New amplifier...possibly..

[fooboo]

Hello there

They look good. The 2nd and 3rd are small enough to fit in the case. But I am going to go the toroid route with this. I have noted down the SMPS details though for future reference.

I have 'knocked up' a scale drawing of the unit with a pair of SMPS PSU's inside and another with a pair of toroids.

Now I have been looking at the ground lift/safety earth issue. Given that I will be running this unit as a dual single rail where do I take the Zero volt line to?

Obviously the earth line itself is to be hard wired to the casing halves. I am also assuming I need 2 circuits to cover each channel.

I have etched the 1562 PCB's today. Some of the finest traces I have had to do so far. One PCB will need visiting with strands of wire, and solder, to just bridge some faults on these fine traces

regards

Fooboo

[danielwritesbac]

One enclosure? Where is the cool air intake? Where is the hot air output?
Capacitors don't last inside of ovens (5000 hours is insufficient).

Real bridge rectifier looks like:
~+
-~

Reference output of the Ground Loop Breaker goes to the power supply board:
To the center tap of a split rail supply, also called zero volt.
Or, to the Negative of a single rail supply, since the car chip is a negative ground model, as is the typical car.

P.S.
There is high variety amongst car chips. Some of them are very impressive, most a are a mid-fi with a pleasant tone but not penultimate clarity, and some are terrible as if a flat raspy low resolution presentation. In this case, if I were you, I'd get one of those up and running and check it out with an SMPS (12.7v to 15v range) before spending it up with luxuries such as a fully dual mono build. Some of that isn't an effective place to spend. Also Toroids don't like to be stacked like that. Personally, I'd use ONE 15+15 dual secondaries transformer and a pair of regulators (one reg per each channel). See Decibel Dungeon about the regulators (or capmulti), which are an audio quality upgrade. You'd just be building two V+ portions to get your dual mono regulated power.

[danielwritesbac]

Originally Posted by fooboo

I am also assuming I need 2 circuits to cover each channel.

Given that the rule for avoiding ground loops is "Once Met" then I think that you are correct to use one ground loop breaker per each channel since Dual Mono is supposed to be electrically identical to Monobloc, with the sole exception the number of enclosures used.
Its yet another expensive enhancement that I've never seen used with a car chip. Before you go much further, please check out the price of a MyRef. The difference in cost between that and a car chip is only a tiny fraction of the total cost.

[AndrewT]

The two transformers give you two isolated amplifiers.
Yes, each amplifier needs it's own connection to Safety Earth.
Now you have dual mono inside a common chassis.

If you can, keep the two isolated amplifiers ALWAYS isolated from each other. The ONLY connection between them must be the two Disconnecting Networks.

Now to discuss the Disconnecting Networks.
1.) the 10r 5W resistor does not pass power. It only passes leakage current to chassis. A 1/4W resistor is sufficient here.
2.) the 100nF capacitor only passes leakage current to chassis.This can be a low voltage type. A small 50V ceramic is sufficient.
3.) one can add a switch across the DN. This switch will be in parallel to the other components in the DN. When the switch is open the DN operates as expected. When the switch is closed the DN is shorted out and you have a direct connection from chassis to the Main Audio Ground. This switch could be termed a "ground lift" switch.

NOTE this ground lift switch does not break the Safety Earth to PE, nor does it break the Safety Earth to Main Audio Ground. The switch is "safe" in both positions.

[carcass]

Originally Posted by AndrewT

Now to discuss the Disconnecting Networks.
1.) the 10r 5W resistor does not pass power. It only passes leakage current to chassis. A 1/4W resistor is sufficient here.

Hi Andrew, I though Rod Elliot suggested the 5W because of leakage currents up to 5 Amps...?!

see here

[fooboo]

Hi All

@Andrew. Thanks for the clarification. Much appreciated.

@Daniel. I know. They are only car amp chips. At this point, there is much cogitation ongoing, I may well save the ali shells for my LM3876TF/LM3886TF. (Rapid no longer stock the **76 nor does Cricklewood but they do have the **86). RS do stock the **76 though.

But the implementation will be much the same bar the single large toroid for the **76/**86. But what I learn here is as applicable whichever way I go.

Cost wise it's not that bad. More importantly for me I am having fun in the practical and picking up some of the theoretical.

Air Vents! Good call. I have changed the wood struts to metal mesh and opened up the base of the front and rear panels. This should allow circulation. I can always add a larger mesh covered opening in the back panel if required.

There is a chance that once complete this pair of amps may segue sideways and be pressed into service in a prop build.

regards

Fooboo

[danielwritesbac]

Originally Posted by AndrewT

. . . A 1/4W resistor is sufficient here. . .

Sir, the average quality of resistors has been dropping dramatically. We have no way of knowing if the 1/4W resistor is of the expected durability. The availability of 1/2w resistors is also affected by this variance. They might wear out, and then the Ground Loop Breaker may not work exactly as expected.

This could cause an inconvenience in the same way that 60/40 solder can cause an inconvenience--looks like it should be working but there's poor or absent conductivity, or later on someday, there's poor or absent conductivity and no visible problem.

So, perhaps a 1w, 3w or 5w resistor? The quality is normally useful.
And, what sort of solder were we supposed to be using?

Thanks!

[danielwritesbac]

Originally Posted by fooboo

. . . Air Vents! Good call. I have changed the wood struts to metal mesh and opened up the base of the front and rear panels. This should allow circulation. I can always add a larger mesh covered opening in the back panel if required. . . .

I'd like it if you maintained the non-conductive front panel. I'm Not in favor of either a voltage or a reference for fingers.

P.S.
Your car chips are single rail. One dual secondaries transformer is what you need. But that should be used for regulated power, with two rectifiers (plus caps) and Two V+ Regulators. Considering the isolation at the transformer and a regulated supply individual for each channel, that is a dual mono power circuit--a very good dual mono power circuit for single rail amplifiers.

Split rail amplifiers would use dual transformers for dual mono, but are you building a split rail amplifier? No. Not with TDA1562Q.
Good spendy audio enhancements include Regulators and Low Impedance caps. Those two things really help car chips. They're awful on unregulated power no matter how high extraneous transformers could be stacked to do the job of an ordinary brick.
Sure, that's a financial aspect rather than purely electrical; but, whatever you do, please remember to use regulated (or capmulti) power with your car chips.

[AndrewT]

The ESP article talks of a "few milliamps" of earth leakage current.

That leakage passing a 10r resistor will give rise to a few tens of milliwatts of resistor dissipation (Daniel, you're away in dream land).

I see the diagram now includes the ground lift switches. As far as I know this is the only safe way to incorporate a ground lift switch. Some combinations of equipment will sound better with the switch closed, while other may better suit the switch open.

I don't mind a master fuse at the main entry into the amplifier, it does offer more safety the closer it is to the input location.
But I recommend that each transformer is fitted with it's own close rated fuse.

Although it's only a drawing, could you show the main Green/Yellow wire from the mains cable going short and direct to the chassis? The DN can still connect to chassis as shown at some remote location. The chassis will pass the Fault Current.

[AndrewT]

Daniel.
Work as expected?
What do you expect of the DN?
It must survive long enough while passing fault current to PE until after the fuse ruptures and the arc extinguishes. It has no other purpose.

The resistor and capacitor have been added at this convenient location to introduce a low impedance route for HF artifacts. This might not be the optimum location to provide that low impedance path.

Now, when or if 30mAac of leakage current passes through the DN. The voltage across the resistor will be ~430mVpk. The diodes might just be starting to pass a few uApk.

If the leakage were to increase to 150mAac then the diodes will pass most of this, because they will limit the voltage drop to ~ 600mVpk and thus the 10r resistor will pass ~ 60mApk at the extreme voltage of the AC waveform. The other 90mApk will pass through the diode/s.
When the AC leakage current drops down to ~60mAac then the diodes pass a very low proportion of the leakage current and the resistor ends up having to pass the majority of this. The maximum dissipation is ~ 0.5V * 0.5V / 10r ~ 0.025Watts. Just 10% of the dissipation capability of a 1/4W resistor.