i've read various discussions in AA saying grid stoppers must be carbon comps. But i've seen response audio using riken carbon film as mod to ASL amps. i've seen other articles (mostly diy homepages) using metal films as grid stoppers.
does that mean these people are doing it wrong? I asked since carbon comps are unavailable locally. that means I have to go ebay usa or places that sell carbon comps (thlaudio, handmade and etc). and I only need 1 resistor.
thank you.
does that mean these people are doing it wrong? I asked since carbon comps are unavailable locally. that means I have to go ebay usa or places that sell carbon comps (thlaudio, handmade and etc). and I only need 1 resistor.
thank you.
Another topic where everyone has an opinion. A grid-stopper is used to isolate/damp any inductive reactance in the grid circuit that could interact with the Miller capacitance (which is itself a form of negative feedback). The inductive reactance can come from the grid wiring itself. Essentially we’re lowering the Q of the resonance circuit that results. An undamped resonance can create peaks in the ultrasonic response, or even oscillation, especially in high gm tubes. So you want to use a low inductance (and low shunt capacitance) resistor for a grid-stopper to avoid making the problem even worse. Carbon comps have their problems when used elsewhere, but here they behave like a resistor out to VHF and even UHF frequencies, which is desirable. For obvious reasons, wire-wounds should not apply for grid-stopper duty. Other resistor technologies such as metal film and carbon film will probably work most of the time, depending on how inductive they are (spiral laser cuts etc.)
Ferrite beads can be used too, since they become lossy and mimic a low-value resistor at RF frequencies, but the ferrites are non-linear, and that gives some people a concern (although the effect happens at many tens of megahertz).
Ferrite beads can be used too, since they become lossy and mimic a low-value resistor at RF frequencies, but the ferrites are non-linear, and that gives some people a concern (although the effect happens at many tens of megahertz).
A ferrite bead (of the kind we're suggesting here) becomes more resistive than inductive over a wide range of RF frequencies. People mistake this part as an inductor when it really acts more like a "short" at audio frequencies and a small resistor at RF (paradoxically). You will see them used extensively in tube test equipment, even tube testers, for this reason. The inductive component of the ferrite bead, especially with a simple straight pass-through of the wire (which makes a single turn) can be quite small.
I'm sure it's possible to find a combination of ferrite bead, tube, circuit and layout that will oscillate, just as it possible to find a combination with a resistor that will also oscillate. There are many variables, and experimentation with grid stopping is always the best proof. If you have a wide-bandwidth scope, run 10kHz (or higher) square waves through the amp while varying the amplitude. Also try sine wave sweeps out to the megahertz region, looking for peaking in amplitude response or bursts of oscillation on portions of the waveforms.
I'm sure it's possible to find a combination of ferrite bead, tube, circuit and layout that will oscillate, just as it possible to find a combination with a resistor that will also oscillate. There are many variables, and experimentation with grid stopping is always the best proof. If you have a wide-bandwidth scope, run 10kHz (or higher) square waves through the amp while varying the amplitude. Also try sine wave sweeps out to the megahertz region, looking for peaking in amplitude response or bursts of oscillation on portions of the waveforms.
The primary property of ferrite beads is resuctance. It does not lend itself well to conventional analysis. But, an assortment of the little guys is very handy for solving all sorts of high frequency and noise problems. Empirical trial and error is the best way to find the right one... the results are often incredible.
I use a ton of these things to keep little signals clean in the presence of big ones.
I use a ton of these things to keep little signals clean in the presence of big ones.
I found a short explanation from Murata that might help:
Murata ferrite bead explanation
BTW, I'm not pushing ferrite beads as audio grid stoppers, just suggesting their possible consideration along with simple resistors. They have the advantage of less audio bandwidth reduction compared to simple grid stopper resistors (especially larger ones), but the non-linearity is worrisome. I haven't spent enough time making listening comparisons to say for sure. I usually use carbon comps myself, as small as I can get away with. Often that means NO grid stopper at all. Keep lead lengths short and grounds beefy and close by. For cathode followers, Tweeker’s points are well taken. In addition, make sure to bypass the plate of a CF to ground with a small, decent quality cap to keep the plate shorted at RF frequencies.
Murata ferrite bead explanation
BTW, I'm not pushing ferrite beads as audio grid stoppers, just suggesting their possible consideration along with simple resistors. They have the advantage of less audio bandwidth reduction compared to simple grid stopper resistors (especially larger ones), but the non-linearity is worrisome. I haven't spent enough time making listening comparisons to say for sure. I usually use carbon comps myself, as small as I can get away with. Often that means NO grid stopper at all. Keep lead lengths short and grounds beefy and close by. For cathode followers, Tweeker’s points are well taken. In addition, make sure to bypass the plate of a CF to ground with a small, decent quality cap to keep the plate shorted at RF frequencies.
Trolling away on the crest of a wave
A spiral cut resistor on a ceramic core is not an ideal way to block oscillation.
Really, the reason to use a cabon comp there is the combined experience of a lot of people. I'm usually not much on that as a reason to do things, but EVERY guy I know who's built a bunch of amplifiers has found this to be true from experience.
Really, the reason to use a cabon comp there is the combined experience of a lot of people. I'm usually not much on that as a reason to do things, but EVERY guy I know who's built a bunch of amplifiers has found this to be true from experience.
Brian Beck said:
A direct link to a 1.6 meg pdf with some helpful information about resistor parasitics: http://rfdesign.com/mag/511RFDF3.pdf
Brian Beck said:
Good! Because I would say that a 100ohms across a few uH or pH would be a rather lossy inductor too, with a self resonant frequency in the low GHz range and a Q of almost nothing...
I'm not saying use spiral cut (film) resistors as grid stoppers, just that the "stock" explanations as to "why not?" are flawed.
Axial composition resistors are of course the best, according to general wisdom, though they are no so easy to obtain. A good (if not superior) alternative is a Surface Mount Resistor, which can be soldered directly on the valve pin.
----------------------------
NB I had not read the Murata article when I wrote the above paragraphs. So it seems that in the low GHz's this could cause a problem...OK
My opinion is, a washer between amplifier and a chair is more significant than any type of resistorfs. And most significant part of this washer that matters is called 'BRAIN". If this part absorbs significant amount of superstitions, special instructions about what kind of resistors to use are strongly recommended.
jarthel said:
No, it certainly does not mean that they are doing anything "wrong". This is yet another bit of BS folk wisdom. Back in the "good ol' days", everyone used carbom comp resistors because that was the only choice for low wattage resistors. These days, we have the better alternative: metal film. Metal film resistors are a good deal quieter than carbon comp, and therefore best to use in low level stages especially.
As for the inductance issue, the extra few nano (or pico) henries that'll make a big difference at 400MHz make no difference at 400Hz. Save your carbon comps for RF projects, and use the metal films for audio projects. As far as stability goes, using good construction practices from the get-go will be of far greater benefit than trying to get a poorly constructed circuit to quit oscillating after the fact.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.