The outerfoil should be connecxted to the part of the circuit that has the lowest Z to ground, not necessarily the lowest voltage.
Easy test:
If there is doubt about which cap lead is outer foil it is easy to find out with an oscilloscope set on high gain. Connect the cap across the input terminals of the scope vertical axis. When the noise grows and 60 Hz hum appears when the fingers are touched to the outside of the cap it is clear that the outer foil connection is hooked to the hot lead of the scope input. If the outer foil is on the ground connection the noise on the screen will not increase. If you do not have an o-scope you could use the mag phono input of your stereo system preamp and listen for hum. Keep that volume control down!
Easy test:
If there is doubt about which cap lead is outer foil it is easy to find out with an oscilloscope set on high gain. Connect the cap across the input terminals of the scope vertical axis. When the noise grows and 60 Hz hum appears when the fingers are touched to the outside of the cap it is clear that the outer foil connection is hooked to the hot lead of the scope input. If the outer foil is on the ground connection the noise on the screen will not increase. If you do not have an o-scope you could use the mag phono input of your stereo system preamp and listen for hum. Keep that volume control down!
Nice one!
This brings memories of my early days in primary school when I used 100-400Hz adjustable signal injector to discover how the amplifiers worked. Not long after that I built my first amplifier with only 4 gain stages in quasi-complementary configuration, and got hooked for life!!!
Extreme_Boky
tested the caps.. and some others
the wimas are a 50/50 mixed bag.
all of the caps I have with the outer foil marked hold true.
... and you know those russian teflons, the little cap symbol that people say tells you which end is which?.. nope, these are a mixed bag as well.
.. thats if it makes any difference of course.
the wimas are a 50/50 mixed bag.
all of the caps I have with the outer foil marked hold true.
... and you know those russian teflons, the little cap symbol that people say tells you which end is which?.. nope, these are a mixed bag as well.
.. thats if it makes any difference of course.
rcavictim said:
Hi
Since it is a coupling cap, both ends show similar impedance (hence similar voltage swing), except for lower frequencies. I tend to agree with above, JP, there is no "lowest potential", correct ?.
In the case of a swinging choke input filter things are different, the input shoul then always be the start of the winding.
best
Earlier I said:
Here is a typical example. Plate coupled to grid of following stage. The plate will have a load R in the neighborhood of usually much less than 150K ohms tied to a B+ rail that has an impedance close to zero. For a small signal stage the grid side of the coupling cap usually has a grid leak/drain R in the neighborhood of 470K- 1 meg ohm to ground (also an impedance close to zero). The plate side of the cap represents the lower Z and as I stated, that is the end that should go to the outer foil of the coupling cap. Any extraneous noise and hum that might become capacitively coupled to the outer foil will develop less noise voltage across a lower impedance on its way to ground.
The operation of this noise reduction technique has absolutely nothing to do with potential, static or dynamic.
rcavictim said:
Here is a typical example. Plate coupled to grid of following stage. The plate will have a load R in the neighborhood of usually much less than 150K ohms tied to a B+ rail that has an impedance close to zero. For a small signal stage the grid side of the coupling cap usually has a grid leak/drain R in the neighborhood of 470K- 1 meg ohm to ground (also an impedance close to zero). The plate side of the cap represents the lower Z and as I stated, that is the end that should go to the outer foil of the coupling cap. Any extraneous noise and hum that might become capacitively coupled to the outer foil will develop less noise voltage across a lower impedance on its way to ground.
The operation of this noise reduction technique has absolutely nothing to do with potential, static or dynamic.
rcavictim said:
Hi
A coupling cap of 1uF has an impedance of say 160 ohm at 1 kHz. This means that the impedances at both capacitor ends are more or less equal, and the Norton equivalent is the parallel value of anode load, plate impedance and grid leak resistor.
Only for much lower frequencies the coupling impedance rises, and the approximation is slightly less correct, though at 50Hz the impedance is 3k, still small compared to 150k or 1M, so one could still speak of a single Norton equivalent value (error <2%)
What helps in practice is to ground the capacitor case, in the situation that both ends are isolated.....
best
Guido,
I cannot argue with your Norton analysis, heck it doesn't even drip oil on the garage floor as nortons are apt to do, but for the sake of Good Design Practice I think my idea is better than a total disregard for capacitor polarity.
As for grounding a shield if one is provided, I think this also deserves consideration based on circuit parameters and layout. Internal capacitive coupling to the shield from the outer foil is inescapable, so leaving the shield to float may put less shunt capacitance to ground on the circuit than by grounding the capacitor. If the capacitor resides in an EMI quiet location to begin with, grounding the shield (case) might introduce unwanted HF rolloff issues.
Although the method I suggested as a quick way to determine which end of the cap was the outer foil connection works well, by modifying it slightly through the introduction of a series R in the lead going to the o-scope input ground, one ought to see the effect of your Norton analysis in operation directly.
As a final thought, you yourself pointed out that the Z of the capacitor will rise at lower frequencies. In audio amplifiers operating on 50 and 60 Hz mains power supplies, the mains fundamental is indeed often the primary source of unwanted noise. Tying the outer foil end to the circuit point of lowest Z will make an improvement measurable by instruments, if not audible, so I think it is still good design practise to do this. Since the phase of the mains noise within the chassis is generally the same everywhere, multiple receptors (stray coupling capacitance to circuit components) in the amp will become noise additive. My suggestion costs nothing to implement.
I cannot argue with your Norton analysis, heck it doesn't even drip oil on the garage floor as nortons are apt to do, but for the sake of Good Design Practice I think my idea is better than a total disregard for capacitor polarity.
As for grounding a shield if one is provided, I think this also deserves consideration based on circuit parameters and layout. Internal capacitive coupling to the shield from the outer foil is inescapable, so leaving the shield to float may put less shunt capacitance to ground on the circuit than by grounding the capacitor. If the capacitor resides in an EMI quiet location to begin with, grounding the shield (case) might introduce unwanted HF rolloff issues.
Although the method I suggested as a quick way to determine which end of the cap was the outer foil connection works well, by modifying it slightly through the introduction of a series R in the lead going to the o-scope input ground, one ought to see the effect of your Norton analysis in operation directly.
As a final thought, you yourself pointed out that the Z of the capacitor will rise at lower frequencies. In audio amplifiers operating on 50 and 60 Hz mains power supplies, the mains fundamental is indeed often the primary source of unwanted noise. Tying the outer foil end to the circuit point of lowest Z will make an improvement measurable by instruments, if not audible, so I think it is still good design practise to do this. Since the phase of the mains noise within the chassis is generally the same everywhere, multiple receptors (stray coupling capacitance to circuit components) in the amp will become noise additive. My suggestion costs nothing to implement.
rcavictim said:
Hi
Norton will be happy to see you agree. I agree with thy design practice thoughts, indeed for mains frequencies thre is one optimum. Note that it will only be benneficial for signals with high dV/dt since we talk high impedance circuits. The anodes of the rectifier tube (if present) are a well known source since they swing several hundreds of volts.
Indeed roll of, when screeniong the cap, may occur. One should ofcourse measure the effect of the parasystic capacitance added.
best
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.