First Lm3886

[PJPro]

OK Andrew. Thanks. I hadn't actually intended to have a switch on the front of the power amp....but it looks like I need one.

Any thoughts on the transformer? Does the one I indicated earlier look OK?

[AndrewT]

Originally posted by PJPro
OK Andrew. Thanks. I hadn't actually intended to have a switch on the front of the power amp....but it looks like I need one.

Any thoughts on the transformer?

not necessary.
You could use a low voltage signal from a pre-amp or a source component to trigger a relay inside the power amp. That relay would replace sw1

What do you have for the relays? voltage and current requirements.
these are the biggest load and they are non audio so there is no benefit to increasing the size of the transformer.
Just remember that the continuous DC current < half the AC current and preferably <=0.25*AC current.

I would expect somewhere between 3VA and 6VA would do the job.
Voltage to match the relays and the chip.

[Stuey]

Originally posted by AndrewT
that EDN circuit is connected directly to mains.
You must take steps to ensure you cannot be electrocuted.

The relay coil is gradually fed with increasing current as the capacitor charges. Eventually it trips.
At switch off the capacitor discharges and eventually the relay releases.
I hate that style of operation.
I think we should have fast trip after a set time delay and fast/immediate release on loss of mains power.

Just to put this in perspective for PJ, 'eventually' means 330 milliseconds for the switch-on versus 100 for the ESP circuit. You wouldn't be twiddling your thumbs waiting...

However, IMO given your current knowledge Peej you'd be better off with a better explained design such as the above, anyway.

Did you look at the SKA softstart module? It's tiny, which is why I recommended it.

Stuey

[PJPro]

Originally posted by AndrewT

not necessary.
You could use a low voltage signal from a pre-amp or a source component to trigger a relay inside the power amp. That relay would replace sw1

I'd prefer not to have to rely on the pre-amp providing the power. If I'm to implement the soft start, it needs to be self contained.

Originally posted by AndrewT

What do you have for the relays? voltage and current requirements.
these are the biggest load and they are non audio so there is no benefit to increasing the size of the transformer.
Just remember that the continuous DC current < half the AC current and preferably <=0.25*AC current.

ESP doesn't state the relays which are used with the circuit. He does say that a 9V / 5VA transformer is required to power the circuit.....and I believe the transformer identified above meets that requirement.

Originally posted by AndrewT
I would expect somewhere between 3VA and 6VA would do the job.
Voltage to match the relays and the chip. [/B]

Yes....that aligns with ESP's statement of 5VA. So, it looks as though the transformer I have selected will be OK.

[PJPro]

Originally posted by Stuey

However, IMO given your current knowledge Peej you'd be better off with a better explained design such as the above, anyway.

Sorry Stuey, which one do you mean? The ESP?

Originally posted by Stuey

Did you look at the SKA softstart module? It's tiny, which is why I recommended it.

Yes, Thanks. I did look at that....but it was a bit scant on the information.

[PJPro]

Originally posted by infinia

The DC block cap is a good idea but it'll never work for a RF bypass or a lowpass filter. RF cuts straight through a large foil cap like a hot knife thru butter. The size of the series DC block cap wouldnt do a thing for RF design, either series or shunt.

OK. So the large cap on its own isn't going to provide a RF filtering solution.

Originally posted by infinia

The solution is a series ferrite bead and a 220pf shunt cap 2.5mm LS.

So I'd need a ferrite bead in series with the large cap? What does LS mean? Sorry for not knowing this.

Originally posted by infinia

PJpro Good luck with your project! I like your chassis esp the nice IEC inlet module, too bad it's in back out of sight.

Thanks. Hopefully, with advice from yourself and others, it'll be safe and sound good too.

[PJPro]

infinia, do you happen to know what the L1 component is which ESP is using to filtering on the output?

[Ted205]

L1 is an inductor. Have alook at the lm3886 datasheet under reactive loading

[anatech]

Hi PJPro,
The circuit shown in the EDN article for discharging the supply capacitors should never have been printed - anywhere! There is no electrical safety organization that would ever pass it. That is one reason we used to see the really expensive switches with the larger DPDT section in front and the smaller DPDT switch mounted further back.

I know you didn't post that, but I just wanted to make sure it was not under consideration.

As far as soft start circuits are concerned, I think they are an excellent idea. The filter caps are protected from large surge currents, as well as the rectifiers. You can safely use fast blow fuses as well. This is normally a relay type circuit as mentioned. You can use a timing circuit, or better yet, allow it to be powered from the raw rail(s). That's easy, cheap and adds another function. If your rails drop due to a shorted rectifier or filter cap, the relay will drop out and insert the resistor(s) in series with the primary again. The resistor(s) will then open in case the fuse doesn't. Added protection at no additional cost.

The question on whether to discharge the rails or not is more complicated I guess. Very often, the bias currents in the output stage will discharge the supplies, depending on the amp section design. If you want to discharge everything within 10 sec (I don't know why you would need that time limit), you can do this in a few ways. I don't think you should be using a relay for this, but you could. The added poles on the power switch are a horrible idea, as mentioned. You could always detect the open power switch contacts electronically, then fire a pair of SCRs in series with the resistors. The SCRs will stop conducting once the voltage dies off by themselves. You could use power transistors as well, but you would need to supply base current for the entire discharge time. It's up to you, but there is probably a way to do this nicely if you really need to. Bleeder resistors will and do work fine for most applications, and they do not need to be sized to kill the charge in 10 sec. I think anything up to 1/2 hour is reasonable, which allows for lower standing dissipation.

One comment on grounding issues. I think it's pretty clear that the safety ground is permanently bonded to the chassis and all exposed conductive parts. I see no wiggle room there for any arguments at all. At most, a 10R to 100R resistor between that and the circuit ground should do. However, if you connect your RCA or XLR grounds directly to the chassis, there should not be any ground currents flowing into your signal ground. You may have trouble with any cable shielding, but that's about it.

-Chris

[anatech]

Hi Andrew,
One comment about removing capacitors from inputs and outputs. I completely disagree with removing any capacitors from any input or output - ever. Situations change and equipment easily ends up in someone else's home. Removing the input caps on a power amp is probably the very worst thing you could do.

So, if you really must remove any capacitors from audio equipment, taking the output caps from the source component is safer. The input capacitors on an amplifier should always remain in place. Yes, some amps do have good DC offset protection, but why burn the contacts needlessly if it uses relays? If the protection involves turning the amp off and on to return to normal operation, the scheme won't work anyway.

You are much further ahead to use the efforts and money to improve your listening room, or the quality of your speakers. You get the biggest bang for your currency in that direction. Removing input capacitors from amplifiers seems to be an irresponsible act to me. Never did understand how that could sound like a good idea.

For others out there, Krell did (and may still) have no coupling capacitors on the output of their preamps or input of their power amps. An acquaintance of mine had a Krell system, and the preamp went DC. Cool! The amplifier did amplify the DC output which caused an almighty bang in the speaker, but at least the amp turned off (eventually). The repair was straight forward, bad offset control IC (DC servo op amp). Mind you, the loudspeaker needed service now as well. That wasn't cheap as they were current PSB Stratus Gold i's. See? Irresponsible and avoidable. The Krell didn't sound that good anyway, so I can't see ho leaving out these capacitors could improve the sound quality any!

-Chris