The AM Forum

The way particular IC's come into and go out of fashion, there is nothing like a gross of same-type tubes to spark the imagination and quell fears of EMP and obsolescence.

Case of 6GH8's
Case of 6J6's
case of 70L7's
3 Cases of 12BQ6's !!

to the point, some time ago, there was a discussion of this and a plethora, or even a cornucopia, of schematics, or copied articles about synchronous detectors.

These were tube construction articles, fairly old ones from before and during the beginning of SSB ham popularity, some complicated and fussy looking, some more simple.

Does anyone recall where this might have been or what topic?

Having recently borrowed an outboard sync detector (little solid state box thingie) and used it with an R-390A, I'm converted to wanting a sync detector and want to study them before deciding what to do.

I don't really need another project but I like to read and think. My boss says extracurricular reading makes me smarter at work.

See http://amfone.net/Amforum/index.php?topic=18379.0;wap2

Thanks! You using a cellphone?

Specifically:

The article by Bill - W3DUQ from the September 1967 issue of 73 magazine is available at the link below.

http://www.amwindow.org/tech/pdf/w3duqsyncdet73mag1967a.pdf

A quote from W3DUQ, the author of the 73 article mentioned:

"Let's get rid of those outmoded, wideband, distorted telephone quality single sideband gizmotchies"

Nuttin' new under the sun....

73DG

I remember that article but did not realize that it was by W3DUQ.   Anyone planning on building it should try to get the demodulator tubes first.   They are hard to find.

It shouldn't be a problem to substitute some of the more readily available sheet beam tubes in place of the 7360 demodulator tubes.  7360 tube supply has dried up because of the commercial rigs that used them but other sheet beam tubes were used in consumer entertainment gear and are readily available at lower prices.

6AR8, 6JH8, and 6ME8 are the most common, but note that they all have different pin wiring, so they are not a direct substitute for the 7360.

The GE YRS-1, circa 1952 or so, is another sync detector that shouldn't be too awfully hard to copy.  There was a project in GE Ham News that was essentially the YRS-1.  The thing uses a passel of dual triodes but I don't think a 6J6 would be usable as the cathode is common in that tube (unless you wanna use each 6J6 as a single triode).

A sheet beam tube moves the beam from one plate to the other, like a CRT moved the beam right to left, but has no similar anodes.

Seems like the 6J6 mentioned or any high quality dual triode could replicate the function to a degree for some purposes.

Hey Patrick
Would there be any stability problems and-or alignment issues using tubes vs IC's and computer magic?
I realize there are some apps that tubes will out do a solid state circuit.
I wonder if anyone built Bill's circuit and how it would compare to today's Sync detector??
Would it give the Almighty Sherwood SE-3 a run for its money??

1929 Style Synchronous Frequency Detection:

http://www.google.com/patents?id=nzhHAAAAEBAJ&printsec=abstract#v=onepage&q&f=false  ;)
 

The Kolozsy circuit looks like a direct conversion receiver with a ring demodulator for a product detector. There is no indication of how the local oscillator "locked" onto the carrier. He leaves out the detail of how the local oscillator is brought into "synchrony" (lock) with the incoming carrier.

The three sheet beam tubes that Pete mentioned, were intended for color demodulator service in TV sets and are plentiful and fairly cheap. 6ME8 - $3, 6JH8 - $4, 6AR8 - $5.

Should anyone be interested, I have created spice models for these as well as the 7360, and LTSpice schematic symbols for all. LTSpice is a free spice circuit simulator available from Linear Technologies.

I may experiment with a Tayloe mixer for the product detector since the L.O. would be at four times the carrier and less likely to get into the I.F. and juice the AGC. Now, to implement the Tayloe in hollow state.

Right, it is unclear how the synchronous detection occurs. It's interesting the mechanical arrangement (fig. 1) for the detection (mixing?) of the desired signal by using the coils of the speakers.
It would be interesting your Spice model for the 7360 but unfortunately I never used the Spice Simulator (my hair and beard are too white?  :-[)

You are never too grey or white haired to learn about something neat, especially when it is free. Sure, LTSpice is aimed at their IC's, but they provide a pretty good set of provided models, and I will gladly share the extra goodies I have accumulated over the years. BTW, one of my associates who learned and uses LTSpice a lot is over 75. It is really good for simple things too, like filters. I used it to tutor my EE son in the workings of the TR-7, especially the front end filters.

I openly offer to help anyone who is interested in getting into using it.

I have a copy of a West Coast handook with a H mode mixer done with 4 triodes.
It was used as a DSB generator if I remember.
Lift the cathodes and add RCs and you get a tayloe.

LT Spice saves me a lot of time and solder, great tool

I can't answer that. There are too many variables. Would you agree that the main issues besides amplifier linearity is frequency stability and freedom from noise?

ICs in many cases have much better specs for stability and noise, and computers control precisely. The desire for tube gear is or a challenge to make it as good as possible. Certainly some of those designs would have been more than adequate.

I like Bill's circuit but am not going to build that one. I prefer to think about one one with a simple AFC circuit and balanced modulators.

This brings up the question of gating. Does one flip the signal polarity abrupltly at zero crossing or does one use a sinewave in a product detector?

Thanks for your offer of help Rob.  Actually I used a couple of programs, like the RfSim99, related to filters, and I found it very interesting and useful. What has piqued my curiosity is your model for the 7360, I'd be curious to see how it performs, as a mixer, (if this is possible) in a program, we can say virtually? I've played in the past with this mixer tube “live” connected to test instruments and in the “real world” connected to an antenna (maybe "someday" I will put a page on my website with some fresh results of these tests).

Besides mimicking the DC transfer characteristics of the tube, the model works in balanced modulator and mixer circuits in LTSpice. Like most spice simulations that involve RF, you have to take the results with a grain of salt. Typical inter electrode capacitance right form the RCA data book were included and you have a perfectly balanced tube. RF circuits with inductors need to have some measure of series resistance included or you may get unexpected results, even discontinuities, in the waveforms. From the 7360, models were derived for cousins 6ME8, and 6AR8. The 6JH8 is too close to 6AR8 to bother to make a model.

GE Ham News 1948 vol 3 nr. 6 - has a locking oscillator adapter for AM and SSB

QST July 1948 has something similar, an exalted carrier demodulator.

These have simple AFC at best, different at least in description from the W3DUQ circuit. I'm not sure they used I and Q back then, not by name anyway.

I stand corrected, to have it all as in W3DUQ's design from "73" September 1967, or these others, means a lot of tubes.

It's all good reading but a big project and no matter how simplified it could be made it is not for a weekend, too much for around here due to other things in the works.

I'd admire it if you guys design or build such a beast though.

I could give those boxes of tubes to the YL.

I confess!

I have given in to the lure of an appliance.. I ordered the rackmount version of the Sherwood Engineering Incorporated SE-3 MK IV adapter today.

This is a good decision based on the amount of time it would take me to build a similar item and the wish to try one out for SWL purposes, which I do a lot.

I asked the gentleman many technical questions in e-mail before deciding. All of them were answered well and after a phone call today I believe it will be very useful with the R-390A and/or the SX-28.

I have used the Rycom R1307A/GR with the R-390A's I.F. output. For elimination of unwanted sideband noises, It works well on AM in either SSB mode as long as the signal is zero beat, and it has an interesting AFC circuit so the BFO can be used in AM. It has no AGC but an input attenuator.
All of this is satisfactory but not at all automatic and takes some fiddling because the AFC is not a PLL and won't lock in over any wider a range than normal drift, so the fiddling comes in on round tables as there is always one person who is rockbound or a bit off. The schematic for the Rycom is a very interesting read (attached).

Once I receive the SE-3 MK IV unit I'll try it out and post in an appropriate topic.

I wish I was retired like some of the members here, with each day mine to do as I please, and I could then try building either the G.E Ham News equipment or the W3DUQ equipment, or even a combo, merged into one equipment and allowing either method of detection. Each has its own advantages. That could also be done at a parts count savings IMHO after studying both.

Thanks for the information and explanations. With particular reference to the 7360 as a mixer, I think these simulators are particularly useful when you vary the different parameters (plate voltage, grid, LO level, etc..) and then you can see improvement or worsening compared to the starting circuit. But I think there could be a substantial difference of measurement results between simulation and reality if one is not careful, as you suggested in your message.

Very interesting the Rycom R1307/GR receiver (and selective voltmeter), very good for VLF/LF reception and excellent as an “IF amplifier” for testing front ends. It would not be very difficult to add an AGC. Have you seen the characteristics of the audio output transformer?  ;)

This one only.. It would not be simple or cheap to replace it.

10KCT
250WV, 75mA DC
sec1 - 600 Ohm 10W
sec2 - 4 Ohm 10W
response +/-0.5dB 30C to 25KC
 (note no power bandwidth given)

The triode 6AQ5 setup can't make 10W, and the 160V B+ won't do it even if the screens were tied to + instead of the plates, but it seems like they considered that someone some day might hook the output stage to the 270V unregulated B+ and tetrode-ize it.

I'm happy with the perfect 1W or so I get!  - but the output stage's phase inverter seems a little weird, is it unbalanced?
Any more needed and I might as well use a 'shared' hi-fi amp.

It would be good to collect operating data on that stage and add it to the "triode connected beam tube" spreadsheet. The sheet has 200V and 250V data and it is very different than the 10K CT winding found in the Rycom data. - the 'book' specs go for 4K to 4.5K CT.

Yes, it would not easy to replace that transformer! That phase splitter seems to be a classic one, maybe not the best but it works. I sometimes connect an external speaker with very good results. This receiver has many interesting circuits, one is the product detector that uses a 7360, although not in its balanced configuration. The low internal noise generated by this receiver makes its use very pleasant. By adding a good SW converter (and an AGC) at its input (or directly to the 2215kHz IF ampli) would make an excellent general coverage receiver.   :D

It's a boon for those with the simplest receivers too as any SW receiver with an IF in the set's range will do.

I don't know why a 7360 was chosen except might might have low noise or few irregularities where linearity is concerned. A byproduct of beam deflection rather than gridwise gating. The 7360 is very interesting.

There's also the E80T by Philips. So are there no Octals? I guess the concept was not thought practical before.

The 6BN6 is a gated beam tube, and somewhere there's a picture of its structure. It is similar in some ways to half a 7360.

The 7360 and the rest act like a display CRT, but it is a signal CRT with just two 'deflection plates', sheet beam tube.. The deflection is more nice and linear, not curved like a grid curve.

The BFO is injected to the control grid, which is decently linear in this tube. That is necessary for the linearity of deflection to be taken advantage of.

It would be very interesting indeed if there were a tube that were a 4 quadrant tube with a pencil beam instead of a 2-dimensional sheet beam tube. Then it could receive balanced input from two sources plus overall control from the grid. constellation demodulator.

Note the 7360 used in the Rycom is not fed a balanced input to its plates but unbalanced, and one plate is at signal ground, like a basic scope.. sort of a waste in some ways.  Am I right on that?

attach: 7360 compare to CRT
attach: 7360 datasheet

Yes, as you said, it’s a waste to use the 7360 in an unbalanced configuration (one plate and one deflector grounded for RF), why Rycom used it in this way? I don’t know. Of course using the 7360 balanced would make life more complicated, a center tapped IF transformer for the deflection plates, a center tapped Audio transformer for the plates……..and more expensive!  But the R1307/GR was a high class professional test Instrument, so why?  ::) This receiver should be from the early sixties (or late fifties) at that time the 7360 tube was in vogue…….  ;D
An Octal beam deflection tube was the US Navy CKR 1636 from RCA that was used in a radar. In the mid thirties a Danish engineer invented the Renode tube which was a sort of beam deflection tube, both derived from the CRT.

Patrick

You will like it.  I actually prefer it over my Flex 5000 SAM performance.  Once one gets over the extreme price, the rest is a pleasurable experience.  I use mine with either my SX-73 or more likely, the SP-600.  The only fault I find is not related to the SE-3 but with the fast AVC of the receivers.  I am spoiled with the adjustable AVC timing on the Flex (and my FT-301). I don't like the extreme noise between transmissions.  Some day, I will play with the AVC timing circuits on these receivers. - yet another post-retirement project.

Al

Wow - the 1636 is  beautifully made tube and I bet it wasn't cheap! I see none for sale with a quick search.

On the AVC and SE-3, The R-390A has a very slow setting but since it is not a SSB rig, my concern is that the AGC will do nothing. I know that when using the BFO on the R-390, the SSB is frequently distorted and I have to ride the RF gain control to properly listen to SSB on that set. This issue also applies to the SX-28. The AGC in that one comes from its own transformer and it seems very narrowly tuned to the place of the carrier. The SX-28 has also an AVC circuit, but it mostly controls the output volume. I'll have to try it all before judging. The guy at Sherwood said he didn't have any built up, so it is taking a bit for mine to show up here but I can wait a few days. He said he would include the schematic upon request with the new unit. I asked about that should I ever need to service it out of warranty. He also said it is all analog unless one considers a PLL to be digital. That is what I like, an analog path for analog signals.

I have not warmed up to DSP products because I don't know, and am not informed by the manuals, what is really being done in each product (I understand the principles), but I like ASP no matter how many tubes (or ICs) it takes.

The SE-3 arrived today. Very nice looking. Need to find a socket for the necessary 24V power as it is a kind of smallish 4 blade Jones connector (I would think a socket would have come with it for the price. Might just use clip leads.) and need to make up a cable for the I.F. The Schematic will take some reading.

It didn't come with a wall-wart PS?  Mine did

I just checked the web site - you need to call Rob up.  It comes with a 20 VAC wall wart and a connector.

Al

Having read that, I searched the box and found the wall wart and with attached connector. I must have missed this before. The packing material is peanuts and the box is quite large because I ordered the rack-mount model so that's why it didn't turn up before. It won't be long before I get to try it out!

WHEW!  Glad I was helpful.  You are going to LOVE this unit.  Sure it is pricey but the quality coming out of that detector is so good you are going to forget about using the receiver's inboard AF.  I'm assuming you have a good amp and HIFI speaker to hook it up to.  Do you work the east coast much and if so, what band?

Al

I saw this PDF article on my Destop.  I must have found it on another hams web site.  It is titled "The Synchronous Detection Process" by W3DUQ.

Open up the attachment and and the third page down (listed as page 6) is the article.

Chuck

I have a great hi-fi, a Stromberg-Carlson 30 watt integrated hi-fi amp in a large wood console with a 12" "Custom 400" coaxial speaker.

I don't get much contact with the coasts. I am told the ol' 4-legged dipole shoots North and South mostly. I am also told that if I swap the legs around, that it will fire more E-W.

A plan drawing is attached that might explain it.

The legs vertically are at about a 45 degree angle sloping down from the feedpoint at 55FT toward the ground. Each is about 63FT long. Before they get there, there is guy rope and each rope's terminus is about a 10-15FT tall pipe. I am guessing the ends of the wires are some 20+FT above the ground.

The feedpoint where the V's of the legs are joined is at the 55FT level. I can't get to it easily and I am afraid to send a weak link like a relay up the tower.

If it is true that switching the electrical direction of the vee would move the 'beams' 90 degrees, then I was thinking of running another OWL up there so there would be two feedlines. Reversing polarity on one would switch the beam. The relay could be in the shack in the ATU enclosure.

Seems simple but will it work? Anyway that's the dope on why I don't do the Coast. I get good reception with the gulf and north through the Dakotas.





That's the W3DUQ design, with many tubes!

The link was posted April 12th, in this same thread. No mystery

Okay.  Yes, there are lots of tubes.
Chuck

I like it, and if I could make the time I would like to try building something like that. Building receiving equipment that gives professional results is much more challenging than building the same level of transmitting equipment.

I'd thought there were some simpler tube type synchronous detectors out there but after a couple months of reviewing the materials and coming to understand the functions involved in the methods, I can only submit the following theorem:

"There is always a positive correlation between the number of tubes in a circuit and the degree to which perfection is approached".

Unfortunately this is not a linear relationship. It is more like P=nT^2.

The W3DUQ unit looks really neat!  Functionally, it looks just like the Binaural DC receivers of late but adds an audio phase shifter so that you can get either USB or SSB without the binaural function.

A phasing type SSB receiver works like a phasing type exciter, but in reverse. You down convert to I and Q audio signals with a pair of mixers that are driven by quadrature LO's. The I and Q signals are then sent through phase shifter networks that have the difference between them be 90 degrees. You then either add the outputs or subtract them to produce either USB or LSB audio. In the W3DUQ unit, the I and Q signals are also made available for stereo headphone output. Spatial ear/brain processing is how the signals are separated....like listening to one conversation in a room of people. W3DUQ also provides a means to phase lock the I/Q LO to the incoming carrier if present. Less clearly understood is the SSB mode where the LO frequency tracks an average DC component from the mixers.

For applications at a fixed frequency, i.e. a detector at the end of an IF strip, you could use a crystal to hold the LO on frequency for SSB. In a PLL, a crystal controlled VCO does allow pulling, but has a very tight lock-in range, as expected. The Signetics PLL ICs of the seventies showed such a scheme in an application note. In the tube world, this type of I/Q demodulator with a crystal stabilized oscillator was used in tube type color TV sets for the chroma demodulation.

 

I remember the TV circuits. The crystal was pulled during burst and was expected to remain close for the rest of the line, so it only had to stay put for 63us. If the oscillator was out of alignment there would be vertical bars of wrong color positioned across the screen.