- This post will describe the modification and optimization of the LPG26T titanium dome tweeter
- The first tweeter modified was one sent to Mark K for measurement, PRIOR to modifying the tweeter- this is an interseting point, considering what he measured and what I measured in the impedance curve, before and after.
- A problem with the 26T which has been noted by others is that it lacks accurate fixed positioning for the tweeter/VC assembly. Other matters to be addressed are the diffuer used for dispersion and frequency resonse shaping, and the relative lack of damping material in the hollow pole piece.
- Read this whole document through in detail before deciding if you want to do this- it does require some fine motor coordination, and some simple test equipment ( a signal generator).
- Understand that I've been testing and modifying drivers from time to time since the the late 70's- I understand this sort of thing may not be everyone's cup of tea, or within their skill set. This tweeter has some good potential, but to realize it this modification and alignment seems quite necessary- I can't recommend using them in the condition they were in as received from the American distributor, Madisound. Not Madisound's fault, just perhaps a slight lack of quality control at LPG, and a subtle flaw in the mechanical design.
Tools and Materials Required
- Hex Key set for removing front panel
- Xacto knife or small jewler's screwdriver for removing diffuser
- Small package of Lamb's wool, Scholl's or Rite Aid, available at most drug stores (foot care)
- Small scissors such as mustache trimming scissors
- Black Silicone RTV glue (optional)
- Adjustable Sine wave Signal generator with 50-100 ohm output impedance, and connecting cables- headphone output from notebook or PC with tone generator software like True RTA is OK. I used my Krohn-Hite Sine generator
- Using a hex key wrench, remove the four screws holding the front panel to the magnet/pole-plate assembly.
- Place the front plate assembly to the side for now, dome with screen side down. Don't rest the plate on the voice coil side, as this may cause deformation of the suspension.
- Prepare a small quantity of the lambs wool, folding it over two or three times in a 1-1/2” length. Cut the length in half, and insert ½ into the pole piece cavity.
- Using the small scissors, trim the wool around the pole piece opening until only a shallow dome of material remains above the opening, and no fibers are in the areaof the magnetic gap where the voice coil sits.
- There are small dots of glue at each of the screw positions. This can be cut away with a Xacto knife, then the grille can be removed from the front panel.
- The diffusor is a thin plasticfilm on the inside of the grille. Pry it off with a jewelers screwdriver or Xacto knife.
- Carefully position the front plate assembly over the pole-piece with gaskets/spacers positioned as when originally disassembled.
- Re-install the screws and screw in until barely snug- you'll still be able to move the front plate around a little, because of slack between the screw body and the holes in the front plate.
Magnet and pole piece assembly with pole piece stuffed with lambs wool
Front Plate with dome assembly; note plastic diffuser donut under grille which should be removed
Re-assembled tweeter, before attaching grille
- The tweeter plate has to be positioned so that the voice coil is centered accurately in the gap. Apparently, they often aren't even from the factory.
- Set the signal generator to 1 kHz output frequency and 0dBV to +6 output (0.775V to 1.5VRMS). The exact level isn't important. 1 kHz will be sensitive to the slope, height, and width of the imipedance hump, which is actually at a lower frequency. (see impedance curves below)
- Connect the signal generator to the tweeter terminals. Rotate the plate as fully as possible in one direction about the central axis; then shift about a little as necessary to get the loudest output without any distortion overtones. This results at peaking of the Z.
- At this point the tweeter VC should be centered fairly optimally in the gap. Snug down the four hex screws, taking care not to move the front plate on the pole piece. If you hear a change in the character of the 1 kHz tone, such as the presence of a low level sharp overtone, the voice coil is not centered and may be scraping. Loosen the screws sightly and adjust the plate position for the cleanest 1 kHz tone and maximum loudness.
- The tweeter grille can be positioned over the dome and will be held on magnetically. Use small dots of RTV near the screw heads to attach more permanently.
This particular tweeter was measured by Mark K, with higher distortion than I or he expected. When returned, I measured the impedance curve PRIOR to this rebuild.
After the modification and alignment, the impedance curve looked like this:
Note the absence of the various ripples in the several kHz region now, and the broader and higher impedance peak at the LF resonance.
Mounted on the small Woodstyle test enclosure, with minimal smoothing, the response at one meter looked like this:
Note that the cabinet boundary and diffraction effects the frequency response measured this way; this has to be taken into account when desiging the crossover. Sometime after I'm back in CA I'll have to do a near field measurement to determine the intrinsic driver response.
~Jon
Comment