Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Fresh_Start on May 31, 2010, 02:59:21 pm
-
Someone here recently advised adding a protection diode to a cathode follower. Rather than explain the theory in shorthand here, please look at Merlin's DC Cathode Follower page (http://www.freewebs.com/valvewizard1/dccf.htm).
Is this more like a seatbelt or a 5-way racing harness for toodling around town?
On a closely related topic, if I use a MOSFET like an IRF820 or a ZVN0545A, is there a need for a 12 volt zener between the gate and the source or is that diode already included in these MOSFETs? Sorry, but I can't remember where that circuit addition came from, but it was in the context of R.G. Keen's suggestion of using a MOSFET to replace the tube triode in a DC coupled cathode follower. (see Keen's MOSFET Follies (http://www.geofex.com/Article_Folders/mosfet_folly/mosfetfolly.htm)
The basic concept is that the grid (gate) of the cathode follower will be at B+ until the cathode of the preceding triode warms up and pulls the voltage down, while the cathode (source) of the cathode follower is at ground potential creating a BIG voltage drop across the grid (gate) and the cathode (gate).
Thanks for your thoughts. I'm polishing up a circuit that got started a long time ago and has been sitting on the shelf for too long. Will post the entire circuit in a couple of days before the final component order(s).
Cheers,
Chip
-
I hesitate to dispute Merlin. But thousands of PA amps didn't have no diode there. It seems silly because the grid-cathode junction IS a diode. No, it is not rated for forward current. But the few mA available for the few seconds does not hurt the tube. IMHO. Heater insulation is another matter, but if V1 plate resistor is much larger than V2 cathode resistor (which is kinda the point) then H-K voltage won't be large (100V in his example).
For the same parts-count, replace 1N4007 with a 0.005uFd cap and let the cathode follower self-bias. IMHO most guitar stages should not be DC-coupled. Leave that for hi-fi guys who never clip.
The 5F6-A tone driver may be an exception. The original has a specific overload, very different if AC-coupled. Many "based on" plans use different values such that there is severe overload, by intent. Find the "tone" then see if tube life is problematic. H-K breakdown is likely to get you before grid-surge troubles.
-
It seems silly because the grid-cathode junction IS a diode. No, it is not rated for forward current. But the few mA available for the few seconds does not hurt the tube. ave that for hi-fi guys who never clip.
I think you may have overlooked the reason for the diode. It is to protect the tube from grid-cathode arcing at start up, before the grid can pass any current itself.
Admittedly, arcing in a 12AX7 is rare, because it does not have a very close grid-cathode spacing. However, if someone plugs in a high gm 12AU7 or something, they certainly do arc (I have destroyed them myself this way!) since the grid-cathode spacing is less than 1/10th mm.
For the same parts-count, replace 1N4007 with a 0.005uFd cap and let the cathode follower self-bias. IMHO most guitar stages should not be DC-coupled. Leave that for hi-fi guys who never clip.
I'm curious to know why you think that? DC coupling has the great advantage of freedom from blocking distortion, and the DC-coupled cathode follower, in particular, is a powerful tonal topology.
(What are you trying to achieve with a 5nF cap?)
-
Thanks guys!
Merlin - your point about different tube types being constructed differently seems to answer my seat belt vs. 5-point harness question. How would a 5751 or a 12DW7 compare in terms of the grid-cathode spacing? Those are the only other tubes I might put in this spot for this amp.
Since I'm trying to mimic a Plexi in a single-ended topology, the DC-coupled cathode follower seems like the only choice.
To the general public: should I post my question about the MOSFET applications in the solid state section?
Cheers,
Chip
-
Merlin - your point about different tube types being constructed differently seems to answer my seat belt vs. 5-point harness question. How would a 5751 or a 12DW7 compare in terms of the grid-cathode spacing? Those are the only other tubes I might put in this spot for this amp.
What is their anode resistance? I believe that's the direct relationship to how close their g-k spacing is. The aU's are low, the aX is high and same for those other two.
-
depends on which section of the 12DW7 that you'll use as the CF. the low mu triode would need protection, the high mu triode optionally. both sections of a 5751 should behave similar to 12AX7.
-
depends on which section of the 12DW7 that you'll use as the CF. the low mu triode would need protection, the high mu triode optionally. both sections of a 5751 should behave similar to 12AX7.
I would use the low mu triode of the 12DW7 as the cathode follower. Guess that answers my air-state question :wink:
BTW the preamp topology is:
Gain
Gain -> CF
Tone stack
Guitar level effects loop with send buffer
Driver triode (NFB return)
Would the higher current output of a 12AT7 type triode (e.g. 12DW7 low mu) be a benefit? In my ignorance, I'm thinking more current would help push the signal through the cables, etc. of the effects loop. OTOH the send buffer on the effects loop probably controls that current, eh?
Chip
-
I'm thinking more current would help push the signal through the cables, etc. of the effects loop. OTOH the send buffer on the effects loop probably controls that current, eh?
Chip
You want a low impedance buffer to be used as send on your loop FX in order to waste the least amount of signal. A 12AU7 used as C.F. would show a very low impedance.
Merlin has an example in his website. BTW, I'm about to finish an amp with it between preamp out and P.I. input.
http://www.freewebs.com/valvewizard/accf.html
Hope this helps
Best Regards
Rzenc
-
Rzenc - wouldn't having the tone stack between the cathode follower and the effects loop "Send" create problems?
Chip
-
> you may have overlooked the reason for the diode
Yes. It was late, I spent the day breaking rocks in the hot sun, brain was fried.
> What are you trying to achieve with a 5nF cap?
Basic AC-coupled CF, such as:
(http://www.freewebs.com/valvewizard/CathodeFollower1.jpg)
I was in-love with the direct-coupled follower for years (especially as Cathodyne). "For years" because I was never really happy with its real-life performance. I think, in most cases, the AC-coupled form can do as well. Even when the extra LF pole is undesireable, the high input impedance means you can get the pole out of the way without an absurd cap value.
> is there a need for a 12 volt zener between the gate and the source
Just do it.
-
Rzenc - wouldn't having the tone stack between the cathode follower and the effects loop "Send" create problems?
Chip
Sorry but I did not understand what you meant by problem here. If I understood what you mean you will have:
GT input, gain, gain, C.F., tonestack, F.X. loop buffer, F.X. loop recovery, power tube grids.
As far as a F.X. loop is concerned, it should not 'color' your tone, although losses thru tonestack may be recovered at F.X. loop recovery stage which will alter and color sound effectively.
F.X. loop input is high impedance, according to the corresponding Merlin's schem and math, around 5Mohms.. which is high enough to be fed from a source such as C.F. which sources a low impedance into a heavy load as tonestack, but I believe F.X. loop input impedance is so high that it may not interfere with gain after tone stack.
I don't know how much this tonestack loads C.F. and how it will source F.X. loop send stage.
My plan is very different from yours. F.X. loop send is after tonestack recovery stage or after O.D. stage. Both ways, source impedances will be less than 10% of F.X. loop send input impedance.
I will let you know how it turns out.
Hope this helps
Best Regards
Rzenc
-
> is there a need for a 12 volt zener between the gate and the source
Discrete audio amplifier basics - Part 2: JFETs, MOSFETs and other circuit configurations (http://www.audiodesignline.com/howto/224200252;jsessionid=TYWD2NF20JCWXQE1GHRSKHWATMY32JVN?pgno=3)
By John Linsley Hood
Because it is theoretically possible for an inadvertent electrostatic charge, such as might arise with respect to the ground if a user were to wear nylon or polyester fabric clothing and well-insulated shoes, it is common practice in the case of small-signal MOSFETs for protective diodes to be formed on the chip at the time of manufacture. These could be either zener diodes or simple junction diodes connected between the gate and the source or the source and drain, as shown in Figure 9.19.
(http://i.cmpnet.com/audiodesignline/2010/discrete_audio_amplifier_basics_fig9.19t.jpg)
Figure 9.19: Diode gate protection.
In power MOSFETs, these protective devices are seldom incorporated into the chip. There are two reasons for this: (1) that the effective gate/channel area is so large that the associated capacitance is high, which would then require a relatively large inadvertently applied static charge to generate a destructive gate/channel voltage (typically >40 V), and (2) that such protective diodes could, if they were triggered into conduction, cause the MOSFET to act as a four-layer thyristor and become an effective electrical short circuit. However, there are usually no performance penalties that will be incurred by connecting some external protective zener diode in the circuit to prevent the gate/source or gate/drain voltage exceeding some safe value; this is a common feature in the output stages of audio power amplifiers using MOSFETs.
Remember this is for modern systems which have no 40V sources; usually everything is on 5V. In your tube amps there's way more than 40V available. You need the diode.
-
Thank you sir!
Chip