diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Analog Line Level (http://www.diyaudio.com/forums/analog-line-level/)
-   -   Getting a PGA2310 to be silent (http://www.diyaudio.com/forums/analog-line-level/221093-getting-pga2310-silent.html)

maxw 7th October 2012 02:23 PM

Getting a PGA2310 to be silent
 
Hi,
I'm using a PGA2310 but having trouble getting it to be silent. There is always some background noise. Not hum or crackle but just noise. Everyone says these chips are silent when implemented correctly but I'm not sure what I'm doing wrong here :confused:

Here is my setup:
USB DAC ---> Passive 50K Pot ---> Hypex Amp ---> Speakers = Silent
USB DAC ---> PGA ---> Headphones = Silent
USB DAC ---> PGA ---> Hypex Amp ---> Speakers = Noisey

The PGA is not in a case and the Hypex Amp does not have signal ground connected to the chassis.

A grounding issue?
Or a impedance issue?

Any input appreciated :)

counter culture 7th October 2012 04:41 PM

I have some experience of the PGA2310 and PGA2320. When implemented correctly they are generally quiet, but you will hear some hiss higher up the gain range in some implementations.

In the simplest general terms, noise is dependent on the impedance that an amplifier sees when looking back out of the input. The higher the (resistive element of the) impedance the greater the Johnson noise, which rises as the squareroot of the resistance.

The PGA series like to see a source impedance < 600 ohms for best performance, so you may need to buffer the DAC. Keep the PGA power supply well up close to the rated +/-15V and well decoupled.

Ian Finch 11th October 2012 03:16 AM

You haven't described the PGA2310 device as more than just the chip. I presume this is an Ebay kit or a breadboard circuit with an adaptor and separate power supply. Perhaps a few details would help to pin your problem down regarding earthing, shielding and impedance. They likely all have a bearing on your set-up, if this is not a ready-built, self contained commercial assembly.

IME all unshielded/earthed audio circuits are noise-prone and digital grounding must be separately maintained to minimise injection from that source. These are great chips and work very well if all the precautions are in place in the application.

jan.didden 11th October 2012 05:33 AM

Quote:

Originally Posted by maxw (Post 3192761)
Hi,
I'm using a PGA2310 but having trouble getting it to be silent. There is always some background noise. Not hum or crackle but just noise. Everyone says these chips are silent when implemented correctly but I'm not sure what I'm doing wrong here :confused:

Here is my setup:
USB DAC ---> Passive 50K Pot ---> Hypex Amp ---> Speakers = Silent
USB DAC ---> PGA ---> Headphones = Silent
USB DAC ---> PGA ---> Hypex Amp ---> Speakers = Noisey

The PGA is not in a case and the Hypex Amp does not have signal ground connected to the chassis.

A grounding issue?
Or a impedance issue?

Any input appreciated :)

Well if the PGA is silent with headphones it can't be the PGA itself; headphones are about the most sensitive to background noise.
Also can't be the DAC Zout for the same reason.
How far do you have the to run down the PGA, what's the level setting?

jan

Bonsai 11th October 2012 11:25 AM

Sounds like radiated or conducted noise from the Hypex unit. Can you post a photo of your set-up. I've built a big high end pre using the PGA device and it is indeed silent - see my website for details. Good luck in any event!

maxw 11th October 2012 08:33 PM

Thanks for the replies everyone!

Quote:

Originally Posted by counter culture (Post 3192872)
The PGA series like to see a source impedance < 600 ohms for best performance, so you may need to buffer the DAC. Keep the PGA power supply well up close to the rated +/-15V and well decoupled.

The output impedance of my DAC is lower than 600 but I have also tried with an input buffer on the PGA.

Quote:

Originally Posted by Ian Finch (Post 3197194)
You haven't described the PGA2310 device as more than just the chip. I presume this is an Ebay kit or a breadboard circuit with an adaptor and separate power supply.

Well, I have actually tested it with two different modules, one I built and designed and a standard one from Ebay. Both show the same result.

Quote:

Originally Posted by janneman (Post 3197269)
Well if the PGA is silent with headphones it can't be the PGA itself; headphones are about the most sensitive to background noise.
Also can't be the DAC Zout for the same reason.
How far do you have the to run down the PGA, what's the level setting?

At -20dB I can hear noise from the speakers when my ear is about 30cm away. Annoying!

Quote:

Originally Posted by Bonsai (Post 3197535)
Sounds like radiated or conducted noise from the Hypex unit. Can you post a photo of your set-up. I've built a big high end pre using the PGA device and it is indeed silent - see my website for details. Good luck in any event!

Could be. I'll take a photo tomorrow. Nice preamp design by the way!

I've just read this application note from Hypex:
http://www.hypex.nl/docs/appnotes/earth_appnote.pdf
So I might try changing the grounding arrangement of the Hypex and see if it makes any difference :)

linuxworks 11th October 2012 10:45 PM

every one of the pga's that I've built (from ebay junker kits, though) were noisy. in most, one channel was bursty-noise and the other more constant. I suspect that they routed the digital and analog traces too close. the chip does seem sensitive to layout. careful about what you put below its core..

I did not pre-buffer my pga's but I think its wise. a lot of people say its not worth the effort but I think its good design practice. I also think that output buffering is a good idea, as well. sandwich the chip from the real world.

I'm kind of giving up on the pga series now that I'm getting good luck with the cs3318. if you need only 2 channels, its a bit overkill but the chip is SO much quieter and more featureful.

maxw 13th October 2012 12:45 PM

1 Attachment(s)
PAH I give up!

I tried all sorts of grounding arrangements including floating, non-earthed chassis, signal GND tied to chassis, different sources like my phone or mp3 player. The noise varied a bit but still not as quiet as my passive pre. I noticed in all the Hypex docs they really discourage the use of unbalanced signals.

I might just give the PGA2310 a miss and take a look at the CS3318.

Linuxworks, do you have any spare? I see they are about 12 here.

Ian Finch 13th October 2012 02:50 PM

PAH?..nahh
 
janneman deduced from your statement that the device was quiet with headphones connected, that the problem can't be with the PGA device. You replied that you could hear the noise at -20 dB with speakers, presumably still with the Hypex amplifier which is not exactly changing anything, unless I misunderstand you.

Logically, that still leaves the amplifier as the noisy or rather noise prone device. The connection and grounding between these units is likely one problem, regardless of the fact that your passive control is fine in this connection. The preamp is not in a case which surely makes it prone to CFL and fluorescent light noise (if you use them) or any background hash from SMPSs which likely you are surrounded by.

Do give the gadget a chance in its properly grounded, metal shield case. I can't imagine you haven't already tried that, but the earthing may not be complete, as I found mine wasn't when leads were bolted to anodised aluminium, for example.

The next feature which I haven't come across in audio DIY yet, is the Vacuum Fluorescent display in what I think I see as the "hi-end PGA2310 kit" on the left side of your nice pic. That uses a high voltage generated by something like a ringing-choke inverter which could cause problems too. I don't see how it logically fits the symptoms and looking at the physical separation, but see if you can disable the display illumuination by disconnecting the ~-70V supply to it. Otherwise, listen if the noise varies as the digits flick over to other readings. This may be interesting.

FWIW, a couple of years ago I bought the cheapest PGA2311 kit and tried it with generally noisy results until I worked out how fussy the assembly was with its case, PSU etc. I improved it but gave up too when a NAC42 clone with cheap pot. proved quieter.

Later I bought a pre-built PGA2311 version because a client was in a hurry for a cheap, working remote preamp. 10 days and voila! Funny; on checking, it was the same chipset and parts with a slightly different layout and display. It was as quiet as a mouse, however, and with whatever I used it. I reluctantly sold it on and bought another for my own use. This newer version was more compact and hum was just evident so I have substituted a shielded, smaller toroid I had on hand and all was well again.

I played it recently in my PC 2.1 setup and can't hear a blind bit of noise at -12dB setting which is unbearably loud at the usual distance.

I tell you this anecdote, not because I have fixed the first kit problem yet either, but because the supplier did it right in his build and I didn't. Had he not cost-cut on the second unit, it would have been a similar result.

Keep at it and better luck! :)

linuxworks 13th October 2012 05:09 PM

Quote:

Originally Posted by maxw (Post 3199954)
PAH I give up!

I tried all sorts of grounding arrangements including floating, non-earthed chassis, signal GND tied to chassis, different sources like my phone or mp3 player. The noise varied a bit but still not as quiet as my passive pre. I noticed in all the Hypex docs they really discourage the use of unbalanced signals.

I might just give the PGA2310 a miss and take a look at the CS3318.

Linuxworks, do you have any spare? I see they are about 12 here.

mostly, I tried junker ebay boards. once or twice I built a pga on perf and I don't remember if I heard the same static bursty noise or not.

what I found weird was that the 3chip pga board from ebay and the single smd pga board from ebay (different layouts, entirely) were both noisy in one channel much more than another.

I didn't get around to trying my opto isolator idea (I am using one on the cirrus board I have, though) but there's a good chance that noise from the cpu could be getting thru, either power/gnd or even the logic signals. I am just not sure, and the opto part is a dollar and good cheap insurance. I guess I'd try that before finally giving up.

note, there are suspected fake chips out there. that could be an issue, too. half of mine came from ebay and only a few came from mouser.

I did try the 5v chips and the 12v chips.

could just not get the noise level down to where its competitive with modern dacs, say. the pga was the noisiest thing in my system.

the cirrus is a lot more complex and its more expensive when you blow it (sigh, I already had a chip burn on me) but the SOUND is so, So, SO much better than pga ;)

and you get 8ch which is great for 'active xo guys' like me. I can allocate 3 pairs for my 2-way spkrs+sub, then allow the last pair for headphones use (switching off the other 3 to engage the last pair for phones).

also, I was never able to get a clean click-free value change on pga. even 32db (binary points) would click and you could definitely hear the clicking in the spkr. not every value and not even every binary rollover point, but some of them.


All times are GMT. The time now is 07:26 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2