Audio Asylum Thread Printer
Get a view of an entire thread on one page
|For Sale Ads|
I'm trying to understand the determination of an appropriate sized cap for a cap-coupled SET. Working with Paul Joppa's update on Joe Roberts' WE91 300B, the schematic shows C7=0.1uF and R10=390k. Using the formula f=1/(2Pi*R*C) am I correct that this results in a corner frequency of 4.14Hz?
I also found Al's thread about replacing R10 with a choke that has the impact of changing R10 to 250k (http://www.audioasylum.com/cgi/vt.mpl?f=magnequest&m=11930). If I incorporate this change with a pair of nice NOS oil caps that measure 0.25uF, am I correct that this shifts the corner frequency to about 2.6Hz (and keeps phase shift below 26Hz)?
Am I calculating things correctly here? Does making this change result in any negative impacts that I am not considering?
Thank you for any input!
A few comments:
When one turns on an RC-coupled amp, the coupling cap must acquire a DC voltage. Part of the charge comes through the grid resistor, and part comes by grid current once the cathode is hot enough. This is complicated by the need to charge up the cathode resistor bypass cap at the same time. Passing grid current while the filamentary cathode is warming up can strip the cathode of emissive power. So there is a tradeoff among the power supply, the filament, and the coupling RC time constants. I learned about this from Jac at EML.
When an RC-coupled power tube is over-driven by musical transients (as happens pretty often with flea-power SETs) the coupling cap absorbs the grid current and develops a DC bias shift. The charge bleeds off by the time constant of the cap and the grid-to-ground resistor. This shifting of the operating point is heard as overload recovery. Normally you want this time constant to be short relative to the human ear's ability to hear it. Almost nobody talks about this issue. I try for a time constant around 30mS, which is around 5Hz, only because it seems to have worked for many designs over the decades.
Chokes, including grid chokes, with metal cores do not have a fixed inductive or resistive impedance. The actual values are functions of voltage and frequency. Among other effects, at low frequencies there is a significant resistive component to the choke's impedance. This fortunately prevents the second-order bass resonance from having a large peak. It also makes analysis of the above effects complicated and difficult.
Paul - thank you for the details! Funny, I was just reading about time constants in an old Army tube handbook yesterday. It's nice to have a concrete example to help illustrate the concepts! It is staggering how much there is to learn every time I pick up a new interest...
It's amazing how many people don't know that they don't know what they don't know.
We have all been there, though sometimes we don't know we've been there.
Here's wikipedia on the darker side of this observation:
Paul - that's just too funny! I make all of my students read the actual Dunning-Kruger article at the beginning of the semester. It sets a good tone for our discussions that follow!
Wow - thanks to everyone for their input! This is a bit more of a complex issue that I had imagined and it will take me a bit of time to digest all of this.
I seem to be having some trouble posing a reply for some reason...
> > > Why does replacing R10 with a choke have the impact of changing R10 to 250k? Damping resistor? I thought a significant benefit of using grid chokes was a dramatic reduction in the grid-to-ground resistance. < < <
Al reported good results replacing R10 with a MagneQuest BCP-16 GC, so I just assumed its impedance was somewhere close to 250k. Paul's original advice was to change R10 from 390k (from original Sound Practices article) to 250k to allow for greater grid leakage than the original WE91.
Given the design of R10= ~250k and that a substitution of R10 with a choke seemed not to cause problems, I was wondering how much latitude exists with the choice of C7 and if it could be increased from 0.1uF to 0.25uF without introducing problems. Or should I stick with 0.1uF for this part?
Please be patient with me, I'm new to tubes and trying to learn.
The BCP-16 GC is listed at greater than 1000Hy of inductance.
1000Hy has a reactance at 1kHz of 6.3 meg ohms.
Have Fun and Enjoy the Music
"Still Working the Problem"
Just to tag onto the comment from Tre a bit, when you put a choke in you now have two things to consider - the DCR (DC resistance, which is very small - not sure, but probably a few hundred ohms tops) and the reactance - think of it as AC resistance, which is very large as Tre points out. So a choke makes a nice low DC presence on the grid of the 300b, but a very high AC presence, also a good thing.
A little more,
the "AC impedance" (known as reactance) of the choke falls with frequency. A 1000Hy choke has a reactance of 6.3 meg ohms at 1000Hz but only 125k ohms at 20Hz.
The other issue is this, when you have a cap feeding a choke you have a 'tank circuit' and that LC tank circuit will have a low frequency resonance.
Depending on the DC resistance of the choke, to keep the resonance frequency below the audio band (and low in amplitude) it takes a rather large coupling cap (3uf to 10uf just as a ballpark number).
The BCP-16 GC has 3850 ohms of resistance and that's not enough to allow a "normal" sized coupling cap (.1uf to 1uf just as a ballpark number).
When the coupling cap is too small you can get an audible increase in the level of the low frequencies. Some use this to "tune" a system.
Have Fun and Enjoy the Music
"Still Working the Problem"
Quote: When the coupling cap is too small you can get an audible increase in the level of the low frequencies. Some use this to "tune" a system.
I am using grid chokes.
I want to make a comment on this because I did tune it to have very low -3db point which looks real good on testing, it but in reality there is NO useful POWER in the added resonance! once I added a larger 5uf coupling, much more useful power in the lower ranges.
BTW there is nothing wrong with larger coupling capacitors as long as you have good ones :)
> > > also found Al's thread about replacing R10 with a choke that has the impact of changing R10 to 250k (http://www.audioasylum.com/cgi/vt.mpl?f=magnequest&m=11930). If I incorporate this change with a pair of nice NOS oil caps that measure 0.25uF, am I correct that this shifts the corner frequency to about 2.6Hz (and keeps phase shift below 26Hz)? < < <
Why does replacing R10 with a choke have the impact of changing R10 to 250k? Damping resistor? I thought a significant benefit of using grid chokes was a dramatic reduction in the grid-to-ground resistance.
Remember also that other things being held constant, adding a grid choke will induce a resonant peak in the bass; increasing the coupling C can help manage the peak.
It has been a while since I looked at grid chokes - I could be way off.
Also consider, it is debatable whether the benefit of reducing phase shift below ~100 Hz justifies the trade-offs, for example: passing sub-bass frequencies.
"Confusion of goals and perfection of means seems to characterise our age." Albert Einstein
Post a Followup:
Post a Message!
This post is made possible by the generous support of people like you and our sponsors: