A
Post Worth Keeping: VoltSecond on Grid and Anode Chokes
Nov 14, 2004 Ver -
Minor format updates 29 Jul 2024
All
choked up on Grid and Anode Chokes
By VoltSecond - November 14, 2004
NOTE: Many thanks to VoltSecond for the excellent work yet again -
ML
1. What does a grid choke actually do, compared to a simple resistor?
A
grid choke allows a much lower DC resistance to be placed on the
grid of a tube for the same or higher audio frequency impedance. The
low DC resistance allows for better bias stability with possibly an
easier load on the driver and the choke offers the potential tuning
of the low frequency performance of an amp.
Update:
When using a grid choke, one must pay attention to the parasitic
capacitances of the choke. At a minimum, there will be capacitance
between the input and output of the choke, a pair of capacitances
from the input wire of the choke to the "core" and to the air around
the winding and a third pair
of capacitances from the Output wire of the choke to the
"core" and to the air around the winding.
On a good grid choke, these capacitances won't matter. The
last two mentioned sets of capacitances can cause the choke to have
a "polarity" where installing one way sounds better than the other
way.
2. What is the difference between a common and differential choke?
The iron in a differential choke must support both the DC current and differential AC voltage across it. The differential measurement to look for is the peak volt*seconds. On a linear inductor the peak V*sec = L * Idc + L * Ipk_ac
The iron in a common mode choke (CMC) does not see the differential (dc bias) current or the differential voltage. The positive differential ampere-turns in one winding is cancelled by the "opposite polarity" ampere-turns in the other winding. The iron in a common mode choke sees the common mode (leakage) current and common mode voltage. A CMC allows the choke to have a much higher inductance for a given amount of DC "bias" current. CMCs are mostly used in noise reduction applications.
When you use a CMC, if you remove it from the circuit, you have to be able to measure an "open" with an ohm meter when you measure across where both windings would attach to ground. A CMC cannot be placed between two points that are DC connected to ground and be expected to operate correctly. Meaning you can't put a CMC on +15 and -15 if the ground before the CMC connects to the ground after the CMC.
There are two main types of CMCs: high leakage and low leakage inductance. High leakage inductance CMCs saturate much easier than low leakage inductance types. High leakage inductance CMC also tend to radiate noise to and pick up noise from adjacent components. Why use a high leakage inductance CMC? The leakage inductance can be used for differential filtering if you are careful and shield the heck out of the CMC. I find it easier to just use a low leakage inductance CMC and use separate iron for the differential inductance.
3. Is there any rules of thumb to judge a single ended transformer on by the specs the maker gives?
A
partial list:
** >=8H per kohm of reflected load measured at near rated output
power. (1kohm/(2*pi*20Hz))
** >=8H per kohm of R_plate measured at < 1/10,000 of rated
output power.
** -3 dB at rated power at less than 20 Hz (i.e. the saturation
limit). 10Hz is better, but you'll need a decent bank account to buy
the part and strong arms to lift it.
** an upper -3dB point of >35 kHz into rated output impedance
when driven by about <1/3 the reflected load impedance. Unless
the output loading and drive impedance is specified, you really
don't know what you get.,
** a primary side self capacitance < 4500 pF/ reflected impedance
(1K reflected Z = 4500 pF, 5K reflected Z = 910 pF). This is to keep
the high frequency load line under control. (Note: 4500 pF = 1 Kohm
at 35 kHz.) Do not calculate this using Self Resonant
Frequency. You have to measure it.
** Low and linear core loss.
** The primary and secondary magnet wire does not come in direct
contact unless both primary and secondary are ground referenced.
(Magnet wire has a high impulse voltage rating but a poor continuous
voltage rating.)
** The coil passes hi-pot and IR. This is for safety.
4. When doing a calculation for a 300B Single Ended, which would you use to get your inductance,
XL=
2 x pi x freq. x L where XL is 3000 ohms say,
or
XL is 700 ohms plate resistance of the 300B
or
XL
is 3000 in parallel with the 700 ohms plate resistance of the
300B? (3K in parallel with 700 is 567 ohms.)
Answer: All of them!
XL >8*3000 ohms reflected impedance at 20 Hz at max output
voltage to have a good load line and XL>700 ohms at 20 Hz with
about 0.1V on the primary for good small signal response. i.e.
items from question 3: since a smaller core normally means more
turns for given inductance, would this mean the hi end would
suffer? *** The high end doesn't always have to suffer for more
turns on a smaller core. There are many trade offs that are made
in the design of a transformer or inductor. For example:
interleaving and winding techniques can offset having more turns
to some extent, but not to extreme levels.
Remember:
1. Touting just one parameter is faulty and preys on the marketing
fad of the week.
2. Every time you improve one parameter, at least one other parameter changes to the worse. ( TANSTAAFL, There Ain't No Such Thing As A Free Lunch. )
Even when you use the inductance per 1000 turns (or nH/Turn)
calculations correctly, does that tell the whole story i.e. what
about core saturation and Low Frequency handling?
Any other 'basics' about the Low Frequency side of a transformer? It all seems a bit confusing. It is easy to get into trouble with just the basics.
For
example, somewhere out there, I showed to cleanly support a 20Hz
pulse, the core needs to support full power at 10Hz. This is
because the "first" half cycle of the pulse is starting at "zero
Gauss" and swinging to +Bsat instead of starting at -Bsat and
swinging to +Bsat like it does in steady state. This issue may not
show up often because I expect the higher power pulses will be
above 20 Hz so full power at 20 Hz is good enough and there aren't
many 20 Hz fast rise pulses out there in music.
Most transformer houses are reluctant to specify all their
parameters for many reasons. These include: having too many
customers that don't understand the trade offs, they don't want to
give away enough internal design parameters that someone could
undercut the price while not offering the same quality or paying
for their development time, customers rejecting parts because they
are a reasonable percentage off from nominal or they measure the
part incorrectly, the raw iron and insulation manufactures not
controlling their nominal parameters tight enough to guarantee a
number (even when they are the best money can buy (note 1)) and
many more reasons.
It is common to see 2:1 variations up and down (4:1 total) in some
transformer parameters . . . My pet peeve parameter this year is
open circuit self resonant frequency (SRF.) On high performance
parts, the design with the lower (SRF) can often be the better
part! With the same coil, if you trade low nickel laminations
(good stuff) for M36 (nasty stuff), the nickel will have a much
lower SRF because it has high inductance at the high frequencies
than the M36.
Play
safe and play longer!
Don't be an "OUCH!" casualty. Unplug it, discharge it and
measure it (twice) before you touch it.
Oh!. . .Remember: Modifying things voids their warranty.
Mike LaFevre liked
this post enough he put it on his Magnequest Technical Reference
page as Article #5.
https://web.archive.org/web/20071013095350/http://magnequest.com/mq_magnetics.htm
Finally,
my standard warning:
Play safe and play longer! Don't be an "OUCH!" casualty.
Unplug it, discharge it and measure it (twice) before you
touch it.
. . .Oh!. . .Remember: Modifying things voids their warranty.
(
New 2024 index page.)
_( Old 2003 index page.)
_( AMP
Second index page.)
( Fancy index
page.)