Sorry the schematic in the photo above is somewhat unclear.
First, I scribbled the original value if R1 on it, as actually found on my PCBs (I checked against different schematic versions). Second, I wrote some recommendations by DieBogener (I believe, IIRC - not sure). The new resistor values are finally written with the computer (as opposed to the other hand written stuff). I should probably have photographed a nicer schematic!
I didn't actually follow the recommendation and change C4 in the tweeter circuit, because as Beobuddy suggests, that'd lower the crossover point for the tweeter, which I don't think is useful. If anything, I would actually have liked to try and shift the crossover point for the midranges upwards, so that they'd reach a bit higher. The new drivers could theoretically do quite well up to 9kHz, although because of the line array arrangement,very much above (specified) 5kHz would probably not be a good idea. In any case, I don't think it really can be done without changing L3 on the midrange board; At least I couldn't do it in XSim.
The modifications to the midrange PCB are 0.68Ohm instead of 2.2 Ohm for R1, because replacement drivers have supposedly (listening, driver data sheet) a tad bit lower efficiency than the originals. According to XSim, 1.4Ohm would have been the optimum value for R1, because lower values seem to create a peak around 2400Hz in the simulation. Listening to both, I however found that I preferred the 0.68Ohm configuration.
One thing to note about my simulations is, I have no data whatsoever (except DC resistance) on the original drivers, so ran simulations with flat 4Ohm drivers. Thus, the simulations are really nothing more but extremely rough approximations! I still found them useful to get an idea of how the crossovers work and what might actually help to attain desired changes, before soldering and listening.
I've also changed the 6.8 Ohm resistors in the boucherot circuits (R3, R4 on midrange PCB) to 7.5Ohms, roughly twice a new driver's DC resistance (measured 3.7Ohm). Something I have not done so far is adapt C6/C7 to the new voice coil inductance. Not sure if I ever will, Im happy with the midranges. If I'm gonna change something at all, it'll be reviewing the 0Ohm in the tweeter circuit.
The new drivers are Visaton FRS8 (4 Ohms). They need a bit of mechanical trimming to fit; I used metal shears which worked a treat. The surrounds have sort of a ring around them on the front. Pay close attention as not to damage it while trimming (I ruined one driver like this). Conveniently, the original driver mountings can be reused for these drivers. It's advisable to have some kind of o-ring between driver and assembly though, as there might result an open sub-millimeter gap between driver chassis and assembly panel, probably when the trimming is not done "round" enough or something. For me, this was a non-issue, my pentas still had (the supposedly original) o-rings. My coworker found it to be a problem though, the o-rings in his pentas had mostly disintegrated.
Considering burnt coils, he had one completely fried (deep brown and crispy) L3, even though the crossovers in his beolab pentas had a 3.3Ohm R1. Interestingly, also L6 on both bass/tweeter boards had ... somewhat deteriorated:
Those L6s were both replaced with iron core versions (didn't find air core ones that large, but they're not located in a signal path anyways), together with the L3s (including the non-burnt one).
Also checked his amplifiers for DC offset and found nothing above few millivolts. I have no idea what happened there. The coils themselfes weren't burnt, also their lower ends still stuck firmy to the PCBs. I somehow doubt that heating of the large coil (located below L6 when mounted) caused this. Maybe it was heat from the amplifiers during heavy duty use? They didn't produce much heat when I had them here for testing, but who knows what the previous owner used them for.
My beovox pentas on the other hand had originally a 2.2Ohm R1 and no coil related problems at all. I'm running those pentas on an Onkyo TX-NR838 with quite a bit of oomph at times, so will probably re-check the L3s after a while and, if burnt or even just preemptively, replace them with suitable air core inductors from intertechnik (they have a somewhat thicker wire). Those inductors are actually not that expensive either, much unlike replacements for the big transformer core coils.
Concerning the "star topology" wiring: The original wiring for the bass drivers is done daisy-chain style; There's a spade connector on the PCB, it has the red wire running to the first driver. The driver actually has two spade connectors per wire color; the wire coming from the xover PCB connects to one of them. The other same-color spade connector has the wire running to the next driver and so forth and so on. This means that the signal arriving at the topmost driver will have passed through a lot of thin wires and (proably even worse) spade connectors. Same applies to the return path (black wire).
I used some high power (32A@230VAC rated) clamping blocks and ran one short 2.5mm2 wire per color (black / red) from the xover PCB to each clamping block. Then, there's 4 seperate 2.5mm2 wires per color running from the clamping blocks to each of the drivers, so that the signal is not looped through numerous spade connectors anymore. In effect, there's two wires per driver now, for the runs along the enclosure.
It shall be noted here that I used absolutely no audiophoolerish $5k/meter speaker wires or somesuch, just regular 2.5mm2 (14AWG) copper speaker wire from what might just as well be home depot. The replacement caps were not super-duper-expensive-copper-tin-silver-unobtainium caps either, but just solid 100V MKT types from Audyn/Intertechnik.
On a side note, maybe the looping-though-bazillions-of-spade-connectors thing is why some people who've done it swear that rewiring makes such a big difference with the pentas. I myself have done it just because I think its good engineering practice to minimize resistance in (relatively) high power paths, while implementing it doesn't cost an arm and a leg either. I'd also not entirely dismiss the notion that, maybe, the bass sounds even dryer and more controlled now than it did before. 😁
