UTC A-20 Blueprint

this is a nice 500:500 transformer,

alternate p/n is GR-815 if shopping around,

pretty flat response, typical of a low ratio transformer with a low turns count,

Inductance vs Frequency

UTC says this thing is good to 15 dbm,

but bench testing proves this to be a puffed up rating,

we had to drop the input level down to about 2 volts to test at low freqs,

otherwise saturation was dominating the graphs at the low end,

even at 2 volts we crap out at 5 Hz,

UTC use to stamp their own lams,

Permeability numbers indicate an 80 Ni alloy or similar,

Hypermalloy, is that their nickname for this lam? it is mentioned in the catalog,

this looks like the same 31 UI alloy lam as used in the A-26,

looks like lower turns (1,060 Series Pri) = higher perm numbers, maybe a DCR thing,

here is the catalog page for this transformer,

testing is nice but for real fun we need to tale a peak under the hood,  ;D

different heat gun, but the song remains the same,

always wear proper safety equipment,

like a good pair of shoes to keep hot tar off your feet,

what you do is take a drill bit and use it to take metal off the mounting holes so that the lead plate can come off without breaking,

took off the standard blue pressboard wrapper that they used to keep the xfmr from shorting to the can, snipped the gnd lead that went from the coil to the can, and flagged all the leads for dissection, now we can ditch the terminal plate,

notice color of lams,

UTC uses wax to impregnate the coil, then black tar to pot the coil/core assy in the can, Mike at Magnequest (Peerless) likes the sound of layered paper and wax,
he has done experiments with different processing techniques and insl.

Sweet!

yes, rather do this than mow the lawn,

they had a layer of cloth tape around the 2 coils which is a real pain to remove as the adhesive is very sticky and if you do not heat it up good, you rip all the wires out of the coil and data is lost for good,

i do not like this tape for another reason, it contains a lot of moisture, lams do not like this moisture, so if wax does not cover everything, you can get rusty lams and other good stuff,

37 ea.  UI lams, standard UTC stack, laced 1 x 1, no loose I bars which is a sign of good processing,

a gnd wire is soldered to one of the I bars, you can see the solder, this connects the core to the can and to the coil shields,

showing a pri splice, a screen splice, and various pieces of insl. which set up capacitance and leakage inductance and thus frequency response,

ran out of wire, must be done,

this thing uses screens made from 101  turns of #37 instead of copper foil,

these screens are reverse wound, do not know why,

color coded insl,

wire is unwound onto spools so that the distance of each wire can be measured and compared with DCR to get ohms/1000 ft spec which can be used to compare old and new wire gauge specs just for fun,

now we get to translate the notes onto the blueprint, what fun,  :-*

keeping the insl which can be measured for thickness
in case we want to replicate this xfmr,

Thanks for the post!
Can you tell if there is extra paper layers before and after the screen wire is wound?

CJ said:

ran out of wire, must be done,

this thing uses screens made from 101  turns of #37 instead of copper foil,

these screens are reverse wound, do not know why,

color coded insl,

Click to expand...

screen insl - 2 layers of about 2 mil poly before and after the screens...

Good work!

CJ said:

UTC says this thing is good to 15 dbm,

but bench testing proves this to be a puffed up rating,

Click to expand...

The bit I see missing here is the A series rate for -2 response, so the rating includes more of the lossy area.  The HA and LS rate at -1 response. 

To revisit, if it makes +15 at 25 Hz, it then makes +21 at 50 Hz, +27 at 100 Hz, etc, etc, no problem.  Practically speaking, we have not so much a frequency response problem with low frequency as a distortion content problem, when such low frequency is present. 

that is right, if i use a higher level when sweeping, the low end dip will appear,

and distortion will be higher as current goes up quickly when the core saturates,

and the older models in the black cans roll off even quicker at low freqs,


here is the print, i had room for a layer by layer drawing due to low turns,

might not be printer friendly, 

Master CJ!