NOS USB Possible Fault Condition - take care Gain Clone users!

[Scroobius]

I am not sure if this is a problem unique to me and my NOS1 but Gain Clone users with no input DC blocking capacitor beware.

With NOS USB remaining powered on I have re-booted my PC from time to time. A few of those times (about 3 in all) there has been a loud hum from one of my speakers - leading me to switch the amp off very quickly!!!!. Today I took the lid off NOS to investigate further and the cause of the problem is that the DC offset on one channel (as measured on the internal volt meters) sometimes goes off scale i.e. a very large DC offset during the time that the PC is powered down until the point in the reboot sequence when the Windows log in screen reappears at which time the DC offset returns to normal.

Obviously this could have serious consequences because a large DC offset on a GC input where there is no input DC blocking capacitor would very likely fry the GC chip.  Indeed this is exactly what happened to me before NOS1 was upgraded & I did destroy a GC chip (lots of smoke!!). This was  also preceded by a loud hum. Although I did not check DC offset voltages (too busy panicking!!) it does seem likely that the same condition occurred then.

Any ideas why this is happening Peter?

P

[PeterSt]

Well, no Paul. And although you could not see me thinking it over while talking about such a possible thing via email earlier on, I really did. However, things have changed a little, because now you can instantly see the offset ...

I can see two reaons :

1. the DC is fed by your gainclones;
2. the DC gets out of order via the USB shield ground being influenced by your PC (via-via-via).

I can't tell which of the both is the most likely, because it is a complex thing. Fact is : the DC Offset itself won't create the hum. That is, not hum as such as you describe(d !) it. A kind of "static" low frequency ticking perhaps, but nothing like hum.

As we talked through already, the ground schematic of the GC should not be underestimated. I mean, many "connections" which will work in normal other situations, will not work here, and we can well say it is all (or always) about oscillation occurring.
Now look at option #1 above and you already see that happening. Thus, first the DC comes from the GC and next that is fed back by the DAC itself. How that becomes "hum" (if real hum at all which I can't tell), I don't know, but in general oscillation *will* be hum. It depends a bit on the speed of the feeback I guess.

Option #2 is far more complex (to sort out) because it is out of our control. It is the PC doing it, and we won't know how the shield is connected in there. One thing I know : the normal USB gnd already will sneak into your PE (protective earth) and from there everything an happen.
Well, maybe I shouldn't repeat all what has been said about it via email already, because you already know (what culprits can be).

In general we can wonder how good specs of GC's are, and then I mean the specs how to build them; That is, I can't imagine any general good specs for them because you will not be able to mimic the lab environment. And the other way around of course.

What you could do is this :
Sw#2 on the NOS1 cuts of the power to the DAC section (not the PSU of it). Switch that off before rebooting and it should help. *Still* that is no guarantee because when the GCs are first to send out "oscillation" (do strange stuff on the gnd, that implied by USB via PE), but at least nothing will fight back.

On this last matter it might be an idea to think over what actually happens when one device (the NOS1) tries to control the DC Offset, while another (the GC) disturbs it. That too is a form of oscillation, although it will be a relative slow one (but still in the mean time all is out of order).

Let me lastly add for now that the NOS1 itself is again totally different on the gnd, DC and DC correction stuff (including the USB relation of course), so if the same is happening to you as before, I find it hard to believe the NOS1 is related at all. But, no guarantee.

Peter


PS: The situations that my own GCs exhibited a LOUD hum are numerous. It can happen by any means of "wrong" (??) connection, up to putting a TVC in front of it with its gnd switch in some (??) wrong position. Mind you, transformers (TVC) lift ground (even signal) ...

PPS: I too can have a louder (and deeper) tick at rebooting at shutown, and a far more shot higher pitched tick when the PC comes up again. This happens when USB gets disconnected and connects, but, this is about a sudden data burst over USB while no circuitry is in there to shut off the data input from the DAC. So, this is a small "all out of order" situation, the DAC thinking it is music, so to speak. But this is nothing like you decsribe. This again is different from disconnecting USB in a live situation when it will happen to my own NOS1 that the out of order state can stay. This *is* about gnd connections, but does NOT influence the DC Offset ...
A test for you might be to shut off the amps, disconnect the interlinks, and next pull out the USB cable (DAC and PC just running). This should not change DC Offset, although you might see a very minor 0.2mV or so brief change forth and back (only a change, and not a permanent different offset). If this is different for you, try to ground mains cabling different, so that this won't happen. Not sure what this may bring to the general "GC problem", and all I know is that all is inherently constructed such that this should not change DC Offset. Thus, if this exhibits differently there could be a general problem (to solve).

[Scroobius]

Since I made my post above I have been rebooting with NOS1 powered on but GC amplifiers disconnected. There was no problem with measured DC offset. Obviously I will have to do many more reboots to be sure but it does suggest that the problem could arise from my GC's.

Interesting that with GC's connected that the problem does not occur at every reboot. Also interesting that there is a hum as you say that suggests oscillation maybe - the fact that one volt meter is showing over maximum does not necessarily mean that it is just a big DC offset but it does show that the voltage is out of control (whatever it is).

The general layout and earthing arrangement of my GC's is totally standard and exactly according to Peter Daniels requirements - as close to a perfect star arrangement as it would be possible to get.

I will make some more investigations.

P

[PeterSt]

Hi Paul - one more remark :

The general layout and earthing arrangement of my GC's is totally standard and exactly according to Peter Daniels requirements

I wonder, since I am merely talking about *outside* of the GainClones. This is PE or *earthing*. This is very different from *grounding*.
Not sure, but it looks like you mix the both up; Towards me ? no big deal. But maybe towards Mr Daniel.

Of course PE is related to the inside of a cabinet as well; just think about whether you connect PE to signal ground or not. Or PE to the cabinet and the cabinet to signal ground. Or no PE at all. Signal ground to the cabinet or just not.
I doubt it whether this is all (all !) prescribed ... because it just can't. You will be dependent on your environment.
Normally maybe not a big deal, but with GainClones one has to be cautious ...

Peter

[Scroobius]

Hey Peter - well I don't think there is any confusion really (except possibly in the way I described it here) - in my amp (bi amp so 4 amps) the main "ground" connection is a star "ground" point that has 4 connections one from each ground point on each amp board plus also the ground point from each of the power supplies (4 x) on the regulator boards - so 8x connections to the star ground point. In my set up the mains "Protective Earth (PE)" is also connected to the star electronic "ground" point and this is also connected to the chassis. So mains PE and electronic "ground" are connected together and that is the way Peter D recommends it.

BUT I previously tried using an earth break arrangement and in that case the electronic "ground" point was connected to the mains earth point via a 100 ohm resistor with a small cap in parallel. That did not work as the mains hum from the speakers was too high. Connecting the electronic ground direct to the mains PE reduced the mains hum significantly.

So I guess I could try disconnecting the electronic ground point from PE or try earth break again but then would it be floating? I seem to remember that NOS1 electronic earth is not connected to PE (except by a capacitor). But then there is that USB connection which could be connected to PE via the computer.

Maybe that could work but I have to say I am nervous about experimenting with earth connections given what has happened before.

Mmmm ideas would be welcome as always.

P

[PeterSt]

Hi again Paul,

No, please don't change anything on my (implied) advise. I just can't have such an advise. The only thing which may be helpful is what I told you per email : you still experience some hum, and this is a first indication something is not right.

"Not right" sounds maybe too hard, but I see no reason why your GCs exhibit hum while mine are dead quiet (with the 115dB sensitive speakers). Hum already is current somewhere which should not be.
"Too hard" because I'm fairly sure everybody will hear at least "something" at full gain and these kind of speakers (which is an easy 20dB more output that already good at e.g. 95dB !). That is a lot ...
From this should follow that your "hum" (remember, I'm talking about any normal situation for you now) is 20dB higher than mine, *if* I would have had hum also. But I have "nothing". So better think in the areas of 30-40dB or something, and *now* you will believe things must be able to change.

Ok, we can keep on talking back and forth for quite some longer (following what we already did by email), but it is and remains that I can't give the general advise. It is interesting (I think) but it won't help I'm afraid.

One last thing :
At this moment my GCs are not connected to PE. NOS1 is though, but nothing for signal is connected to that (thus, NOS1 chassis is connected to PE). But, PC  connected to PE (hey, I actually never tried without), and USB shield is connect to NOS1 chassis. Thus, NOS1 chassis is connected to PE while PE's "ground" connects to PC as well. Still GCs are not connected to PE which makes all prone to let travel noise through the interlinks because GC is floating. NOS1 is floating just the same, and now it will be a matter of how large the difference in potential is between the both. This is out of our control and the PC adds to it.
While something like the GC is not PE connected, it still is connected to neutral, while neutral virtually is bonded to PE. This implies that anyway PE is drawn to some other state then really "neurtral". It is here where your mains sine will stop to be that, because things start to be out of phase.

Shall I continue ? Better not. It is totally out of real control. Notice though that the not-so-nice-sine already is a source for (mains) hum at frequencies other than 50Hz (in your UK case). But, now it depends on PSU's and stuff to what degree this is covered for.

[Scroobius]

Peter - thanks I appreciate your help and also that this is not your problem. My mission - no hum even with ear one cm from speaker   

Watch this space

P

[Scroobius]

Ha ha he he  -   I took the plunge and changed the grounding arrangements in my GC. So electronic ground point (GP) is now disconnected from PE and two GC stereo amps are totally separate (ie the GP's not connected) - normally I would not be so reticent to experiment with earth arrangements but recently I blew up a  GC chip (due to offset voltage during reboot and possibly (?) due to earth currents (?) but I guess I will never know that for sure because as Peter states elsewhere earthing is complex and beyond the comprehension of mere mortals.

The result - almost total silence from my 98db speakers - so quiet in fact that I thought something was wrong until music played through them. If I place my ear right next to the speaker cone it is really difficult to hear any hum but there is the very faintest hum - really it is at an incredibly low level. Difficult to imagine it could be improved upon.

So thanks Peter and also thanks Nick for your help via PM.

I will post a picture of my final GC amp layout for anyone interested.

Thanks Peter - your help was over and above the call of duty and I do really appreciate it.

P

[PeterSt]

Ahaaa ...

Well Paul, never mind that noise or hum being almost out of the way now. Just try to "interpret" how this influences the music playback. Not that you would hear that noise/hum through the music before, but you know what I'm always saying ...
(no guarantee for a change though).

In the mean time keep caring your actual problem, because this doesn't tell that will be out of the way now. But who knows ...

Regards,
Peter

[Scroobius]

Well I have now had a chance to sit down and have a proper listen and there is a noticeable improvement. Particularly noticeable at the top end there is more detail or at least it is clearer and better delineated tidier I guess would be the word. Tiny tiny details (very fast little transients) can now be heard better - in fact to properly hear what is being reproduced requires really quite background and possibly a glass of scotch (but I have not tried that just yet).

This really highlights just how much detail is being produced by NOS1 and just how much attention needs to be paid to everything that is downstream for it to be heard.

I was exchanging emails with Nick today and he always turns his amp and DAC off when he reboots because he has had problems with off set voltages in the past. I rebooted today (with amp turned off but not disconnected) and had a big offset voltage again on one channel. I will continue to try and find out what is going but in the mean time NOS Insurance will stay in place http://www.phasure.com/index.php?topic=1734.0 and the amp will be turned off during a reboot.

The next upgrade is to have my ears syringed 

All the best

Paul

 

[Scroobius]

A note for Peter - since "floating" the earth ground arrangement in my GC (i.e. disconnecting electronic ground from PE and separating the ground of the 2 left amps from that of the 2 right amps) I (so far) have not had any problems at all rebooting - I have rebooted about 15 times with the GC and NOS1 powered on & no problem with offset's (also checked on the little volt meters). Of course I may have been lucky so far but previously about 1 in 5 reboots would have the offset problem. I will report back again after more re-boot's.

But also the sound quality is just so much better that I have to wonder if there was another underlying problem. Can't be sure? and I will probably never know but I have to wonder if just reducing background hum could make that much difference or maybe there were other current flows between equipment that clobbered performance or something else. Music now comes out of an inky black silent background. Sometimes I still wonder if the amps are actually switched on and have to check. Interestingly the amount of "hiss" or white noise from the tweeters reduced orders of magnitude along with reduced 50/100hz hum. But to get this in perspective whilst there was hum before it was not that much and about the same I have had in the post from commercial amps.

Man does it sound good now  - maybe for the first time I am hearing this system working properly. I sit down listen to music and just do not want to get up. My wife is asking for music to stay on because she is enjoying it (that is a first!!). To get this in perspective it sounds as though someone has added a sub woofer and an ultra high frequency ribbon tweeter to my system such is the improvement in the bass and the shimmering detail now heard on symbols - can that level of improvement be down to reducing hum?

Sounds stonking good now and maybe I am at last ready for Mani ("old critical ears") to come down for a proper second opinion.

Paul

[manisandher]

Hi Paul. Well I've just got to come down and take a listen now, don't I? PM sent with possible dates...

Mani.

[Scroobius]

Just had to re-boot - result big offset voltage on one channel leading to panic switch off of GC amps.

OK looks as though this one not solved even though everything is working OK and sound quality wonderful.

So care needed during re-boot - I can live with that but as it is a potential amp killer great care to be taken in future. I am sure this will only happen with amps that do not have capacitors on the line inputs.

P

[christoffe]

I am not sure if this is a problem unique to me and my NOS1 but Gain Clone users with no input DC blocking capacitor beware.

Hi,

in some poweramps, such as Dartzeel, the DC compensations should be shut OFF for a superior SQ . During all the booting and shutdown a larger deep “crack” came up for 3 times and the last one  blew the fuse of the right channel of the poweramp.

It is strongly recommended with a disabled DC offset,
- to start the poweramp AFTER the booting of the PC
- and shut off the poweramp BEFORE  a shutdown of the PC is ahead.

Joachim

[PeterSt]

Joachim, excuse me, but what does this suddenly come from ?

You propose me something, I ask you questions back ny means of three emails this monring, and without further notice or answers, this is what you come up with ?

FYI : It is not even clear *what* you come up with.
So what are you saying ? you first shut off some "DC correction" in your main amps to next have a fuse blown ?

Well, if that is the message indeed, then I suppose your advice is OK too. But maybe the real advice is to not shut off that DC correction ... (I'm not sure)

Peter