AnalogSteph
Major Contributor
Isn't that how it often goes? I am reminded of a certain line of 2000s era TEAC CD + SACD players that decidedly were IEC Class I with their PE prominently screwed to chassis. And no, they had traditional mains transformers, if maybe of the toroidal variety. Even magazine review measurements showed the effects of a ground loop...
On the schematics for either L-05M or L-07M, I see no indications of a shield winding being present in the transformer. Power amp and larger integrated amp transformers tend to have substantial pri/sec parasitic coupling well in the hundreds of pF, so you may see dozens of Vac between chassis and the outside world if left floating. Hence hum on an open unbalanced input. Coupling is not necessarily the same for both ends of the primary winding so you may observe a difference when live and neutral are flipped (I don't think you can easily do that with Australian plugs, while here in Schuko and Euro plug land you can just turn the plug around).
A shield winding connected to secondary-side ground is an effective means to combat parasitic coupling, but you might have guessed that it makes the transformer more complex and more expensive, and definitely special order (though I think most of them are anyway, except from small boutique manufacturers).
The issue becomes even worse when SMPS-style mains filtering is involved. The classic Yamaha *60 series amps (760 and up) with their primary-side-regulated transformer supplies are one such case (it's sort of a dimmer circuit, so presumably quite noisy). If memory serves they have about 2.2 nF (Y class) caps going to both L and N from secondary-side ground in the 220-240V version, leading to substantial hum on open inputs if the unit is just sitting there by itself. The leakage currents were possibly expected to be dumped via the tuner and its antenna system at the time. (JBL 104s are like that, too. Trying to plug the audio cable into a PC is accompanied by scary levels of buzz, but once it's in they're dead silent save for a bit of internal hiss.)
Fortunately these can be modified relatively easily by just resoldering a few things (the US version gets a bit more complex, but the 220-240 V one has a convenient jumper wire on the power supply board). Basically replace the power cord by a 3-wire job, remove connection of Y cap midpoint to secondary-side / chassis ground, hook up PE wire to Y cap midpoint, hook up a ca. 4.7 nF Y class cap between PE and secondary-side ground. This way filter leakage is dumped straight into PE and what remains from transformer parasitic coupling is knocked down by the new Y class cap without creating undue potential (pardon the pun) for ground loops or hazards in case of a badly miswired outlet.
At some point this area needs attention anyway as RIFAs abound, resulting in the famous recall from several years back (I have to wonder how many people would actually want to send in their priceless relics from 40 years ago instead of fixing them for a few bucks in parts, but it was a good PR move for Yamaha if nothing else).
Anyway, back to the L-05M - if that gets unusually noisy, my primary suspects would be Ce3 (a 100µ/35V) and Ce6 (a 470µ/6.3V). You can't rule out a noisy input FET, of course, which would be a bit more tricky to address. Either way, it won't matter much for the OP's very insensitive speakers - at 84 dB/2.83V/m, even 100 µV equates to -5 dB SPL @ 1 m. This is still quieter than almost any integrated amp.
The aforementioned approach to knocking down hum from parasitic coupling could be replicated for the L-05M by grounding its GND terminal via a ca. 4.7 nF, preferably Y class capacitor.
So basically bi-wiring? It would be great if one could attenuate signal going to the mid-treble amp by about 2 dB in order to make up for the driver sensitivity mismatch, but dialing in such a small amount with no ill effects is easier said than done. A 5 or 10 kOhm passive attenuator in a box would do decently enough, or a spare headphone amp. Or a preamp, of course, but the traditional hi-fi variety with a 2-prong mains cord that doesn't spell instant ground loop if there's a PC feeding the DAC, not a Schiit Saga S.
Last edited:
