attitube

• MC vacuum tube phono preamplifier  – “The Punishment”

The phono preamplifier ‘journey’ was the most time consuming and painful diy trip I ever performed. I tried several topologies for MM phono over the time: silicon chip ones, FET transistors, plus several vacuum tube topologies. None of them were satisfying sonically speaking, or they don’t have enough gain for my needs (the fet one sounded good to my ears if battery powered, but to find 4 fet devices as pairs is not an easy task, as you need like hundreds of units to test them to match few similar pairs, their values being spread too much). The mistake I’we made at the time was that the record player – constructed in parallel with the phono stage – was far from perfect, suffering all kind of upgrades and changes; It wasn’t good enough to serve as a base for serious sonic testing. Don’t make the same mistake, it will hurt and it will be a time consuming process. Be sure you have a good record player unit first, and then attempt the construction/testing of a phono stage! Thanks to my friend Virgil, after failing to build on my own a decent sounding phono pre, I follow his advise and start to test the topology he suggested: a SRPP – passive RIAA network – SRPP architecture, by using the tubes ECC803s for the first stage and E188CC for the second stage. I want to mention some other projects you could try, depending on the components you may have or afford:Triode Dick’s –  PhonoDude , and Stuart Yaniger – His Master’s Noise . My first taught was to build a passive LR riaa, but the cost was way too much for my budget. It’s the most elegant way to build a passive riaa  phono stage one could afford: it’s very easy to get the correct riia curve, no matter the topology you used, as the only think you could easily adjust is a resistance, which is easy to get in less than 1% precision after finding the proper value for a perfect curve. Here is the Punishment diagram. The problem with a vacuum tube driven passive riaa preamplifier is that is hard to obtain a correct riaa reproduction curve because of the influence of the internal impedance of the first valve, the input Miller capacitance and input impedance, seen in each case from the riaa. So, after realizing the stage and the riaa network, one should inject a sweep signal trough a reverse riaa circuit and watch with a spectrum analyser, then to correct the values of the riaa network  in order to get a curve as closed as possible to the standard riaa curve. For the first stage we need a high amplification, low noise triode with a minimum internal equivalent resistance  – as seen from the riaa network. We have chosen ECC803s, which was a good compromise for our needs. The second stage needs the biggest input impedance as seen from the riaa network, and a decent output impedance in order to drive the next stage – the final amplifier I my case. As it was supposed to drive the 6s45PI valve, the second tube choice was E188CC SQ, cause it was producing enough gain for that, they sound terrific and I had a pair of them around. The passing capacitor C1 should be 220-330 nF to minimize its influence on the riia network; I choose an 470nF/ 630 Volt  Auricap one. As for the exit capacitor, a military Russian < PIO K40Y-9 > 680nF/ 1000Volt was picked for the task. You may find better capacitor combination for your project, but these two make me happy at the time. RIAA network is the most important spot, here precision is required. Capacitors with good precision 1 – 2.5% are hard to fund at high AC voltage rating, so silver-mica high precision were used for C2 and C3, and 1% Vishay resistors for the R7 and R6.  As the riaa network components are not resisting to high voltages, we took the signal from the first stage trough a capacitor, and the riia network is direct coupled to the next stage. Someone could reverse this topology like SRPP – riaa – Capacitor – SRPP, but then he must use appropriate voltage ratings for C2,3 and R7,6. I was using for the first SRPP the R4 and R3 the 1 kOhm value (mean approx. 1.5 mA trough the Ecc803S) but more skilled engineers like Morgan Jones recommend a higher value like 1.5K ( as result the current trough the first valve is 0.8 mA), thus contributing to a lower noise floor for the first stage. You should check this on your own.

Also, Virgil calculated he proper value for R5 to be 46.5 K, R7=6k7 and R6=220 for a value of C2=16nF, if R8 is 2Mohm (he checked his montage with reverse riaa method). The values I used are those from the diagram, except that R10 and R9 = 330 Ohm. A simple online passive riaa calculator is here at kab, and all the theory ‘in extenso’ on tube powered passive riaa could be found here at platenspeler.
The PSU used for this project is again a passive one; but this will be changed in the near future for a stabilized modern one. I used a CLCL configuration, by using two chokes of about 15 H. The rectifier bridge was made from four discrete diodes BYW96E fast soft-recovery controlled avalanche rectifiers. A mains filter was used before the mains transformer, which is a toroidal type one. For the filaments of the vacuum tubes a soft start slow rise output voltage based on a LM 350 T chip, which gives the full filament voltage after 1 minute , thus in the hope of  increasing tube’s lifespan! The filament psu from the diagram was used in the past, being replaced now. The montage was hardwired in a ‘controlled chaos’ manner, star grounding technique being used, the signal and psu  stages being hosted by a two piece extruded aluminum box decorated with a front wooden plate.

Vibration wise, the tube sockets were mounted on a sandwich plate consisting in three layers: 1mm thick black Bakelite, HQ double sided scotch tape, and 0.3mm copper foil. Also, the front plate of the box was doubled with lead, the lower section was damped with rubber, and the upper aluminum part was covered with 0.2 mm tape inside. The box is resting on three aluminum spikes, and serves well from a wooden stand full of sand, with a floating stone over. There were no issues with vibrations or acoustic feedback when playing really loud music. Regarding the electronic generated noise, a good advise is to use high quality grid-stopper resistors, and for the R1 the best choice is VISHAY PRECISION GROUP – Y078550K0000T9L – RESISTOR, 0.01% 50K (from mouser or farnell), which costs a fortune. This is not a joke for the hardcore DIY–er .Ferrite rings are implemented on the input RCA unbalanced sockets, as well as on B+ rails for the input stage. I am using just four thin twisted wires from the MC cartridge to the output sockets of my record player (having no ground screen around, cause I am using a linear tonearm), and I just wanted to be sure I will not start chasing radio ham guys overnight, instead of listening to the music. The input impedance matching transformers I am using are 1:9+1:12.5 ratio professional studio condenser microphone transformers, which perform at least as the Sowter ones (highly recommended for their craftsmanship, measuring data and their sonic touch). As Bob  from Bob’s devices stated, their best is possible only without loading them lower than 47k. The sound of  Punishment phono: well, I still have to work on bringing the psu noise floor down a bit, but the phono stage sounds quite good: with a hudge soundstage, deep controlled bass, extended silky highs and natural present voices and medium frequencies. I have used for about three years a Van Den Hull DDT II Special: scary deep bass and hypnotic left-right soundstage, fast sounding. Now I am using a Lyra Delos: the dynamics are tremendous, as the precision of sound sources/instruments/voices L to R and  front to back is surgical. Still, the base fundament is not that solid like for DDT II, but Lyra handles large scale music with ease and, were VDH was a bit congested. Still, the timbre reproduced by DDT II was more pleasing and tasty, with more air near the fingers of the guitar player :] . With DDT you cold tell if on the stage is Zappa, Hendrix or The Edge, as with Delos you could tell which was the recording hall or the brand of the concert piano :] .

I assume you understand that if one could feel such subtleties between two different decent cartridges, then the phono stage is decent in itself. I would highly recommend  this project which,  maybe with minor changes, could mate any very good combo record player – mc cartridge from 3K up. (of coarse, with a high quality separate PSU, not like mine ).

© attitube 2011