transformer question

[mac]

dumb question of the month....

If we can wire say a 10k:600ohm output transformer backwards as an input (600:10).... then why do manufacuters build identical spec "input" and "output" transformers? (like for example the edcor XSM10k:600 and XSM600:10k).


Mac

 

[gyraf]

Per Lundahl told me that this is because of different optimum winding-geometry

 

[mac]

Hey Jakob,
Yeah that would make sense - I guess the copper has to start and stop somewhere around the core for each tap.

So in theory, when "waiting" for transformers to come along for projects, I probably should try to get the "right" ones for each spot in a circuit and not just assume that using one backwards will sound and react the same way in a circuit as using it as it the way it was designed.

Mac

 

[PRR]

There is an AMPLIFIER in the middle.

Logically, the input has small levels, the output has large levels.

Inputs may want to be shielded, but they don't feed more than a few inches of line. Inputs taking very weak levels should have very low losses so as not to degrade system noise figure.

Outputs may have to pass significant power to charge long lines or multiple loads. Outputs may be allowed considerable loss... you just use a slightly larger-power amplifier.

You can sometimes run NFB around an OT; this is rarely practical on an IT.

> output transformer backwards as an input

You may or may not regret the little/no shielding or (slight) shifts in frequency response. Although transformer designers agonize over this stuff, in many practical cases the difference is insignificant.

> edcor XSM10k:600 and XSM600:10k

On this series, the input starts on Pin 1. Your low-paid assembly workers do NOT know 600 from 10K, but they can be told pin-numbers. Your semi-skilled technicians may know 600 from 10K, but wiring the supermarket attic on a (northern hemisphere) hot day like today, it really is better to use the specified part and then just squint pin 1 and be done rather than be sweating all over an impedance pinout table.

On this series I think Edcor does just wind the bobbin the other way to make the other part. This is not over-agonized iron. It works great in the 99% of apps that don't depend on some fancy tricky iron.

And if your tech staff can think straight, then the one 600:10K part can be used 10K:600 by getting the wires on the right pins. You would only stock both parts when it streamlined your assembly or installation process more than the cost of another bin and inventory time/capital.

 

[emrr]

I recall the Edcor resistances being different for each type. 

The hilariously popular and mysteriously desirable RCA OP-7 Portable Mixer uses the 5 of the same 250:50K transformer; 4 for mic inputs, one for a summed -50 mic level output. 

 

[PRR]

> resistances being different for each type.

Probably just wound primary-over-secondary (or vice versa). For the same turns-count (1:1) that will give 10%-20% difference of DCR due to the longer length of turn. Even at 10K impedance, with good cores and good design that gives excellent "utility" performance. And is a heck of a lot cheaper than elaborate splitting and interleaving such as telco hybrids or huge-NFB amps require.

> uses the 5 of the same

I think the Shure boxes do too. Level-Loc and that old small mixer.

 

[emrr]

uses the 5 of the same

I think the Shure boxes do too. Level-Loc and that old small mixer.

Forgot those; you're right.