
Tonearm/Cartridge Analysis
From the first inspection of this turntable, I had my reservations regarding this tonearm. It follows Dual’s low mass tonearm ideology, which was brought to the extreme in their later ULM series turntables. The reasons are easy to understand, especially after reading last speed stability chapter and witnessing how tonearm resonance is folding to audible tones. There are couple ways to shift this resonance up in frequency and lower it’s audibility. Decreasing tonearm effective mass is one of them. But as always the case in engineering – there are no free lunch. Making tonearm lighter inevitably will compromise it’s structural integrity and invite new resonances. In our particular case, this removable plastic head-shell will make things even worse. How worse ? Let’s find out. Here is spectrum of pink noise track 3 side one, compensated for 3db/oct fall.
We see only one cartridge resonance at 15kHz. This can be predicted empirically.
Modeling Cartridge Resonance
Let’s build our basic model by measuring Shure M75MB Type II cartridge parameters and adding loading capacitance (only wires to measurement preamp in our case) and resistance.
Here is the resulting frequency response.
We see a big discrepancy between measured and simulated frequency responses. It seems, that our electrical resonance at 7kHz is pretty much fully damped by wires capacitance and loading resistor. So why we measured this 15khz peak ? Well, reality can’t be that much simplified. Cartridge is a mechanical device and has it’s own mechanical resonance, which can be modeled as 2nd order LP filter. Also it’s inductance is not linear with frequency, so it’s beneficial to divide it in to two parts and add shunting resistance to model losses. There are much more caveats, but that’s a whole different can of worms so let’s leave it for separate article. Here is more advanced cartridge model and it’s frequency response.
That’s more like it. Now loading capacitor and resistor can be selected for flattest response with good compliance to reality.
It seams that 30kΩ with 500pF is optimal here. Let’s check this setup for real.
Much better, we lowered resonance peak some 3dB. Overall I feel that pink noise measurements show a good correlation with real life tonal balance of the cartridge. So it always pays of to have it as flat as possible. Unfortunately it doesn’t show any structural resonances of tonearm assembly. There is just not enough energy density to excite them.
Measuring Tonearm Resonance
For this purpose we must use slowly sweeping sine tone, and the slower – the better. This way all resonances will have time to build up to their maximum. This resonating or “flexing” of the tonearm and it’s assembly should then produce peaks and dips in frequency response. HFN test record has 20 seconds 20Hz – 20kHz sweep, but it’s a logarithmic one, which means there could be again – not enough excitation for resonances to build up. At least in high end spectrum, where sweeping speed increases logarithmically. So it’s not ideal, but let’s try and see. Here is a spectrum of this sweep, compensated for 3db/oct fall.
Something is seriously wrong here. I expected some frequency anomalies, but this is beyond what is reasonable. First thing that stands out, is this repeating pattern of peaks at 20-30, 200-300 and 2k-3k Hz. They are spaced too far apart to bear any harmonic relationship and are too pronounced to be resonances. This must be defect of the test record. Quick check on other turntable confirms this suspicion.
HFN strikes again. Well, it’s really time to buy a new test record it seems… So the real resonances will stay undetermined this time, and we all know whose fault is that. Lesson learned. Next time we will be more prepared, with better test record and will adapt vibrometric approach to the problem.