Swordfishy/ASPEN FETZILLA power amp

[mikelm]

Originally Posted by Woody1911a1

now measuring with my BK precision dmm set too 200mv scale , amp on .1mv ac . pre amp turned on same . cd player turned on - same . speakers were still connected if that matters .

If the room where you measure is very quiet it does not matter.

If the room is not quiet the speakers will act like microphones and send a signal down the leads to the amps o/p.

The amp's feedback will attempt to null the o/p if the i/p is silent but this will not be sufficient to prevent a reading on your DMM.

So it may be your amp is even more quiet than you measured.

Normally to test an amp's noise you would short the i/p and disconnect the speakers.

[RonR]

Originally Posted by mikelm

Normally to test an amp's noise you would short the i/p and disconnect the speakers.

Thanks Mike, I'll measure again tonight with these conditions.

@Andrew:

Dead quiet was what I heard, not what I measured. Maybe there's a signal at a frequency I can't hear? I hope not, but it's possible.
Please also bear in mind that the amp and case are mostly disassembled, with 2 transformers piled one on the other and generic bridges rather than discrete diodes, as I need to make changes for the various tests.

BTW, I'm interested to know the threshold level of noise for you to hear anything from your speakers? I know it's dependent upon sensitivity/impedance, but a ball park figure would be nice.

@Hugh:
Testing resumes tonight, so I hope to have some more info then.

Cheers,
Ron.

[AndrewT]

For me, just audible at 1mVac and clearly audible at 3mVac using 88dB speaker.
If I were using my Tannoys @ 92dB or my headphones then the numbers for audibility would be much different.

[JKC]

Here are noise measurements from my Fetzilla amp. I have used Agilent U3401A bench multimeter on 500mVac scale with 10uV resolution. The amp and multimeter was powered up for 30 minutes before measurement. Measurement was done with input cables connected to DAC's output but DAC was not powered up. I have measured at the end of speaker wire with speakers connected and disconnected. There was no difference in measurements in both cases. Since you have read this far then probably you are curious about results. Here you are, the measurement was: 000.00 mVac. Is that good in your book Andrew T?

George

[AKSA]

Thanks George, that's impressive......

I guess the earthing system on this amp is OK, then. Didn't you have a different power supply, George? However, I'm confident it uses a separate bridge and caps for each rail, preserving the original intent of the design?

Hugh

[JKC]

Hugh,

As you remember I have designed different power supply for my Fetzilla amp. It is old fashion design with two transformers, 4 plain vanilla 35A bridges and some good filtering caps in pi filter configuration (I have posted this schematic here). I think that the key to have quiet amp is actual execution of design and planing ahead. As you have said before good grounding is black magic. Partly it is because you will have to figure out the best way to do it during the building process and think about external components that will be connected to the amp too. Even when you manage to get it right and your amp alone is dead quiet you can screw things up when you connect to it equipment that has grounding problems.

George

[AndrewT]

Originally Posted by JKC

Here are noise measurements from my Fetzilla amp. ............... Here you are, the measurement was: 000.00 mVac. Is that good in your book Andrew T?

That is better than good. I can't measure to that sensitivity.

[JKC]

Andrew

I could not believe myself during my first measurement, so I left multimeter connected to the amp for 3 hours and was checking results quite often. All the time the reading was the same and DC offset during this period was changing between 3mVdc and 4mVdc. This multimeter can measure frequency up to 500KHz when set to AC voltage.

George

[AndrewT]

My bench DMM has a 50000 count that auto ranges down to 499.99mVac.

If the reading is >=100.01mVac it displays correctly
If the reading is <=99.99mVac the display instantly changes to 0.00mVac. I can see it is wrong.

My hand helds do not do this. They are not as accurate, nor are they as sensitive, being only 2000 count down to 199.9mVac, but I can see they agree with each other and from experience I know the display is in the correct ballpark.

[AKSA]

I so enjoy it when measurements truly describe the physical world. This rarely happens with distortion measurements, of course.

Hats off, George.

Hugh