Like all projects & tinkerers, yes I've done some work, but I haven't finished it. And of course since you asked the question then I'm not currently working on it, nor do I have access to my notes at this time.
That being said though, I found some interesting results that I remember off the top of my head.
First was that I found a couple sources of information on the internet, and drug up tech manuals that had block diagrams, wiring pin outs, etc. And sadly none of them really agreed. What turned out to be most correct was a pinout/block diagram of the LS671 internals that I found in a tech manual. In the end, I traced everything and experimented with it on my own to validate the results. I never cracked open the unit, but what I observed matched the block diagram I was seeing.
The unit certainly runs best on 24V. It will run on 12V but it won't drive the speaker as hard compared to 24V. I didn't characterize the reduction in power however. I could have given the fact that the raw speaker pins are brought out to the main connector. I'm guessing there is probably a point between 12V and 24V where the loss starts to occur, which probably matches up to how they built their voltage regulators inside.
The volume knob/mute switch adjusts the volume on both the speaker and handset. When I first looked at the block diagram it looked like they might be independent which I thought would be cool. However, they are tied together. As it turns out though, they setup the gains correctly so that its works nicely switching from handset to ears only. The LS671 internal circuits also handles cutting off the radio feed back when you transmit so the speaker doesn't create a feed back loop. Note: this feed back is produced by Military radios as an indication that the radio is transmitting.
The main connector( J1 ) provides alot of nice signals. It appears that it observes some sort of "line level" behavior. However I recall that its version of Line level doesn't match up with the modern definition of Audio gear. It seamed to be alittle larger, maybe +-2V instead of the +-1V of modern gear. These results along with the block diagram indicate that there are active components inside the unit. The handset for instance uses a dynamic mic which only puts out something like +-10mV, yet the audio from J1 is +-2V.
I used both an Oscope and Signal generator to do some testing on my unit. Everything looks clean and sounds good. I did this sort of testing since I got my surplus and it was shipped AS-IS, UNTESTED. As it turns out I got lucky and the unit appears to be fully operational.
I made an adapter cable that would go from J1 to the DB9 pinout commonly used in ham radio for TNCs. With this I then used my existing TNC cables to connect the LS671 directly to a couple radios. I tested against the PRC119, a TMV7A, and an older Kenwood HT. For the quick test I was doing I simply used a POT to correct the voltage levels going in each direction. All of the radios gave me the same results. RX Audio( Speaker ) came through but there was always the same background tone in the audio. I never did debug it further. I did order some 600:600 transformers to insert some isolation to see if that was the problem, but they still sit on my work bench unopened. TX Audio worked fine. This is where I really had to knock down the voltage levels and insert a DC blocking cap. TX audio was shoved through a standard H250 handset and my voice. My lovely assistant and I took turns on each end of the radio link and the audio sounded fine.
I did find a document some where( I can't remember which one ) that did say that the J1 interfaces were all 600ohm busses.
I did try and test fit the LS671 in my daily driver and its a tad bigger than I was expecting... or perhaps my daily driver was a tad smaller than I thought. So I'm still working out how I want to mount the unit.
In another experiment I got a hold of some LS454 NOS speakers. After tweaking their U229 plug so that it'd work on the PRC119 directly I was surprised to see that the unamplified LS454 sounded just as loud as the LS671. Its not truly a fair comparison, just an interesting observation.
So that's where I stand. I still see the speaker sitting on my work bench and want to pursue the project further, but life gets in the way. Remaining items would be tracking down the background tone I'm seeing and then building a proper interface unit. At minimum I'm going to have to build a circuit that will adapt the various voltage levels between what my radios produce and what the LS671 expect. At the higher end... this adapting may happen in a larger multi radio system or intercom system. Though to be fair and realistic, I'll probably build the minimum at first to work out the bugs and then build something bigger.
A final note on how these units where originally used. The LS671 was always driven by the mounting plate/structure for the SINGARS radios when installed in vehicles. This was done for a couple reasons. 1, the switch on the LS671 controls the main power distribution within the vehicle mount. 2, most installs were in smaller vehicles that didn't have intercoms, as such the LS671 was the main interface to the radio system. It was also common to have a long cable on the LS671 so it could be removed from the vehicle and used at a remote location. This is actually the most common use I've seen for it. This being said... there will always be some sort of adapter/middle man component between the LS671 and radios when you're trying to use it in a custom application.