[aos]

geom_tool, try using a cheap regulator such as 78L12 (or 78L12 and 79L12 in your case, given separate rails, or TO-220 7812/7912 if using higher power buffers such as EL2008), and keep the capacitance. You may be surprised at the amount of improvement you'll get for just a buck or two.

 

[geom_tol]

ppl: I'm hooking up the + pole of the one pack to the + rail, and the - pole to the ground.
The other battery pack is hooked up with - pole to - rail and + pole to ground.

I will try and find a couple of those regulators! Seems a really easy tweak.

 

[tangent]

The META42 does apparently couple power supply noise into the audio stage. I don't yet know where this is happening, but the real moral of the story is, the cleaner the power supply, the better your META42 will sound.

 

[eric343]

True

Awhile ago I put an OPA2132 into a pretty much stock META powered off the Elpac linear supply, with .1uF per input rail filter caps... I swear, that amp had enough air to win a skijumping contest. FROM AN OPA2132!!!

 

[geom_tol]

Quote:

Originally posted by tangent The META42 does apparently couple power supply noise into the audio stage. I don't yet know where this is happening, but the real moral of the story is, the cleaner the power supply, the better your META42 will sound.

Any theories?

My theory is the physical closeness between all the components and traces.

 

[morsel]

It would be interesting to see if the reverse is true, that is, do other amps couple power supply noise into the audio stage and thus sound just as noisy if a cheesy power supply is used.

 

[geom_tol]

Eeyy. My ps is not cheesy.
16aa's, 3000uF caps (Elna cerafines) no ground drivers or anything else interfering with the batteries.

And how do you know it's only ps noise? There could be other noise also.

 

[eric343]

Something to consider - apparently you can lower the ESR by doing powers-of-ten value picking; such as 2200uF, 220uF, 22uF, 2.2uF, .22uF, etc. I know I've gotten great results with 2200uF/220uF per rail.

 

[tangent]

My comment about PS noise doesn't apply to your config, geom_tol. I've only had this happen with cheap unregulated power supplies -- I mean the $5 pieces of junk. I've never heard PS noise from batteries, decent switchers, or linear power supplies.

 

[wilsonjm]

All noise i've experienced is when something isn't grounded properly or the noise is actually coming from the source

 

[ppl]

Some time ago i talked about my methods of Removing AC noise present upon the Supply rails wetther from a poor Power supply or the Back EMF produced by Driving a reactive load. This is not unique to the META42 or any other Audio Power Amp. The Headphones are basicaly inductors (Coils) and inductors store current then Dump it back into the Source and in this case the Headphone amps output stage. the magnatude of this back emf is the same and somtimes larger that the Curent flowing through the inductor. All Power supplies have an Finite impedance and as sutch some of this back EMF will be present upon the supply rails. It is accepted practace in High End Audio Power Amp design to allow all that Dirt to be on the Output stage rails since no voltage gain is available so the problems are not as noticed. this is why some High Priced High End Audio power amps will use an unregulated supply for the output stages and use separate regulation and somtimes separate Transformer windings for the Voltage gain stages. This keeps the Voltage supply for the Voltage gain stages Cleen and free of this type of noise. Now Voltage regulation is not Practical in a battery powered amp since the prospect of the regulators droping out of regulation is very real with declining Battery Voltage. The real problem is keeping the Noise on the main rails from getting into the Opamp and being amplified by the Full Open loop gain - the PSRR ( Power supply rejection ratio) of the Opamp. My solution was the capacitence Multiplier ? current sourcing i used on my Orig. Amp. It works i have measured the difference Improves the PSRR of the opamp by about 30-60 dB. But the Current sources will not pass AC so any AC upon the rails will be removed and the cap multiplier futher improves this and is a intergrail Part of the scheme i used to deal with this problem.

Hopefully in the not to distant future tangent, morsel and myself will introduce my Pocket amp that deals with this problem in a unique way using some of the concepts used in My orig. amp Stay tunred.

As far as using the LM78 and 79 series of three term regs. I think that AA series connected NI-MH cells will have a more stable voltage and for sure alot less High frequency noise than these jelly bean Three term regulators, so if you insist upon using IC regulators at least use good ones like The LT-1033 and LT1085 types but evean these are only marginaly better than the far less expensive LM series. so your damed if you do and damed if you don't. is true in this context. kind of like pick your poision.

IN conclusion i like my method it works and was the result of me investigating this issue at great lenght. Sure no voltage regulation is used but you can just use batteries with enough current to drive the load without voltage drops.

Alternatly you can use a dedicated ultralow noise Regulator like the Jung, Schultz or Kevins but then the Portability options is no longer available.

 

[goon-heaven]

Thanks PPL, very interesting.
Please can someone point me to more info on PPL's amp (my searching cant be up to scratch).

 

[aos]

ppl, I disagree. Both listening test and oscilloscope test showed significant superiority of cheap regulator to no regulator. Also, Kevin's solution didn't really work for me (I tried LM6171 not the super power buffer though). Second, today's "78xx" are a cut or two above those used two decades ago. Just look at the datasheets, some of the replacements are order of magnitude better. Yes, some older stuff - and still manufactered dirt cheap - does have close to 1mV of noise which is totally unacceptabe and really gross, but some others have only about 40uV to 100kHz which is nothing to snear at. If you use 2 7.2V NiMh and a 12V regulator you shouldn't have a problem with out of regulation when the battery is low (besides NiMh has a rather sharp voltage drop unlike alkalines), and then there are low dropout models as well. And the size is very small. Using current sources on voltage amplification stage is a very proper thing though and should certainly be used if space allows, I don't doubt that. But don't discard regulator that easily (price difference between cheap and bad ones and much better ones is 5 times).

 

[morsel]

I am open to opinions on the best voltage regulators to use. We should evaluate all options carefully.

As an aside, PPL, Tangent, and I are waiting for the next release of PCB123 software to see if it fixes some of the more serious current problems before moving forward on the amp project.

 

[aos]

I am for my part going to add them to my PDAC PCB. I'm also considering using specialty polymer caps (which I already use in digital section) to bypass the buffers as well instead of Panasonic FC. This can be done if voltage regulation is used so you know the voltage on rails (since capacitor range is very limited, 8V max). For comparison sake, FC 100uF 10V capacitor has 800 or 500 milliohm resistance at 100kHz (depending on size), while 100uF 6.3 or 8V polymer has only 12milliohm. I wonder what that'll do to the sound...