SSB

Just a guess but try opencore.org

Not im my ham days (70's and 80's). Even at large hamfests it was next to impossible to find a crystal filter with decent shape factor for under $100. But I was able to whip up a good phase shifter for under $10. Of course that's all different in a production setting since combining (and finding...) the correct capacitor combinations in my limited stock took half a day.

Mechanical filters were absolute rarities even back then. You had to come up with some serious dough to lay your hands on one of those. And then you still wouldn't know whether it had been damaged. I don't remember but they exhibited some weird pathologies and really did not like to be bumped around in a trunk.

pricey.

Well, maybe it was that those guys could afford Collins gear, so a $100.00 mechanical filter was a drop in the bucket. ;-)

But I seem to recall seeing them in the Allied catalog for way less than that, albeit for the era, the price might have been equivalent to $100.00 at today's prices. :-)

And I've _always_ wondered what's inside those 2Q4 plug-in phase shifters!

Cheers! Rich

But the manufacturer's cost has never been the retail price. Over thirty years ago, looking for a crystal filter, I contacted the local Heathkit outlet to check the price of one for one of their SSB rigs. You could buy the parts for replacement purposes. The price I was quoted was a significant percentage of the cost of the rig itself, which is something fairly constant. So the manufacturers were getting a good price on filters, be they crystal or mechanical, because otherwise the rigs would be far costlier.

I suspect you saw the Japanese made mechanical filters, that were in the catalogs and showed up in some of the early made in Japan ham equipment, and from the ads I saw (later) seemed relatively cheap.

Hams also saw Collins mechanical filters because they appeared on the surplus market, either by themselves or inside surplus eqipment.

Not much. Something like four capacitors and four resistors. But then, they weren't wideband, they were intended for voice-bandwidth. And users could get by with a limited suppression of the unwanted sideband. It was the relatively early days of SSB for amateur radio, and once they got rid of the carrier, getting rid of the opposite sideband was icing on the cake. So they could live with not that great opposite sideband rejection.

And as I said in a previous post, those phasing networks were mostly seen in transmitters, where the signal they saw could be limited by a audio filtering (if the voice itself wasn't actually limited to 300-3000Hz or so), so the phasing network didn't have to deal with a wide range of signals.

Amateur SSB was dominated by the phasing method in the early days, because it was doable. Circa 1947, there wasn't much else, and it was relatively cheap to get a few matched resistors and capacitors. Unless one started with LC filters at a very low frequency, and then worried about image rejection as it was heterodyned up to the desired frequency. But the late forties or early fifties brought the mechanical filter, so that started changing things. And then the mid to late fifties, hams caught on that they could use surplus crystals in the 455KHz range to make crystal lattice filters. There were a lot of crystals lying around from WWII in that range, because they'd been multiplied up to a desired frequency, and since they were in FT-243 holders that could be unscrewed, if the frequency wasn't quite right, one could always grind the quartz a bit.

I think those proceeded the introduction of commercial crystal filters, but my memory of the details may be wrong.

Then a few years later, they were able to make useful crystal filters in the HF range, both by hand and commercially, which bumped things up a notch.

ONce there were filters, commercial or home made, that put the phasing networks aside.

It was only in more recent years, I guess it's about twenty now, that phasing was looked at again seriously. SOlid state helped, since fancier audio phase shift networks didn't come at the cost of more bulky tubes.

ANd I suspect even more interest has been driven by digital processing. SOme have used it for the audio phase shift networks, but it just seemed as commercial interest rose in that area, it was like a pointer saying "look at all you can do with phase shift networks".

Michael

pricey.

This was in Europe where stuff from the US is usually more expensive than here.

AFAIK it was described in an ARRL Handbook from the early 90's. Don't remember what the phase error on those B&W shifters was but I believe over 1%. That would be ok but IMHO not much to write home about. In med ultrasound we had to do better than that.

The circuit given (after correcting the top left opamp wiring) if rather similar to phase shifting circuits used for those old quadraphonic (QS? SQ??) phase-method 4-channel hifi systems decades ago. It will probably give a bit of amplitude variation through the audio range, but not too bad (do a spice simulation to check). Is it for voice or "hifi" audio??

Mark A

John Lark> The schematic he posted is trickier than the standard active allpass,

On a sunny day (14 Feb 2007 00:38:00 GMT) it happened snipped-for-privacy@FreeNet.Carleton.CA (Michael Black) wrote in :

I build a receiver with the XF9A xtal filer (9MHz) early seventies IIRC. Those were German made Xtal filters, XF9A and XF9B, and less then a hundered guilders then (came with upper and lower sideband xtal), so that would have been less then 25 dollars in those days currency. These were really good filters. Several designs with these were published ine the Veron amateur radio bulletin around that time. My litte receiver was based on one of those, with 2 JFETS as RF pre, a dual gate MOSFET as mixer, then the XF9A, a JFET source follower, a CA3028 first IF, then one as balanced mixer / product detector, and a CA3020 audio 'power' amp .... with a transformer of course. And varicap tuning.... JFET LO.

Exactly, too much drift with ECC85 tubes.... hehe.

pricey.

It doesn't matter. The crystal used for the final mixer determined the mode, not the filter. Of course, this was a SSB filter, not for AM, so there is no "Center" frequency. A 6 KHz wider filter was used for AM reception.

I paid $20 for it, in the original package at the Dayton hamfest in the early '80s. There were always a few floating around for sale at a show that size.

These will mil spec, not those Japanese units sold by Lafayette, and other importers. BTW, its kind of ironic that one of the first US companies to import boat loads of early Japanese electronics was put out of business by cheap imported crap, isn't it? :)

pricey.

But he's alluding to the fact that in some cases, one could get matched sets of mechanical filters, one for upper sideband, the other for lower. The BFO crystal would be in the middle, and never change.

The tradition is normally to have a single filter and switch the BFO frequency. And in some cases (maybe many?) the filter passband is assymetrical, so there's a tradeoff there.

And of course, once you have a USB and LSB filter in place, it doesn't take much more circuitry to allow for independent sideband, ie transmitting and receiver both sidebands at the same time, but with different information on each.

Michael

...

Is that the basic idea for "stereo AM"?

Thanks, Rich

Later Japanese electronics were (and are) really good though. I have several products from Japan Radio Corporation (JRC) here in the lab. Top notch workmanship. Their circuit design is IMHO not quite as innovative as in US products but the mechanical quality is better. Then again after some not-so-good experiences I am not sure at all that the younger generation of engineers could live up to the level of design skills of our retired generation. Academia doesn't seem to breed guys like Bob Pease or Ulrich Rohde anymore. Heck, most can't even solder when they are handed their degrees.

Is that still around?

NOt originally.

It was so you could send two channels of information, and the real user would have been military. Send teletype on one channel, and voice on the other. IN amateur radio, the only time anyone had the need to talk about ISB (Independent sideband) was for slow scan tv (which took 8 seconds to build up the picture), where being able to talk at the same time as sending the picture made sense. Though, it was never common, since it either required an ISB transceiver (and none were commercially made for amateur use, so it would have to be homemade or bought as military surplus), or two complete commercial transceivers with suitable splitting and summing between them and the antennas.

As you might recall, there were various schemes for AM stereo. ONe of the schemes did use ISB, but it wasn't the one that won out.

Michael

True but the C-QAM standard ia partially compatable. See:

Phooey, and i was going to suggest using direct digital translation using a Spartan 3. He beat me to it with only a Spartan 2. Well no, he implemented a digital Weaver system, not the same as direct translation.

What do you mean by direct digital translation?

I suppose that an ideal ssb modulator has some impulse response, so one could map that directly into a single fir filter, given enough macs.

John

By doing directly in time domain the equivalent of successive DFT, translation up frequency, mirror image elimination, and DFT back to time domain.

But that's a batch process, isn't it? Not suitable for continuous speech. I think there are "rolling" FFT algorithms... I may even have one around here somewhere.

John

It is likely that less processing would be needed to do the phase shifters as a bunch of IIR filters. A DFT needs a lot of multiply-adds