By Pete Basel    9-9-2008
The Classic SPICA TC-50 Loudspeaker
Restoration and Modifications
The Author's TC-50 Pair - Serial #12990 - Service Code H3.8-12
Many years ago, around 1984, my brother brought the SPICA TC-50 loudspeaker to my attention after reading about it in Stereophile and noting that it was a linear phase design. The Spica Enthusiasts web site covers much information about both SPICA as a company and the TC-50 in particular. There is also the Yahoo SPICA Users Group (SLUG) for Spica Enthusiasts. This delay derived, approximate linear phase design, involved a new crossover invented by the designer, John Bau. The owners manual lists it as having a 4th order Bessel low pass woofer response, an approximate 1st order tweeter response, and time delay to provide a linear phase response. The TC-50 design was derived empirically through computer simulation. This paper, published in 1983, provides an analytical treatment of similar crossover topologies. I am not sure if John Bau drew from this paper when developing the TC-50:
S.P. Lipshitz and J. Vanderkooy, "A Family of Linear-Phase Crossover Networks of High Slope Derived by Time Delay", J. Aud. Eng. Soc, vol 31, pp2-20 (1983 Jan/Feb.).
The TC-50 crossover design is also mentioned on this web page: John K. on the TC-50 refer to text around Figures 13 and 14
The TC-50s were on my mind for years and I had intended to purchase a pair at some point. I still had not found a pair by 2004 but the schematic had been posted on the net which allowed me to simulate the design in CALSOD based on published data for the drivers. This simulation showed good agreement with data published by SPICA and it did show a reasonably flat phase response in the passband: TC-50 Design Analysis by the author.
Finding a Pair of TC-50s
I began looking for TC-50s in my area since I do not like to ship speakers, however they are fairly rare and do not show up often for sale locally. I found a clean pair recently not too far from my home and purchased them in June of 2007. They looked very clean, and are pictured above, these are the later production series which are easily identified by the small cutout in the felt pad for the tweeter; the earlier version had a much larger rectangular cutout. This later version is sometimes referred to as the TC-50i. The tweeter screws looked as if they had been removed and these systems probably had the tweeter domes replaced at some point. The felt pad was probably removed to do the work and had not been reglued. The felt can be seen resting on the woofer frame in the picture at the top of this page. It was clear that the binding posts had been replaced with what look like a solid copper type:
Initial Listening Test
Initial listening tests were very discouraging as the image was diffuse beyond belief, and the bass very weak. I measured the frequency response and found a large notch in the response around 3 kHz; this notch was larger in one system. I removed the tweeters and opened them to have a look at the domes. Looking closely the voice coils were clearly not at 90 degrees to the mounting surface. It seems that the felt rested on the dome for a long period causing a shift in the alignment. Input impedance measurements showed that the dome was scraping. There does not seem to be a good solution to this problem. The tweeters in the TC-50i are a later version than those used in the TC-50. The part number marked on these tweeters is: DTW100T25F 4* (4* indicating 4 ohms). Replacement domes are not available, however 4 ohm dome/voice coils are available for the next revision of this tweeter (AAC/Audax TWO25A2) which have a catenary shaped dome and fit the same magnet structure. I purchased a pair from Madisound and installed them and the tweeters now tested normally concerning their T&S parameters. Note that the polarity markings are incorrect when this dome is used in the old magnet structure. The staff at Madisound offered a warning about this and I confirmed it with a 1.5 volt battery.
I also noticed that the pressure from the felt pad resting on the edge of the woofer suspension had caused it to begin to separate from the metal frame. Both woofers were reglued paying careful attention to maintain the proper voice coil centering. I believe that these were leaking causing the diminished bass output.
I modified an input terminal cup to accept dual gold plated basic binding posts and wired the drivers directly to the posts so that the crossover network could be out-boarded. This worked out nicely, allowing modifications to be tested and crossover transfer functions to be easily measured.
I also measured the Thiele and Small parameters of the woofers which passed a sanity check. Detailed impedance characteristics were measured for all of the crossover components; these are needed to refine the CALSOD simulation and to select proper replacements.
Listening Test Round Two
Listening again with the new tweeter domes and repaired woofers revealed the magic associated with these speakers. The TC-50s now disappear producing a very solid center image. Head movement from left to right showed that there was clearly a point where the image came into focus, more so than with most speakers. The system seems a bit forward and bright which might be due to the newer tweeter being more efficient. I'll have to test an original type tweeter to confirm this, however this was also reported in several test reports.
A TC-50 was placed in the center of the room, back to the floor facing up, and a Neutrik 3382 measurement mic placed 1m away on the axis 7.5" up from the bottom of the enclosure as suggested in the TC-50 manual. A MLS type frequency response measurement was taken using Liberty Audio Suite (LAUD) system. The time markers were adjusted on the impulse response to remove the time of flight delay. This is required to obtain a true representation of the system phase response. The measured amplitude and phase response are shown below with no smoothing; note that the responses are not valid below 400 Hz due to the MLS time windowing. This is a typical plot, not of this pair, measurements of this particular pair will be taken as time permits. Agreement with both simulation and published specifications is good:
Fig. 1    Typical TC-50 MLS Frequency and Phase Response Grille On (not this Pair)
The measured crossover electrical response is shown below. It can be seen that the peaking in the amplitude response at approximately 8 kHz (actual is marker #1 at 7131 Hz) is real, also of interest is the deep notch at 3.8 kHz (these are service code H3.8-12). The baffle step compensation is clearly seen between 100 Hz and 1 kHz. The notch depth is dependent on the ESR of the cap in the notch circuit, and similarly the peak height is dependent on the ESR of the 3.1 uF capacitor. These were not known when I performed the CALSOD simulations and the agreement in these areas will be improved by refinement of the model taking ESR into consideration:
Fig. 2    Measured Woofer Crossover Response, Note Peak at 7.1 kHz
Fig. 2    CALSOD Simulated Approximate Woofer Crossover Response
Standard Cup Modified to Mount Two Industrial Grade Gold Posts for Bi-Amping
Drivers are Wired Directly to the Posts Allowing the Crossover to be Out-boarded
More to come as time permits.