Yep that is a chinese one for domestic use there (220V, in Europe 230/240V are the standards).
In this case I would just connect the secondary windings loosely in parallel and attach a DMM to both combined wires. Just a very short (think of a few seconds) power on and you'll see either 15 to 20V AC or near 0V.
It is even better to feed the primaries a way lower voltage than 230V like 24V from another transformer (if you have one) and then swap the secondary wires. That is safer too.
In this case I would just connect the secondary windings loosely in parallel and attach a DMM to both combined wires. Just a very short (think of a few seconds) power on and you'll see either 15 to 20V AC or near 0V.
It is even better to feed the primaries a way lower voltage than 230V like 24V from another transformer (if you have one) and then swap the secondary wires. That is safer too.
Last edited:
Hi and thanks a lot for the very kind and helpful advice.
On this basis i have decided to use just one secondary and leave the other open. In general i am not comfortable to parallel things. I have this feeling that there is nothing similar to something else.
Like for transistors parallel. I will never find 2 really identical and one will work always more than the other.
This device just need 0.5A, half the current of one secondary.
But i am sold on the R-core also after your recommendation.
Actually i would replace all the toroids with this kind of transformers.
I am sure there could be significant benefits in terms of isolation and noise suppression. 100% sure.
Thanks a lot again.
Kind regards, gino
P.S. by the way i have come to the conclusion that if power is limited a AC umbilical solution can work just fine, maybe keeping the umbilical as short as possible and with wires of good section ?
Just connect them in parallel as half the current is not much headroom.
Regarding your question: forget AC as cabling will pick up nasty signals easier. Best is to rectify and filter as short to the transformer as possible and go on with DC. As a sidenote: DC is in fashion, also for high voltage energy distribution. Back to the future !
https://en.wikipedia.org/wiki/High-voltage_direct_current
It would be terrible if Thomas Edison would be "right" after all. We all like Nikolay Tesla better for his marvellous work on AC, don't we ?
https://en.wikipedia.org/wiki/War_of_Currents
Regarding your question: forget AC as cabling will pick up nasty signals easier. Best is to rectify and filter as short to the transformer as possible and go on with DC. As a sidenote: DC is in fashion, also for high voltage energy distribution. Back to the future !
https://en.wikipedia.org/wiki/High-voltage_direct_current
It would be terrible if Thomas Edison would be "right" after all. We all like Nikolay Tesla better for his marvellous work on AC, don't we ?
https://en.wikipedia.org/wiki/War_of_Currents
Last edited:
Sorry but i do not understand. Do you mean that an AC umbilical acts more like an antenna than a DC umbilical ? why should it ?
I could always used a shielded umbilical in case.
But now that i think better i could go with let's say +/- 18 VDC to the internal regulators that are +/- 15VDC ?
I would just need a 3 wires umbilical.
More complex by the way.
Oh oh dear mr. Ginetto you lack basic knowledge on electrical matters. It makes me wonder that you want answers on how two do things right and you also work on devices but you don't have enough experience/knowledge/electrical background to see the bigger picture (not meant degrading but facts are facts). Maybe you must lessen the tight wishes and demands and start with the basics instead of high end/hifi. I know the feeling, it is very frustrating to walk against walls when starting with this hobby. It is a question of putting a lot of time in it and making mistakes really. Anyway, make sure you always work safely !
The "more complex" method is better indeed and there are many audio devices following that path. Please think of dropout voltage of regs so with LM317/LM337 and 78XX/79XX you need more than 3V difference. LDO regs need way less and are preferable. Check what regs you have and then calculate.
The "more complex" method is better indeed and there are many audio devices following that path. Please think of dropout voltage of regs so with LM317/LM337 and 78XX/79XX you need more than 3V difference. LDO regs need way less and are preferable. Check what regs you have and then calculate.
Last edited:
For sure i lack knowledge but just i do not understand why AC umbilicals should be more prone to pick up noise than DC ones. 😕
I understand that both can act as antennas but a shielded cable for the umbilical should cure that.
Regarding the local regulators on board are indeed of the 7815/7915 family.
I agree that some more headroom in current is a nice thing.
Not that the unit came with a better one. Actually the contrary.
Quite poor actually. I do not think it is even the original one.
Thanks again, gino
What is being said is DON'T put the noisy AC to DC rectification and smoothing in the box, leave that remote and just supply DC down the umbilical cable... It is nothing to do with noise pick up its just common sense... the noisy AC in the remote box feeding clean DC ,through a common mode ferrite and possibly a pi-filter where it enters the main board.
do the DC as twisted pair cable, shielded or not, I wont go into shielding as it is a complex subject, just read Henry Ott or similar.
do the DC as twisted pair cable, shielded or not, I wont go into shielding as it is a complex subject, just read Henry Ott or similar.
It looks to me that two long narrow C cores have been clamped together.
I would call that PSU transformer a "double C core".
It also looks expensive. The ends of each C need to be ground very accurately so that when clamped to it's neighbour, the "gap" in the core is minimised.
A winding around the "gap" helps minimise leakage. That's why C cores have the join inside the windings. And this is what makes a C core much better for some performance parameters than an EI.
I would call that PSU transformer a "double C core".
It also looks expensive. The ends of each C need to be ground very accurately so that when clamped to it's neighbour, the "gap" in the core is minimised.
A winding around the "gap" helps minimise leakage. That's why C cores have the join inside the windings. And this is what makes a C core much better for some performance parameters than an EI.
Hi and thanks a lot.
This is exactly how this commercial unit is built.
The audio circuit is powered with +/-15VDC.
I could arrive with +/-20VDC unregulated (the 15+15VAC transformer should provide that) with a 3 wires umbilical and keep the final regulation stage already on board.
However I see only 1000uF before and 1000uF after the regulators when I would like to see 4700uF before and 2200uF after.
uFs are never too much. The more the better.
I guess all this filtering must be placed before the final regulators ?
I understand. I will try to keep the DC umbilical as short as possible to minimize the antenna effect and be able to use a normal unshielded cable.
I have come the conclusion that the quality of the power entering the circuit can do a lot indeed.
For instance I was even thinking to put bypassing caps directly on the output transistors in order to create a local power reserve 😱
I mean placing uF as close as possible to the output devices.
I have seen some amps using this solution.
I have already some 1000uF Panasonic but I could go also for 4x2200uF of very good quality like Nichicon FG or KZ.
This should lower noise and increase dynamics.
I did it already in the past with another preamp and liked the result.
Thanks a lot again, gino
Filtering is ALWAYS done as near to the cable entry as possible otherwise the cable can introduce noise into the box especially if it is shielded properly.
If you don't use local supply rail decoupling, your amplifier will be much more susceptible to ringing and instability.
Read what Leach did and how he cured it in his early iteration of the Lo Tim.
He had some local decoupling but only the HF part. He omitted the MF part and the Lo Tim oscillated. He added the MF decoupling and the problem went away. Prof Leach got it wrong initially.
There are a lot of single wires in that assembly................
...............
Where are the twisted pairs?
Current flows around circuits to get back to the source.
If the Flow current is twisted with the Return current in a low loop area pair then this massively reduces interference.
Both types of interference are attenuated: the emi that is emitted by the single wire aerial when a changing current flows and the emi that is picked up by the single wire aerial.
Hi and thanks a lot again.
I see the issue clearly now.
Moreover I am understanding that I am walking blindfold.
For a really serious approach I would really need that blessed Spectrum Analyzer.
It would show me almost any incoming noise 🙁
I should stop everything and focus on instruments.
If not it is like driving in the dark without lights 😱
Thanks a lot again, gino
Hi and thanks a lot again.
I have to check but I guess that there is already local decoupling.
I just want to add more uF directly on the output transistors ... like 1000uF from the rail to ground.
imle every time I added uF I liked the sound better.
I will look for this article for sure. Thanks for the advice.
For me caps are very interesting parts. It is where energy is stored.
I like caps a lot.
Thanks a lot again, gino
Hi and sorry ... I just upload the pic to show the concept of local uF instead of the usual two big ones 😱
I mean, the shorter the distance between caps and output devices the better I think.
The release of current from the caps will be instantaneous.
And moreover the higher the Voltage the faster the delivery.
Like in tube amps.
There is a huge amount of energy in the tube power amps caps.
Energy goes with V^2.
this means that also many expensive commercial amps are far from ideal.
In others I indeed see the power wires twisted.
I will keep it in mind. Because I intend to modify many amps.
Especially in the power supply.
Usually it is where the major cuts are carried-out.
Thanks a lot again, gino
A good read if you want to get deeper into the digital side of power supplies is Eric Bogatin, Signal and power integrity. Going through the PDN section now and for a while, pretty heavy on the maths....
The supply shown as Andrew has pointed out is worrying, prone to noise pick-up and no care taken to optimise the PSU or cater for the AC loops withing the PSU.
Hi ! you have hit my weakest point ... maths ! 😱
Actually one of my sources of frustration. I have an immense admiration for math and for person who are good at that.
Especially when i read that the ability at math is a measure of intelligence.
I tend to respect a lot intelligent people.
Anyway, from this i do not even try to design anything.
2nd choice is to biuld something. Also i am very bad.
3rd choice would be to test something ... this i guess i could do with the right instrument.
What i mean is that i should start with a decent instrument and measure everything i have at hand to see their performance.
Then reading here and there try to smooth out some corners ...
without instrument i am just wandering in the dark.
I am not so naive to think that i can hear the effect of a dc filter or different diodes.
Unfortunately the instrument is expensive.
A spectrum analyzer.
a very complex issue i see. I will try only to beef up the uF for now.
As i said there are 1000uF before and after the regulators.
I have just received some nice 3300uF Panansonic to place before and i have already 2200uF Nichicon FG to place after the regulator.
Unfortunately i will only be able to check by ear for any improvement
It is mainly to get used to desoldering and soldering for now.
I should receive a cheap soldering station soon.
Thanks a lot again, gino
- Status
- Not open for further replies.