Exactly my point Andrew. Sounds like the German maker is still specifying transformer primaries the same way as they did 20 years ago (when Europe was 220V) and has not done any safety testing! Just put the Mark on.
It might be possible to change the transformer inside the unit.
It might be possible to change the transformer inside the unit.
Based on what I thought and what Richie seems to be confirming then this German manufacturer is breaking the law by selling product that does not comply.
You have a few options:
Report to the authorities.
Threaten to report and ask for full reimbursement of repair costs.
Just do nothing and let this German manufacturer continue breaking the law.
Did you arrive at a solution to make yourself a decent DC blocker? I'm in a similar situation and I'm no electronics guru so I'm a bit boggled with info at the moment even though it looks like I need to build quite a simple, if high current circuit?
I thought it was a simple DC blocker. The caps see less than 3 volts between terminals.
I would worry about the mains voltage with respect to the shrink plastic covering of the lytics. If the can of the lytic is at the potential of either terminal, then you would be relying on the shrink plastic to support the peak line voltage as well as any transients.
I also recommend a fuse in the blocker chassis as well.
jn
Hi, as I say. I'm a real noob. Why would there not be 240v between the capacitor terminals?
I guess I am looking for a nice easy to follow diagram which I can then make and put in a little case, like the Ava hum dinger. With an AC in and an AC out, which I can connect before my airlink 2kva balanced power supply. I've gathered from this pcb from snowstorm:
Sjöström Audio - DCT02 The DC trap, high-end style
that I need some capacitors and diodes, but after that I'm clueless!
I guess I am looking for a nice easy to follow diagram which I can then make and put in a little case, like the Ava hum dinger. With an AC in and an AC out, which I can connect before my airlink 2kva balanced power supply. I've gathered from this pcb from snowstorm:
Sjöström Audio - DCT02 The DC trap, high-end style
that I need some capacitors and diodes, but after that I'm clueless!
Rod's circuit puts diodes across the capacitors.
When the currents are low, there is very little voltage across the capacitors. When very high currents are being drawn, the diodes will begin to conduct, and they will limit the voltage across the caps. If you look at the way the diodes are wired, you can see that there will be a maximum of two diode drops across both capacitors. Typical 35 amp bridges will have less than 1.5 volts across them even at 25 or 35 amperes.
edit: But if you notice, the capacitor terminals will be at line voltages, so the case to ground isolation is important, as well as fusing the input to this assembly.
jn
It's not about the current which flows through the blocker/transformer - it's about the asymmetry between the AC voltage's half waves - this is the reason for the DC offset in the mains. If you look just for the DC Blocker, I can offer a solution, however I'm afraid I shouldn't advertise myself directly in the thread.
Yesterday I was pretty sure it was DC being dumper by my plasma TV that was causing the issue. How since a reply on another forum I'm not so sure. It seems like it might be the current draw characteristics of my plasma instead:
Building a Balanced Power System - Page 2
Building a Balanced Power System - Page 2
You are addressing a statement which I did not make.
At low AC draw, the voltage across the caps will be current dependent. As the AC draw increases, the diodes are there to shunt past the caps to prevent destroying them with ripple currents.
This is important specifically for turn on inrush currents to the end use load. Once the transient is past, the diodes will essentially get out of the way assuming the caps are large enough.
As for DC offset in the mains, the vast bulk of measured DC offset is a result of inaccurate measurement. Specifically, how the meter actually measures. The best way to measure it is by using a high bandwidth DCCT, it does not affect the measurement by it's insertion yet measures DC current.
In the USA, mains offset at the load will be a consequence of DC IR drop through the neutral, a 100 foot length of #12 will have 17 milliohms, and 10 amps DC drawn from another load on the branch will produce a neutrally induced DC offset of .17 volts. When the primary sees this, it'll buck it but can still pull significant asymmetric current which can bias the core, forcing saturation at line rate frequencies. edit: remember, unloaded AC current will depend on the inductance of the transformer, DC current will depend only on the primary winding DC resistance.
jn
Last edited:
Well, if the inrush current is bigger that (let say) 50A, it makes sense. But I still consider the main reason of diodes is to turn on when blocked DC over the caps becomes larger than 2.5V (or the sum of turn-on voltages of serial connected diodes).
As I said, it is both. If one does not consider the real application, then one can ignore inrush (I do not recommend that).. However, the diodes must be specified to withstand actual use with actual equipment. I did not state that the diodes are there just for inrush, but that they must survive it. We are not arguing here.
What the caps can withstand is defined by the caps themselves. Without an inrush specification nor single pulse specifications, it is impossible to select a reasonable cap to withstand the application. I've not seen a surge specification on a capacitor which is consistent with surge ratings of diodes.
jn
Just calculate the capacitor impedance at your mains frequency and from that predict the maximum AC current that can pass without exceeding the voltage protection limit set by the diodes. The voltage protection limit is Vf at the turn on voltage of the cold diode, about 400mVf. If the diodes have been passing, then this Vf will be lower due to the higher Tj and could be as low as 300mVf.
That AC current is the maximum current that the DC blocker can pass.
Any extra current drawn while the amplifier is suffering start up transients, or abused, or overloaded, or because the selected capacitors are just plain too small, will turn on the diodes and the DC blocking action is over-ridden.
The DC blocker works when ALL the AC current demand, flows THROUGH the capacitors.
The Diodes are there for overload protection.
That AC current is the maximum current that the DC blocker can pass.
Any extra current drawn while the amplifier is suffering start up transients, or abused, or overloaded, or because the selected capacitors are just plain too small, will turn on the diodes and the DC blocking action is over-ridden.
The DC blocker works when ALL the AC current demand, flows THROUGH the capacitors.
The Diodes are there for overload protection.
Last edited:
I've built one of the esp DC blockers for my f4. Although it hasn't stopped my transformer humming it seems to sou d a tiny bit clearer.
Anyway, I just wanted to ask. Brief question about thee ripple current.
I have used 2 22000 35v caps that were left over from the psu. The esr is 0.024 and 0.018 (120hz,20khz) and Amax 5.48A and 6.85A at 120hz and 20khz.
From the artical it appears these would be ok.
I have wrapped some cable ties around them to avoid the plastic wrap touching the base plate or the earth box I've used as a blast shield. I also earthed this.
I used a kbpc3504 (35a, 400v, 280v RMS) bridge from maplin which is bolted to the base plate with a little thermal paste.
Are the caps or bridge likely to get hot?
I inserted the DC blocker between the soft start and the transformer (500va).
Can anyone see any major issues with this?
Thanks
Anyway, I just wanted to ask. Brief question about thee ripple current.
I have used 2 22000 35v caps that were left over from the psu. The esr is 0.024 and 0.018 (120hz,20khz) and Amax 5.48A and 6.85A at 120hz and 20khz.
From the artical it appears these would be ok.
I have wrapped some cable ties around them to avoid the plastic wrap touching the base plate or the earth box I've used as a blast shield. I also earthed this.
I used a kbpc3504 (35a, 400v, 280v RMS) bridge from maplin which is bolted to the base plate with a little thermal paste.
Are the caps or bridge likely to get hot?
I inserted the DC blocker between the soft start and the transformer (500va).
Can anyone see any major issues with this?
Thanks
The soft start resistor and the DC blocker are in series with the transformer primary. They can go in any order, so you're OK.
If the capacitors are continuously passing near their ripple rating then your capacitors will be running hot. You will feel the outside as warm.
This is another reason for getting the capacitance big. The lower impedance of a bigger capacitor will heat less.
Keeping the capacitors and all these Live components away from the chassis is important.
What about gluing a sheet of thick plastic to the chassis over/under/to the side of the DC blocker?
I have started using sheets cut from yoghurt containers, the big 1kg rectangular size.
If the capacitors are continuously passing near their ripple rating then your capacitors will be running hot. You will feel the outside as warm.
This is another reason for getting the capacitance big. The lower impedance of a bigger capacitor will heat less.
Keeping the capacitors and all these Live components away from the chassis is important.
What about gluing a sheet of thick plastic to the chassis over/under/to the side of the DC blocker?
I have started using sheets cut from yoghurt containers, the big 1kg rectangular size.
Last edited:
Thanks. I've just realised that as long as it's inline with the soft start it will limit the inrush through the blocker anyway. Doh. But after was more convenient for wiring.
I think the caps should be well under they're limits. Esp reckoned 2A and 500va. I can't quite remember what my boards draw but was either about 80w a side or 160w a side but I did significantly over spec the transformer. Am running it in stereo mode into 8ohm loads.
I did think about lining with plastic and had a look round to see if I had a sheet from something broken I could use. I'll insert something.
I did power on with light bulb tester and the transformer wasn't humming.
If the transformer is humming (without lightbulb) because of DC presumably this lightbulb limiter was stopping sufficient current flowing and it no longer saturating.
If the transformer hums due to something else would the lightbulb still stop it or merely attenuate it.
Having tested the DC circuit on the bench it seems it can only block about 500mv. Above that it just reduced the offset.
Could it be possible that my mains are above 500mv DC offset?
Would a second blocker help. (Might play with the circuits low voltage on the bench to see).
I think the caps should be well under they're limits. Esp reckoned 2A and 500va. I can't quite remember what my boards draw but was either about 80w a side or 160w a side but I did significantly over spec the transformer. Am running it in stereo mode into 8ohm loads.
I did think about lining with plastic and had a look round to see if I had a sheet from something broken I could use. I'll insert something.
I did power on with light bulb tester and the transformer wasn't humming.
If the transformer is humming (without lightbulb) because of DC presumably this lightbulb limiter was stopping sufficient current flowing and it no longer saturating.
If the transformer hums due to something else would the lightbulb still stop it or merely attenuate it.
Having tested the DC circuit on the bench it seems it can only block about 500mv. Above that it just reduced the offset.
Could it be possible that my mains are above 500mv DC offset?
Would a second blocker help. (Might play with the circuits low voltage on the bench to see).
the mains is an AC signal. It is a very distorted sinewave.
There is no DC on the waveform.
There is an asymmetry on the waveform that varies from second to second. The iron core of the transformer accumulates this asymmetry as a flux that is not equal for both halves of the signal. It is the larger flux in one direction compared to the other that gets closer to the saturation. This draws extra primary current and makes a noise.
The DC blocking capacitor somehow reduces the way the iron core accumulates the unbalanced halfwaves.
The primary current into the transformer should be an sinewave, but the secondary loading changes this significantly.
If you have a capacitor input filter then your primary current is not a sinewave. It is pulses of fairly short duration (1ms to 3ms) with idle periods of 7ms to 9ms.
The current during a pulse has to provide the same energy as the rms equivalent current out of the secondary plus a bit for transformer efficiency.
That pulse peaks at values roughly 3times to ten times the rms current you would measure into the primary.
The pulses have to pass through the capacitors.
There is no DC on the waveform.
There is an asymmetry on the waveform that varies from second to second. The iron core of the transformer accumulates this asymmetry as a flux that is not equal for both halves of the signal. It is the larger flux in one direction compared to the other that gets closer to the saturation. This draws extra primary current and makes a noise.
The DC blocking capacitor somehow reduces the way the iron core accumulates the unbalanced halfwaves.
The primary current into the transformer should be an sinewave, but the secondary loading changes this significantly.
If you have a capacitor input filter then your primary current is not a sinewave. It is pulses of fairly short duration (1ms to 3ms) with idle periods of 7ms to 9ms.
The current during a pulse has to provide the same energy as the rms equivalent current out of the secondary plus a bit for transformer efficiency.
That pulse peaks at values roughly 3times to ten times the rms current you would measure into the primary.
The pulses have to pass through the capacitors.
I'm afraid there is only DC or AC. The DC may wobble around, or use other descriptive words such as flicker or fluctuate or unbalanced - but that is a DC signal, as it is not a periodic signal.
If you view the core flux swings as being unbalanced then it is the DC component of the signal. The transformer inductance impedes the transient change in the DC levels - maybe that is what you are trying to illustrate by the verb 'accumulates'.
I was going to use the word "integrates" but that was probably just as inaccurate.
The area of the waveform above and below the zero volts varies.
It's this unbalance of upper and lower areas that adds up to a net unbalance of flux.
I would not call that DC. In my head I see that as wobbly AC.
But whether it is wobbly DC or wobbly AC, it seems that a DC blocking/AC passing capacitor of sufficient size for the current demanded does work at reducing the transformer flux sufficiently to make it much quieter.
The area of the waveform above and below the zero volts varies.
It's this unbalance of upper and lower areas that adds up to a net unbalance of flux.
I would not call that DC. In my head I see that as wobbly AC.
But whether it is wobbly DC or wobbly AC, it seems that a DC blocking/AC passing capacitor of sufficient size for the current demanded does work at reducing the transformer flux sufficiently to make it much quieter.
Last edited:
- Status
- Not open for further replies.