[Collins] Transformers

[Dr. Gerald N. Johnson]

On 8/27/2013 10:26 AM, antqradio at sbcglobal.net wrote:
> "The harm is in the poorer regulation avoided with the bucking
> transformer and many an early Collins rig is rated at 115 volts, not
> 117."
>
> As I understand it, the only way to get any degree of voltage
> regulation in a power transformer is to operate it near saturation.
> Otherwise, the transformer's secondary voltage is determined solely
> by turns ratio.  Near saturation, with the flux density in the core
> at maximum, any increase in input power just heats copper and core
> material because there can be no increase in flux density.  If the
> transformer was designed to operate at these higher temperatures then
> no harm is expected.  But if this is not so, then the transformer
> will just cook itself to destruction, most commonly resulting in a
> shorted turn.

Not quite. When you saturate a core solidly, the instantaneous impedance 
goes very low, the slope of the B/H curve and so the current spikes. You 
have to include some series impedance to keep from drawing hellacious 
peak currents. And that's not how a Sola works. The sola includes a 
resonant winding and a loosely coupled winding. When the applied voltage 
rises that resonant winding's core moves towards saturation reducing its 
inductance and moving its resonant frequency away from 60Hz (or whatever 
design frequency) and that compensates by lowering the coupled voltage 
to the load winding so keeping it constant. Its not possible to change 
the working frequency of a Sola by changing the capacitor, I've seen 
engineers try it and fail. You have to change the core stack, the 
winding and the capacitor.

Otherwise you improve transformer regulation by reducing its internal 
impedance. Less leakage inductance and lower winding resistance, while 
avoiding saturation in the normal operating range. One way to accomplish 
both is with higher permeability core material, one I used more than 50 
years ago was Selectron E IIRC, it was a tape wound grain oriented 
silicon steel. Then the tape winding was shaped into a pair of U's and 
cut apart. Because it had higher permeability and a higher allowable 
peak flux it could operate at more volts per turn than standard E-I 
laminations and so fewer turns for lower R and leakage inductance. A 
typical pole transformer with that for core has 1 to 2% impedance 
instead of 4 to 5% for conventional transformers. But when you drive it 
into saturation the input currents rise more than hellaciously, more 
like astronomically because the B/H curve turns FLAT.

Speaking of transformer impedance. That is conventionally measured by 
shorting the secondary (ies) and then applying low voltage to the 
primary raising that low voltage until the rated primary (and secondary) 
current is reached. The percentage of rated voltage applied is the % 
impedance. What it means is that with a super powerful primary supply 
that doesn't vary with load that putting full load on the transfomer 
will drop its secondary voltage(s) by that percentage. E-I transformers 
typically have 4 to 5%, tape wound grain oriented silicon steel are 
typically under 2%. Losses are smaller too with the better core so 
utilities like that design but don't like the initial cost. I have wound 
transformers with it.

So that means the inherent line impedance in your house is a few percent 
of the total power capability of the distribution transformer supplying 
the place. For a 25KVA transformer that can mean the normal load 
resistance for full load at 240 volts is 104 amps, and that is an 
impedance at 240 of 2.3 ohms so a 4% impedance transformer has a series 
impedance of .09 ohm. If you try to run a transformer load into 
saturation that's the series impedance of the power system. The primary 
wiring contribute very little to the impedance seen at the house 
voltage. So little its not worth computing, I have on occasion but it 
was a waste of time. The last transformer impedance is very predominant. 
Its impedance is higher than the secondary wiring in most cases.

In my PHd research I was working on communicating from load to 
substation using LF energy. I worked a couple years on trying to drive a 
LF current against that impedance and to find it at the substation. 
After burning up many a transistor and SCR, I gave up on the idea and 
drew a sequence of half cycle currents instead that I was able to detect 
at the substation. I had both substation, 7200 volt distribution, and 
distribution transformers in my basement lab. I supplied the substation 
transformer from 120 volts through a variac and current transformer 
hooked to my scope so I was able to observe core saturation very clearly 
while looking for my signal currents accompanied by ordinary load 
currents. I used a current transformer on the substation 7200 volt 
secondary with filtering to notch out 60 Hz and harmonics and to detect 
my 15Hz signaling by my sequencing of half cycle currents. Four half 
cycles per bit, one phase or the other with respect to the line voltage. 
Coding would have to have allowed for detecting the initial phase which 
has been solved for phase shift keying communications circuits so i 
didn't have to invent that too.
>
> Any of the two methods already mentioned to back off from saturation
> by reducing the input voltage, will obviously reduce the inherent
> voltage regulation that comes from operation near saturation.  I have
> no problem with the bucking transformer method.  It is just that
> finding room for it in the equipment is problematic.

I think it should not be in the equipment but should be one bucking 
transformer for all the vintage equipment in the shack. The bucking 
transformer can be fairly small because its only handling the change in 
voltage, not the entire load VA.

I know one area ham and electrician (now retired) who installed a 
bucking transformer for his whole house and I think his shop and office 
years ago. He's moved and I don't know if he moved the bucking 
transformer or not.
>
> Cost is also an issue, new filament transformers are not cheap.  And
> finding one with a non standard secondary voltage is next to
> impossible, so one is stuck with a compromise unless you can rewind
> the bucking transformer to suit your needs.  That said, power
> resistors are obviously less expensive then a bucking transformer and
> can be easily sized according to need and they can be made to fit in
> smaller spaces by series connecting several to arrive at the optimum
> value.

Standard bucking transformers stocked by electrical distributors all 
over come in two types. One has two 12 volt secondaries and two 120 volt 
primaries. The other has two 16 volt secondaries and two 120 volt 
primaries. So with 120 volt primary you can get buck/boost voltages of 
6, 12, or 24 with the first type (half power when the primary is 
connected for 240 but running on 120) or 8, 16, and 32 with the second 
type. 32 just happens to be very handy in industry when needing to run 
equipment made for 208 volts on 240 or visa versa.

Then there are universal rectifier transformers with multiple primary 
and secondary taps, getting less common but they still exist.

And the bucking transformer wired with the low voltage in series with 
the "primary" allowing applying 125 volts to the sum rated at 132 or 
more moves it further from saturation where its energy efficiency is 
enhanced.
>
>
> Saving power transformers from self destruction is the issue, not the
> best method for doing so.  Power transformers will benefit from
> operating at a lower temperature, regardless of the method used to
> get there. Jim

Yes, but in the long run the bucking transformer will run up fewer KWH 
on the electric bill and produce less heat in the hamshack while giving 
the radio circuits better voltage regulation.

73, Jerry, K0CQ, Technical Adviser to the Collins Radio Association.
>
>