not have to move, or use extra markings. Modern rigs display a calculated number with the correct offset.
And when looking at sideband inversion, it normally occurs when the VFO is higher than the incoming frequency
and that scheme is also used. But the same maths apply - the VFO and BFO have to move if using only one filter.
This comment has been removed by the author.
DeleteYour last line: "And when looking at sideband inversion, it normally occurs when the VFO is higher than the incoming frequency
ReplyDeleteand that scheme is also used. But the same maths apply - the VFO and BFO have to move if using only one filter." That's the problem. DeMaw and others were claiming that with a single 9 MHz filter and a 5 MHz VFO you can have LSB on 75 and USB on 20 WITHOUT moving the BFO. Do you also think that is possible?
No offense, Bill, but I had to draw it out on paper myself to believe that old W1FB got it wrong. You are correct--there's no inversion in either the 20M or the 80M case. The same 5-5.5MHz VFO and 9MHz IF filter can be used, but it won't work without moving the BFO from either above or below the filter.
ReplyDeleteI don't know if anyone has ever built this receiver, but it looks to me like DeMaw designed it so it could work, his erroneous explanation notwithstanding. First of all, he's selecting 20 or 80M with switchable BPFs at the very front end (i.e., he's not relying on inversion to change bands), and he does provide 60pF of variable series capacitance to adjust the BFO frequency. It also looks like he designed in a deliberately-wide xtal IF filter (one xtal at 9.000 and the other at 9.0015) within which a sideband could rattle around. Which sideband? The only one that would be present depending on which front-end BPF was switched in: upper-sideband if for 20M and lower-sideband if 80M.
I don't know how well his scheme would work, but I bet it would. If I didn't have a crap-ton of other projects already underway, I'd build it just to see. It's almost like a design for a cheap commercial receiver with minimum parts. It's even more bare-bones than the Barbados. 73, Todd K7TFC
Todd: I had that thought too -- I wondered if DeMaw got this thing to work just by keeping that filter very broad. I guess he could, but he clearly thought that with this scheme one sideband would invert while the other would not. As you note, that is just incorrect. And that trimmer on the crystal - I think that was just to tweak the placement of the BFO re the passband. That's what you'd have if you thought that both bands would be passing the same sideband to the filter. In several other articles we see DeMaw repeating this mistaken notion. Oh well, nobody's perfect. And he was very good. 73 Bill
ReplyDelete