UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Vintage Tape (Audio), Cassette, Wire and Magnetic Disc Recorders and Players (https://www.vintage-radio.net/forum/forumdisplay.php?f=35)
-   -   Cathode follower line out - best practice (https://www.vintage-radio.net/forum/showthread.php?t=159651)

unitelex 10th Sep 2019 12:36 am

Cathode follower line out - best practice
 
Working on a Ferrograph 631 which had a line output fitted not so long ago.
The installer did a neat job and added an ECC82 configured as a cathode follower. There is a grid to ground resistor (470K) and a cathode resistor of 2K2 with a DC blocking capacitor on the output.

Anode current is about 4mA taken from C20. Such anode current is pulling down the HT voltage for the rest of the equipment by some 50V causing other problems, so not ideal.

To reduce anode current I was considering increasing the cathode resistor, but mindful of the good practice of a low output impedance to drive line levels.

Apparently this additional line out is used to drive the line in for some modern transistorized/digital equipment with a short cable so not convinced it needs a very low impedance but I don't have the equipment to try it. So I am aiming for a reasonable compromise.

What is a reasonable line output impedance?

Considered a few options:
1. A series chain in the cathode with overall higher resistance, taking the output from the lower resistor to keep impedance down at the sacrifice of some attenuation. With the preceding stage driven harder the signal level can be recovered back to what it was but this simple cathode follower is not biassed to support a large input signal. A spice simulation showed it close to clipping.

2. I have seen better biassing arrangements with the grid resistor tied back to the cathode chain. That way it handles a bigger signal without clipping however I don't want to drive the preceding stage too hard either to minimize distortion.

What is best practice for a cathode follower to get sufficiently low output impedance without suffering high anode current?

Nuvistor 10th Sep 2019 7:11 am

Re: Cathode follower line out - best practice
 
Can you supply the HT from another source and have its own smoothing RC. Say from the screen grid of V4.
I have never worked on Ferrograph machines so apologies if not practical.

Andrew2 10th Sep 2019 8:39 am

Re: Cathode follower line out - best practice
 
I don't have the cct in front of me, but it sounds like the HT supply for the new stage is being taken from a fairly high resistance part of the supply, maybe the part that supplies the low-level stages?
I'd suggest going directly to the main smoothing cap and installing a dedicated R/C decoupled branch for the cathode follower.

Oops, Frank beat me to it!

rontech 10th Sep 2019 8:49 am

Re: Cathode follower line out - best practice
 
I would imagine the power supply in a Ferrograph recorder to be conservatively rated and would supply the extra 4 mA

GrimJosef 10th Sep 2019 8:53 am

Re: Cathode follower line out - best practice
 
Quote:

Originally Posted by unitelex (Post 1175033)
... I was considering increasing the cathode resistor, but mindful of the good practice of a low output impedance to drive line levels ...

The output impedance of a cathode follower stage is not usually dominated by the value of the cathode resistor. The action of the valve is that it can source or sink a good deal of current out of or into its cathode for a relatively small change in the cathode voltage. So the valve itself behaves as if it has a cathode 'impedance' which is in parallel with the cathode resistor, and that impedance tends to be low and therefore, in a parallel circuit, to dominate.

The formula for the cathode impedance is a little complicated, but in the approximation that it does dominate and that the external resistance in the anode circuit (including the HT rail's effective impedance) is small, it is approximately 1/gm, where gm is the transconductance of the valve. I don't know what operating point your ECC82 is running at, but the valve's typical gm is 2-3mA/V in which case your circuit's output impedance would be 500-330ohms in parallel with the cathode resistor. Changing the cathode resistor will affect this, but the major change might come from its effect on gm rather than its parallel contribution.

EDIT: You might think this gets you off the hook and that you're free to lower the standing current in the valve to make life easier for the HT supply. But as others have said it sounds like your HT rail is not very 'stiff' to DC load. If it's also not very stiff to AC load (because there isn't a substantial smoothing capacitor) then your cathode follower may induce an AC signal on the HT rail as it tries to drive significant current into the output cable. This would be A Bad Thing. So beefing up the HT supply as far as you can would still be a good idea.

Cheers,

GJ

Nuvistor 10th Sep 2019 8:58 am

Re: Cathode follower line out - best practice
 
Hi Andy,
What you say appears to be the cause, R26 is the feed resistor to C20, R26 is 15K. G2 of V4 is connected to the main HT line straight off the main smoothing cap C33.

It just seemed the simpler way to sort it out rather than change the cathode follower that appears to work ok.

kalee20 10th Sep 2019 9:13 am

Re: Cathode follower line out - best practice
 
4mA sounds a lot of current.

What signal voltages are involved? And what cable impedances (I presume mainly capacitive)? Knowing these it is possible to calculate signal currents, and then decide on DC current. Following that, it's just a matter of biasing the valve.

barrymagrec 10th Sep 2019 9:19 am

Re: Cathode follower line out - best practice
 
I`d be inclined to fit another resistor (10k or so) from the hot end of R26 and decouple with another 16 uf, feed the extra valve from this.

unitelex 10th Sep 2019 9:55 pm

Re: Cathode follower line out - best practice
 
Thanks for all the great responses. I am going to try the improved HT supply... The owner of the machine has confirmed the external load to be a Tascam unit which I believe has a 10K input impedance

GrimJosef 10th Sep 2019 10:11 pm

Re: Cathode follower line out - best practice
 
There's a rule of thumb which says the output impedance of the source should be less than 10% of the load impedance. So a few hundred ohms from an ECC82 should do nicely if you can sort the HT out.

If you have 2nF of cable capacitance (that's a lot !) then its reactance at 20kHz (the upper limit of anyone's hearing) would still be 16kohm. So you wouldn't have any trouble driving that either.

Cheers,

GJ

TIMTAPE 11th Sep 2019 3:49 am

Re: Cathode follower line out - best practice
 
I can understand the wish to preserve the added cathode follower cct but if it was me, for digitising of old recordings I'd just use the existing external speaker output, with the internal speaker, or a suitable resistor, connected as a load of course. The Ferro power amp is of good quality and unless pushed into overload is unlikely to degrade the signal off tape.

Tim

kalee20 11th Sep 2019 1:03 pm

Re: Cathode follower line out - best practice
 
I'd be tempted to use a much greater cathode load, nearer 10kΩ and take the grid resistor not to 0V, but to a potential divider from HT+. You should be able to get away with much less than 4mA. You'll have to feed the signal in via a capacitor of course, but that is not a problem.

Ripple on the anode may cause hum, though the PSRR of a cathode follower is pretty good (you can simulate it or there is a formula of course).

Output impedance of a CF is, as GrimJosef says, approx. 1/gm. Tapping down on the cathode resistor may actually give you a source impedance to the line of greater than this, depending on the cathode resistor and where it is tapped.

Quote:

Originally Posted by GrimJosef (Post 1175270)
There's a rule of thumb which says the output impedance of the source should be less than 10% of the load impedance.

Then that would be violated within almost every IC; it would be violated in almost every single-ended pentode output stage feeding a loudspeaker; it would be frequently violated between stages in a valve AF amplifier; it would be violated in RF and wideband scenarios where a cable is both-ends-matched (50Ω cable fed from 50Ω source impedance and terminated by 50Ω load, which will absorb reflections either way). It's like most rules-of-thumb, it has areas of applicability, but shouldn't be applied slavishly. Though I do agree with you in this instance!

G8HQP Dave 11th Sep 2019 1:58 pm

Re: Cathode follower line out - best practice
 
It is a useful rule of thumb for audio, when designing a source which may be asked to drive any reasonable load.

Dropping the quiescent current in a cathode follower doesn't just raise the output impedance (by reducing gm), it also reduces the available current to drive the load which may raise distortion. However, a 10k load is not too bad. A volt of signal will require 0.1mA of signal current so the quiescent current just needs to be a few mA.

PJL 11th Sep 2019 3:46 pm

Re: Cathode follower line out - best practice
 
You can use a low value cathode resistor by lowering the anode voltage with an RC anode feed network. The cathode resistor just needs to be high enough to support the required audio signal and the anode voltage just enough to prevent the vale going into saturation.

From the ECC82 datasheet a suitable operating condition might be:
Va=85V
Vk=4V
ia=2mA
Rk=2K2

Not sure on your HT volts but if say 200V
Rfeed=(200-85)/0.002 = 56K

When I had a similar requirement I made a simple JFET buffer powered by doubling (might have been tripling) the heater supply. This way you are not adding to the LT load.

kalee20 11th Sep 2019 4:02 pm

Re: Cathode follower line out - best practice
 
You can. But there is little virtue in a low value of cathode resistor.

The lower it is, the harder the valve has to work. Consider if you got the DC conditions right to allow you to use 10Ω, then the voltage gain of the CF would be fairly low and although output impedance would be super-low (a bit less than 10Ω) the distortion would be rather higher.

Really, the cathode resistor needs to be high in comparison with 1/gm. Then the gain approaches 0.9 and distortion approaches zero.

PJL 11th Sep 2019 4:55 pm

Re: Cathode follower line out - best practice
 
Agreed, too focused on the desire to reduce the anode current! and as Dave says, reducing the anode current will reduce the drive capability.

barrymagrec 11th Sep 2019 5:17 pm

Re: Cathode follower line out - best practice
 
4 mA will be of no consequence to the Ferrograph PSU so long as it is taken from an appropriate point and not the decoupled supply to the front end EF86s.

Michael Maurice 11th Sep 2019 6:27 pm

Re: Cathode follower line out - best practice
 
What I would possibly have done is to fit a 100K resistor between the HT and the triodes anode but decouple the anode with a 10uF capacitor.

That way DC wise it would behave like a normal triode circuit but AC wise, like a common cathode.

Try also looking at the circuits for the Revox A-G36 (I don’t have it to hand) that has common cathode outputs as standard

unitelex 11th Sep 2019 8:34 pm

Re: Cathode follower line out - best practice
 
1 Attachment(s)
Took a look at the Revox G36 as suggested. Reproduced the relevant parts in the attached.


Seems they used an ECC83 CF to drive the VU meters. 220V HT, with 2.2K and 10K in the cathode load. Vk at 10V, DC block output cap 40uF. For the VU meter his did not need to be super linear.

Also another ECC83 CF on each of the ext. line outputs. 190V HT, with 1.2K and 56K in the cathode load. Vk at 45V, DC block output cap 180nF
There is some feedback to the earlier stage via 10K

Both using the self-bias method to the tapping point between the cathode resistors.


All times are GMT +1. The time now is 8:58 am.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.