I see the PM-3 as being so efficient, it doesn't need the power increase available with amps that offer more power with balanced output than with single-ended output.
The PM-1 is even less efficient than the PM-3 and I hear no improvement in bass control or dynamics when I run the PM-1 with 2000mW into 32-Ohms from the HA-1's 4-Pin XLR jack vs. 500 mW into 32-Ohms from the HA-1's TRS jack. There are others who agree. 500 mW is more than enough power to bring out the PM-1's best performance.
The PM-1 behaves much the same way with my portable iBasso PB2. I can't say for sure how much power we're talking about there, because it varies with the op-amps I've selected, but when using LME49990s in the input voltage gain stage and dummy buffers (no current gain), the PM-1 again sounds just as good on the TRS jack as it does on the balanced output (that offers twice as much power).
I think that's why several people have (incorrectly) said the PM-1 doesn't scale well to more power. In truth, it does scale to more power, just like an HE-6 or LCD-2, but you have to start out with a very weak amp, like that of a Sansa Clip+ or Sony Walkman, where the PM-1 will sound woolly and uncontrolled in the bass, to appreciate how the PM-1 can scale to more power. It's just so much more efficient than an LCD-2 or HE-6 that you will not hear it "scaling" when comparing 500mW to 2000mW, as with the HA-1 TRS vs.4-Pin XLR jacks.
Not able to test the even more efficient PM-3 with balanced cables on either the HA-1 or my portable iBasso PB2, I have to conclude from my PM-1 observations that going balanced with the PM-3 would offer little or no audible improvement, unless we're talking about using a balanced-capable amp with unusually weak output at its TRS jack and only modest output at its balanced jack. I don't know if any such amp actually exists. Even the surprisingly under-powered Sony PHA-3 DAC/amp might have enough power at its TRS jack, with 100 mW into 32-Ohms, for the PM-3 to audibly sound just as good it would on the PHA-3's balanced output, 320 mW into 32-Ohms. I don't know of any other balanced headphone amp that puts out only 320mW (or less).
Aside from the "potential" benefits of more power offered by balanced connections, there are some experts I admire and trust, like Dr. Jan Meier of Meier Audio, who can explain how the shared ground of a single-ended connection allows impedance fluctuations caused by one driver to influence the frequency response of the other driver - a "problem" not suffered with balanced headphone connections, but planar magnetic headphones are almost purely resistive, and thus not as vulnerable to these cross-channel impedance fluctuations with unbalanced connections.
Frankly, even with dynamic headphones, I have to say that I, personally, have never heard any benefits from running balanced vs. single-ended when using an amp that offers sufficient power at both jacks. Seriously, I don't even know what to listen for when attempting to hear these, no doubt real but apparently difficult to discern, cross-channel impedance fluctuations. And I've never read a single Head-Fi post where someone was saying, "Oh man, the cross-channel impedance fluctuations were really bad on the TRS jack, but now that I've switched to balanced cables, I can't hear them any longer!"
Lastly, one has to consider that there are many very expensive headphone amps out there that offer only a single-ended TRS jack, and yet, their typically very experienced owners are perfectly content to run their flagship headphones without balanced cables.
The lust for balanced cables has a life of its own.
Mike
