What was wrong at my question about #33?
I read now only about the weaken and the royal way for rectifier design and the input requirements of swtich mode power supplies (SMPS) - have a look to the headline of this thread
Nevertheless very interesting, unfortunately the wrong subject.
If you want to find this at later times, nobody looks quite certainly about
" Variations of DC Main Filter against buzzing Toroid Transformers - what is the right? "
I read now only about the weaken and the royal way for rectifier design and the input requirements of swtich mode power supplies (SMPS) - have a look to the headline of this thread
Nevertheless very interesting, unfortunately the wrong subject.
If you want to find this at later times, nobody looks quite certainly about
" Variations of DC Main Filter against buzzing Toroid Transformers - what is the right? "
Last edited:
I work in a place where there is a big lab and a lot of SMPS type work is going on. If you look at the voltage and current waverforms from all sorts of equipment connected to the mains, you see it is very distorted with lots of harmonics (think about triac dimmers, motor speed control, power supplies getting switched on and off - hopefully using FWBR!). These are some of the reasons you get DC on the mains. DC on the mains means for a period of time the positive and negative halves of the supply are not closely balanced - ie.e don't cancel out. Result is a net DC on the supply lines. Thank fully, this normally does not last for too long.
I would aldo suspect that if there is DC on the mains that it would usually be accompaied by lots of second harmonic. So here is an interesting question. We can block the DC by the diode-capacitor means we have discussed here, but that will not do anything about the second harmonic. What will the presence of the second harmonic do to the transformer? We will have an asymmetrical waveform there, with equal area on top and bottom, but different peak voltage values and correspondingly different half-cycle times.
Might it interact with core magnetic nonlinearities to re-create some DC current flow?
What will the presence of significant second harmonic do to the rectification process on the secondary side of the power transformer?
Cheers,
Bob
Do you think that this post is offtopic?
http://www.diyaudio.com/forums/soli...id-transformers-what-right-4.html#post2105213
Essentially I'm telling that even a pair of back to back diodes and a single electrolytic capacitor are very effective to make toroids quiet, and I'm demonstrating it with waveforms captured from oscilloscope.
It was usually done that way in old live chassis television receivers so that the internal ground could be connected to neutral. It made them a little bit less dangerous to service without an isolating transformer and would also help combat capacitive hum pickup. Usually a non-isolated switching or linear regulator was used to feed the horizontal deflection circuit directly. Auxillary power supplies could be derived either from the flyback transformer in the horizontal deflection circuit, the inductor of the switching regulator, if it used one, or from a small line frequency transformer.
Modern TV:s don't use half wave rectification as they have switched over to isolated SMPS power supplies. It's hard to make a live-chassis receiver when you need AV inputs and outputs. Usually the only interconnections on live-chassis ones were a capacitor isolated RF input and sometimes a transformer isolated headphone output.
Last edited:
No, I don't think this regarded your posts - in the attachement I summarize it a bit.
I'm trying as I can replicate in my simulation program (Circuitmaker) different conditions from the mains (parasitics like unwanted dc components through rectifiers for half power mode). If I have success, then I can investigate exactly all mains DC filter topologies.
Who knows about modelling of toroidal transformers, especially the behaviour variations of different materials of the core? I want an equivalent circuit network of resistors, capacitors and inductors therefore.
Last edited:
this I have found today:
DIYCable.com : Intro » Home » Power Products »
a combination of DC filter and such for RF/RFI
DIYCable.com : Intro » Home » Power Products »
a combination of DC filter and such for RF/RFI
Example one should do the trick (there are several threads about that particular DC blocker here). No need for lots of components. Just a nice simple way to prevent DC
Hi hope you can help me I've checked out the example one
how can I know the values I should use ?
or could you tell me wich to use please?
the ac where I live is 120-127
Thanks.
Hard to say. Also by my own circuit idea about post #33 for a bit more reliability.
If I know more about the exactly variations range of the DC components in your available main wall connection.
If you have a possibility, check your mains curvature at different times (morning-evening both working and weekend days) especially at that times, where your transformer produce a high level of buzzing/hum noise.
Very interesting to see would be various pics.
In this case also this URLs are of interest:
http://sound.westhost.com/cgi-bin/s...ttp://sound.westhost.com/articles/xfmr-dc.htm
http://www.diyaudio.com/forums/solid-state/2080-dc-filter.html
Last edited:
Hi,
the capacitor passes the normal operating AC current.
These caps must be sized to drop as little voltage as is tolerable during normal working conditions.
The diodes pass overload current during short term operating conditions.
eg. start up current for the mains transformer, extreme overload current if the amplifier is put on full power test signals.
A single diode pair allows ~500mVpk to 550mVpk across the capacitor/s without significant diode current.
A double diode pair allows ~1000mVpk across the capacitor/s without passing significant diode current.
What must not happen is allowing the capacitors to develop sufficient voltage drop so that the diodes are passing during normal operating conditions.
The diodes will run hotter than necessary. The capacitors will suffer severe ripple current stress. The DC blocking action of the capacitors is defeated if the diodes pass.
Let's take a 120Vac example.
300VA transformer has a maximum operating current of 2.5Aac.
During charging of a capacitor input filter the crest:average ratio can be quite high. During normal music reproduction it is sensible to assume that the peak current (at the crest) is ~maximum AC current peak, i.e. ~3.5Apk.
Assume a 10mF pair of caps in series. The effective capacitance is 5mF, the frequency is 60Hz (USA).
The impedance of the caps is ~ 1/2/Pi/F/C ~0.53ohms.
3.5Apk will develop ~1.8Vpk across the pair of series caps.
A double diode will pass current when the peak primary current >~1.9Apk
For a 300VA transformer on 120Vac with 10mF+10mF and a 4diode bridge, the diodes will pass a little peak current when the SPL is turned up. But most domestic reproduction will probably just keep the diodes from passing.
I tend to adopt 10m+10mF for 240Vac with double diodes. I actually use 10off 3m3F 16V in series parallel for ~8mF effective
I think 110/120Vac needs at least double the capacitance for medium power draw transformers.
the capacitor passes the normal operating AC current.
These caps must be sized to drop as little voltage as is tolerable during normal working conditions.
The diodes pass overload current during short term operating conditions.
eg. start up current for the mains transformer, extreme overload current if the amplifier is put on full power test signals.
A single diode pair allows ~500mVpk to 550mVpk across the capacitor/s without significant diode current.
A double diode pair allows ~1000mVpk across the capacitor/s without passing significant diode current.
What must not happen is allowing the capacitors to develop sufficient voltage drop so that the diodes are passing during normal operating conditions.
The diodes will run hotter than necessary. The capacitors will suffer severe ripple current stress. The DC blocking action of the capacitors is defeated if the diodes pass.
Let's take a 120Vac example.
300VA transformer has a maximum operating current of 2.5Aac.
During charging of a capacitor input filter the crest:average ratio can be quite high. During normal music reproduction it is sensible to assume that the peak current (at the crest) is ~maximum AC current peak, i.e. ~3.5Apk.
Assume a 10mF pair of caps in series. The effective capacitance is 5mF, the frequency is 60Hz (USA).
The impedance of the caps is ~ 1/2/Pi/F/C ~0.53ohms.
3.5Apk will develop ~1.8Vpk across the pair of series caps.
A double diode will pass current when the peak primary current >~1.9Apk
For a 300VA transformer on 120Vac with 10mF+10mF and a 4diode bridge, the diodes will pass a little peak current when the SPL is turned up. But most domestic reproduction will probably just keep the diodes from passing.
I tend to adopt 10m+10mF for 240Vac with double diodes. I actually use 10off 3m3F 16V in series parallel for ~8mF effective
I think 110/120Vac needs at least double the capacitance for medium power draw transformers.
Last edited:
Here comes the questions...
I'm having mechanical transistor humm, none at audio output and the humm varies with time...
The conclusion I've come upon is that it could crap in the AC, which is why I'm considering filtering it. I've seen lots of circuts, but since it's in AC-land I would prefer a pre-built unit with a metal housing for safety reasons. So I wonder if this unit would do the magic needed?
https://www1.elfa.se/data1/wwwroot/assets/datasheets/gyFKH_data_e.pdf
The circut is:
L = 2x4 mH
R = 1M ohm
Cx = 68 nF
Cy = 2.2 nF
I'm eying the 2 amp non-medical unit.
Thanks in advance,
Olle
I'm having mechanical transistor humm, none at audio output and the humm varies with time...
The conclusion I've come upon is that it could crap in the AC, which is why I'm considering filtering it. I've seen lots of circuts, but since it's in AC-land I would prefer a pre-built unit with a metal housing for safety reasons. So I wonder if this unit would do the magic needed?
https://www1.elfa.se/data1/wwwroot/assets/datasheets/gyFKH_data_e.pdf
The circut is:
An externally hosted image should be here but it was not working when we last tested it.
L = 2x4 mH
R = 1M ohm
Cx = 68 nF
Cy = 2.2 nF
I'm eying the 2 amp non-medical unit.
Thanks in advance,
Olle
Do you mean mechanical transformer hum?
The filter you have shown is to attenuate RF and spike interference coming from the mains wiring. It does nothing for airborne interference and cannot solve a noisy transformer that may be suffering intermittant saturation due to unbalanced AC waveforms, i.e. DC on the mains.
The filter you have shown is to attenuate RF and spike interference coming from the mains wiring. It does nothing for airborne interference and cannot solve a noisy transformer that may be suffering intermittant saturation due to unbalanced AC waveforms, i.e. DC on the mains.
EVA, I can completely back your findings with pretty much the identical measurements I made recently, in fact extending them with frequencies below 50Hz, not restricting the test to "pure DC".
One or two back-to-back diode sets shunted with not too big capacitors usually do the trick. Too large capacitors pass too much of sub-50Hz components (there are lots of subharmonics and beat frequencies on the mains).
Another important variable is the often too low primary inductance of cheaper toroids in an attempt to reduce copper cost. Quite often one can see a steep exponential rise of idle current starting at 220V and lower. At 240V/50Hz there often are already significant signs of saturation and, depending on the construction, corresponding mechnanical hum. Under such conditions even smallest amounts of DC and sub-50Hz stuff make things go rapidly worse.
Once I spec'd toroids to maintain low and stable idle current up to 250V@50Hz (5..10% more turns for all windings) the transformers became much quieter in general and also react much less severe to DC. Together with the diode blockers this issue now is completely settled, for me.
Last edited:
because the costs the air-coils for crossover networks is increased significantly by Intertechnik (Kerpen, Germany) in the last years, I am assuming that the price of copper was also higher. I have the impression that this annoying transformer hum behaviour in the amp devices just for newer models therefore can be observed currently more often than before.
In order to effectively remove, one can also reduce the secondary voltage by introducing an additional transformer.
Hi
Few years ago I built and still use DC Filter + RFI/EMI Filter found somewhere in internet. Here is the schematic:
- diodes STTA1206
- elcaps 22000μF/16V BC Components
- film caps X1, X2, Y2 Evox Rifa
- varistor S20K300
- chokes RN222-4/02 4A/250V, RN112-4/02 4A/250V Schaffner
- resistors 3W
Few years ago I built and still use DC Filter + RFI/EMI Filter found somewhere in internet. Here is the schematic:
- diodes STTA1206
- elcaps 22000μF/16V BC Components
- film caps X1, X2, Y2 Evox Rifa
- varistor S20K300
- chokes RN222-4/02 4A/250V, RN112-4/02 4A/250V Schaffner
- resistors 3W
Attachments
Oooo... you just don't *happen* to have pcb drawing as a file, or as black/white image such that one can print and etch some boards oneself?
The grateful (greedy) Olle
EDIT:
Or even better: it seems that pre-built EMI filters are abdunant, but does someone know a decent DC filter buyable through say mouser?
Last edited:
