Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1] 2 Next   Go Down
  Print  
Author Topic: BC-453 Multiband Converter  (Read 9387 times)
KB1WSY
Member

Posts: 714




Ignore
« on: February 07, 2012, 10:04:14 AM »

I'm a newbie in the Boston, MA area: taking the Technician and hopefully also the General test at the end of this week. Am renewing with this hobby after dropping it more than 40 years ago, when I was 12 years old -- I passed the British test back then but never actually acquired a callsign because teenagerdom intervened and my interests shifted! For the time being I am going to stick entirely with CW, with tube technology, and entirely homebrew (with a little help from vintage surplus equipment).

I am starting with the receiver, which will be a converter to stick in front of a BC-453 that I picked up to use as a tunable IF (190-550kHz). Initially the converter will be a one-band or two-band crystal controlled thing, very simple. But I would eventually like to make it multiband for all the ham HF bands. The BC-453 has only 360kHz bandspread and a receiver that would cover all of the HF ham bands would have a large number of crystals, if I stuck with the "one crystal per injection frequency" system (and even if I made use of some useful harmonics).

But I had another idea. How about using a 100kHz high accuracy crystal (like the ones used in calibrators) and picking up the harmonics that I need? As far as I know the challenge would be to get strict filtering at each injection frequency to avoid birdies every 100kHz, but I am prepared to build the required filters to prevent this. If it works, this system would be very flexible and would need only one fundamental oscillator and one crystal. Do y'all think it would work? What is wrong with my idea? (Remember that I am sticking to tube technology circa the 1950s, thus cannot use a solid state synthesized frequency generator or anything like that!) For what it's worth, I have read a gazillion vintage ARRL and other articles about BC-453 conversions including the famous Stoner article on the "Novice Q5er" and numerous others.
Logged
WB6BYU
Member

Posts: 13112




Ignore
« Reply #1 on: February 07, 2012, 10:51:22 AM »

The modern approach would be to use a phase locked loop that divides the hetrodyne oscillator
down to 100 or 200kHz and compares it to a reference at that frequency.  The divider chain
is then selected for the desired output.  There were some professional synthesized receivers
from the tube era, but the circuits aren't as simple as they are with specialized ICs.

But I think I've seen circuits using tubes where you would apply a reference frequency
to an oscillator and it would lock on multiples of the reference frequency when the oscillator
was tuned close enough.  (The Wadley triple-mix scheme is another example, though not
perhaps the ideal approach for a beginner.)

I would suggest you start with a modular approach - for example, with the oscillator section
in a plug-in board that would allow you to start with a couple of crystal converters, then
change it out to something more elaborate.  You'll need good front-end selectivity if you are
feeding a 500kHz IF if you want to avoid images. (Building a second IF strip for 9 MHz or so
would be another good approach for the higher bands.)

I've seen a lot of FT-243 crystals available surplus at cheap prices, because they aren't
in the ham bands.  But you don't care about that - in fact, you want them about 500kHz
away.  For example, a 4 MHz crystal would cover 80m (with optimum image rejection in
the CW range around 3.5 MHz.)  There are lots of crystals in the 6 and 8 MHz range that
could be used for 40m - the FT-243 holder has the advantage that you can take out the
quartz and grind it to a higher frequency if needed.  I'd recommend you plan to use the
400 - 550 kHz tuning range for the most important part of the frequency range on each
band due to image rejection.

For the lower bands you don't even need to use a crystal oscillator - an LC oscillator may
give sufficient stability, especially if you have some way to calibrate it periodically.
You could use a switched inductor or capacitor (or bank of mica trimmers perhaps) to
set the oscillator frequency on each band, then use the BC-453 for fine tuning.


I'm on vacation and don't have my reference books handy, but I'd start by looking
through G3VA's Amateur Radio Techniques[/i] and see if he has something useful
in there that gives you some ideas about oscillators. 

But the main thing is to lay out the receiver/converter so it is easy to try new ideas,
circuits, etc.
« Last Edit: February 07, 2012, 10:54:17 AM by WB6BYU » Logged
G3RZP
Member

Posts: 4442




Ignore
« Reply #2 on: February 07, 2012, 11:01:55 AM »

The BC453 will give you image problems on the higher bands. The 3 to 6 MHz version (BC454) would be better: you could then come off its mixer plate at 1415kHz, and mix again with a 1 MHz crystal to get the BC453 selectivity. Change the 12 SK7 1st IF in the BC454 to another 12K8 and feed the output to the BC453.
Logged
KB1WSY
Member

Posts: 714




Ignore
« Reply #3 on: February 07, 2012, 11:37:37 AM »

Both WB6BYU and G3RZP mention the desireability of a 2nd IF stage in the multimegahertz range, either internal, or through the artifice of adding a 3-6MHz BC receiver in the chain. In fact I was originally thinking of doing that, even though it complicates this project -- which is only my second homebrew effort. (A 3-tube regen superhet is my first project.) Adding the 2nd IF would have the additional advantage of being able to include a sharp-skirted filter for CW.

But I got the impression that in that case, and even with the relatively limited tuning range of 360kHz edge-to-edge in the BC-453, I would need to make the 2nd IF stage tunable, which introduces all sorts of issues with capacitor ganging and so on. As opposed to just having a single variable tuner i.e. the BC-453.

I hope this is making sense. As of now I'm writing based entirely on theoretical knowledge, not real life experience and I'm probably getting things mixed up to some extent.

I like WB6BYU's suggestion to do it in a modular fashion and gradually get more complex. In fact I already had plans for a tube version of a "breadboard" (surprisingly hard to do because you have to make most of its yourself) and that may be the first thing I build, before putting actual circuits together.
Logged
KB1WSY
Member

Posts: 714




Ignore
« Reply #4 on: February 10, 2012, 09:37:40 AM »

>>The BC453 will give you image problems on the higher bands.<<

When would that set in? Above 20m for instance? I am thinking of starting out with just 80m/40m/20m. I can see the advantage of a second IF but I haven't built any tube stuff for 40 years, so would prefer to leave that until I have more experience. Best to get a single superhet working first. Also, staying below 20MHz will allow me to use my stash of NOS ceramic coil forms that have "red" cores -- according to the 50-year-old printed insert that came with these parts, that frequency is the cutoff above which "green" forms are best and I don't have any of those handy.

Many thanks for the advice.
Logged
WB6BYU
Member

Posts: 13112




Ignore
« Reply #5 on: February 10, 2012, 10:35:35 AM »

The upper frequency depends on the selectivity of your front end and the acceptable level
of images in your receiver.

When your IF is 500kHz the image is 1 MHz away.  Tune the BC-453 down to 200kHz and
the image is only 400kHz away.  That's 25% or 10% of the receive frequency when tuned
to 4 MHz, or 7% and 3% at 15 MHz.  Then look at the response curve and see what the
attenuation is that far from the center.

The other issue is how many strong signals are in the image range.  If 2.5 and 4.5 MHz are
quiet, then you shouldn't have much trouble using the high end of the range for 3.5 MHz.
But if you use a 3.7 MHz oscillator and tune the BC-453 to 200kHz to listen on 3.5 MHz,
you'll also hear a lot of strong SSB signals from 3.9 MHz.

Similarly, trying to tune 20m may not be an issue with low side injection if there aren't
many strong stations around 13.5 MHz, but with high side injection you'd have a lot of
problems with the SW BC stations around 15 MHz.

I suspect you could get by up to 20m with careful selection off injection frequencies, but
that would be the place where I'd think you'd start wanting to use a higher IF as a later
improvement.

One possible approach is to use the IF transformers and filters designed for 10.7 MHz FM
stereo receivers to make an IF strip that is 100 to 200kHz wide, then use the BC-453 to
tune across that range following a down converter.  That should give you sufficient
selectivity in the first IF to reduce images and improve stability, while still using the
BC453 tuning knob to tune across the band.
Logged
KB1GMX
Member

Posts: 739




Ignore
« Reply #6 on: February 11, 2012, 01:31:49 PM »

Ok,  the BC453 is a start but if you use it with a converter the highest Id' go with it is maybe 40M.
otherwise image rejection will be pretty bad.  the general rule is to keep the first IF at not less than 5-10% of the receive frequency.  So that at 5% makes it to 3.8 mhz at the low end and 11mhz at te high.  It would be less than satisfactory for most above 80/75M.

Back in the day, the 453 was popular as a replacement for a radios weak IF at 262, 455 or 500khz
and was the primary radio for Q5ers.

The solution if using that radio is desired is to use it as a fixed IF at say 500Khz and do a dual conversion with a first IF at 3mhz then a converter that is tuneable with an output at 3MHZ.

I'd suggest finding a old copy of the ARRL Radio Amateurs Handbook from the mid to late 50s
they can be found and often cheaply from used book sellers.  Also on the net check:

http://www.qsl.net/n6ev/Glowbug.html 

or

http://home.comcast.net/~btse1/vintrad/hb/hb.htm

As both feature old tech glow in the dark radio.

Allison
Logged
KB1WSY
Member

Posts: 714




Ignore
« Reply #7 on: February 11, 2012, 03:09:20 PM »

All this advice has been very useful, leading me to the conclusion that a single-stage conversion based on the BC-453 should be kept to the simple principles of the original "Novice Q5er" and stay with 80 and 40 meters (and perhaps 30m, which looks like it might be good hunting ground for CW). At least, that is what I will build first. When I progress to a  more ambitious multiband receiver I may or may not stay with the ARC-5 series; it probably makes more sense to build something completely from scratch, although that is quite a daunting undertaking. Simultaneously I will be building a transmitter, probably the "Novice Special" from the 1950s/60s. Plus negotiating with my landlord to allow some kind of antenna.

I did a certain amount of homebrewing 40 years ago as a 12-year-old but in those days I was so keen to complete the projects that I never really learned enough about how the whole thing worked, down to the individual component level, although I understood the general theory. Ironically, I never actually got a callsign although I did pass the British test. This time, I am determined to understand it well enough to be able to make design decisions of my own (and to troubleshoot if things don't work right).

I have a large and growing pile of vintage publications from ARRL and other sources, which have been a great help, and so has this forum.

The other motivation to go for simplicity is that I'll get on the air faster! I passed my Technician and General tests today so should have a call sign within days. But even at the simple end of the "tube homebrew" spectrum it will take months to get on the air. I don't have a shack set up, I haven't procured test equipment yet, and have only just started obtaining the right tools. What I *have* been doing is building a stash of vintage components for the RX and TX. And learning CW!!! Am using the Koch method starting right in at 15WPM. I've done the first 8 characters and am copying them more or less "reflexively" which is fun. I'm expecting it will take a few months to finish the full character set though....

73 de Martin, callsign pending
Logged
AA4PB
Member

Posts: 12766




Ignore
« Reply #8 on: February 11, 2012, 05:38:51 PM »

As I recall, the original concept of the BC453 Q5er was to use it as a fixed IF at 455KHz and connect it to the 455KHz IF output of another receiver. The VFO in the other receiver was used to tune in the signals, converting them all to 455KHz.

When you use the BC453 as a tunable IF for a fixed converter you will run into image problems at some point, as explained by others. If you've got strong broadcast signals in band (such as on 40M) you are likely to hear them as an image. The front end filter in the converter is what keeps the image out. That's hard to do when the images are so close to the band that you want to pass. In early designs they used dual gang capacitors, one to tune the VFO and the other to track a front end filter to the receive frequency.


Logged
KB1WSY
Member

Posts: 714




Ignore
« Reply #9 on: February 11, 2012, 07:14:09 PM »

I understand the concept you are describing, but I don't think it's the one at the core of one of the best known BC-453 "Q5er" circuits, the one by Don Stoner, whose article was published in the late 40s or early 50s and is available here:

www.w7ekb.com/glowbugs/rx/NoviceQ5er/NoviceQ5er.pdf

Stoner's circuit is a simple Xtal-controlled local oscillator converter for 80m and 40m feeding straight to the BC-453 used as a *tunable* IF. Stoner even stresses how useful this is because it takes so many turns of the dial on the BC-453 to get from one end of the ham band to the other end. (And there is a lot to be said for that very slow tuning dial.)

Now, there were of course lots of other conversion circuits, most of them published after Stoner's, and I don't know which one ended up being dominant in the '50s and '60s (it probably wasn't any single one). I researched this at some length and was bewildered by the profusion of converter circuits, although most of the ones I saw were single-stage.

There were also some more elaborate multistage ones. The most complicated I found was in the February 1961 QST: "The BC-453 as a Tunable IF in a Multiband Receiver" by Carl H. Ericson, W2PPL. That one was so relatively complex that it even included a mechanical coupling so that the BC-453 tuning dial could be ganged together with the earlier stages in the added front end. (I cannot post a direct link to that article but if you are an ARRL member you can find it in their online archive.)

I was all set to do something very simple like Stoner's circuit, to get me started with ham radio, when I came across an article in the September/October 1960 issue of "GE Ham News" which you can find here:

http://n4trb.com/AmateurRadio/GE_HamNews/issues/GE%20Ham%20News%20Vol%2015%20No%205.pdf

This is actually an article about creating a mobile (automobile mounted) BC-453 installation and it includes a multiband converter (all bands between 80m and 10m albeit with limited coverage on 10m). I got quite excited about this because the circuit is almost as simple as Stoner's but covers all those bands, albeit with far greater mechanical complication because of the numerous coils and their attendant 5-position, 5-deck rotary switch (that is one reason why I posted a separate eHam thread about homebrewing 1/4" slug-tuned coils). As for the local oscillator, the design limits the number of crystals by taking advantage of harmonics, but you still need several crystals of course.

Now however I have been persuaded, by the cogent explanations on this forum, that such a design would likely be flawed because of image response on higher bands. So I will be pulling back to the earlier plan of just doing the Stoner circuit or something like it. If I have time I may experiment with the GE Ham design as well just for kicks. I suspect that, in the context of a mobile rig in the early 1960s, things like image response weren't considered too big a deal -- it was enough of a miracle to cram all that gear under your dashboard or in the back seat!

73 de Martin (callsign pending)
Logged
K4FMX
Member

Posts: 17




Ignore
« Reply #10 on: February 15, 2012, 02:30:08 PM »

I have a BC453 project that I started many years ago. Every few years I drag it out and do some more work on it.
I have added a product detector, IF driven AGC, audio amp, all band converter and power supply.

Look at the converter article in Stoners SSB handbook printed by CQ magazine. It has an all band converter that uses a single 3.5 MHz crystal. The converter is double conversion on all bands except 80 meters where the first converter acts straight thru as an additional RF amplifier.

All of the bands above 80 meters are converted down to the 80 meter band which helps the image problem. The 3.5 MHz crystal will let you tune from 3700-4000 on 80 meters. on the higher bands the first converter works the same and in addition there is a harmonic multiplier to make use of 7, 10.5 etc for the other converter.
The first IF has a bandwidth of 300 KHz (from 3.7-4.0) which the BC453 tunes across.

while the concept is neat it doesn't work quite as well as Stoner says it does. If I remember right he claims that you just add the dial on the 453 to the crystal injection frequency to read where you are. But if you use the injection harmonics that he shows in the article you will be tuning backwards on some bands. There are also problems not only with imiges spaced by twice the IF frequencies (both the 200-500 KC and the 3.7-4.0) but also problems with the harmonics of the 3.5 oscillator causing reverse imigaes to be present. Think about high side injection and low side injection at the same time due to the crystal harmonics!

My latest change was to add a different, second, crystal for the higher bands that operated at higher fundamentel frequiencis so that the harmonics of it don't cause high side and low side injection at the same time.

I am also working on adding gang tuning of that variable 3.7-4.0 IF to eliminate imigas on the conversion to the BC453. That seems to work pretty well but it adds another knob to peak in addition to the front end ganged tuning.

It is a fun project and can be made into a very nice sounding, very stable receiver that you can directly read out the frequency on.

By the way, using an 80 meter command receiver as the front end is not great as they were never very stable and dial readout/tuning is something left to be desired compared to the BC453.

Mine is all apart now in a box where it will probably stay for another couple of years until I get the bug to work on it again.

73
Gary  K4FMX
Logged
N2EY
Member

Posts: 3875




Ignore
« Reply #11 on: February 15, 2012, 02:31:10 PM »

I have actually done the BC-453/converter thing several times, going back to about 1970. It is a very good combo for the complexity. Puts most simple tube receivers to shame.

Here is what I've learned:

1) The idea appeared in QST at least three times, and was in CQ and PE as well. The W6TNS implementation looks nice but is unduly complicated, using unobtanium coils and ganged RF/mixer circuits.

2) The biggest single problem is image rejection. You want at least two good tuned circuits before the first mixer for frequencies above about 5 MHz.

3) You can go as high as 20 meters if you're willing to add even more front-end selectivity.

4) For a first try, use the W6TNS circuit, but spread it out and use separate variables for the RF and mixer. You can wind coils on plastic pill bottles to do the job - #26 enameled is a good choice. An RF gain control in the cathode of the RF amp is a good idea. With 365 uuf AM BC variables and good coils you can cover from below 80 to above 40. You may even be able to cover 160 through 40.  

5) At 350 kHz per xtal, it takes a lot of them to cover from, say, 2 to 8 MHz. But there are some tricks, such as using both the high and low side, and using harmonics.  

6) There are FT-243 crystals available outside the ham bands fairly cheaply. A fellow on eBay sells them for $1.29 apiece, and has a pretty good selection.

7) If you are like me you soon get used to doing the mental arithmetic of seeing, say, 500 on the dial and knowing it's actually 3550 because you have a 4050 xtal in the converter.

Cool The big advantage of the idea is simplicity. Try for just 80/40 at first. (One sure route to failure is to attempt a too-ambitious project on the first go).

9) Get the BC-453 working right first. Remember that even a mint one is probably over 60 years old! The metal-can capacitors may be bad (or they may be perfect), and the resistors should be checked.

10) The really good news is that the manuals for the Command sets are available free for the download, so you can do by-the-book alignment, see the actual schematics, etc.

11) Join the Glowbugs reflector. You'll be glad you did.

-----

A Story:

Some years back, ARRL added a special Field Day bulletin to the W1AW schedule, as a way to earn bonus points. Copy the bulletin and send it in for 100 bonus points. First transmission of the bulletin would be 10 AM on Field Day Saturday. (They have since added earlier transmissions).

I was going on Field Day with a local club, and discovered a few days before FD nobody had made plans to copy the bulletin. Knowing this group, they'd not have anything ready to go at 10 AM Saturday. This was a group that was all about the latest sand-state stuff....

So I tossed the 453/converter combo, a dynamotor, a pair of cans and some wire in a box and brought it to the site with all the rest of the stuff. About 9:45 I set up the receiver and dynamotor on a picnic table next to the car, tossed a bit of wire up in a tree, clipped onto the car battery and flicked the switch. The dynamotor whined, the rx came to life, and soon W1AW was pounding in.

Copied the special FD bulletin on 80 CW with no trouble, and made the first 100 points for the club. Got a lot of questions, too.

73 de Jim, N2EY
Logged
N2EY
Member

Posts: 3875




Ignore
« Reply #12 on: February 15, 2012, 03:30:00 PM »

while the concept is neat it doesn't work quite as well as Stoner says it does.

I can understand why!

W2PPL had a better idea. He had two receivers in QST in the late 1950s and early 1960s.

In the first one, he moved the tuning range of the BC-453 into the BC band. Not a good idea! But his ideas about adding a product detector and other things are worth a look.

In the second version he left the BC-453 tuning range alone. He added two conversions in front of the BC-453, with a tunable IF around 3 MHz. The first conversion is crystal-controlled and brings the bands down to about 3 MHz, then the second conversion does the rest. The tunable IF has a gang-tuned capacitor that is string driven off the BC-453 dial nut. The article is well worth a look.

IMLE the BC-454 (80 meter ARC-5) is plenty stable; the problems are that it's not very selective and the tuning rate is way too fast (100 KHz per turn!).

The tuning rate can be greatly improved by replacing the '454 tuning cap with one from a '455. Of course this means a complete realignment, some added padding, and reduction of tuning range. But all you really need is about 600 kHz anyway. But it doesn't solve the selectivity problem.


73 de Jim, N2EY
Logged
KB1WSY
Member

Posts: 714




Ignore
« Reply #13 on: February 16, 2012, 09:00:31 AM »

What a lot of great advice!

First of all, I will be renovating my BC-453. I deliberately bought a beat-up one that had already been converted by a ham many years ago. I'll probably be replacing all the old electrolytics and making various other changes. I don't feel this is "vandalism" because of the already poor condition of the original.

Then I'll build the simple Stoner circuit for 80/40 only. I have tracked down, not the original Miller coils (which are indeed very close to unobtainium) but close substitutes from other makers. I am not expecting fantastic results but I suspect it might be quite a lot better than another project I am starting to work on, which is a 3-tube regen superhet (from QST 1963).

I was unaware of the Stoner SSB book until reading the above posts. I have tracked down and ordered a copy. Will also check out the various suggestions made about other circuits from QST etc. -- I've already read quite a lot of them but clearly not all! I had already checked out the W2PPL (the second one, with the capacitor ganged with the BC-453 tuning dial) and decided that was way too complicated for me at this point.

Once I have these simple projects up and running I will figure out what to do next. Clearly, two of the most fruitful routes involve involve beefing up the initial RF stage and adding a second level of IF conversion. Another idea, to control image response, would be to add an external preselector; there are quite a few circuits for those in QST and the ARRL handbooks. Of course this adds yet another button that you have to twiddle and tune, but isn't that the point of these glowbug projects anyway?

I have tried twice to joint the glowbug reflector, at a two-week interval. In both cases my email was rejected: "host unknown." (That's the address at uidaho.edu.) If someone has the updated address I would be grateful....
Logged
G3RZP
Member

Posts: 4442




Ignore
« Reply #14 on: February 18, 2012, 03:39:38 AM »

I too would appreciate the glowbugs reflector URL
Logged
Pages: [1] 2 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!