Need help fixing my amp - 60hz noise.

[owdi]

I bought what I thought was a decent used amplifier, a Niles Audio SI-275. However, the damn thing is noisy. When I connect my speakers to it, I hear a distinct 60hz hum, as well as HF noise.

I know it's not a diy amp, but I want to fix/improve this amp on my own, which I hope is diy'ish enough to post here ;-)

So where do I start?

SPECIFICATIONS
Design Principle: Linear voltage/current amplification.
Continuous Average Power Output
* 75 watts per channel RMS at 8 ohms.
* 100 watts per channel RMS at 4 ohms.
Input Impedance: Approximately 15K ohms
Input Sensitivity: .985V for 8 ohm rated output
Overall Voltage Gain: 30 dB
Frequency Response: Bandwidth Limited from 10 Hz to 100 kHz
Total Harmonic Distortion:
8 Ohms - .05% THD from 20 Hz to 20 kHz @ full rated power
4 Ohms - .06% THD from 20 Hz to 20 kHz @ full rated power

Here are some pictures of it's guts. I can post much higher res pictures upon request.









This is what the noise looks like when recorded in Arta.





And finally, a link to the manual

Any advice is appreciated.

Dan

[AndrewT]

Hi,
welcome.
your query is certainly DIYaudio.
Nice output stage.
Shame about the power supply.
The smoothing caps could usefully be doubled, or tripled and my brute force design method says +-40mF to +-50mF is needed for each channel.

That three pair is capable of a lot of current even with a hot heatsink. It is capable of at least 140W/4ohms with a highly reactive load (probably nearer 180W if it was not strangled by lack of current supply).

But that has nothing to do with your hum and/or noise problem.

Can you check resistance from the input RCA ground to audio ground and to safety earth?
Can you post a pic or two of how the grounding has been arranged?

[owdi]

Using my DMM I measured no resistance between input RCA ground to audio ground. The power connector for this amp only has two prongs, so I guess there is no safety earth. Could that be part of the problem?

Nice output stage.
Shame about the power supply.
The smoothing caps could usefully be doubled, or tripled and my brute force design method says +-40mF to +-50mF is needed for each channel.

That three pair is capable of a lot of current even with a hot heatsink. It is capable of at least 140W/4ohms with a highly reactive load (probably nearer 180W if it was not strangled by lack of current supply).

Are you referring to the three chips labeled A1386? There are a total of 12 of these (four groups of 3). The power supply has 4 x 10,000 uF 63v capacitors, are those the caps you recommend upgrading?

I'll take more detailed pics later today, as soon as the wife returns with our digital camera. For now you can see a high res version of the first pic I posted.

2560 x 1920 pic of internals - 2.5 MB

Dan

[owdi]

Here is a link to a page with 6 high resolution pictures of the guts of this amp. It's about 13mb of images, so 56k beware.

Link to pictures of internals. 13mb of images

In case it matters, my source is my computer. The soundcard, an M-Audio Audiophile 2496, shows no signs of noise when doing loopback tests.

Dan

[Bigred]

Originally posted by owdi
Are you referring to the three chips labeled A1386? There are a total of 12 of these

Maybe an oversight but its complimentary pairs 2SA1386 and 2SC3519 (3X 1386 & 3X 3519 per channel)

[AndrewT]

Originally posted by owdi
Are you referring to the three chips labeled A1386? There are a total of 12 of these (four groups of 3). The power supply has 4 x 10,000 uF 63v capacitors, are those the caps you recommend upgrading?

I am not recommending that you change the capacitors. The other changes to allow that to work may be extensive. I just wonder why they have specified a super output stage that cost signifcant money and PCB space and heatsink area and yet skimmped on the transformer/rectifiers/smoothing caps. It all comes down to cost. The way they have done it will minimise warranty returns. That may be very important to them.

The output stage is as Bigred identified.

Download the datasheet from Sanken, these are nice 130W 15A transistors. Can you identify the gain selection letters on them?

[djk]

The power supply is fine, a pair of 10,000µF per 75W channel.

He has a ground loop with his computer.

Everyone needs one of these in his toolbox:

http://www.edcorusa.com/products/mat...2s10k-10k.html

[jaycee]

It might just be me, but I cannot see anywhere where the mains safety ground connects to the amp chassis, even though a 3-pin IEC socket is used.

The manual also suggests a 2-prong cable, so the chassis is not earthed. The amplifier might well be picking up hum in this case.

[owdi]

Originally posted by AndrewT
Download the datasheet from Sanken, these are nice 130W 15A transistors. Can you identify the gain selection letters on them?

Sanken 2SA1386 datasheet. Pardon my noobness, but how do I find the gain selection letters?

Originally posted by djk
The power supply is fine, a pair of 10,000µF per 75W channel.

He has a ground loop with his computer.

Everyone needs one of these in his toolbox:

http://www.edcorusa.com/products/mat...2s10k-10k.html

I have a similar device, designed to isolate components. I opened it up and it's nothing more than a pair of transformers, one for each channel. I tested it using RMAA, and SNR and stereo separation became about 20db worse. I'd rather not use it, if possible.

Originally posted by jaycee
It might just be me, but I cannot see anywhere where the mains safety ground connects to the amp chassis, even though a 3-pin IEC socket is used.

The manual also suggests a 2-prong cable, so the chassis is not earthed. The amplifier might well be picking up hum in this case.

The socket only has 2 pins.

Dan

[djk]

"I opened it up and it's nothing more than a pair of transformers, one for each channel. "

How else could it work?

Since the hum is 60hz, it cannot be the supply caps (that would be 120hz).

How expensive are your transformers?

The ones in the units sold for computer and car stereo use are the size of a pencil eraser and have limited bass. The one I recommended is much better. You want the best, go Jensen.

http://www.jensen-transformers.com/iso_aud.html

A good read (all the white papers and app notes are too):

http://www.jensen-transformers.com/an/ts_guide.pdf

"and SNR and stereo separation became about 20db worse. I'd rather not use it, if possible."

The SNR sounds like a measurement error. You care about separation? That's a red herring. A good phono sounds much better than most digital, and separation on a good cartridge seldom measures much past 20dB.

Do you have any cable lines hooked up?

Try the cheap transformer you have. It may not sound the best, but unless you don't also have a cable line in it should fix the problem. If it fixes the problem then consider a better transformer.

The Edcor units have enough core size for good low end, but have M6 steel cores for low cost. The most expensive units have high nickel cores to reduce H2 distortion. The small amount of H2 added by M6 steel cores is actually part of the 'sound' of some expensive pro equipment. Replacing M6 with high nickel in these products may result in loss of the 'warm, lush' sound that people buy them for.

Excerpts from a john curl thread:

"You people worry too much about superficial things."

"Listening is more important than measuring."

"Absolute THD is not very important to me, (because) it is usually below what most references would consider audible."

"I would prefer a transformer to an IC chip. I once designed out (removed from the design) the best IC chip that I could find.... when it failed critical listening tests."