Earthing question.

[Diabolical Artificer]

Just built two "monoblocks" with a separate PSU, so there is no mains in the amps, do they need a mains earth? The amps are wooden boxes with a ali plate to mount all the gubbins.

The voltage rails are connected by yon big military connectors, and at present a 3ft earth cable bolted to each ali plate, problem is I have an earth loop hum & N increases from 5mV to 50mV and sounds awefull.

Andy.

[Radio Wrangler]

You have several aspects to consider:

Safety
Legality
Sound performance

Relying on a separate earth cable is dodgy. There is the possibility of the power connector from the power supply being fitted with the earth cable left off. This must not create a dangerous condition. So you wind up having to have an earth connection included in the power supply connector. Voila! instant hum loop.

It's easy to say that it's my amplifier and I'll do it as I choose, or that no-one else will use it and I know to set it up properly. This may well be true, BUT if something does go wrong and someone gets hurt, you can wind up having to convince a coroner/FAE that EVERYTHING you did was reasonable and that you took all necessary precautions. They'll want to see documentation from the time. This is a very unlikely situation, but the difficulties it would drop you into are the same as for a commercial product. Time machines don't exist, you can't just pop back and change what you did, so you have to be confident that everything you do is right at the time. You can try to keep visitors, kids, and pets away from the risky bits, but the clueless have a natural affinity for going straight to the dangerous bits, defeating various safety measures on the way.

For homebuilt stuff, you don't have to jump all the safety testing hoops, but you still carry the same responsibility should the worst ever happen. So guard your back ahead of time.

So safety consideration pretty much force you into earth loops and hum as a consequence. You can't do much to the Earthing that makes any improvement, so you attack the signal paths. Use a balanced differential signal path, and make sure that the gear on the receiving end has plenty of common mode rejection. The BBC used to use isolating transformers in signal feeds as a matter of general housekeeping, and the same is true in most good quality studios.

So, what do you do?

First of all, amplifiers with integral power supplies give you appreciably more control over interconnections and how they are arranged. This makes the safety aspect a bit easier, and makes signal quality issues a bit more stable and repeatable. All upsides so far. Is there any downside? Yes! separate hunky power supplies are seen as more impressive by many audio aficionados. Whether or not their ears hear any difference, their eyes see a difference and so they want/expect to experience a better sound from the more impressive looking gear.They seem to overlook the worse hum...

Secondly, why monoblocks? They again multiply the number of earth loops and do they really give you any better sound than the same designs sharing a common case would have given? They do give a benefit when the total power input level starts to exceed the capabilities of a 13A plug, say 3kW, then you might want separate mains cables. At lower total powers, if having a common mains cable and plug creates any real disadvantage, then there is something wrong with the design of your amplifier.

The single power cable amalgamates several safety aspects and reduces the number of loops you have to handle. You have to get alll the loops fxed before the hum goes away. loops interact so fixing one loop can undo all your work on one you did earlier. Reducing the number of loops makes a big difference in difficulty.

So, the old Leak Stereo 20 got several big structural things right. No doubt someone somewhere has sawn one in two to make monoblocks and is holding court somewhere on the internet about his 'mods' to the old lump. Maybe he went further and hoiked the power supply stuff off onto a separate chassis, duplicating it for the other channel.

Bet it hums!

There ARE ways to get monoblocks working well, even with external power supplies, but it's an awful lot of work and involves decisions that have to go in early in the design process.

The advantages accrued from the pure monoblockishness of a good monoblock amplifier compared to a good stereo are essentially in the mind of the beholder. They got there via the eyeballs not the lug 'oles. But the disadvantages like hum and fridge clicks are quite real.

The setup in my lounge has a separate preamp and power amp with 600 Ohm cable between them. The power amps have differential inputs, though very much unbalanced. Two power amps live in one cabinet. It has one mains cable. There are separate big power transformers for the two channels, and they share a small transformer with 4 identical and separate. secondaries for some ancillary supply rails.

Overall, there is a little hum, but you have to stick your head right in front of the bass driver to hear it. It does have an earth loop, but it's planned for and managed without signal transformers.

Both the preamp and power amp have solid connection of chassis metal to mains ground and take a 25A PAT bonding test with ease.

David

[Hartley118]

600 ohm cable between preamp and power amp?
Pretty massive lounge?
Those pesky termination mismatch reflections can really wreck audio transparency.

Martin

[Radio Wrangler]

...still one ought to please any passing audiophool.


with apologies to M Flanders esq.

David

[ms660]

I like pictures.

A link to the whole article (if it works):

https://books.google.co.uk/books?id=...cience&f=false

Lawrence.

[knobtwiddler]

Signal transformers are truly wonderful things. Modern ones from the likes of Jensen and Lundahl have 3rd harmonic that's below -90dBV even at sub-50 HZ levels (there is one Lundahl type that uses a 'negative impedance driver', which gives 3rd harmonic at below -110dBV @ 20Hz, albeit at a slight compromise @ 1KHz).

If you don't want to pay for new transformers, they're easily found on Ebay. The LL1517 uses the negative-Z driver, and it's often seen in broadcast surplus items.

Transformers will give you your cake - and you'll be alive to eat it on account of not worrying about safety earth creating loops!

[GrimJosef]

Just make sure the signal transformers aren't close to a power transformer leaking a lot of flux at 50Hz, 150Hz, 250Hz ... Otherwise the shielding has to be very good if you want to avoid the hum coming back.

Cheers,

GJ

[PJL]

I assume the monoblocks are earthed through the PSU on the same 3 wire cable/connector that provides the power and there is a single earth connection to mains earth from the PSU?
This is not the normal cause of earth loops which occur when equipment that are separately earthed are connected together with a shielded signal cable.
Am I right in thinking that the aluminium plate is not used as signal ground? If so this is probably pickup and not an earth loop.

[Herald1360]

Originally Posted by ms660

I like pictures.

A link to the whole article (if it works):

https://books.google.co.uk/books?id=...cience&f=false

Lawrence.

Rabbit hole time!

[GrimJosef]

It might be a 4-wire cable (2 x heaters, HT, 0V/GND).

A ground loop can arise because signal ground is connected to 0V/safety ground in each monoblock and is also connected again at the audio source. The 0V/safety grounds from the monoblocks are connected at the psu. So the ground loop consists of the audio cables' signal grounds and the power cables' 0V/safety grounds.

The possibility for a different, more usual, loop involving the signal ground(s) and the mains supply safety earths to source and psu also exists, of course.

Cheers,

GJ

[stuarth]

If the PSU for the amp had an appropriate transformer, eg a separate primary bobbin, then does the HT 0V need to be connected to mains earth? The whole thing might then have to meet double insulated standards though.

Stuart

[GrimJosef]

Originally Posted by stuarth

... does the HT 0V need to be connected to mains earth ...

In principle, no. But if the amp is not of the 'balanced' variety then the HT 0V is very often also signal ground. For safety reasons, accessible metallic parts should be connected to safety earth (unless the equipment is Class II, which used to be called 'double insulated'). This can conflict with using a phono socket as the input connector, since those tend to have the outer, which is generally accessible, connected to signal ground. If negative feedback is taken from the same output transformer secondary as drives the speakers, one of the speaker terminals tends also to have to be connected to signal ground. Since speaker terminals are also accessible it could be argued that again signal ground has to be safety earthed.

One solution can be to connect signal ground to the safety earth via an 'earth lifter'. This will have sufficient impedance to reduce the current circulating in any ground loop to a level low enough that it doesn't interfere with the audio signal. It's difficult to find a compromise between, on the one hand, safely allowing foreseeable fault currents to trip whatever circuit protection is in place and, on the other, still maintaining the required high impedance. The best solution is a pair of very high-current silicon diodes wired in anti-parallel. When the voltage between signal ground and safety earth is less than 0.6-0.7V, as it can often be in ground loops, the diodes' impedances are very high and little noise-inducing current flows. If a fault should develop, though, which would normally raise signal ground to a dangerous voltage above safety earth then one or other (or both, with AC) of the diodes will conduct and limit the voltage difference to <1-2V, passing whatever fault current is needed to do this. If that fault current is high enough it will blow the fuse or trip the RCD as it would if signal ground were solidly connected to safety earth. The important thing is to ensure that the diodes are robust enough to survive the fault current for long enough for the fuse/RCD to act.

Cheers,

GJ

[Diabolical Artificer]

Thanks all, some clarification.... The power supply consists of a toroid with several secondaries, after rectification etc OP's are -

+500v
+360v
+300v regulated
0v ground
-100v
these go to the "monoblocks" via a 6 pin connector, the 500v goes separately via it's own double insulated cable and connector. Separately there's a heater connector consisting of two 6.3v AC. The amps ground is grounded in one place, as per standard practice, next to the IP valves base.

At present the mains earth enters the PSU via an IEC and is bolted to the PSU top plate, 2mm ali. Off the same bolt I've run two earth wires to the two "monoblocks" where each is bolted to the amp top plate. See pics.

From all the replies the answer is yes I need a mains earth. There's a few things I can try like "lifting" the ground using back to back diodes/a 10r resistor/ 100n in parallel. There is also provision by the use of an XLR 3 pin IP to use balanced IP's, though here I'd need a unity gain balanced buffer.

Two other things occur - I didn't ground the PSU ground to the PSU chassis and I was using a SMPSU to power the delay start relay; I have a suspicion the hum/noise I was getting may be in part due to it's use,though not sure quite how at present.

Thanks for all you suggestions and idea's, Andy.

[ms660]

Originally Posted by Herald1360

Rabbit hole time!

Sometimes Google book links don't always get to the right pages, did the link work?

Lawrence.

[Diabolical Artificer]

The link works Lawrence, I think Chris was referring to chasing down hum & N is a rabbit hole. I'll read the article when I get a mo, thanks for the link.

Andy.

[Herald1360]

Actually I was referring to the other magazine articles after the hum control bit

[GrimJosef]

Originally Posted by Diabolical Artificer

... I think Chris was referring to chasing down hum & N is a rabbit hole ...

Too right ! The worst part (with 50Hz hum) is when you spot a change which should make the hum better, so you spend an hour implementing it only to find that things have got worse !

The problem was that there were multiple ways the hum was getting into the circuit and, depending on where and how each one was working, their phases could be opposite. So you've spent an hour eliminating one of them, but it turns out that one was largely cancelling another one. Before the change you only had the difference between them, which might have been small. Now you're left with the whole of the other one, which might be large. It's enraging .

You'll get there in the end though. If there's a baseplate don't forget to screw it back in place before making the measurements. It can make a difference too.

Cheers,

GJ

[ms660]

Originally Posted by Diabolical Artificer

The link works Lawrence, I think Chris was referring to chasing down hum & N is a rabbit hole. I'll read the article when I get a mo, thanks for the link. Andy.

Cheers, got it now, was only familiar with rabbit hole and a certain type of knot.

Lawrence.

[Radio Wrangler]

Originally Posted by GrimJosef

The problem was that there were multiple ways the hum was getting into the circuit and, depending on where and how each one was working, their phases could be opposite. So you've spent an hour eliminating one of them, but it turns out that one was largely cancelling another one. Before the change you only had the difference between them, which might have been small. Now you're left with the whole of the other one, which might be large. It's enraging

Leaving Alice chasing the white rabbit, we've moved into the world of Laurel and Hardy... that scene with them shoving a drawer in on a chest of drawers only for another one to be pushed out by it... and on and on...

David

[joebog1]

I read this post yesterday, and left it till today to comment ( NOT answer!! ). In my build I have thought about it and come up with this idea, dont know yet if it will work.

Mains comes in through standard IEC socket ( steel chassis) and is safety earthed at the main earth point, which is also the 0V earth point of my HT power supplies. Power supply is dual mono, as is the amplifier.

Umbilical cables ( yes, two of them ) are exactly the same length, and run through two separate output sockets to the main amp, which has a common bus wire ( 2mm sqr ) running the length of the amplifier chassis. Its power supply earth and signal earth, and connected to the amp chassis at the inputs. Umbilical cables have 2X 12 volt heater supplies. One for the output bottles, and one ( elevated to 50 volts ) for the input bottles. Neither is earthed YET. I will determine which way I will earth them when I finish both the the amp and the power supply, and switch it on.
Umbilicals are identical and as discussed 2X12 volt AC heater supplies. (four wires 14/.2mm
one HT line ~ 425 volts
one HT return line ( also common earth buss inside the PSU, and inside the power amp )
one negative 105 volts bias supply, which uses the common earth buss as its return.

The 12 volt AC supply for the 5B/254's is centre tapped and "may" be returned to earth via the CT, OR I may elevate it also to reduce hum if I have any.

The 12 volt heater supply for the input bottles, as explained is raised about 50 volts positive to protect the cathode follower drivers. ( 12BH7A ).

Why this long explanation ?, my power supply is safety earthed and my power amp is safety earthed through the common earth buss inside each chassis.
That covers the safety aspect.
Its also ties together all the signal earths via the same earth buss inside the amp, and via the umbilical cords to the power supply common earth in the power supply.

In my mind it looks like one only earth point from the power supply 0V to the signal earth.

Joe