Valhalla - AKSA 55

andyr · « **Reply #120 on:** 9 Oct 2003, 12:33 pm »

Ginger, hi there.

I have just finished building my new AKSA 55N, which I modified after our postings to run an earth wire from each PCB 0V tag to a "secondary earth" on the back-plate. Unfortunately, I found this gave me a very big HUM ... when I disconnected the secondary earth wires from the earth bolts, the hum went away and I could listen to the amp.

Now, I appreciate that you said "this is what worked for me" but could you do a quick experiment ... plug your AKSA power cord into a different electrical circuit in your house (maybe you'll need an extension cable) and see whether you get hum.

You see, my setup is that I have my source components up one end of the room (on one electrical circuit) with my power amp up the other end of the room, next to the speakers but plugged into a different electrical circuit.

I think the secondary earth wire that you suggested, which effectively links the PS "star earth" point to the chassis and thus (because the IEC earth is also connected to the chassis - albeit in a different place) to the mains earth, causes an earth loop because my AKSA is on a different electrical circuit to my preamp.

If you could simulate my setup by plugging your power amp into a different circuit, I would appreciate hearing the result.

Thanks,

Andy

ginger · « **Reply #121 on:** 10 Oct 2003, 01:06 am »

Andy,

You'd think an EE who started in electronics as an apprentice almost exactly 30 years ago would be smart enough not to start a discussion on earthing.

As you suggest your problem is almost cetainly an earth loop.

Please check that your RCA connectors are FULLY isolated from the chassis and the return sides (the outers) are wired to the amplifier Input (ground side) Pins (NOT the 0V spade lug) and NOT connected to chassis locally except via the Ceramic Caps if you've fitted them. You should measure 10 Ohms from the RCA outer to 0V Spade Lugs for that amplifier channel.

If your still having problems - email me direct
[email protected]

Cheers,
Ginger

stvnharr · « **Reply #122 on:** 10 Oct 2003, 04:40 pm »

Clarification please.
Star Earth - the grounding bolt on the power supply pcb.
0V referred to in the preceding discussions - Is this the Earth spade on the amp module pcb?

What is the difference in running a wire from either Star Earth or amp module earth as they are connected by wire already?

AKSA · « **Reply #123 on:** 10 Oct 2003, 09:47 pm »

Hi Ginger,

There might be some confusion about your description of Star Earth as 0V (SRP). Do you mean Star Earth on the pcb M3 bolt?

If yes, I think Steven's question is answered.....

I'm glad you too have your frustrations with this stuff. I think it is very difficult to get your head around and I've probably spent as much time on the earthing of the AKSAs/GK-1 as any other single design aspect.

The point raised by Andy concerning running amps/preamps from different power sockets is valid; this might well create earth loops if any connection is made from Star Earth to Mains Earth. I suggest the option in the instructions of trying a connection between Star Earth and Mains Earth; sometimes this reduces hum, sometimes it increases it. This circumstance usually applies to precisely Andy's setup where amp and preamp are taken from different power sockets (or circuits!).

Hmmm. Earthing is a PITA to get right, the last frontier........

Cheers,

Hugh

stvnharr · « **Reply #124 on:** 11 Oct 2003, 04:47 am »

With nothing better to do today, I decided to do this new grounding scheme, as I did hear some hum with volume at full blast.
New ground bolt near input. Cap connected from ground on rca to bolt. Then new ground wire from earth spade to new bolt.
Didn't really change anything though. However, the hum I hear at full volume is from the cd player, not the amp.

andyr · « **Reply #125 on:** 11 Oct 2003, 05:13 am »

Quote from: stvnharr

With nothing better to do today, I decided to do this new grounding scheme ...... However, the hum I hear at full volume is from the cd player, not the amp.

Steven,

Could you please just amplify your comment .... just so we all know we're on the same page!

How did you decide the hum which you hear at full volume is from the CD player? Did you disconnect the CD player ICs from your pre-amp, so you no longer heard hum from your AKSA?

Regards,

Andy

stvnharr · « **Reply #126 on:** 11 Oct 2003, 01:36 pm »

Andy,
My cd player has both rca and xlr outs as it is a balanced player.
Hum with rca's at full volume is really there. Hum thru the xlr's is significantly reduced.
Yesterday I received a new sacd player, solid state. It is dead quiet at full volume.
Interesting to make conclusions here. As I have preferred the music through the cd player's rca outs rather than the quieter xlr's.
And with the high resolution of the Aksa and it's very clean sound, I'm only hearing a slight difference between the much higher resolution of sacd as compared to my 24 bit cd player.

My Aksa's are monobloc identical twins. And nice to know that I have no amplifier hum.

Steve

PSP · « **Reply #127 on:** 11 Oct 2003, 07:24 pm »

ginger,
Thanks for your aside on "why Rikens sound so good"... that their resistance is slightly voltage dependent and that this leads to small amounts of second order harmonics...

This suggests the possibility that one should substitute Rikens into the AKSA and GK-1 with care... taken too far or used in the wrong places, is there the possibility that "too much of a good thing" could degrade AKSA sonics?

Thanks,
Peter

AKSA · « **Reply #128 on:** 11 Oct 2003, 09:21 pm »

Hi Peter,

I would agree; this is certainly possible.

VDRs introduce asymmetric distortion, which is essentially second harmonic, and euphonious. Any more than about 0.8% will sound overly lush and cloying, and is to be avoided. However, the most influential in this connection are R1 and R9; it's hard to imagine any others would have such profound effect so you are probably OK.

Cheers,

Hugh

ginger · « **Reply #129 on:** 15 Oct 2003, 04:16 am »

G'day guys,
I've been "missing in action" for a bit, up to my ar.. er elbows in GPS an Inertial Navigation Systems for work.

stvnharr - appologies for the inexact wording, the "star earth" and the OV bolt on the power supply pcb(s) is the same thing. The reason for connecting to the amp 0V spade terminal rather than the 0V bolt on the power supply PCB(s) . (Yes they are connected by a bit of wire) is that the "bit of wire" carries power supply return currents and speaker return currents and so the amplifier 0V spade terminal is likely to be the "quieter" end. I tried both and in fact ended up leaving my 2nd chassis earth wire connected to the power supply 0V bolt purely for convenience since I did'nt notice any difference. Then my "bit of wire" is 3 inches of 12 AWG (25 Amp rated) - not enough resistance to develop much noise voltage across.

andyr - tried putting my sources onto a different power circuit - did'nt notice any difference in background hum.
Much as I hesitate to suggest it and ONLY as the last resort to get rid of your earth loop simply "float" the power amp.
That is:
DO connect mains earth to chassis ground adjacent mains input socket
DO use the caps from the RCA outers (the "ground" side) to local chassis earth adjacant the RCA input connectors.
DON'T connect the 0V power supply to the local (adjacent to the RCA connectors)chassis earth (or anywhere else).

I've always believed that this compromises safety to some degree - BUT I'm reliably informed by our safety standards/EMC/EMI expert that its legally acceptable and is not uncommon practice in commercial audio equipment. Without an Amplifier 0V to chassis ground connection the risk is that if ever a transformer primary to secondary insulation breaks down (which rarely happens) then the amplifier circuit can become "live" and fault current can flow back through the input interconnects to your other equipment, or to you if you happen to be hanging on to the wrong thing at the wrong time.

This "attitude/opinion" is reinforced by the fact that power transformers used in "double insulated" (no earth connection to either the chassis or to the electronics boards) equipment must comply with a more rigorous safety standards than ordinary power transformers.

Having said that I recently did a circuit trace on my 845 Single Ended Triode Power Amp which runs 900 Volt rails. There simply wasn't even a connection to the IEC mains socket earth terminal - a fault I quickly rectified by adding a chassis earth and an amplifier earth much the same as described in the posts above. In this case the point I chose to connect the 2nd chassis earth to, was the 0V point where the grid resistor of the first (input) valve was connected.

Cheers,
Ginger

SamL · « **Reply #130 on:** 29 Oct 2003, 01:00 am »

Quote from: Malcolm Fear

We couldn't find Riken Ohm in Australia.
We bought from Parts Connection. They also do Black Gates. They are fast, cheap and reliable.

It appear that Parts Connection is not equal to www.parts-connection.com
or www.partsconnection.com.

Can someone kindly pass the URL please.

Ta,
Sam

andyr · « **Reply #131 on:** 29 Oct 2003, 01:03 am »

Here y'are, Sam:

http://www.partsconnexion.com/

They are prompt and helpful.

Regards,

Andy

ginger · « **Reply #132 on:** 10 Nov 2003, 02:46 am »

Spent yesterday (Sunday) doing more mods to my 55N.

As I was poking around inside checking running temperatures of various components I noticed that in mucking about, that C5 (the Siver Mica which sets the amplifier dominant pole and hence high frequency roll off), the VAS transistor T3 and the bias set pot P1 had been all squashed hard up against each other.

To make it look nice (I don't know who is going to be sneaking a peek inside my amplifier) I prodded these parts with my finger to separate them physically. To my surprise I had an immediate top end improvement.

This is not so surprising in retrospect. The Dominant pole Capacitor is ONLY 47pF (standard) or 33pF (Nivarna) and so an extra 5 to 10 pF of stray capacitance due to these components touching each other can make an audible difference.

Take a look in your amp and make sure T3 is standing up nice and straight and prod C5 until its sitting evenly between T3 and P1 with equal gaps either side of it. I would guess that the T3 to C5 gap is more important than P1 to C5 gap.

There you go - who said you never get anything for nothing.

CAVEAT: I'm running a modified VAS Stage on my 55N and so the effect may be greater on my unit than it will be on yours - BUT the cost of this "mod" is the right price.

Cheers,
Ginger

andyr · « **Reply #133 on:** 10 Nov 2003, 03:10 am »

Interesssting (as they used to say on that great 70's show, "Laugh In").

I will certainly follow your advice, Ginger, but I would be interested in your comments on the following:

Should you make sure all capacitors sit off the PCB - ie. not pushed right down to the surface which is the "neat" and commercial way of mounting them?

IE. is the "stray capacitance" you referred to caused by:
* C5's inherent capacitance being affected (increased!!) by being very close to the mass of T3 & the bias pot P1 ... which would indicate that the mass of the PCB will have the same effect?
OR
* is it because the close physical proximity of these particular other components - an active device and a resistor - caused "stray capacitance" which affected C5 ... in which case, C5 being in proximity to a PCB will not have any affect?

The reason I ask is that, when measuring capacitors on my old Topward TI 111 digital LCR meter, I have noticed from time to time that bringing my hand near to the capacitor affected the readout!

Regards,

Andy

ginger · « **Reply #134 on:** 10 Nov 2003, 03:34 am »

Sitting caps up from the board increases lead length and hence stray inductance and resistance and possibly even capacitance. My view has always been to mount caps (make that ALL components) against the board with shortest leads practical. The ONLY variation to this is for resistors (and transistors) which dissipate a reasonable amount of power which I sit up a bit to allow airflow (plus the extra lead length acts as a heatsink).

Since Capacitors are made up from parallel plates separated by a dielectric anything nearby can effect its value. This is ONLY critical for very low values of capacitance. Adding 5 pF to a 1000pF capacitor obviously is Not as critical as adding 5pF to 33pF.

In this case I think the culprit is that T3 has a large adjacent metal area acting as another paralleled plate with air being the dielectric. Thats why I suggested that the T3 to C5 gap was probably more important than the C5 to P1 gap.

Stray capacitance is inevitable in any electronic circuit. Keeping leads short as possible is always the best strategy in managing strays whether it be stray L, C or R
Cheers,
Ginger

andyr · « **Reply #135 on:** 10 Nov 2003, 04:00 am »

Thanks for the explanation, Ginger .... I'm sure everyone will appreciate it, not just moi.

Regards,

Andy

andyr · « **Reply #136 on:** 19 Nov 2003, 08:07 am »

Quote from: ginger

PS: the theory above on picking the 0V point to ground also applies to working out which end of a shielded cable's shield to ground when you only want one end connected to avoid ground loops - always connect it at the "quietest/least noisy" end which invariably is the end with the smallest currents flowing through or passed it.

Cheers,
Ginger ...

Ginger, sorry to go back over old ground but a post on AA caused me to go looking at this thread again, to see what you had actually written.

Surely, on an AKSA amp, if you have put caps between the earth tags of the input RCAs and a nearby chassis earth - to 'bleed off' HF hash to ground - then it stands to reason it is the AKSA end of the shield on the IC which is the correct end to be earthed (ie. connected to the RCA plug's ground tag, together with the '-ve' signal wire)? IE. any HF hash picked up by the IC shield immediately gets shunted to earth at the input to the amp?

Thus it is the SOURCE end of the IC shield which is the floating end?

Regards,

Andy

stvnharr · « **Reply #137 on:** 19 Nov 2003, 03:20 pm »

Andy,
I know you directed the question to Ginger, but I'll put in a couple words since your question reminded me of a couple things.
In general, an Interconnect with a shield will have the shield connected to the rca connector at only one end. This end will then be connected to the source, leaving the "noise, rfi, etc. to just "float" in the shield. This is generally the reason that interconnects would have arrows for direction on them.
Got that from Kevin @ diycable.com.
But, with the rca grounded in the amp, the IC could be connected "backward" with the rfi and noise, if there is any, being grounded right there in the amp. Although, that would tend to negate the shield.
I use braided IC's to cancel any noise. Also, have grounded the rca to a chassis bolt after reading Ginger's post.

stvnharr · « **Reply #138 on:** 19 Nov 2003, 03:42 pm »

Just read that cable post on AA.
I just repeated the standard reasoning on cable shielding.
In audio everyone seems to try everything every which way and then decide which is best. Certainly seems that way with cables.
However, I have never read a plausible reason for cable directionality other than the floating shield. An electrical signal in a copper wire shouldn't really be directional. But then again..............

andyr · « **Reply #139 on:** 19 Nov 2003, 08:03 pm »

Quote from: stvnharr

Andy, ....
#1. In general, an Interconnect with a shield will have the shield connected to the rca connector at only one end.

#2. But, with the rca grounded in the amp, the IC could be connected "backward" with the rfi and noise, if there is any, being grounded right there in the amp. Although, that would tend to negate the shield.

#3. I use braided IC's to cancel any noise. Also, have grounded the rca to a chassis bolt after reading Ginger's post....

Steven,

Re 1: Yes, I agree an IC with a shield should only be connected at one end. (If it is connected at both ends then the braided shield becomes part of the 'return/earth' path ... and you now have multi-stranding on this path.)

Re 2: I understood that the whole point of Ginger's recommendation to put in a 'secondary earth' point was to force any RFI and noise picked up in the shield to ground at the input RCA, so it did not enter the electronics of the amp. This does not 'negate the shield' ... it uses the shield for what it was designed to do.

My question to Ginger, in effect is asking "if the shield is grounded at the source end, does this get rid of any HF/noise picked up along the length of the shield?"

Re 3: If your braided ICs cancel noise, I would have thought there is no benefit in using the Ginger-recommendation of a ceramic cap between the RCA earth tag and the secondary earth bolt. (Coz you shouldn't have any noise to ground!)

Regards,

Andy