I suggest you ask Cary. I could find no circuit data for the AES AE-3 DJH but I would be concerned about excessive loading, especially for a headphone.
There's an easier way. There are splitters that have one male RCA to two female RCA. You can either buy the type that is a little solid adaptor:
or get the type that has a little bit of cable. Just google "male RCA to two female RCA splitter" to see your options. It will be cheap and won't affect the originality of your preamp.
Good comments above. As Kal indicated, loading is the concern. And using an external splitter or y-adapter, such as Viridian suggested, will result in loading effects that are identical to those that would result from the internal modification you proposed.
To assure that no sonic degradation occurs, the parallel combination of the input impedances of the headphone amp and the main power amp should ideally be 10 or more times as great as the output impedance of the preamp, at the frequency within the audible frequency range for which the preamp's output impedance is highest. If that criterion is not satisfied there may or may not be adverse sonic consequences.
The preamp has a specified output impedance of 560 ohms, which is most likely at mid-range frequencies. I wouldn't be surprised if it increases to 2K or 3K or so at 20 Hz, due to the output coupling capacitor it probably uses. So ideally you would want the parallel combination of the two load impedances to be 20K or 30K or more, or some rolloff of the bottom octave, and possibly other frequency response irregularities, may occur.
The combined input impedance of the two loads is equal to the product (multiplication) of their individual input impedances divided by their sum. For instance, if the headphone amp has an input impedance of 50K and the power amp has an input impedance of 100K, their combined input impedance would be (50 x 100)/(50 + 100) = 33.3K.
You would also want to avoid cable lengths between the preamp and the headphone amp and power amp that are particularly long, and/or use cables that have low capacitance per unit length, to avoid possible upper treble rolloff as a consequence of the interaction of cable capacitance with the preamp's output impedance. The TOTAL of the capacitances of the cables to BOTH the headphone amp and the main power amp will affect the signals that are seen by BOTH of those components. In other words, the capacitance of the cable to the headphone amp, as well as the capacitance of the cable to the power amp, will affect the signal seen by the power amp. And vice versa. If either or both of the cable lengths are particularly long, post back with the lengths, and the cable makes and models if known, and the relevant calculations can be performed.
Very informative reply Al, greatly appreciated. When I plug in the input impedances of my amp and and headphone amp to the equation I get 13.5k ohms for the combined input impedance. So its definitely more than 10x the listed 560 ohm output impedance of the preamp, but not if in fact sometimes increase to 2-3k as you mentioned. What do you think?
Al, the DJH does not have coupling caps. It's direct coupled. I don't know how that influences impedance, but I wanted to point it out.
There is very little in the way of documentation on this preamp. I had one for several years and loved it. You should contact Cary if you have any technical questions. They are good about answering.
Your headphones' impedance could be too low even all by itself. Many headphones have impedances like speakers', in the 4 to 16 ohm range. (Putting a high impedance amp in parallel to a pair of headphones would not make a lot of difference.)
Presumably, you would unplug the headphones while listening through your speakers, so the preamp should be OK then. But when listening through the headphones, the low impedance of the headphones could result in damanding more current from your preamp than it can supply without distortion, as others have already aluded.
So, why don't you plug your headphones into the output of your amp instead?
James is correct, aside perhaps from the last sentence, but I and I suspect most of the others have been assuming that the second RCA output would go to the input of a headphone amplifier (as I mentioned), rather than to headphones directly.
A preamp having a nominal output impedance of 560 ohms will be unable to directly drive nearly all conventionally designed headphones, with any semblance of good results.
As far as connection to the amp output is concerned, without knowing the specs of the components that would be involved I suspect that you would be putting way too much signal into the phones unless the volume control is set extremely close to the bottom of its range.
Sounds like a plan! Given the 24x ratio, and the info Mingles provided that there is no output coupling capacitor, and the response from Cary, I think you'll be fine.
Concerning splitters, opinions tend to be divided. FWIW my own opinion is that it is very likely that using an external splitter would be sonically indistinguishable from modifying the preamp to do the split internally. So I think that Viridian's suggestion is a good one. Here are some additional splitter alternatives, from Audioquest (the second and third of the three items shown).
If for any unforeseen reason you find that having both signal paths connected at once compromises the sound quality of either, HifiHarv's suggestion of a switchbox would be a reasonable fallback approach, IMO. The DB Systems DBP-2JAU/5 would be a good choice, IME/IMO. See the nearby links labelled "photo" and "review" for more information on it. You would connect it in "reverse," with the preamp going to the jacks labelled "Output," and the two amps going to two pairs of "Input" jacks. But again, I don't think that a switchbox will be necessary.