Tuning the Ultradyne.

 A diary of getting it working properly and how it performs in the Real World.

14 Dec 2005 -

From a 2005 perspective one can see some of the problems with a design such as this one as it applies to modern conditions.  The most obvious is the low frequency IF, in my case 60kc.  Most modern sets, in fact most broadcast sets after the 1930s use a 455 kc IF.  With 60 kc there are images will fall 120 kc away from the desired frequency, versus 910 kc away in the 455 scheme.  It takes extremely sharp front-end selectivity to reject sufficiently signals that are so close.  Moreover, with the 455 scheme there are essentially no images falling inside the BCB except from stations transmitting on 1450 kc or higher.  In the mid-late 1920s there were only a relatively few stations operating, mostly with fairly low power and often with schedules of only a few hours a day.

As mentioned on the main page, a tuned loop antenna can provide much better selectivity than the internal coil and in the original articles an external antenna tuner is discussed for applications using an outdoor wire antenna.  The loop has the additional advantage of rejecting undesired signals by its directivity.

The Wirco antenna coil which I am using leaves a lot to be desired in its Q.  A better coil could be fabricated with larger litz wire but this really wouldn't eliminate image problems.  An external loop is the best option.  Although my version doesn't have the jack for an external loop one can still be connected via a coupling link configuration.

The other nemesis that haunts me on most 1920s radios at my location is a nearby broadcasting station, 5 kilowatts on 1370 kc.  The presence of this strong signal created dozens of undesirable "birdies" and mixing products.  I used a small trap made with a ferrite toroid and trimmer cap hidden under the antenna block to reject 1370.  The radio quietened down considerably after doing so.  This wasn't such a prevalent problem in the 20s with the limited number of stations but even still there existed all sorts of magic trap-like gimmicks on the market to achieve this same result.  Nowadays in an urban area with dozens of strong local stations this could indeed be an insufferable situation.  Again, an external loop antenna would go a long way in minimizing the effects.

Another bugaboo is the harmonics from the local oscillator.  Imagine that if you are tuning a station at 600 kc.  The oscillator is supposed to be running at 660 kc.  When doing so it is also sending out a second harmonic at 1320 kc.  Bad enough that the receiver when set for 600 kc will also tend to receive 720 kc (other-side injection) but it will also tend to receive 1380 and 1260 simultaneously!  If any of these stations are particularly strong then additional mixing products will be created.  More indication for the need of front end selectivity.

Ok, time to trim the sails....

First, and easiest, is to get the rf input tuning range established.  This will help later in sorting out the images and oscillator harmonics.  With my tuning capacitor of  ~475 pf I had to remove about 3 turns from the Wirco coil to get a tuning range of about 550 to 1550.  Its not particularly sharp at the top end and I can still receive a station on 1620 with the input tuner set at minimum.  I made a little chart using the signal generator that can aid me in tuning the desired signals.

25 Dec 2005-

In spite of my best attempts I hadn't been able to get regeneration.  The original Ultradyne calls for a tuned input on the first IF xfmr.  Any attempt to tune my 60kc Victoreen xfmr would swamp and kill the oscillator.  I suspected the regeneration problem to be interrelated.

Nothing like a bit of trial and error.  I wasn't comfy with the size and number of turns (24) on the Wirco tickler so I wound a new one with 40 turns and a slightly larger form.  That seemed to do the trick...at least partly.  I had been subbing in some small value caps across the primary just to see what would happen.  It just so happens that I had left a 150pf tacked in when I tried the new coil.  Removed the cap and no regeneration.  So I played around a bit and it still didn't seem to want to solidly regenerate.  I tried different values down to 47 pf with varying degrees of success.  100 or less seems to be the magic number.  I left 100 pf in place for the moment and I get a nice regen edge when rotating the tickler.  The antenna tuning cap interacts with it somewhat but thats to be expected.

This tickler is slightly larger - 1-5/8" instead of  1-1/2" and has 40 t instead of 24.

So, I think I have licked that problem.  The downside is that the extra capacitor is seen by the oscillator tuning so my previous trimming of the osc coil has now gone to pot.  I had also changed the osc tuning cap to a one with better SLF characteristics.  It tunes well but is a tad floppy mechanically so I may change it again.  Remains to be seen how much this added cap is going to bugger up the top end of my tuning range.  Those second harmonics from the osc may come in handy after all!

How much improvement is there with the regeneration working?  WOW!  I was tinkering while using a lashup with diode and phones connected across the output of the 2nd IF xfmr....that means osc, mixer and 1 rf amp stage and no attempt to match the phones.  Without the regeneration I could just barely hear a semi-local station on 840 kc.  When the regeneration kicks in its at comfortable headphone listening level.  That old Wirco antenna coil got decidedly sharper too!  By the way, I measured the Q on it and its not as shabby as I thought.  Its better than most small basketweave coils.  So I think the regen will go a long way in sorting out the images and make a marked improvement selectivity-wise.  I might also add that there's not a bunch of howling and squawking with this regen scheme.  It acts more like a smooth rf gain control until you hit the edge where it drops out.

*     *     *     *     *

21 January 2006

The set has now been rewired with brass rod, 120kc Ultraformers have been fitted and ready for testing.  Initial results were quite pleasing.  The higher IF certainly cuts down on the images.

Two 'bad' things were noticed, however.
1.  There seems to be much more pulling of the local oscillator when the regeneration kicks in.  Regen also seems kind of 'rubbery' which aggravates the problem.  The rubbery aspect may be because I had previously increased turns on the tickler but the pulling aspect when tuning the tickler or antenna condensor is on the verge of being objectionable.

2.  Bandwidth has increased proportionately.  Thats great for nicer audio on a sparsely populated band but I'm not sold on it.  My first thought was that the Ultraformers are overcoupled and considered going back to the Victoreens and a 60kc IF.  That would be a major chore since the terminals on the two types of IF xfmrs do not match up for an easy replacement.

So, I tried lowering the IF freq on the existing Ultraformers.  I went down to about 77 kcs and indeed the selectivity sharpened up nicely.  The quantity of images and mixing products came back too.  (I'm doing this in the daytime when my nearby station is running 5kw and the trap is not removing it as well as at night).  This confirmed to be that the relationship between IF freq and bandwidth was predictable.

I then tried the IF at 100kc.  This seems to be a good compromise.  For comparison, at 120kc a 60 mile distant station at 810kc was about 5 divisions wide on the 0-180 tuning dial.  A much stronger station at 840 could be heard in the background although not objectionably strong.  At 77kc the same 810kc station on the dial was barely 2 ticks wide, nothing heard from 840.  The decrease in audio quality was noticeable as well although that doesn't stand out as an issue.  At 100kc the 810 signal was about 3 ticks wide and sounded pretty good.

I'm going to run with 100kc for a while and get a better feel for it around the band.  I do think these older IF transformers are too overcoupled at higher frequencies for todays use so there's another gotcha in playing with old superhets.  It would be fairly simple to fashion an xfmr to modern standards for better performance but that seems to be going too far when wanting to experience an old 1924 circuit and its inherent shortcomings.

At this point I think its time to clean up the old loop antenna and rig it with fresh litz wire.  The shop cats have now matured and don't seem to have much interest in it anymore!

To be continued....

Back to main Ultradyne page

21 Jan 2006