advertisement


NAC82/NAPSC recap question

C

Cesare

pfm Member
I recently got hold of a somewhat poorly NAC82 (thanks Tony!) which is in need of some TLC. Basically, it produces some rather weird effects which i'm in the process of tracking down.

So i'm going to stick my progress and questions on my journey of discovery in this thread.

First up, the unit looks to be very clean, and no obvious problems or signs of previous work, so it looks like a pretty clean unit. The serial number puts it in 1999 so a later one with the POTS-8 board.

Sticking a 1Khz test tone through and adjusting the output to -12db gives me this sort of exciting trace:

Screenshot 2020-03-08 at 17.14.56 by Cesare Ferrari, on Flickr

This sort of 'glitch' happens all over the place. The amp is fitted with NA523/1 boards, and weirdly, removing these makes this problem go away. The problem is associated with the left board - only fitting the right and all is fine, so i'm assuming there's an issue with this board which i'll need to investigate.

Further testing, with the boards removed showed a very weird issue - this is the same 1Khz/-12db test tone with the fundamental and some 50Hz bleed masked out showing the noise floor:

Screenshot 2020-03-08 at 17.19.03 by Cesare Ferrari, on Flickr

Those spikes in the noise floor don't look right, and on a hunch I plotted the same test against the NAPSC power supply input voltage:

Screenshot 2020-03-08 at 17.04.32 by Cesare Ferrari, on Flickr

So i see spikes in the NAPSC voltage at the same time as the spikes in the noise floor - hmm, not sure if this is cause or effect, but my instinct tells me to recap the NAPSC and see what happens. On inspection the NAPSC 2200uF capacitor looks a bit swollen, so it probably needs doing anyway...

So, is this likely to fix the problem? And any advice about where to go next if not to resolve this issue? Also, if there's any thoughts about what components are sensible for a NAPSC/NAC82 recap, i'm all ears as this is not something I routinely do, but i'm perfectly able to solder some components without lifting tracks :) Actually, i'm probably tempting fate by saying that...
 
http://www.acoustica.org.uk/t/naim/powersupplies.html
The napsc is decribed here under point 3.
Is said to only power the digital circuit (remote control, switching, lights) and probably has no connection to the audio circuit.
The napsc is feeding 18V DC into the 82, further regulation is cared for there.
Might be a damage in the napsc could result in damage to the regulators in the 82 ?
But if you'd hear anything of this is beyond me.

Regarding the phono cards, you could take them out and connect the line pins with a wire,
thus making them a line-in...if the problem is gone, it's the card that's knackered & the pre is ok,
if it's still there the pre has a prob on that channel and the card can still habe a prob, too.

or you crocodile-clamp the left board to the right channel pins and see if your wheirdo scope trace
is going with the card or with the channel.
 
Before recapping it would be best to try and find the actual fault first.
Check the power supply from the transformer the noise could be a iffy rectifier or regulator, could be some oscillation of the rectifier, check all caps in the power supply.
Check the power supply voltage to each 523 board.
Trace you 1K signal through the circuit until it goes bad
If you have one good and one bad 523 board remove them both and carry out a series of checks comparing one board to the other looking for a difference in the left hand board?.
Carry out a diode check on all transistors & Diodes, carry out a resistance check on all resistors, ok you are not going to get the correct resistance in circuit but you are looking for a difference between the 2 boards, carry out a capacitor test again you wont see the correct value in circuit but there may be some unusual readings or short circuit

Alan
 
FWIW the voltage I measured from the NAPSC was a little under at about 16V without load. I did mention this in the sale thread, but no idea if it is a factor or not.
 
According to Naim’s Power Supplies manual, the NAPSC secondary supply (off load, presumably) should be 18.5V. However the manual is a recent version and that might be for a NAPSC2.

Edit: I’ve just disconnected my (old style) NAPSC from my 82. The Vsec at the connector was 16.68V.

I have a (new style) NAPSC2 on order, I’ll measure that when it arrives.
 
Whilst the power from the NAPSC is only for non-audio supplies, it still has a common 0V with the rest of the amp which could be an issue.
 
Ok, thanks for some pointers. I'll make some better measurements on the NAPSC, but I think i'm convinced now that it can't possibly cause the sort of glitches i'm seeing, and it's most likely that something is affecting the 0v and i'm just measuring it in two places. The glitches aren't equal on both the left and right channels though, some are correlated, but the left channel seems to get it worse:

Screenshot 2020-03-09 at 11.34.10 by Cesare Ferrari, on Flickr

So the next question is what sort of fault can lead to these sorts of spikes on the output - is this just a dodgy capacitor, and a recap will fix it?

Of course another option is it's a measurement mistake and there's nothing actually wrong as my setup isn't exactly an RF clean environment...
 
'as my setup isn't exactly an RF clean environment...'

@Cesare - while thinking about your problem, easy check would be to leave the kit all set-up as you had it for teh above, and short your test probes - record that output and see if its anything like the above. Yes, it could be a compact florescent light, or fed-through CM noise if your measurement kit is on an SMPS plugtop supply on one socket with the device under test plugged-in elsewhere on an available socket, not the same socket via and adaptor/4-way lead etc... Yes- this stuff is a nightmare.

Secondary thoughts - there's not much in the way of useful supply rail decoupling on the phono boords at all (adding some is one of the easiest ways to get a little more out of them) So I am thinking you could be seeing glitches from the phono boards merely reflecting something fed-through from elsewhere.

If your measurement kit has an SMPS, the fact the NAPSC shares a 0v bond with the 82 analogue side might just be enough. If it's a laptop, try measuring while running off the internal battery; if its a USB-like 'scope' , try off a 5v powerbank... or all of the above. HTH.
 
Thanks Martin - you're thinking along the exact same lines as me, that something external is getting in and causing problems. I've just had the setup downstairs away from everything likely to cause problems, and measuring on a laptop running on batteries, I get exactly the same results as before, so I can basically rule out this as being an external effect and it's probably real. The left channel is still noisier than the right, so that's totally consistent with the problem as seen before.

The NAPSC does measure a little low, around 16.9v in circuit powering the preamp.
 
Will your rig allow simply recording the voltage across the reservoir cap of the NAPSC? That's where I'd go next.
 
The digital circuitry in the 82 has 2 regulators inside, IIRC they're set to 12 and 12.5v. One is for the front panel electronics and the other for the POTS board, so any NAPSC voltage above 15 or so will have enough headroom there.

But the crap noise may be from the NAPSC after all. An often mis-quoted fact is that Naim specifically test the Supercap regulators over the Hicap ones. Not true. All the 317s are tested and those that are quiet enough, and generally that's no shot noise, get used. The steel rejects ended up in the NAPSC and the 52/Supercap digital supplies. But the reason for their rejection isn't noted. So you can end up with one that's just a little noisier than usual (I'm talking about 10uV variance instead of 5uV over time measured with an AC millivolt meter) or you can have one that's bouncing all over the place, but still gives a 'steady' DC value. I have many supercap boards I can send on for £20 each - they've got 2 decent LM317K's on them. PM me.
 
I don't have access to scopes anymore, since I retired......

I might even have to go out buy one!
 
Update: I've recapped the NAPSC, and this is measuring exactly the same as before, and this hasn't made any difference to the amp behaviour, so although it was probably due, it wasn't causing any issues.

Next up, i've been looking at the board a little more closely, and I see a 150 ohm resistor which is clearly getting hot leaving a bit of a black shadow on the PCB. It's in the middle of this photo below the two capacitors.

DSCF4001 by Cesare Ferrari, on Flickr

Measuring with the amp on, it's got 7.5v across it, hence it's heating up - this doesn't seem right... This is pointing to that 10uF being bad, but i'm not sure what the circuit is (other than the -ve side appears to be connected to ground). A schematic would be a great help if there was one available...
 
Oh, 10uF axials like that - replace with extreme prejudice. They cause all sorts of mayhem in the NAP250 (on the reg boards)
You can find equiv parts widely, and cheap.

And while the cap is out, re-measure the warm resistor, and replace with the declared value (bet its way-high, given the slow cook exhibited)
 
Question for Mark of WH, just for the sake of interest.

We now appear to have three "old case" NAPSCs with Vsec measuring around 16.5 V (and my two are 1999 & 2002 vintage), and one "new case" (2006 vintage) measuring around 18.5V. Was there a change of output voltage for later models?
 
You must log in or register to reply here.


advertisement


Back
Top