Regen Receiver, Heterodyne Oscillator. - UK Vintage Radio Repair and Restoration Discussion Forum

ms660 · 26th Jul 2015, 11:44 am

Just came across this video:

https://www.youtube.com/watch?v=27yoGmXtqQE

The idea is not new, a brief description and schematic is given in this old article:

http://kearman.com/images/W3LW-regens.pdf

Any thoughts as to what is actually going on in the detector? Has anyone tried this method for SSB reception?

Lawrence.

G8HQP Dave · 26th Jul 2015, 5:01 pm

As I understand it (from reading the article - not seen the video) the idea is that an oscillating regenerative detector is quite fiddly (e.g. sensitive to fluctuations in supply voltage, very narrow bandwidth). A non-oscillating regen is much less fiddly, so easier to adjust. For CW (and SSB) you need a local oscillator, so instead of making the regen oscillate you provide a separate osc instead.

ms660 · 26th Jul 2015, 5:17 pm

Looking at the article again it looks like it's a fore runner to what today would be called a direct conversion receiver.

Lawrence.

G6Tanuki · 27th Jul 2015, 9:01 am

Can be viewed as a direct-conversion receiver combined with a signal-frequency Q-multiplier.

Has the big advantage that the "local oscillator" can be designed to be stable and isolated from external effects by a buffer, and is not going to have its frequency pulled all over the place by the antenna swinging in the wind which is a problem with the classic 0-V-0 or 0-V-1 type regens.

[Am I the only one who always mentally models a direct conversion receiver as a superhet with an IF of 0 KHz?]

ms660 · 27th Jul 2015, 9:51 am

That's what I reckon too.

Always fancied building a hybrid direct conversion receiver, eg: Valve RF..Double balanced mixer (SBL1 etc)...DDS VFO...AF filters...Valve audio.

One day maybe.

Lawrence.

G8HQP Dave · 27th Jul 2015, 11:58 am

Quote:

Originally Posted by G6Tanuki

Am I the only one who always mentally models a direct conversion receiver as a superhet with an IF of 0 KHz?

No, for that is precisely what it is.

G6Tanuki · 27th Jul 2015, 7:52 pm

It gets a bit more mentally-complex when trying to visualise the resultant I- and Q- components.

Some decades back I built a 1.8MHz "direct conversion" receiver using an EF97 RF amp and ECH83 as oscillator/detector. Yes, 12-volt-HT valves: it worked surprisingly well to resolve top-band SSB stations and 'fish-fones'. Tuning AM signals was a problem - getting the VFO to within a few Hertz of the radiated carrier. I could happily do this because my receiver was powered with stable DC from a pair of 120A/H lead-acid batteries - but doing so revealed the Kilohertz driftiness-between-overs of the stations I was listening to.

G8HQP Dave · 27th Jul 2015, 9:30 pm

OK if you can think in complex numbers or vectors. If not, I and Q seem a bit abstract.

Phil G4SPZ · 2nd Aug 2015, 11:01 pm

The problem with DC receivers is not that the IF is zero, but that the image is twice zero away from the wanted signal! I've only ever used a DC receiver for CW, where the image is less of a problem than you might expect.

Skywave · 2nd Aug 2015, 11:21 pm

Quote:

Originally Posted by Phil G4SPZ

The problem with DC receivers is not that the IF is zero, but that the image is twice zero away from the wanted signal!

Pardon? How can something be "twice zero away"?

Can you clarify, please? But I've always had a bit of a job getting to grips with d.c. receivers and their consequent image, so perhaps I'm missing something obvious. (Not unusual!

)

Al.

Skywave · 2nd Aug 2015, 11:25 pm

Quote:

Originally Posted by G6Tanuki

It gets a bit more mentally-complex when trying to visualize the resultant I- and Q- components.

And even if I could, in what way would that help me design and / or successfully operate one?

Al.

ms660 · 3rd Aug 2015, 12:46 am

The only IF I can think of in a DCR is the demodulated audio frequency at any given instant in time, that being a frequency intermediate between DC and the RF signal frequency.

There's no image frequency in the conventional superhet sense in a DCR, not in the ones I've built anyways.

Not built an I & Q jobie yet, they work well apparently but not the same challenge like wot a twice the bedlam receiver is.

Lawrence.

Synchrodyne · 3rd Aug 2015, 7:11 am

In a direct conversion receiver, effectively the sidebands are images of each other. So, for example, if the received frequency is 819 kHz, then an upper sideband of 829 kHz and a lower sideband of 809 kHz both demodulate as 10 kHz. That’s fine if you’re receiving AM, but could be awkward with an SSB signal, and even more awkward with ISB, unless perhaps the receiver has very sharp RF selectivity, perhaps unlikely if it has variable tuning. Hence the need for demodulation along both the I and Q axes.

The I axis demodulator responds to the net AM from both sidebands.

The Q axis demodulator responds to any net PM. And there will be PM in the event of any sideband asymmetry, whether deliberate, as in the case of SSB, ISB or VSB, or incidental, as in the case of a selectively fading AM signal. That can be seen in the simple vector diagrams that are often shown in the literature for the AM, SSB and SSB cases. The carrier vector is usually shown vertically, with the AM rotating sideband vectors at equal angles and equal lengths each side of it. So, the resultant is also vertical, meaning no PM. If the sideband vectors differ in angle and/or length, or one is missing entirely, then the resultant deviates from vertical, indicating PM.

By phase-shifting and sum-and-difference matrixing, USB and LSB can be recovered from the I and Q signals.

Cheers,

Phil G4SPZ · 3rd Aug 2015, 5:37 pm

Quote:

Originally Posted by Skywave

...How can something be "twice zero away"?

Can you clarify, please?

I was speaking somewhat tongue-in-cheek, Al. I've only ever built DC receivers for CW or SSB purposes, and to be honest I didn't think they worked on AM as one sideband would be inverted in frequency, surely? But tuning a DC receiver is much like operating the BFO control on a superhet. Either side of zero beat, the same signal appears. With CW it matters little, but with SSB the audio gets inverted and becomes unintelligible if the local oscillator is the wrong side of the signal carrier frequency. The key (pardon the pun) to making a good DC receiver is the tailoring of the audio passband, not difficult for CW but more complex for good quality SSB.

I and Q demodulation is outside the area of my limited experience!

Skywave · 3rd Aug 2015, 8:46 pm

O.K.: thanks Phil. Your explanation has helped my comprehension. Yes, I've read books and various articles that explain DC receivers, but like most technical study, unless it is backed-up by putting what you've learnt into practice, (and preferably soon after), you don't really get 'the feel' as to what the theory translates to in practice. So it looks like yet another entrant to my round-to-it job list: build a DC receiver!

Quote:

Originally Posted by Phil G4SPZ

I and Q demodulation is outside the area of my limited experience!

Ditto. Have heard of those terms; never taken things further; remain ignorant of their meanings.

Al.

Herald1360 · 3rd Aug 2015, 11:00 pm

Quote:

Originally Posted by Phil G4SPZ

I was speaking somewhat tongue-in-cheek, Al. I've only ever built DC receivers for CW or SSB purposes, and to be honest I didn't think they worked on AM as one sideband would be inverted in frequency, surely? But tuning a DC receiver is much like operating the BFO control on a superhet. Either side of zero beat, the same signal appears. With CW it matters little, but with SSB the audio gets inverted and becomes unintelligible if the local oscillator is the wrong side of the signal carrier frequency. The key (pardon the pun) to making a good DC receiver is the tailoring of the audio passband, not difficult for CW but more complex for good quality SSB.

I and Q demodulation is outside the area of my limited experience!

On AM you have identical upper and lower sidebands so if the DCR is tuned to zero beat with the carrier (and stable enough to stay there) it is on the correct side of each sideband for everything to work just fine.

ms660 · 4th Aug 2015, 11:05 am

I & Q, A good video on the basics:

https://www.youtube.com/watch?v=h_7d-m1ehoY

From the same person, a circuit walk through for a DCR/SDR job:

https://www.youtube.com/watch?v=BG0rmMGpwMc

I've watched several videos by this person, I've always found them to be good.

Lawrence.

ms660 · 4th Aug 2015, 11:28 am

Part 2 to the first link posted above:

https://www.youtube.com/watch?v=5GGD...2yv6HzOZx8wfNj

Lawrence.

Phil G4SPZ · 4th Aug 2015, 10:57 pm

Useful information, thanks for that. The I & Q principle is of course embodied within synchronous detection, a process I'd all but forgotten about before reading this thread!

AlanC · 23rd Aug 2015, 4:16 pm

I've built some really good direct conversion receivers using rf preselectors ahead of an SBL1 mixer, with 'brick wall' audio filters and audio AGC. The performance of these gives any commercial radio a run for its money, and has less audio distortion than you get from an IF signal that's been through a crystal filter or two. But of course the image is always there. I still use one now that has a DDS synthesiser for the VFO, and you can set this to zero beat with an AM transmitter, and it will stay there all day! The DDS is not hard to build- there is loads of info and software for a PIC microcontroller on the web.

It occurred to me that in a direct conversion receiver with a traditional variable VFO, it ought to be possible to phase lock it to the carrier by DC amplifying the output of the ring mixer and feeding it to a varactor in the VFO.

I have a Sony ICF2001D which has a synchronous AM mode, there is a chip in it that in effect locks a VCO to the IF (there is a panel LED that tells you when it's locked) and the audio is demodulated in a balanced mixer. It eliminates the distortion caused when the carrier is reduced relative to the sideband(s) by selective fading, resulting in what looks to a diode detector like an overmodulated signal. It actually works spectacularly well.

Cheers
Alan

26th Jul 2015, 11:44 am	#1
ms660 Dekatron Join Date: Apr 2011 Location: Cornwall, UK. Posts: 13,454	Regen Receiver, Heterodyne Oscillator. Just came across this video: https://www.youtube.com/watch?v=27yoGmXtqQE The idea is not new, a brief description and schematic is given in this old article: http://kearman.com/images/W3LW-regens.pdf Any thoughts as to what is actually going on in the detector? Has anyone tried this method for SSB reception? Lawrence.

26th Jul 2015, 5:01 pm	#2
G8HQP Dave Rest in Peace Join Date: Sep 2008 Location: Solihull, West Midlands, UK. Posts: 4,872	Re: Regen Receiver, Heterodyne Oscillator. As I understand it (from reading the article - not seen the video) the idea is that an oscillating regenerative detector is quite fiddly (e.g. sensitive to fluctuations in supply voltage, very narrow bandwidth). A non-oscillating regen is much less fiddly, so easier to adjust. For CW (and SSB) you need a local oscillator, so instead of making the regen oscillate you provide a separate osc instead.

26th Jul 2015, 5:17 pm	#3
ms660 Dekatron Join Date: Apr 2011 Location: Cornwall, UK. Posts: 13,454	Re: Regen Receiver, Heterodyne Oscillator. Looking at the article again it looks like it's a fore runner to what today would be called a direct conversion receiver. Lawrence.

27th Jul 2015, 9:01 am	#4
G6Tanuki Dekatron Join Date: Apr 2012 Location: Wiltshire, UK. Posts: 14,007	Re: Regen Receiver, Heterodyne Oscillator. Can be viewed as a direct-conversion receiver combined with a signal-frequency Q-multiplier. Has the big advantage that the "local oscillator" can be designed to be stable and isolated from external effects by a buffer, and is not going to have its frequency pulled all over the place by the antenna swinging in the wind which is a problem with the classic 0-V-0 or 0-V-1 type regens. [Am I the only one who always mentally models a direct conversion receiver as a superhet with an IF of 0 KHz?]

27th Jul 2015, 9:51 am	#5
ms660 Dekatron Join Date: Apr 2011 Location: Cornwall, UK. Posts: 13,454	Re: Regen Receiver, Heterodyne Oscillator. That's what I reckon too. Always fancied building a hybrid direct conversion receiver, eg: Valve RF..Double balanced mixer (SBL1 etc)...DDS VFO...AF filters...Valve audio. One day maybe. Lawrence.

27th Jul 2015, 7:52 pm	#7
G6Tanuki Dekatron Join Date: Apr 2012 Location: Wiltshire, UK. Posts: 14,007	Re: Regen Receiver, Heterodyne Oscillator. It gets a bit more mentally-complex when trying to visualise the resultant I- and Q- components. Some decades back I built a 1.8MHz "direct conversion" receiver using an EF97 RF amp and ECH83 as oscillator/detector. Yes, 12-volt-HT valves: it worked surprisingly well to resolve top-band SSB stations and 'fish-fones'. Tuning AM signals was a problem - getting the VFO to within a few Hertz of the radiated carrier. I could happily do this because my receiver was powered with stable DC from a pair of 120A/H lead-acid batteries - but doing so revealed the Kilohertz driftiness-between-overs of the stations I was listening to.

27th Jul 2015, 9:30 pm	#8
G8HQP Dave Rest in Peace Join Date: Sep 2008 Location: Solihull, West Midlands, UK. Posts: 4,872	Re: Regen Receiver, Heterodyne Oscillator. OK if you can think in complex numbers or vectors. If not, I and Q seem a bit abstract.

2nd Aug 2015, 11:01 pm	#9
Phil G4SPZ Dekatron Join Date: Apr 2005 Location: Bewdley, Worcestershire, UK. Posts: 4,748	Re: Regen Receiver, Heterodyne Oscillator. The problem with DC receivers is not that the IF is zero, but that the image is twice zero away from the wanted signal! I've only ever used a DC receiver for CW, where the image is less of a problem than you might expect. __________________ Phil Optimist [n]: One who is not in possession of the full facts

3rd Aug 2015, 12:46 am	#12
ms660 Dekatron Join Date: Apr 2011 Location: Cornwall, UK. Posts: 13,454	Re: Regen Receiver, Heterodyne Oscillator. The only IF I can think of in a DCR is the demodulated audio frequency at any given instant in time, that being a frequency intermediate between DC and the RF signal frequency. There's no image frequency in the conventional superhet sense in a DCR, not in the ones I've built anyways. Not built an I & Q jobie yet, they work well apparently but not the same challenge like wot a twice the bedlam receiver is. Lawrence.

3rd Aug 2015, 7:11 am	#13
Synchrodyne Nonode Join Date: Jan 2009 Location: Papamoa Beach, Bay of Plenty, New Zealand Posts: 2,944	Re: Regen Receiver, Heterodyne Oscillator. In a direct conversion receiver, effectively the sidebands are images of each other. So, for example, if the received frequency is 819 kHz, then an upper sideband of 829 kHz and a lower sideband of 809 kHz both demodulate as 10 kHz. That’s fine if you’re receiving AM, but could be awkward with an SSB signal, and even more awkward with ISB, unless perhaps the receiver has very sharp RF selectivity, perhaps unlikely if it has variable tuning. Hence the need for demodulation along both the I and Q axes. The I axis demodulator responds to the net AM from both sidebands. The Q axis demodulator responds to any net PM. And there will be PM in the event of any sideband asymmetry, whether deliberate, as in the case of SSB, ISB or VSB, or incidental, as in the case of a selectively fading AM signal. That can be seen in the simple vector diagrams that are often shown in the literature for the AM, SSB and SSB cases. The carrier vector is usually shown vertically, with the AM rotating sideband vectors at equal angles and equal lengths each side of it. So, the resultant is also vertical, meaning no PM. If the sideband vectors differ in angle and/or length, or one is missing entirely, then the resultant deviates from vertical, indicating PM. By phase-shifting and sum-and-difference matrixing, USB and LSB can be recovered from the I and Q signals. Cheers,

4th Aug 2015, 11:05 am	#17
ms660 Dekatron Join Date: Apr 2011 Location: Cornwall, UK. Posts: 13,454	Re: Regen Receiver, Heterodyne Oscillator. I & Q, A good video on the basics: https://www.youtube.com/watch?v=h_7d-m1ehoY From the same person, a circuit walk through for a DCR/SDR job: https://www.youtube.com/watch?v=BG0rmMGpwMc I've watched several videos by this person, I've always found them to be good. Lawrence.

4th Aug 2015, 11:28 am	#18
ms660 Dekatron Join Date: Apr 2011 Location: Cornwall, UK. Posts: 13,454	Re: Regen Receiver, Heterodyne Oscillator. Part 2 to the first link posted above: https://www.youtube.com/watch?v=5GGD...2yv6HzOZx8wfNj Lawrence.

4th Aug 2015, 10:57 pm	#19
Phil G4SPZ Dekatron Join Date: Apr 2005 Location: Bewdley, Worcestershire, UK. Posts: 4,748	Re: Regen Receiver, Heterodyne Oscillator. Useful information, thanks for that. The I & Q principle is of course embodied within synchronous detection, a process I'd all but forgotten about before reading this thread! __________________ Phil Optimist [n]: One who is not in possession of the full facts

23rd Aug 2015, 4:16 pm	#20
AlanC Hexode Join Date: Jul 2006 Location: Sheffield, South Yorkshire, UK. Posts: 312	Re: Regen Receiver, Heterodyne Oscillator. I've built some really good direct conversion receivers using rf preselectors ahead of an SBL1 mixer, with 'brick wall' audio filters and audio AGC. The performance of these gives any commercial radio a run for its money, and has less audio distortion than you get from an IF signal that's been through a crystal filter or two. But of course the image is always there. I still use one now that has a DDS synthesiser for the VFO, and you can set this to zero beat with an AM transmitter, and it will stay there all day! The DDS is not hard to build- there is loads of info and software for a PIC microcontroller on the web. It occurred to me that in a direct conversion receiver with a traditional variable VFO, it ought to be possible to phase lock it to the carrier by DC amplifying the output of the ring mixer and feeding it to a varactor in the VFO. I have a Sony ICF2001D which has a synchronous AM mode, there is a chip in it that in effect locks a VCO to the IF (there is a panel LED that tells you when it's locked) and the audio is demodulated in a balanced mixer. It eliminates the distortion caused when the carrier is reduced relative to the sideband(s) by selective fading, resulting in what looks to a diode detector like an overmodulated signal. It actually works spectacularly well. Cheers Alan __________________ Always expect the unexpected...