the 'sound' of a CF

[themagicmanmdt]

hey everyone, don't post too much but I do when there's something to add/say.

well, i've now dealt with and explored CF's and the tonality of them, when properly set up (or even improperly!), and i can't say that i've heard them impart something pleasing to the circuit.

i've bypassed them in eico amps and dynaco's; maneuvered around them in guitar amps and effects loops; opted for an extra gain stage and a transformer instead of a CF transformerless output; monkeyed with 12AU7 and 6SN7 at several different voltages and tube types for many different uses...

...and it just looses something in the audio. i'd say, to my ears, it has the effect of a 'balloon' - the tone feels a bit bloated and not clean, lean and extended in both directions - whenever I get around them or find a way to not use them, all of a sudden the 'hologram' comes back!

further thoughts on the matter, everyone?

i suppose i'm asking for the vet's 'sum experience' and overall love/hate/apathetic with CF's in a variety of uses.

in short, i've seemingly always found a better solution, whether or not it's been simply doing without a feature or just using iron instead if it's an impedance issue.

discuss.

for the record, usually i'm building microphone preamps and hi-fi recording equipment, diving a little into the playback/power amp aspect, but mainly in the sourcing end of audio...yes, everything, no matter how much we try, will always come down to a certain amount of 'color' and preference...

i just never prefer the color of a CF, and fight other ways around the reasons why i'm thinking of using one.

[Tony]

a lot of chinese tube buffers being sold......

[ttan98]

Hi,

I am building a buffer based on using off the shelf Ebay Chinese parts, I will investigate what you are saying is true, I have a few ideas on few variations to this CF circuit, e.g.

1. CF based on tube like 6DJ8 or Chinese 6N1 with the cathode bootstrap
2. Replaced the Boot-strapped resistors of above with a FET current source
3. Replaced the CF tube with a FET(2SK170BL) and retain the current source FET
4. If 1,2,3 fail then replace them with high current OPamp.

I am sure one of them will sound fine.

[regal]

The CF performance is very load dependant. Take a typical CF 6DJ8 output on a line level stage, you will usually find that if you put a 10K resistor across the output (increasing the load) it has lower distortion. Same thing happens for an SRPP. So what you find is that for a 50k input amp feeding it with a anode-follower-cathode follower you have less distortion if you change the load to 10k. If you added the CF for better damping/drive ove the anode follower alone, changing the load just stole that current drive and defeated the purpose. My personal preference is to avoid cathode followers and SRPP as the output stage of line level amps, the flexibility isn't there, they can be useful if designed specifically as a dedicated source matched to its amp.

[ChrisA]

Originally Posted by themagicmanmdt

well, i've now dealt with and explored CF's and the tonality of them, when properly set up (or even improperly!), and i can't say that i've heard them impart something pleasing to the circuit.

That is what you SHOULD expect. Ideally they should not add any sound of their own and with gain close to one.

If you think they do something else try and rule out your test method

A CF is more then just a stage. It lowers the load on the preceding stage and drives the following one, It surely does have a large effect on the surrounding circuits, even if the CF itself sounds like a short length of wire.

I'm wondering how to devise a test to separate the effect the CF has on it's surroundings from the sound of the CF itself. And how to verify the test works.

It could be a simple as that you happen to like the sound of triode type distortion and the amps that use a CF buffer lower it.

[freddi]

sometimes an anode-follower is a better compromise - might not be "provable"

[Bigun]

I remember Broskie talking about the limitations of the performance of a CF when you consider the load you place on it. The issue as I understand is that the classic output impedance calculated from Ri / mu is only telling the story close to the operating point - as the signal grows, the ability of the CF to sink current to the load is no different than common cathodes ability to drive current to the load.

[Tony]

output impedance calculated from Ri / mu

=gm.....

[SY]

Originally Posted by Bigun

The issue as I understand is that the classic output impedance calculated from Ri / mu is only telling the story close to the operating point - as the signal grows, the ability of the CF to sink current to the load is no different than common cathodes ability to drive current to the load.

That's right- this is the reason why it's generally hopeless to use something like a 12AX7 to drive a long interconnect. The small signal source impedance might be low, but it can't source the current needed to charge and discharge the capacitance.

A properly designed cathode follower will do exactly what it's supposed to do- have a high input impedance, low output impedance, and be sonically transparent.

[hidnplayr]

themagicman, have you tried using SLCF?
Such as shown on the schematic of the FVP5A on Vacuum State - High End Hifi Equipment