Resistor, capacitor across output transformer AKAI M-7

[fyzygy]

There is a resistor (100 ohms) across the secondary, and a capacitor (0.001 uF) across the primary, of the output transformers in my AKAI M-7 mono blocks.

What do these components do, and should I keep them in place if rewiring for another circuit?

Specifically, I'd like to make a pair of RH84 SE mono blocks out of these.

[Rod Coleman]

The 100R is probably a safety-load, to prevent the output load of the transformer from ever being open.

If the speaker is disconnected or open, and you put some music into the amp, the primary voltage would increase until the insulation breaks down (RIP OT).

Same way - if you use a speaker with a huge peak in the impedance somewhere.

The cap may be an attempt to prevent transients from blowing the insulation, or it may be trying to roll off ultrasonics before they get into the feedback loop. Either way, it is under high stress, and may be better for a new one (eg 1500V rated LCR PC/HV/S).

You can use antiparallel EHT diodes (cathode to EL84 anode; anode to GND) to snuff transient overvoltage threats to the OT. (EDIT: sorry - that's for PP. ignore for SE)

[fyzygy]

Interesting. The original circuit has largish 8R shunts to ground on the switched output jacks, in case of a disconnected speaker. So, the 100R would be an extra safety precaution?

I like the idea of a cap to roll off ultrasonics ... is that a common practice?

[Rod Coleman]

I might be wrong, if there are some dummy loads as well.

It doesn't seem likely that any frequency compensation is the aim, since the speaker would swamp it.

Some speakers do have large impedance peaks at crossover frequencies, or at the ends of the audio range, so it might be trying to compensate that effect.

In vintage equipment, I've seen RC networks across the primary very frequently.

[costis_n]

Originally Posted by fyzygy

Interesting. The original circuit has largish 8R shunts to ground on the switched output jacks, in case of a disconnected speaker. So, the 100R would be an extra safety precaution?

I like the idea of a cap to roll off ultrasonics ... is that a common practice?

the AX84 P1 project does it, to name one (5n/1500V across the primary)

[richwalters]

Originally Posted by Rod Coleman

In vintage equipment, I've seen RC networks across the primary very frequently.

Yup.....I'm using the method to a mild effect in my 200W amp....to control square wave ripple created by the o/p transformer leakage inductance with rising frequency as this has considerable energy. Other designs instead have an R/C on the secondary which I don't like as one often forgets to change this with the taps. The bigger the transformer the more of a problem it is.

A pic with the 200W 10Khz square wave at the 50W level.
One cannot get a better squarewave than this, but as power increases the linearity between the squarewave harmonics and the output transformer parasitics falls apart which shows up as less square with humps and ridges. It is a wicked test for HP amps with large o/p transformers which are inherently HF deficient.
One must avoid too high R/C values across the whole primary which absorbs considerable dissipation for the output stage. A wirewound resistor is the only type of resistor for this location and the energy which I have calculated for the output stage of the 200W amp; is around 2.5W rms which is quite low, but the high dv/dt spike energy will immediately strip metal oxide & film types which cannot survive.
Worse is to come; This "across the primary" snubber interaction adds a pole in the Bode plot, in conjunction with the step R/C (if fitted) in the global nfb loop and with the 1st stage zobel makes the approach rather awkward as there aren't two output transformers that perform repetitively as the HF cutoff characteristics are determined by the winder and if any varnish has seeped in.
In my method, which differs from the traditional loop gain approach is the Bode plot and phase shift analogy combined, that can determine the maximum trim capacitance for the o/p step global nfb, at the same time seeking the minimum capacitance for the 1st stage Zobel. This is the proper approach for min thd per rising frequency.
This is the biggest stumbling block in knowing what to fit in 3-4 stage amps with global nfb with perhaps poorly designed o/p transformers with single sectioned tapped secondaries ....Nothing is easy, but one is on the right stability track with any amp when the global neg feedback resistor can be reduced to a total of around 35dB before instability sets in depending on transformer quality. For doing squarewave measurements, be very aware where the scope earth is placed, if other equipment also earthed to the same chassis, this can interact with the fast risetimes giving misleading waveform spikes and humps.

richy

[fyzygy]

I left the resistor and cap in place when building my RH84 into the Akai mono block enclosures. Everything works well, sounds fine, see no reason to remove them.