DC filter
Hi,
I use two diodes and a 4700µF 16 V cap... works fine. There is no need for a high voltage cap since the maximum DC they can "see" is about 1 V.
\Jens
Hi,
I use two diodes and a 4700µF 16 V cap... works fine. There is no need for a high voltage cap since the maximum DC they can "see" is about 1 V.
\Jens
Re: DC filter
I imagine that higher voltage caps are a little bit more heavy duty but in this case I have made room for just about about any cap.JensRasmussen said:
Phew... finally quiet 1KVA toroids 🙂 used two 2200uF caps, two diodes, one varistor and a 0.33uf cap. Combined with a slow-start circuit ( http://mitglied.lycos.de/Promitheus/delay_circuit_for_toroids.htm ) it made my amps ready for case placement (ok, after I change only a couple of things I have planned 🙂
line filter effect?
It could be just my ears and my brain that's playing tricks on me, but I also noticed a tonality change, also present after I switched to 4x4700uf caps instead of 2x2200: the bass is more thin than before and the mids and highs are somehow clearer (better decay & less 'background' noise), but everything's not as punchier than before. Did some more reading around here, and that seems to be the general effect of power line filters/conditioners (higher impedance on trafo primaries?
).
It could be just my ears and my brain that's playing tricks on me, but I also noticed a tonality change, also present after I switched to 4x4700uf caps instead of 2x2200: the bass is more thin than before and the mids and highs are somehow clearer (better decay & less 'background' noise), but everything's not as punchier than before. Did some more reading around here, and that seems to be the general effect of power line filters/conditioners (higher impedance on trafo primaries?
).
maybe the missing bass and improved mid resolution is due acoustically to the transformer hum/buzzing being gone ?😉
mlloyd1
mlloyd1
Ehm, no. I recently applied a dc filter to my GC which never had any hum at all. The sound definitely changed and it took me a long time and many switches back and forth to finally make up my mind that the change is after all for the better.
Of course there is no perceptible change in the impedance on the primaries, at least not at 50 Hz.
mlloyd1 said:
200W Slone OPTIMOS amps + large floorstanders; bass is flat down to the low 30Hz region (-1dB in specs - not in room 🙂 ... so I don't think so 🙂
The toroids did not buzz at all times, but when a neighbour decides to plug in god-knows-what-crappy-device-that-makes-me-nervous-because-there-is-DC-offset-on-the-mains-and-things-buzz.
However, there's definetely a change in the overall sonic character
Found out the culprit: it was a 330nF MKT mains rated caps connected across the primary for 'HF' filtering... grr...
I absolutely hate metallized polypropylene or MKP or MKT caps ... I seem to get gutless, reduced dynamic, congested, constipated sound - even cymbals sound thin with those wherever I happen to (use) them: mains filter
, DAC PS bypass, amp PS bypass, speakers... they're totally 
for music.
Had far much more success with mylar caps for bypass/decoupling sonic-wise.
I absolutely hate metallized polypropylene or MKP or MKT caps ... I seem to get gutless, reduced dynamic, congested, constipated sound - even cymbals sound thin with those wherever I happen to (use) them: mains filter
, DAC PS bypass, amp PS bypass, speakers... they're totally for music.Had far much more success with mylar caps for bypass/decoupling sonic-wise.
Eva,
You said, two 1000uF and some diodes, which diodes rating are you using and in what configuration?
You said, two 1000uF and some diodes, which diodes rating are you using and in what configuration?
How much DC on the line can cause a transformer to hum?
Mine hums, but I only measure -53mV with my DMM (yes, always -53mV even if I swap the wires)
Mine hums, but I only measure -53mV with my DMM (yes, always -53mV even if I swap the wires)
Raka :
I tested two 1000uF 16V capacitors in series back to back, with two strings of four 1N4007 diodes in paralell, one string on each direction, to allow up to about +-2.5V DC blocking
Bricolo :
DC levels on mains cannot be measured by directly connecting the multimeter to the line. You have to place a low-pass filter between the line and the DMM. I use a first order filter made of a 100K 1W resistor and a 22uF non-polar capacitor for that purpose
Non-gapped transformers must be operated on 'pure AC'. Even 50mV of DC may be enough to see signs of saturation on toroids
I tested two 1000uF 16V capacitors in series back to back, with two strings of four 1N4007 diodes in paralell, one string on each direction, to allow up to about +-2.5V DC blocking
Bricolo :
DC levels on mains cannot be measured by directly connecting the multimeter to the line. You have to place a low-pass filter between the line and the DMM. I use a first order filter made of a 100K 1W resistor and a 22uF non-polar capacitor for that purpose
Non-gapped transformers must be operated on 'pure AC'. Even 50mV of DC may be enough to see signs of saturation on toroids
Eva,
thanks for your reply. I'm going to try this today. What would be the current capability of the caps? Will standard caps be enough, or should I use parallel+antiparallel?
thanks for your reply. I'm going to try this today. What would be the current capability of the caps? Will standard caps be enough, or should I use parallel+antiparallel?
I used Panasonic 1000uF 16V low impedance capacitors rated at 1,44 Arms ripple current, but beware that standard 1000uF 16V ones are rated at much smaller values like 100-300mA so you'll need to get high ripple current capacitors, to use several in paralell or to use higher capacitance values rated at higher ripple currents in order not to exceed their specs
About capacitive impedance issues, when the output of the transformer is rectified, most current consumption happens at 300Hz [6*F] and also at 100Hz and 500Hz [2*F and 10F], so you should calculate the voltage swing amplitude across the capacitor taking into account the peak current [roughly 3 times the average current] and 300Hz
Higher current output from the transformer requires higher capacitance [lower capacitor impedande] to reduce voltage swing across the capacitor since this voltage swing may get high enough to cause the diodes to clip and may ruin the DC blocking capability of the circuit
I calculated 1.000uF for 200W maximum output @230V input [I tested a 750VA toroid but I had nothing at hand to dissipate more than 200W]. For higher power output it would be also a good idea to use diodes rated at a higher current than 1N4007 since they will have to pass all the load current in case of capacitor failure or excessive voltage swing
About capacitive impedance issues, when the output of the transformer is rectified, most current consumption happens at 300Hz [6*F] and also at 100Hz and 500Hz [2*F and 10F], so you should calculate the voltage swing amplitude across the capacitor taking into account the peak current [roughly 3 times the average current] and 300Hz
Higher current output from the transformer requires higher capacitance [lower capacitor impedande] to reduce voltage swing across the capacitor since this voltage swing may get high enough to cause the diodes to clip and may ruin the DC blocking capability of the circuit
I calculated 1.000uF for 200W maximum output @230V input [I tested a 750VA toroid but I had nothing at hand to dissipate more than 200W]. For higher power output it would be also a good idea to use diodes rated at a higher current than 1N4007 since they will have to pass all the load current in case of capacitor failure or excessive voltage swing
The lower voltage Nichicon UHE series are good for this application; besides having good ripple current ratings, the impedance (ESR) which I measure is usally quite a bit better than specs, where as I've found Panasonic FC series caps are worse than spec.
Measurements are done using an HP4192 network analyser.
You can get 560-820 uF caps with ESR in the 30 milliohm range.
Due to current rectification occuring primarily at the peak of the waveform for capacitor input rectifiers, the current stress on the caps is higher than you might expect. Capacitor input rectifier systems have very poor power factor, which puts more stress on everything in the system.
Incorporating a DC filter should be standard IMO, with toroidal transformers. Also, if one wishes to use common mode filter chokes to improve noise isolation for the power amp, then I think this DC input filter is very necessary to prevent saturation or magnetization of the common mode baluun. This point is often overlooked.
Regards,
Jon
Measurements are done using an HP4192 network analyser.
You can get 560-820 uF caps with ESR in the 30 milliohm range.
Due to current rectification occuring primarily at the peak of the waveform for capacitor input rectifiers, the current stress on the caps is higher than you might expect. Capacitor input rectifier systems have very poor power factor, which puts more stress on everything in the system.
Incorporating a DC filter should be standard IMO, with toroidal transformers. Also, if one wishes to use common mode filter chokes to improve noise isolation for the power amp, then I think this DC input filter is very necessary to prevent saturation or magnetization of the common mode baluun. This point is often overlooked.
Regards,
Jon
Eva,
Just to confirm, is this what you mean? Sorry for the quality of the picture.
If the concern is the current rating, could I use some 63V caps? I have some 1000uf FC series that goes to 2.7Ams.
I intend to use this circuit for a preamp, and a power amp that is biased at 250ma the output stage.
Just to confirm, is this what you mean? Sorry for the quality of the picture.
If the concern is the current rating, could I use some 63V caps? I have some 1000uf FC series that goes to 2.7Ams.
I intend to use this circuit for a preamp, and a power amp that is biased at 250ma the output stage.
Alternative layout of DC trap
Have successfully built this, removed 99-100% of all my DC noise from my amps.
Am no electronics expert, but is this similar to earlier schematics ?
Norway, I use 230V mains.
If needing more DC removal than in this schematic, can I just add more diodes in serial to the existing 2 ? Do I also need more capacitors or bigger ones ? Can someone draw me a updated schema with more diodes so I will be 100% sure of what to do ?
Have successfully built this, removed 99-100% of all my DC noise from my amps.
Am no electronics expert, but is this similar to earlier schematics ?
Norway, I use 230V mains.
An externally hosted image should be here but it was not working when we last tested it.
If needing more DC removal than in this schematic, can I just add more diodes in serial to the existing 2 ? Do I also need more capacitors or bigger ones ? Can someone draw me a updated schema with more diodes so I will be 100% sure of what to do ?
Forgot to post the values used in the schema above:
Axiell ellyt 2200uF/25V
and
LT10A04 diode 10A 400V
Axiell ellyt 2200uF/25V
and
LT10A04 diode 10A 400V
Re: Alternative layout of DC trap
When you choose the capacitor and the value of it you must know how much current you expect to have AND how much current the capacitors can take continuously. You must check the cap datasheet for this.
The idea is this:
Insert a capacitor in series with the primary winding.
Make sure that the caps won't get too much reverse voltage. This can be done in two ways, connect two in series or add a tranzil or two diodes across.
If you have two caps in series you ought to have some transient and inrush current protection (I used plain diodes but tranzils are also OK)
This circuit is rather pointless, because if you are into this design you can do as Lars Clausen has done, a signle capacitor plus a 4.3 V tranzil. The same result is achieved. The extra capacitor in your design won't do any good at all.LydMekk said:
When you choose the capacitor and the value of it you must know how much current you expect to have AND how much current the capacitors can take continuously. You must check the cap datasheet for this.
The idea is this:
Insert a capacitor in series with the primary winding.
Make sure that the caps won't get too much reverse voltage. This can be done in two ways, connect two in series or add a tranzil or two diodes across.
If you have two caps in series you ought to have some transient and inrush current protection (I used plain diodes but tranzils are also OK)
nasty slow-start
This is a horrible design. The relay contacts will close slowly as the voltage across the coil builds up, most likely causing sparking, fizzing and nasty noises, and rapidly burning out the relay contacts.
This should be much better: http://sound.westhost.com/project39.htm
lucpes said:
This is a horrible design. The relay contacts will close slowly as the voltage across the coil builds up, most likely causing sparking, fizzing and nasty noises, and rapidly burning out the relay contacts.
This should be much better: http://sound.westhost.com/project39.htm
