MyRefC build guide

[westers151]

Originally Posted by jonclancy

Now that would be handy.... nah, dual mono all the way!

I thought you needed 160VA per channel?

I asked about this back on page 9 and this is the reply that AndrewT gave me

a 100W amplifier can operate quite well with a transformer rated from 100VA to 200VA. any bigger becomes uneconomic.

A 60W Chipamp can use a 60VA to 120VA transformer.

Unfortunately a very low VA transformer has very high regulation.
I recommend that the transformer should be >=160VA.

A two channel amplifier of 60+60W would require 120VA to 240VA and add in that "avoid high regulation guidance" and you get 160VA to 240VA.

I don't know where 300VA came from. It is icing on the cake. An uneconomic way to improve an amplifier. But some builders do hear a sound quality benefit and consider the bigger transformer as offering good value for their budget.

I would consider a 300VA as good for a 100W + 100W two channel amplifier.

Unless I've understood this in completely the wrong way, our 2 channel amps (being around 60W per channel), could be run from a single 225VA transformer with 24-25V dual secondaries.

AndrewT, could you confirm that I've understood you correctly.

Thanks.

[AndrewT]

A two channel amplifier of 60+60W would require 120VA to 240VA and add in that "avoid high regulation guidance" and you get 160VA to 240VA.

Originally Posted by westers151

I asked about this back on page 9 and this is the reply
Unless I've understood this in completely the wrong way, our 2 channel amps (being around 60W per channel), could be run from a single 225VA transformer with 24-25V dual secondaries.

I thought I made it clear.
60+60W can be run economically on a transformer rated from 120VA to 240VA.
The condition of avoiding a high regulation transformer raises the lower limit from 120VA to 160VA.
The 225VA proposed is right in between the suggested lower and upper limits, i.e. 160VA < 225VA < 240VA.

Did I confuse some Members?

[AndrewT]

Originally Posted by jonclancy

Now that would be handy.... nah, dual mono all the way!

I thought you needed 160VA per channel?

I think you will find that a two channel amp will run quite well from a transformer where the VA = total maximum output power.
However better performance is usually achieved by using a higher VA transformer. All it costs is money and space and weight. some go as far as saying that 300VA is best for a chipamp. I don't disagree with that opinion.
But there is a downside to using the minimum size of transformer.
Low VA transformers have a much higher regulation.
The effect of this is that off load they produce a high DC voltage. This potentially damaging voltage must not overload the chipamp.
As the load increases the available DC voltage falls (sags). All transformer rectifier smoothing cap PSUs suffer from this characteristic.

It has been found that avoiding the smallest transformers helps improve performance by reducing the sag in PSU voltage with increasing load.
I have suggested that the lower limit be set at 160VA. That has absolutely nothing to do with whether the chipamp is capable of working or not working with a smaller transformer. It is an economic decision that gives better performance than saving a few $/£ on a 80VA or 120VA transformer. It comes down to good value.

[westers151]

Nope, that's what I thought you meant - i.e, that only one transformer (with dual secondaries) rated between 160VA and 240VA would be sufficient to power both of the channels.

[abraxalito]

Originally Posted by AndrewT

I think you will find that a two channel amp will run quite well from a transformer where the VA = total maximum output power.

Certainly its possible, but please be aware that such a transformer will be seriously overloaded when delivering the maximum output power and substantially overloaded well prior to that point. That its possible to get away with this without burning out the transformer is because of the peak to average value of music. Classical music aficionados will get away with it more readily than head-bangers.

In order to see how overloaded a transformer will be at maximum output, take into account both the efficiency of an average class-B amplifier at full power (of the order of 70%) and the non-linear load presented to the transformer by the rectifiers and capacitors.

Transformers' VA ratings are specified with a resistive load, this draws current throughout the mains cycle. However, rectifying and smoothing the voltage from a transformer results in current only being drawn at the peak of the mains cycle. Typically under load, the transformer will be supplying current for 20% of the time, the remaining 80% of the time the amp is supplied by the reservoir caps.

A rough estimate then is that the current pulses from the transformer will be five times greater than the peak current into a resistive load. This results in five times greater heating (heating is proportional to the square of the current) than with a resistive load.

I don't disagree with Andrews analysis, this is merely to point out the potential hazards of going to a transformer whose VA rating is equal to the rated output of the amp. It would be a no-no commercially as CE marking requires certain sustained power output tests without overheating.

[AndrewT]

Originally Posted by abraxalito

this is merely to point out the potential hazards of going to a transformer whose VA rating is equal to the rated output of the amp. It would be a no-no commercially as CE marking requires certain sustained power output tests without overheating.

this is precisely what commercially is done.
Putting in a smaller transformer that can only just do the job of giving out a music signal without catching fire or failing in the first year of operation to save a couple of £/$

I normally recommend a 20dB of overhead between maximum output power and average listening power.
Some may go as low as 10dB. That would require 6W of average output power in a 60W amplifier.
Look at the chipamp datasheet. How much heat/power does it dissipate when delivering 6W, most will be around 10W. That's a total 16W. Two channels of headbanging music would consume ~32W. That is just 27% of the smallest transformer (120VA for 60+60W) that any sensible builder would use.
If 160VA were used for 2channels then the consumption reduces to 20%.

I don't see the problem.

If the builder were to do some continuous maximum power testing, I believe that same builder would be very aware that temperature is critical to reliability. He would do the testing in an appropriate manner to avoid damaging any of his equipment.

[jonclancy]

Hi All,

Thanks for the detailed responses. I intended to over-spec my trafos, but not by that much! I must have picked up the larger recommendation (160VA per channel) from somewhere or the other. No worries here, though. As I have four of these, and 8 channels of MyRef (four Ultimate BOM), I can try all configurations to see what suits me and my system.

The irony is that with the two trafos in my chassis, I cannot use the 2 big 'sinks I have planned (one for each amp), but could configure 2 'sinks and 1 trafo, or 2 trafos and 1 sink in each chassis. I bi-amp here.

Cheers

Jon

[abraxalito]

Originally Posted by AndrewT

this is precisely what commercially is done.

How would they pass CE then? You're saying that the manufacturers are fraudulently self-certifying? Or evading EU directives?

Look at the chipamp datasheet. How much heat/power does it dissipate when delivering 6W, most will be around 10W.

That would imply the amp was around 38% efficient at 10% output. Its considerably worse than that.

Have a look here for an efficiency table for a class B amp.

A figure of 20% efficient is more like it and that figure is optimistic as it takes no account of PSU regulation.

Two channels of headbanging music would consume ~32W. That is just 27% of the smallest transformer (120VA for 60+60W) that any sensible builder would use.

I note you've ignored the power factor issue I explained above. Do you think I'm wrong in that? If you do I'm interested to learn why.

If the builder were to do some continuous maximum power testing, I believe that same builder would be very aware that temperature is critical to reliability. He would do the testing in an appropriate manner to avoid damaging any of his equipment.

As far as I recall, CE testing does not require any continuous maximum power testing. So yes, I agree with you, its prudent to do such testing very carefully.

[regiregi22]

I think Andrew is implying that CE markings are not strict enough. They guarantee that the product will be safe, not that it won't fail soon or that it is made of crap. Rubbish can be safe

[westers151]

If I have my power supply in a seperate enclosure then I would be wanting to rectify this at the transformer and only send DC across the power umbilical.

How would I connect that to the board as the board assumes the rectification is done by a bridge onboard?