[R-390] R390 Posts: Excess line input voltage Noise Figures

[Bob Camp]

Hi

Rather than specifying the noise figure of the radio Collins (or more 
correctly the military) specified the sensitivity of the radio with 
specific IF bandwidths. For what ever reason this was the more common 
approach in that era. The "correct" approach to making the measurement 
uses a resistive pad between the generator output and the radio input. 
You measure a signal to noise ratio on the output of the radio. The net 
result can be loosely translated into noise figure since you know the 
impedance of the pad and the bandwidth of the filter.

The whole line voltage thing has been hashed over a number of times. 
The general conclusion is that the radio works a bit better with a 
higher line voltage. One of the several reasons we always see the 
radios beat the military sensitivity specifications is the higher line 
voltage. As long as you keep the ovens turned off the total amount of 
heat in the radio is well below what the military considered acceptable 
when the radio was designed.

Cooler is always better when it comes to reliability so more heat is 
never good by it's self. There does not seem to be any significant 
evidence of major heat issues on the R390A. There are some issues on 
the R390 not an A in the vicinity of the high voltage regulators. 
That's a thread unto it's self and the short answer seems to be to use 
a fan on the 390 not an A.

The "normal" power line load from a R390A is quite a bit less than the 
rated 240 watts if you have the ovens turned off. Roughly half of the 
power into the radio goes into the ovens if they are turned on. There 
are some "interesting" ratings listed on the transformers that make 
this a bit hard to figure out. Like a lot of things on the 390 the 
origin of the transformer markings has had several threads devoted to 
it over the years.

A 10% change in line voltage should increase the power into the radio 
somewhere in the 10 to 20% range. Since most of the power is going into 
constant current loads or constant power loads (tubes) the number is a 
bit closer to 10% than it is to 20%. Net heat rise may be another 10 
degrees or so. That's not insignificant, but you can get more change 
with mounting the radio tight in a case versus well spaced out in a 
rack.

The US government bought a *ton* of tubes for the 390 back in the 
1980's. They then proceeded to take all of the radios out of service. 
The net result is that there are probably enough tubes out there to 
keep the R390's running for another hundred years or more. The only odd 
exception to this is the ballast tube. I have yet to see a reasonable 
explanation of why they didn't also buy a ton of these as well.

Most of the heat change in the radio when you change the line voltage 
shows up in the tubes themselves. Plate current does not go up a lot at 
the higher voltage so power in the dropping resistors does not go up 
very much. Obviously this does not apply to the dropping resistor in 
series with the voltage regulator tube or to the ballast tube.  However 
most of the impact will be on the tubes. Since we are awash in tubes 
this probably is not a terribly bad thing if it makes the radio work 
better.

Higher voltage also puts more stress on the capacitors in the radio. 
The capacitors that we hit the hardest are the good old paper insulated 
parts and the electrolytics.  I have yet to see anybody recommend 
keeping these in the radio. The obvious conclusion is that you should 
replace the caps with ones that have a higher voltage rating than the 
originals. The mica and ceramic parts all seem to be significantly over 
rated in terms of voltage so they do not appear to be an issue.

Running MOV's in front of the radio is not a bad idea at all. On a 
radio with solid state diodes in it there is a possibility of damage. 
This is especially true if you are running the original fuses. (With 
the ovens off you can run smaller fuses). The issue is that as you have 
observed you need a fairly low voltage rating on the MOV's to have them 
do any good. A 250 volt rated part on a 120 volt line does a nice job 
of protecting the diodes. It also will take a *lot* of hits on a 
typical power line. There is a significant body of evidence that 
running parts this way eventually causes them to fail. When they fail 
they pop open. You need to be sure they are mounted in a fashion that 
when they have enough room around them to explode. If they don't then 
you wind up with a fire ...

Depending on the brand of constant voltage transformer you have running 
hot may not be all that unusual. One of the basic regulation techniques 
is to saturate the core of the transformer. You then depend on the 
stability of the saturation loop in the core to give you the 
regulation. Regardless of the load this kind of transformer is always 
running the core in a "hot" state. That's not to say that there aren't 
problems that will make it hotter than it should be, only that they 
tend to be hotter and louder than a normal transformer.

One thing that we do not seem to have an infinite supply of is the 
micro switch based power switch on the radio. It obviously works a bit 
harder when you have a higher line voltage. The weird thing is that the 
switch seems to fail when you turn the radio off rather than when you 
turn the radio on. Some line regulators have a nasty habit of being a 
bit slow to react. They tend to boost the line voltage as you drop the 
load. This may not be a good thing as far as the micro switch is 
concerned. The fix is obvious, but not terribly easy: Drop the front 
panel, pull the power switch and make sure the microswitch is nice and 
tight on the assembly.

Nothing is ever simple ....

	Take Care!

		Bob Camp
		KB8TQ






On Oct 10, 2004, at 5:50 AM, face at netunltd.com.au wrote:

> First, thanks to Bill,Barry and Cecil for their replies and advice..
> Its nice to feel wanted.  Aint the net great!!
>
> Aplogies too, for inadvertently replying to Barry alone  and not via 
> the
> administrator.  In my inexperience
> I simply repied to Barries email, thinking it would also go to mailman.
> Wrong again, John!
>
> So I shall reply to you all by re typing and re submitting  this
> condensed version and sending it correctly ..... serves me right !
>
>
> EXCESS LINE VOLTAGE TO R390A
> Barry replied by suggesting I use a variac in front and twiddle the 
> knob
> to maintain line constancy.
> (tongue in cheek, maybe?).
> I replied that I had solved this problem by (first) using a servo
> feedback motorised variac with a front panel screwdriver adj.
> When this broke down on me I used a constant voltage, ferroresonant
> transformer, (CVT) with sine wave output option
> (an Advance Model CVN230A)
> This gave a very well regulated, transient free  230v AC constant 
> output
> to run my R390A and allowed me to remove the line bucking transformer
> previously installed..
>
> Barrys  reply was that he uses a similar CVT, but that it 'ran hot'  
> and
> was noisy, which I take to be lamination buzz.
>
> REPLY:  Barry, if your CVT is not actually faulty, thats usually caused
> by not running the CVT at, or close to, its rated output current.
> These things are designed to run under a load close to the rated 
> maximum
> output and literally shake themselves to bits when they arn't. !  Not
> only do the darn things then run hot, they add a whacking big amount on
> your electricity bill !!!!
>
> Mine has a max rated output of 230W with a power factor of 1.0.
> The R390A is rated at 225W which matches it pretty well.  My CVT then
> runs quiet and cool . You wouldnt know it was there.
> One of the big advantages of running a CVT in front of the R390 is its
> ability to suck out damaging line transients.
> I made up a box adding silicon carbide voltage supressors to go across
> inputs and outputs  when I first used this thing, but removed them as
> they wern't needed, even in the electrically noisy environment I was in
> before moving  house (was on a main industrial line feed 50 ft behind a
> pole mounted line boost transformer).
> Also, they were only 275 v rated and was scared they would smoke out on
> the frequent high line peaks I had then (275v RMS measured when quiet
> !!!).
>
> Barry also replied that servo controlled variacs often fail because of
> brush wear tracking carbon dust over a small segment of the winding due
> to motor hunt when stabilizing.     A valid and sensible point , Barry!
> I will check mine out next time I find it. (Its buried neath a near ton
> of gear forced into a garden shed.  Moving from a five bedroom house
> with three large rear sheds to a 2 bedroom plus garden shed forced the
> loss of a pile of good gear..but i'm sure thats still around.
> Methinks a servo'd Variac in front of my CVT would make a 'cool'
> combination for my R390A.
>
> Not everyone out there's got that sort of gear to hand, though.
> Nor does it solve the over high HT when changing to Si diodes.
> =============================================
> Bill suggested that high HT does not impact on R390 performance much.
>
> REPLY: That may well be so. Probably because of the R390 remote cut of
> valves used in the RF/IF  stages, necessary for AGC.
>              I remember, as a student, being given the task of 
> designing
> a VTVM, though transistors were plentiful enough then, but   expensive.
>   If my memory is reliable on this, I found that the sharp cut off twin
> triode (12AX7) initially used, DC drifted a lot, but a wider grid base
> 12AU7 fixed the probem. (same pinout)  Seems that sharp cut off types
> had a 'u' more dependant on small changes of anode current than the
> others.   Could be that pentodes have the same characteristics and the
> remote cut off types used in the RF/IF stages of the R390 don't worry
> too much about stability of HT supplies.  Bill could be quite right
> there.
>
> BUT: ....they will run hotter !!! (as will all the dropping resistors 
> in
> the HT chain)
> We should be kind to our bottles and keep dissapation down.  Hence I
> think keeping HT at the recommended level by Collins is
> probably a neat thing to do.  Less stress on the filter caps too.
>
> Bill suggested that my suggested mod of placing a resistor in the CT
> return of the HT transformer would increase stress  on the (winding)
> insulation.
> REPLY: The voltage across this resistor to chassis is  going to be
> around 20 to 50 volt, at a guess.  (depends on line voltage and R390
> settings and condition).  The transformer insulation is factory tested
> to some 2,000v, usually for one minute.  An increase of 50v is not 
> going
> to cause a surgeon generals health warning here.
> But,  the measured voltage between the end of the resistor (chassis) 
> and
> the outer end of each secondary winding is still the same.
> There's no algabraically added voltage between transformer winding and
> core.
> All we have done is add a few ohms exrtra to the secondary winding
> resistance.
> The ac voltage drop across it isnt added to the AC secondary voltage, 
> it
> becomes part of it.  There is still the same voltage as before across
> the secondary windings and between them and core.
> The stored charge on the capacitive input filter used on the R390 takes
> care of worsened transformer regulation due to increased resistive
> losses in the xformer winding.
>
> Its the relatively low voltage across this proposed dropping resistor
> which seems so attractive to me for regulating the HT voltage to a 
> fixed
> setting, independant of AC line and R390 load current (within reason,
> anyway).
> If the needed voltage drop acroos this 'R' is more than , say . an
> LM317/337 max rating, we just add a couple of rated  Zeners in series 
> to
> drop it to a safer value to use
>
> Further, with good, working vacuum bottles getting scarcer to get
> (certainly the case in Aussie)   it does seem that preserving tube
> working life would be enhanced by ensuring operation inside specs.
>
> So far as my R390 is concerened, all this is armchair theorising on my
> part.  I havn't had to do a thing to it since buying it.
> So until I get my hands dirty and actually do the above mods, I will be
> taking all your advice thus far  to heart and thank you all for
> contributing.
>
> I end with a question.
>
> Does any of you know if Collins specified a noise figure test on the
> R390 series ?  Ie recommended test circuit.
>
> I was going to use a carbon resistor across the co ax input (known 
> noise
> voltage). to determine overall receiver noise floor.
> (when I get my hands on a replacement set of good bottles)
>
> But I have seen 'dummy antenna' used to evaluate more realistic noise
> figures when connected to a 'real' antenna.
> Example:  LCR network used to evelaute noise performance of LF
> receivers... such as seen in Termans and Langsfords Smiths books)
> We used such LCR networks across antenna  terminals years ago for
> lw/shop evaluating probable S/N floor under actual operations in 
> remote,
> desert  areas when using a resonant long wire.(2 harmonically related
> 'ops' frequencies)
>
> I dont own an active, calibrated noise source and wondered whether the
> LCR dummy load would give a more realistic figure than just the 
> resistor
> (The LCR circuit can be tuned to the resonace of the 'real' antenna, or
> to its receiver entry  impedance at a working frequency and thus give a
> more realistic result than the broad banded carbon resistor approach.)
>
>
>
> Regards to all , particularly the guy that got all this running on the
> net!!!
>
>
> John Byers
>
>
>
>
>
>
>
>
> ______________________________________________________________
> R-390 mailing list
> Home: http://mailman.qth.net/mailman/listinfo/r-390
> Help: http://mailman.qth.net/faq.htm
> Post: mailto:R-390 at mailman.qth.net
> Unsubscribe: http://mailman.qth.net/mailman/options/r-390
>