Understanding Record Repro

Andrew, you have sufficient electromagnetic disturbance for RCA caps to make a difference to one piece of equipment, but the same electromagnetic disturbances are magically not picked up by unprotected wires advertising that they're an antenna with a great big neon sign telling the electronmagnetic disturbance that the antennas are there, ready and waiting and ripe for picking.

All I know is I have to work like **** to get a good performance out of what I design, and it all has to comply with all the laws of nature or it sounds crap.

This has been an interesting thread, thanks one and all.

Cartridge loading and more about cartridge frequency response

I recently read an article on another respected site which showed graphically that the cartridge response started to rise from around 200Hz and then continue to rise at +6dB/octave (20dB/decade) to where it peaks with its capacitive load; and a simulation model of a cartridge had been derived from it.

If the cartridge has this response, then when equalised to (constant velocity) RIAA there would be boost from 200Hz instead of the 50.5Hz specified by RIAA. I have a feeling the website has it wrong. Also the independent evidence of Mr Galo suggests it is wrong. Then again, I and Mr Galo could be wrong... I suppose?

So if we are, then the lower frequency part of the constant velocity curve of this topic is wrong and RIAA records are recorded with some 200Hz cut to compensate, which is news to me. It would also mean that the following simulations and results are wrong, but designs based on them sound right, and if there is cut at 200Hz it should surely be immediately audible, but it isn't.

Anyway, we will press ahead. The magnetic cartridge (MC; MM or MI) is an inductor which has resistance (like a speaker voice coil), and we can find these values from manufacturers' data.

The equivalent circuit is shown below. The cartridge is a signal source (generator) in series with an inductance and a resistance (R, L and G). We also have an arm cable and arm internal wiring capacitance which is more often than not 100pF, but here we simply call it C arm. Then in the phono stage input we have the load capacitance: C load; and the load resistance: R load.



It bears a striking resemblance to a crystal radio set!



(more info can be found here: https://www.electronics-notes.com/articles/radio/radio-receivers/how-does-crystal-radio-work.php)

In fact, the Denon DL103 moving coil cartridge inductance and the tone arm capacitance have similar values to that needed in a crystal or simple transistor AM radio.

The next image shows the difference between a MM and a MC cartridge into the arm cable and a phono stage input having 100pF C load.



You will note that both have a rising response which all magnetic cartridges have, and is evidenced by Mr Galo's article as well as my own research.

You will note the moving magnet rises to around 16kHz then falls off, and the top is rounded, and when equalised, this causes a tiny bit of peaking below 16kHz and then a gentle roll-off. The 47k load is instrumental in damping the response to give the rounded portion. This was OK when high fidelity meant 30Hz - 15kHz, but now it's 20Hz - 20kHz, and to push the high frequency response, the stylus cantilever is 'tuning fork' tuned to resonate provided the tuning isn't 'dulled' by fluff on the needle, when it falls back to the electrical response shown.

The moving coil although having a tenth of the output, keeps rising until it reaches about 1.6MHz, and forms a spike due to the capacitance of the arm cable and C load. Obviously there is no audio output up there, but any interference in the form of AM radio landing on the wiring will do a rough job of tuning into the upper end of the AM band. And with some BJT input phono amps faint distant radio stations are sometimes heard. Even if AM radio isn't detected by the phono amp, its frequency will have to be handled by its circuitry, and it will mix with high frequency harmonic distortion, and this by another radio type function, will result in coloration of the sound.

This next image shows the result of various loads with a moving coil cartridge (Denon DL103: 56uH; 43 Ohms).



The spike with 47k load and just the arm cable capacitance centres on 2MHz. The spike is rounded off by loading it with 500 Ohms. This reduces the amplitude but widens its selectivity to AM radio. By making the capacitive load heavier: 4n7 which is 4700pF, the peak is brought down to around 300kHz but still peaks into 47k. By reducing the resistive load to 500 Ohms the peak is only brought down slightly. 100 Ohms rounds it off and its effect on the audio band is only a reduction in amplitude (not frequency).

The sound quality with moving coil is considerably improved by loading the cartridge with 4700pF and either 500 Ohms or 100 Ohms, and this is because the phono amp isn't having to deal with superimposed AM radio frequencies for which it isn't designed. When called upon to deal with them the frequencies mix with its high frequency harmonic distortion and coloration of the sound is the result.

Also, the radio frequencies which can be easily amplified by any audio quality op-amp, develop across every load in the circuit, and that load signal is returned to the power supply. The signal can easily swamp that of the cartridge, and the op-amp (or a discrete circuit) doesn't have much if any power supply rejection at such frequencies. Therefore the power supply electrolytic bulk storage capacitors have to deal with it. They may have good ripple current capability at low frequency, but not at radio frequencies, and that means they heat up, and a hot electrolytic capacitor doesn't last very long.

So is MC such a good idea?



Edited by Graham Slee - 24 Sep 2017 at 1:35pm

Fatmangolf wrote: ... does increasing the load capacitor to say 100pF change the graph much please?

I'm sure Graham will be along soon to discuss MC loading to suit various cartridges.

In the meantime, with respect to the capacitor question, Graham's graph shows MC slopes with 500R or 100R load resistors both of them with 4n7 load capacitors. 4n7 (n is for nano-Farad) is 47 times larger than 100pF (pico-Farad).

Geoff