Dear All,
I am curious about the possibility to use a choke as grid leak resistor or as a load for a driver, but I did not find documentation on good and bad of this approach, and how to choose the right value of inductance.
Morgan Jones does not even mention the existence of grid leak chokes.
Any input ?
Thanks,
Davide
I am curious about the possibility to use a choke as grid leak resistor or as a load for a driver, but I did not find documentation on good and bad of this approach, and how to choose the right value of inductance.
Morgan Jones does not even mention the existence of grid leak chokes.
Any input ?
Thanks,
Davide
I'm on the road, so will try to fill in the blanks later, but briefly:
1. For most tube common-cathode voltage amps, the grid is held at ground, negligible current. The exception is grid leak bias, but that's an unusual way to do things. So... no such thing as "too low" for the DCR.
2. To calculate the inductance you need, you'll want to consider the coupling cap, the source and load resistance, and desired LF cutoff. Bottom line: it will be a LOT of inductance.
3. Grid leak resistors work perfectly well. An inductor may look cool on the chassis, but it's hugely more expensive and bulky and quite a bit less ideal as a component.
That's what makes it fashionable, I suppose.
1. For most tube common-cathode voltage amps, the grid is held at ground, negligible current. The exception is grid leak bias, but that's an unusual way to do things. So... no such thing as "too low" for the DCR.
2. To calculate the inductance you need, you'll want to consider the coupling cap, the source and load resistance, and desired LF cutoff. Bottom line: it will be a LOT of inductance.
3. Grid leak resistors work perfectly well. An inductor may look cool on the chassis, but it's hugely more expensive and bulky and quite a bit less ideal as a component.
That's what makes it fashionable, I suppose.
On the Silk website there is one of the item I was referring to SACThailand
I don't manage to link the exact page, it's under the transformer section. They claim 7000 H.
D.
I don't manage to link the exact page, it's under the transformer section. They claim 7000 H.
D.
I presumed by "grid leak" he was refering to a "grid leak resistor" and "grid leak biasing".
Otherwise shouldn't it just be a "grid resistor" since the bias is generated either from a cathode resistor, or from a bias supply?
With low DCR, how can one develop the proper bias voltage if it is a grid leak bias circuit? Low DCR implies higher grid current which is counter to a Grid Leak Bias circuit.
Otherwise shouldn't it just be a "grid resistor" since the bias is generated either from a cathode resistor, or from a bias supply?
With low DCR, how can one develop the proper bias voltage if it is a grid leak bias circuit? Low DCR implies higher grid current which is counter to a Grid Leak Bias circuit.
Sadly no first hand experience but a friend of mine with thrustful ears bought some Silk grid chokes a month ago. Since then he simply cannot stop gushing about the improvements they brought to his amps. I think he tried them on the output tubes and the drivers and eventually in the phono stage. Last i know he was ordering a box of chokes.
Last edited:
I am a huge fan. And you get a doubled headrom which even a CCS cannot provide.
Of course SY would probably say they perform nearly as good as a 5c CCS
Here is my take on plate Chokes.
A plate choke with sufficient inductance looks like a CSS and lower the distortion of a triode stage. It has the additional advantage of being able to output ac at up to twice the supply voltage. Chokes are far from ideal, and potentially have issues at both low and high frequencies. I believe the rule of thumb for a plate choke is the plate choke should be greater than 5 Henry's for every 1K of Rp. A 6DJ8 would need a 15H or larger plate choke.
Grid chokes have the promise of having near infinity impedance at audio frequencies, lowering distortion of the previous stage. There is also a possibility that it reduces the time that DC blocking occurs. Real grid chokes are very non-ideal with significant capacitance, and resistance. Grid chokes are also very expensive compared to resistors.
As far as I am concerned, both are tools that may be useful in some situations.
HTH
Doug
A plate choke with sufficient inductance looks like a CSS and lower the distortion of a triode stage. It has the additional advantage of being able to output ac at up to twice the supply voltage. Chokes are far from ideal, and potentially have issues at both low and high frequencies. I believe the rule of thumb for a plate choke is the plate choke should be greater than 5 Henry's for every 1K of Rp. A 6DJ8 would need a 15H or larger plate choke.
Grid chokes have the promise of having near infinity impedance at audio frequencies, lowering distortion of the previous stage. There is also a possibility that it reduces the time that DC blocking occurs. Real grid chokes are very non-ideal with significant capacitance, and resistance. Grid chokes are also very expensive compared to resistors.
As far as I am concerned, both are tools that may be useful in some situations.
HTH
Doug
leak and load
Blocking distortion as I know it is caused by rectification of the grid-cathode "diode" when the grid is driven positive wrt the cathode. A grid choke does not prevent this.
I'm with Gimp here, could we be careful to use the phrase "grid leak" only to refer to grid leak bias, which *depends* on a tiny amount of grid current to create a negative charge on the grid. The 220K to ground or whatever should probably be called simply a grid resistor.
Similarly, the use of the term "choke loading" is a little misnomer since the choke is expected *not* to present a significant parallel load. A grid choke is not a "load" (nor is an anode choke really but we seem to have a common definition calling it that).
A choke acts as a sort of bandpass filter where the in-band loading is of very high impedance but the out-of-band loading may result in lower impedance than a properly sized resistor, with the added features of phase shift and possible core saturation.
I have used anode chokes on relatively low anode resistance tubes with excellent results but have not used a grid choke, prefering to direct couple
Michael
Blocking distortion as I know it is caused by rectification of the grid-cathode "diode" when the grid is driven positive wrt the cathode. A grid choke does not prevent this.
I'm with Gimp here, could we be careful to use the phrase "grid leak" only to refer to grid leak bias, which *depends* on a tiny amount of grid current to create a negative charge on the grid. The 220K to ground or whatever should probably be called simply a grid resistor.
Similarly, the use of the term "choke loading" is a little misnomer since the choke is expected *not* to present a significant parallel load. A grid choke is not a "load" (nor is an anode choke really but we seem to have a common definition calling it that).
A choke acts as a sort of bandpass filter where the in-band loading is of very high impedance but the out-of-band loading may result in lower impedance than a properly sized resistor, with the added features of phase shift and possible core saturation.
I have used anode chokes on relatively low anode resistance tubes with excellent results but have not used a grid choke, prefering to direct couple
Michael
A choke load is advantageous in situations where a lot of volts have to swing, e.g., the driver stage for low mu output tubes. Otherwise, it's a very large, very heavy, very expensive, non-ideal way of loading a plate to no particular advantage (other than fashion). If I had to swing 200V, it's definitely something I'd explore.
Based on experience I would say for good linearity and LF flatness you need a lot more than 5H per K of rp. (I'm almost comfortable with anything > 8 - 10X this) I'm currently working with 100H choke loaded D3A with LED bias and an estimated rp of about 3K (20mA, 12ohm dynamic impedance) and have measured FR down more than 1dB @ 40Hz and -3dB @ 20Hz, linear swing available was down to ~80% of midband at 40Hz, and somewhat less than 50% at 20Hz.
Two thoughts come to mind here, one is that the 5H rule is insufficient, and second that my chokes are not making 100H @ 20mA of standing plate current. Reducing Ip did not appreciably change the situation and I cannot currently directly measure any inductance over 20H - so I have no idea whether these expensive boutique chokes are actually meeting their specs. At some level I hope they are not. They are also quite capacitive, killing the HF response above 60kHz (-3dB) which is quite a bit worse than I expected given the low rp. (Maybe the rp is higher than I think?)
Combine the above mediocre FR performance with a grid choke having appreciable capacitance, and the tube's miller capacitance and it might take a miracle to make 20kHz.
I've ordered a set of 1:1 IT to evaluate next. These are Lundahl LL1635 and I expect they will outperform the choke/grid choke combo I was contemplating by a wide margin. I am going to use fixed bias on the D3A which should reduce rp below 2K and the Lundahl are minimum 30H at 20mA so we shall see. I'm still worried about LF extension.
Last edited:
Not below 20Hz or wherever the choke stops having sufficient reactance. A shorter discharge time of the coupling cap is likely.
I am not a proponent of grid chokes, mostly because i haven't heard one and they seem too complicated. What if substantial grid current flows and saturates the choke? What if there is an obvious bass resonance? Plate chokes seem much simpler.
Well, if you're designing so that the choke reactance causes the f3 to be higher than 20 Hz (say), you're back to time constant is a time constant- an RC coupling can have an arbitrarily high time constant as well. If you are comparing an LC with a 20 Hz f3 to an RC with a 5 Hz f3, that makes no sense.
Dave Slagle has done a ton of simulation work on CL coupling stages, much of it on AA. Of course the benefit is supposed to be the effect of the low choke DCR on blocking distortion recovery while maintaining very high AC impedance in the audio band. Spice sims however show potentially extreme low frequency peaks from the LC resonance with typical cap values, suggesting the use of much larger coupling caps to rectify, large enough to swamp the RC recovery benefits. However squared, measurements don't typically confirm what simulations predict. Where Spice says should be a big peak, the meter often says marginal bump.
Upshot: it's still a grey area that requires measurements to get right. One thing I can say is Hammond 150H chokes, for example, always add significant measurable distortion harmonics over a CCS.
Upshot: it's still a grey area that requires measurements to get right. One thing I can say is Hammond 150H chokes, for example, always add significant measurable distortion harmonics over a CCS.
There have been several available for many years that are for all intents and purposes ideal.
Audiophile amps using chokes to load triodes are fashionable in the sense that polyester leisure suits are fashionable. They've been around nearly that long. I would certainly recommend to the beginner choke loads or, even better, transformer coupling over CCS because of circuit simplicity and ease of construction with guaranteed satisfying results. Leave the complex circuits to EE geeks.
John
Essentially plate chokes make for lower distortion compared to a resistive stage. And higher power supply noise rejection. Also the full mu of the tube is reached. Unfortunately the inductance may limit bass performance and the inter-winding capacitance limit the highs. So get a good (expensive) choke for the job.
Grid chokes have the advantage of potentially allowing less rubbish from your ground leaking into your grid. But again it might create reactive problems.
I believe disadvantages such as high cost, size and weight have been mentioned. I don't consider those things to be a disadvantage if the best amp possible is your aim.
I mean why does cost and weight even come into the equation? You're not building the amp for your wife are you? Then again cost might be very important for you?
Grid chokes have the advantage of potentially allowing less rubbish from your ground leaking into your grid. But again it might create reactive problems.
I believe disadvantages such as high cost, size and weight have been mentioned. I don't consider those things to be a disadvantage if the best amp possible is your aim.
I mean why does cost and weight even come into the equation? You're not building the amp for your wife are you?
Then again cost might be very important for you?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.