RIAA phono stage completed

FET RIAA Phono Stage

Finally completed today. It seems impossible from me to get a project finally boxed properly. I’ve done it this time with the RIAA phono stage built recently thanks to the help and guidance from Rod Coleman. This MC stage has a gain of about 70dB for MC cartridges and currently running my DL103a with a 200Ω input load. The beauty of the folded cascode (or shunt cascode as Rod refers it to) is that Miller capacitance is not a problem thanks to the fixed voltage at the drain of the FET input stage. This helps us to have a low input capacitance stage. The second stage is op amp OPA637.

I need to take final measurements of this Phono stage but this is how it looked when I was initially testing it on the bench:

Nice RIAA compliance to +/- 0.1dB. LF noise and gain levels made it difficult to capture the LF response below 100Hz.

The distortion is also really nice with less than 0.025% @ 1kHz and nice harmonic distribution thanks to the CCS at the output which is forcing the operational amplifier to operate in class A:

This stage was well received at the recent London Audio Circle Meeting. It has a nice clean and detail sound. I tried it with multiple MC and even low MM cartridges and has a nice response across the whole bandwidth. Great bass and delicate treble.

I think it is a very quiet phono stage given its high gain. I really like the overall sound and response and I guess that the negative side of it is the 18V battery pack made up of PP3 rechargeable batteries. It needs charge every two weeks and may be a pain in the back. I will try a DC supply, but will have to be really quiet to substitute my battery pack.

IRIAA test box

Testing RIAA stages is a real challenge. However, with current testing gear becoming affordable to us thanks to the development of PCs and sound card interfaces, now you can test your RIAA stage without further excuses. There are software-based solutions to implement IRIAA transfer curve for looking at a flat frequency response of our RIAA stage, but they are not free or easy to implement unless you are skilled on data manipulation on your PC. You can also build your own analogue IRIAA stage. In fact, Morgan Jones suggests a great way of implementing this in his great “Building Valve Amplifiers” (Second Edition) which was recently released. I recommend the book highly, albeit I will not fall in the trap again of buying the e-book version. I was highly disappointed with the e-book release of 4th edition of “Valve Amplifiers”. I guess I will just use my Kindle for reading plain books instead.

The circuit proposed by Morgan Jones is inspired on Hagerman’s paper (thanks Gary B for providing the link) which includes an IRIAA stage for same purposes:

Extracted from "On Reference RIAA Networks" by Jim Hagerman — Extracted from “On Reference RIAA Networks” by Jim Hagerman

Morgan Jones changed Hagerman’s design component values to fit to the 3.18uS Neuman pole instead of 3.5uS and combined 1% and 0.1% resistor tolerances to improve the accuracy of the IRIAA response. You can get the essence of the circuit from the above diagram.

I modified the circuit proposed by Morgan Jones slightly to fit my requirements and in particular I added the additional resistors to simulate DL103’s output resistance of 40Ω.

I had a nice strong mints sweet metal box from Marks & Spencer and used it ast the housing for the IRIAA stage. A BNC adapter also was added to the input to ensure I could hang the IRIAA box from either the waveform generator or the SC interface:

Inside the IRIAA box

MC and MM outputs

IRIAA box

Input connector

Inside box 2

Inside box 3

Building this box takes just a couple of hours and you will be amazed with the results you can achieve. I wish I could post the diagram but I don’t have Morgan Jones’ permission so I hope this post encourages you to buy the book if you weren’t still convinced to do so.

This IRIAA stage has two outputs. One for MC and the other one for MM. You can see the ideal frequency response below:

Continue reading “IRIAA test box”

JFET RIAA Preamp Build (Part II)

I made the following changes to the RIAA stage to improve its accuracy:

Measured all caps
Adjusted 75uS network to compensate for the capacitor variance and the cascode output impedance
Adjusted 318uS and 3180uS networks to compensate for the capacitor variance

All resistors were adjusted with a pair of 0.1% SMD resistors to avoid series connection to reduce noise level.

As shown below, the response is really good matching at less than 1% for frequencies above 50Hz. the error at low frequency is due to the rumble low pass filter, the hum and the soundcard LF response which introduce a slight distortion at very low frequency in the MLS measurement:

Frequency (Hz)	Test (dB)	RIAA (dB)	Error
20	18.80	19.27	2.44%
100	13.00	13.09	0.69%
1,000	0.00	0.00	0.00%
10,000	-13.70	-13.73	0.22%

JFET RIAA preamp build

Last week I started with my JFET RIAA stage. This stage is designed for my current MM cartridge (Ortofon 3e) which has a relatively high output at about 4mV. The stage is a variation of the Werner Ogiers’ circuit, inspired by the John Curl Vendetta. Rod Coleman helped me out here to get this design optimised, so thanks Rod again for the contribution.

A two stage RIAA stage with only about 42dB gain is not ideal. I’m looking at replacing my cartridge soon and hopefully buying an DL-103 MC cartridge. I will then easily tweak the gain and RIAA network to use the Denon one instead.

I built a breadboard of one channel and tested it:

Here is a first test of the RIAA response: Continue reading “JFET RIAA preamp build”

JFET RIAA preamp project started

Now is the time to upgrade my current RIAA stage. A bit unorthodox approach for a valve enthusiastic but you will see me now with my hands on sand. A very simple JFET input stage with a shunt cascode pair to drive the passive RIAA network. The final gain stage is an OPA604 with remaining part of the RIAA network. The Op Amp is forced into class A by a nice pair of BC847C forming a CCS.

Given the shunt cascode bipolar is an PNP Darlington formed by 2 BC860C low noise and the CCS has also SMD devices, I started by soldering these into 4 PCB adapters for SOT-23 devices

More to follow…(just a start)