diVine Audio

A very small “bookshelf” using the RS100S-8 and ND20 tweeter. Designed around the Hammond 1590E project box to keep the package as small as possible. This is intended to work with a sub, such as the up-firing RS265HO I have. As such, this is a sealed design with a fairly high F3.

Both the 8ohm and 4ohm RS100 exhibit some broad peaking in response when measured in-room, in-box – with the RS100S-8 it’s not a particular problem, but with the RS100-4 it’s quite an issue. You can see it in the response graphs for the optimized RS100-4 variants below.

For the RS100-4 I have two variants posted – there is a lower parts count version, though I’m still a little wary of the power handling implications on the tweeter here since this pushes the crossover point to 3kHz – a bit lower than perhaps ideal. I also have a version (TM and MTM both) that’s optimized – it’s almost perfect LR4 tracking with a 5kHz crossover point – but the parts count goes up quite a bit (12 parts including a pretty big cap.) It may be worth it in the end if you’re set on the RS100-4.

I will probably be revisiting the RS100-8 design since the shielded version I used is no longer available. I may also try out a couple other tweeters since this was originally a $5 tweeter and is now in range of some other superb (and more flexible) options – not to mention some other interesting options if I use the 5kHz target.

Box design

As noted, the main box is a Hammond 1590E project box measuring ~7.4″ x 4.75″ where the baffle attaches. I have epoxied in bracing of 1/4″ ply (with one 1/2″ ply brace used to allow adequate room to mount the crossover with screws.)

RS100S-8 + ND20, Full BSC

RS100S-8 + ND20, On-wall variant

RS100-4 + ND20

This should work well unless you push it, though I haven’t verified just how much the tweeter can take with a 3kHz crossover. Some contouring of the response is being done with the zobel here – that’s the key to dealing with the peaking in the RS100 response. Unfortunately, that peak also forces the crossover point.

RS100-4 + ND20 – optimized response, high(er) parts count

This is a near perfect LR4 example, using a combination of tweaks to reach this. The parts count is up, but the result should be worth it. This is a 5kHz crossover point to the tweeter.

RS100-4 + ND20

This has NOT been built and is based on measured data from the TM. It features the 3kHz tweeter crossover point and may have power handling issues as a result. Build at your own risk.

RS100-4 + ND20 – optimized response, high(er) parts count

This again is a near perfect LR4 example but at the cost of more parts. Same 5kHz crossover as the TM. Also tentative based on the TM data, but should be good.