DISCLAIMER: Vacuum tube circuits work with dangerously high voltages. Do not attempt to build circuits presented on this site if you do not have the required experience and skills to work with such voltages. I assume no responsibility whatsoever for any damage caused by the usage of my circuits.

All rights of photos and text reserved. Usage of photos or text from my blog on other websites or for any other purpose only with prior permission. If you want to use any material from my blog please contact me by email.

Thursday, April 25, 2013

The Octal Phono Preamplifier, Part 1 : Circuit


I have often been asked if the phono section of my Octal Preamplifier can also be built as a stand alone unit without the linestage part. This is not feasable since it has a split RIAA, which means the RIAA EQ circuit is split into two parts, one after the first stage and another one after the second. This means the second RIAA part's accuracy depends on the loading resistor that follows after it. The frequency response of a stand alone phono stage should not be influenced by the load it drives. Another problem would be the  high output impedance of that phono section.

After I introduced a stripped down version of the 6CB5A amplifier. I thought it would be a good idea to also develop a matching simplified version of the Octal preamp. This time the concept will be such that the phono section can either be integrated into a full function preamp or used as a stand alone unit.

In this post I will introduce the circuit of the phono stage:

As you can see in the schematic, this version uses different tubes as the original Octal Preamp circuit. Instead of the 6SL7 and 6N7 this phono section uses 6SF5 and 6J5 which I presented in the two recent tube of the month articles. Let me explain this tube choice. Since this phono circuit should be usable as stand alone unit, it needs a bit lower output impedance. But I wanted to keep it as a two stage circuit. So a output tube is needed which still provides some gain at a reasonably low output impedance. I chose the 6J5 as output tube since it is extremely linear and availability of the tube is good. The 6J5 has a lower amplification factor as the 6N7, hence the need for a different input tube which has a higher mu as the 6SL7 to compensate for that. I chose the 6SF5 again for it's superb linearity. I also wanted to keep the Octal theme. Another benefit of these two tubes is their availability as metal tubes which have shielding built in. Being a stripped down version this circuit is supposed to be used with a power supply in the same chassis, so any shielding is of advantage to keep hum low.

Let's go through the circuit in detail. Both tubes are used in plain vanilla RC coupled configuration. The 6SF5 has a 250k plate resistor. With a 2.7k cathode resistor the tube biases such that the plate voltage divides about equally across the plate resistor and the tube. The same is reached with 22k at the plate and 1k at the cathode of the 6J5. The 6SF5 idles at around 0.6mA and the draws around 5mA.

Both cathode resistors are bypassed with capacitors to maximise gain and to avoid an impact on the output impedance which would rise with an unbypassed cathode resistor under the 6J5. The output impedance of this circuit is not terribly low. It falls between 5 and 6 kOhms. This is a compromise which was taken to avoid a third stage and to keep the circuit simple. This output impedance is ok to drive short cable lengths and input impedances of 47kOhms or higher which is typical for tube line stages. In case of integration into a full function preamp it would typically drive a volume pot, for which 100k would be a suitable value. If used as stand alone unit, this phonostage should be complemented by a proper active line stage. It is not meant to drive a passive linestage which is usually followed by long cable runs to the power amp. It might be ok to drive an integrated amp if it does not have too low input impedance and will not be placed too far away from the phono. But gain might be a bit low for that, depending on the sensitivity of the amp.

Some explanation of the remaining circuitry: In the schematic the 6SF5 is shown with a 47k grid to ground resistor, which is standard for MM inputs. It could be raised to 100k if needed and of course could be lowered if the cartridge needs a lower load. Loading capacitors can also be added in parallel if needed for the cartridge. Of course a MC step up transformer can be put in front of the 6SF5 to use it as a MC phono stage.

Next let's talk about the RIAA EQ, a vital part of any phono stage. As mentioned above a composite RIAA is chosen. This is formed by the 200k series resistor which connects the 6SF5 plate to two RC networks to ground. The first one, 10nF and 32k Ohm is responsible for the low frequency corner at 50Hz and the turn over frequency at 500Hz (3180 and 318 uS time constants). The 3.3nF does the high frequency roll of above 2122 Hz (75uS). The 1k Ohm in series to the 3.3nF stops the high frequency attenuation at some point to provide an approximation of the 4th time constant (Neumann). Usually this point is set at 50kHz but it is not well defined in any standard.

The 47nF blocks the DC of the first stage from the grid of the 6J5 which is connected to ground through a 1M resistor. Since the RIAA circuit is ahead of the blocking capacitor, the RIAA caps see high DC voltage across them. Caps with suitable voltage rating need to be used there.

And finally the 680nF couple the signal from the 6J5 plate to the output. The 1M resistor across the output avoids any static charge build up in case the phono stage is not connected to a load.

The 47nF and 680nF capacitors determine the low frequency roll off. With these values the low frequency roll off will start at around 10Hz. The values of the caps can be lowered if some earlier roll off is desired to avoid low frequency rumbling. The high frequency response is extremely good. I measured the -3dB point way beyond 100kHz with a signal feeding the phono through an inverse RIAA network.

The power supply can be very simple. About 300-350V well filtered DC are needed. The first 4.7k forms a separate decoupling network for each channel together with the 220uF cap. The input stage is further decoupled by the 10k and 47uF network. The 100k resistors in parallel to the B+ caps ensure that the caps are drained after turn off, even with the tubes unplugged. In addition they are a crude form of passive shunt which provides a little stabilisation for the high voltage.

The actual power supply will be covered in another article.

That's about it. Nothing fancy. A straight forward circuit. In case you are wondering why there are no grid stopper resistors: They are not needed here. These tubes are fairly low in transconductance and are not probe to self oscillation. Of course you can put some in if you feel the need for them.

I will show the initial prototype build in the next part and also some measurements.

Best regards



  1. Hello Thomas,
    I've been thinking about building this circuit for a while now, it certainly seems to be among the more intelligent phono pre-amp designs floating around out there. I have a question though; approx. what level of gain does this little phono preamp put out? I'm guessing that real world expectations would be somewhere between 36 and 38 db, no?
    Again, this is a wonderful design and I can't wait to try it out.
    B. L. Moyer

  2. Hi!

    Yes you are right with the gain estimation. It is meant to be used with an active line stage.


  3. Have you used the 'SF5 as a straight gain stage (non-equalized)? Any opinions regarding its susceptibility to microphonics, or its tonal characteristics? Have you reached any conclusions regarding the merits of the GT vs metal versions?

    1. Hi! I only used it in the circuit above. It works well there and microphonics are not an issue. I don't practice tube rolling, so I don't have an opinion on metal vs glass tubes


  4. Hi,
    I can expect it should roll off at 20khz by approximately -2db due to Miller effect...
    How it was in reality?


    1. the 3.3nF and 1k resistor in the EQ network can be trimmed such to counteract roll off.