The Caresser- for superb sounding phono and line stages

[Paul N]

Hi Everyone

Some time ago I tried out a modification to my both my line and phono stage that Paul Hynes suggested- with superb results. I thought it was about time I shared it with you!

See the first diagram, which shows the circuit of a standard cathode connected triode stage- as used, for example, in Yeo’s 5687 preamp. Valve purists- and yes reader, I was one- tend to stick with this arrangement, though it’s far from perfect. Distortion increases with input level, and output impedance is high, unless you use a low rp valve (such as the 5687; but it’s still higher than we’d like); and susceptibility to power supply noise is high.

One obvious improvement is to replace the plate resistor with a constant current source (CCS).  Some use a plate choke (the ultra purists) or a triode based active CCS (the purists), but neither approach is ideal. (For example the choke’s impedance falls at low frequencies, and must be carefully made to operate well in this testing environment- i.e. expensive; and the active; and it raises the output impedance a little).

Paul suggested a simple circuit using a couple of FETs- one a rather exotic high voltage IXYS device- to replace the plate resistor- see the second circuit. This performs two functions- it provides the triode with a constant current, and it lowers the output impedance of the stage enormously.  In effect it “caresses” the triode into giving its best performance- hence the name, the Caresser!

As it’s Paul’s design, and as he’s now offering it as a small pcb- small enough in fact, to retrofitted to most pre-existing circuits- I don’t feel I can share any details of the circuit, except- changing one resistor on the board sets the current- Paul recommended 3mA minimum, and I’ve used up to 10mA. And too work properly, the circuit needs 50V or more across it. (Paul also offers a complete phono stage using this topology. I haven’t tried this myself, as I constructed my own using different triodes etc.) See-

http://www.audiocircle.com/index.php?topic=86440.0

The caresser boards are the smallest upright pcbs, either side of the valve- the ones with the tree resistors above one another. (The other pair of boards are voltage regs for the HT supply; I’m still using choke filtered supplies in my amps).

The circuit has an additional advantage- see the third circuit. Replacing the cathode resistor and cap with a suitable forward biased diode (one or more 1N4148, or a suitable LED) provides fixed voltage bias too! The diodes are selected so their forward voltage drop equals the required bias, at the constant current provided by the Caresser (exactly the same arrangement which Monica originally use). And their “dynamic impedance” is very low, and so there’s no need for a cathode bypass cap! (The latter are, IMHO, at best a necessary evil, and best avoided if at all possible).

The sonic effects are staggeringly beneficial, without any downsides I can hear at all!

The choice of triode I leave to you. This circuit is a great leveller- it allows the valve to operate at its best.  One thing to watch though.  For optimum performance, use a triode needing no more than 2- 3 V bias. Any more than this, and the number of diodes required for bias will increase, and so will their dynamic impedance. (Though this may not matter in practice- I haven’t tried ).  For triodes, mu (its voltage gain) rises as required bias falls, and thus suitable triodes will have a mu of 20-30. If the following power amp has high sensitivity, this will result in audible hiss. The triode I use has a mu of ~ 30 (the caresser CCS loading results in the stage having this gain too), and hiss from it is too high when used with e.g. Charlize. It’s fine with my valve power amp though (2 V sensitivity).  (Note this isn’t a problem of the caresser circuitry- the same would apply when using any high gain pre amp with any high gain power amp. It’s not a problem when the caresser is used in a phono amp- high gain is required).

Best wishes

Paul  N

[yeo]

hi paul

so this is what you are up to lately! 

this is rather timely as i haven't played with tubes for soooo long.

hi paul hynes

any application note? price?

yeo

[Paul_Hynes]

Hi Yeo,

I use these CAR1 modules in my phono stage. They can be used as a pure current source or to fabricate a hybrid Mu Follower working as the anode load for the valve. The module uses depletion mosfets in a cascode configuration. There are a variety of options regarding current and voltage ratings of the devices used for construction depending on the application. I normally use Takman Carbon Film resistors in my phono stage current sources, as I like the sonic signature of these.

I sell the modules with the Takman resistors for :

The CAR1-120/5 costs £24. This uses a SOT223 surface mount mosfet to buffer the current source section and has a maximum breakdown voltage of 240 vdc and I would recommend this as the maximum HT voltage for this module. The 120/5 refers to the operating parameter limits for a maximum dissipation in the SOT223 device that sees the HT supply, of 0.6 watt in free air. This means it can be used with 120 volts across it and with a maximum current of 5 ma. Voltage can be traded for current as long as the maximum breakdown voltage and the 0.6 watt dissipation limit are not exceeded. For example it could be set up to operate with 60 volts across the module with a current setting of 10 ma.

The CAR1-200/10 costs £30. This uses a TO220 package depletion mosfet to buffer the current source that has a maximum breakdown voltage of 500 vdc and I would recommend this as the maximum HT voltage with this module. The 200/10 refers to the operating parameter limits for a maximum dissipation in the TO220 device that sees the HT supply, of 2 watts in free air with a clip-on heatsink. This means it can be used with 200 volts across it and with a maximum current of 10 ma. Once again voltage can be traded for current as long as the maximum breakdown voltage and the 0.6 watt dissipation limit are not exceeded. For example it could be set up to operate with 100 volts across the module with a current setting of 20 ma.

Once the valve operating parameters are set up the HT voltage is shared between the valve and the current source load. It is possible to predict these relationships reasonably closely from the operating curves for the chosen valve. I like to discuss the projects with customers before selling modules to ensure the modules are applied correctly and safely. This may be a good time to set up the thread we talked about, for PH module application and discussion as a resource for your customers. I will write an application note if enough interest is shown and you want to stock the CAR1 modules. In the meantime, if you have a specific project in mind we can work out the operating parameters and set the CAR1 up for you to try. One thing you have to watch, as PaulN said, is the gain of the valve stage with the CAR1 module, which is essentially the same as the Mu value for the valve. For example the 12AX7 has a Mu of 100 and the CAR1/12AX7 combination has a gain of 100. For line preamps you need a low Mu valve.

Regards
Paul

[yeo]

hi paul

dang... what else you have under your sleeves?

as much as i like plate choke or ccs loading, one thing we still cant' avoid is the coupling cap. and this fella, some times cost more than an arm and leg.

yeo

[Paul_Hynes]

Hi Yeo,

You would be surprised at what I have under my sleeves.

Regards
Paul

[vt4c]

Paul,
  I'm very eager to know! 

[Paul_Hynes]

Here is one thing I have under my sleeve.

Not a lot of people know that the PR3 regulator topology can be used for valve HT supplies up to 450 volts with suitably rated components. A larger module is required to house the larger high voltage components but the same high performance and ultra-low noise operation is available for valves. Using two of these high voltage modules in series gives enough supply line rejection (around 150 dB up to 100 KHz) and is quiet enough to use with a valve head amplifier for moving coil cartridges, where extremely low noise is mandatory. 

Regards
Paul