Before you wonder that you already have a RIAA/phono input in this machine: Well, this RIAA sounds really bad. Uses EF86 which may debated if good or bad, but I think the community already agreed that the ECC83 sounds far better. Then there is the fact, that the circuit with EF86 uses active RIAA (so the filter circuit is built in the negative feedback loop). There are some other solutions instead of active RIAA, like the Hybrid, but I am convinced, that the passive RIAA is the only solution if we wish a perfect sound. Then we have to notice that the input of the phono has no stable impedance as it sees the active feedback (this feedback should be between 1st and 2nd tube but this poor RIAA uses just one pentode) So the standard 47K closing for an MM cartridge is nowhere in this design, securing poor sound.

A solution I received the schematic from “Tunerman”, who is a well-known face of the hifi scene. This new RIAA is a passive one (Kondo design, riaa network from M77). Of course, I used Philips Mustard + some quality capacitors for the 100uF in the realization. I separated the power supply for left and right as well, complete schematic and pictures are below. PLEASE TAKE CARE: The EF86 (B1) heating circuit pins are differing from the ECC83 (B2), so please copy the connection from the 2 outer tubes into the inner. The ECC82 (new B2) and ECC83 (new B1) have matching pinout.

Please measure RIAA caps with something like Der ee de5000( after self-calibration) to secure the correct curve. I bought 20x30x more components to allow selection for the best values. I also made emulations and I think the 10nF should be 9.6nF. So just buy +-10% 10nF and you will find such due to the tolerances. The 47K should be really as close to this. I bought 40pcs 47K till I found a properly matched carbon resistor pair… Same with 1nF, 1M… At the rest I took just care that resistors are matched on left/right, so amplification or characteristic is the same.

To minimize humm noise: Most probably you already replaced the big electrolytic caps, which most probably have no galvanic connection to the housing. Please create a new central ground from the second capacitor (C2,C2a) to the chassis and connect to here the preamp capacitors (C3, C3a, C3′,C3a’ see in below new drawing) All the rest of the GND marks mean the closest GND post (C7,C8, C107, C108 as example connect such way), but try to keep the first tube GND post clean, so only with it’s own circuits, to avoid humm noise. Disconnect the Phono IEC GND pin from the middle pin (if you already use RCA, then just keep GND – so signal GND – separated for right and left and from chassis!). To set filament caused noise to a minimum, disconnect all inputs, connect 230V phase properly in the less attached way (see transformator section), do not connect protection earth GND, shorten the input with it’s own signal GND. Reduce humm with an oscilloscope to a minimum with R41, R141 (use a plastic screwdriver or isolate a normal one to avoid connection to the chassis). And take care of phono cables! (see below)

BOM

Tunermans RIAA upgrade (same RIAA):

https://philipsag9015.home.blog/2019/08/07/tunermans-ag9018/

And here is some result video to explain the differenece between old and new RIAA:

Phono cable!!!

Please take care to use phono cables with low capacity if you are using MM cartridge.

Reason is that typical impedance of cartridge is high – 47K. If you enter this value here, then you will realyze that above of 169pF high frequencies will start to cut away. (169pF is 20kHz at -3dB) (If cartridge supplier has recommendation, typical 150-200pF, follow the lower. )

http://sim.okawa-denshi.jp/en/CRlowkeisan.htm

Let’s see from where the game starts with the ECC83. The miller capacity of the tube is (so this is the input of the tube):

Cgk = 1.6pF + 0.7pF stray = 2.3pF
Cgp = 1.7pF + 0.7pF stray = 2.4pF
A = 56 (for the above circuit) (Calculated here: https://www.ampbooks.com/mobile/amplifier-calculators/cathode-capacitor/)

Therefore the total capacity is:

Cin = 2.3pF + (56+1)* 2.4pF = 139.1pF

Well, not much room left till 169pF (typical transistorized riaa input is similar)! So if you buy an 50pF/meter phono cable and you sourced it in 2 meter length, then you can add 100pF to the above calculated 139.1pF, which results 239.1pF and with the typical 47k impedance of the cartridge you land below 15kHz… ( http://sim.okawa-denshi.jp/en/CRlowkeisan.htm )

In reality cartridges have also some inductance, but detailed data is rarely shared to well simulate it… And you have to calculate with other capacitances (tonearm, cables inside the riaa corrector…)

MM cartridges are poor in high frequency anyway (bigger weight), but loosing due to wrong cables additionally from what remained is really bad.

I recommend to keep cabeling as short as possible and never use cables above of 100pF/m! I personally using an 36pF/m measured 80cm cable and just sourcing an VDH Tide Hybrid (44pF/m) for internal cabling.

Please do not use the AG9015 factory internal audio cable at the phono. These are 300pF/meter cables, which is way too much even on this short 40cm (0.4*300 = 120pF adder)

Nice page, detailing the issue more: https://www.mysonic.space/phono-cartridge-impedance-loading cached version:

MC cartridges:

MC’s are not affected due to their low impedance. You can use these with the above-explained riaa, but you will need most probably an MC step up transformer if the output is below ~2mV. Example this: http://www.72audio.com/mc-trafok/

Heater:

The AC heater is better you would think! It is an DC elevated AC heater, with potentiometer (R41, R141) to find perfect sweet spot. R44, R154 splits the 210V to ~50V DC, which is flattened by the C135 to create the elevation. More here: http://www.valvewizard.co.uk/heater.html

Measurements:

You may use Ortofon test vinyl-s to check amplitudes over the audio frequency range.

Or you may build up the below anti riaa and measure values:

bass reponse must be increased about 20dB at 20Hz

high frequencies must be decreased about 20dB at 20kHz.

More practical is to measure in audible ranges: 100Hz and 10kHz where about 13.6dB of increase / decrease are necessary.

For anti riaa I used 3x 1.2nF as 3.6 nF and an 1nF Philips mustard + carbon composite resistors. I used below schematics: