No bass....

[Bassterror]

Hi,

yesterday i compared my deskgainclone to a temporary MDF Plate setup with one of veterans 4780 boards.

The 4780 was running with an 20V Toroid and 2x 2200µF Elkos, Panasonic FC. No Snubber.

The two 3886ers where running with an 18V Toroid, 2 older 16000µF Elkos "Sprague Powerlytic" i got very cheap with others from ebay and 100µF//100nF at the pins. Snubber as described at chipamp.com.

I knew my 3886ers had no good bass, but they ran with my desk speakers, 2-way, with little Visaton W170-S Bassdrivers.

Now I tested both with my old Canton speakers, with 12inch woofers.

Result:
The 4780 gave good bass... little soft, but OK.
The 3886 gave no bass at all.

So i enabled my software equalizer and turned the 100Hz button to max boost -> makes absolutely no difference with the 3886. Simply no bass responce at all.

I did no futher tests, but i suspect these old sprague cans...
Any Idea how to check them?

By the way:
Do u know any better Snubber PSU design for about 10000µF than the one from Chipamp.com?

[tiltedhalo]

What on you input for capacitance and resitance? sound like roll off on the lows. For a 1K input with 22K to ground a minimum capacitance IMHO is 2.2uf I go with 4.7uf.

[AndrewT]

Hi,
are you running the 3886s as inverted or non-inverted?

The input impedance varies a lot between them. That in turn will change the high pass roll-off frequency of the DC blocking capacitor.

[Bassterror]

Hi,

The 4780 has 2.2µF and the 3886 nothing.

Both are non-inverted, standart circuit like this one
http://www.audio.webd.pl/pdf/lm4780_sch.pdf

No capacitors in the feedback. And the 3886 has no 1K at the input.

[AndrewT]

Hi,
what value of DC blocking capacitor is fitted to the source?

Are you confirming that the 3886 is DC coupled at both the input and at the negative feedback?
and no RF filter either?

But it does seem odd that although both the amps are assembled similarly and yet they appear to perform differently.

Is there any possibility that Zin resistors R1 & 3 are the wrong values?
The major difference left to examine further is the PSU.
Maybe your original thoughts pointing to the old smoothing caps holds the key.
Would it be possible to swap over the PSUs?

[Dxvideo]

Hi I am usign both LM4780 and LM3886 gainclones also. They works in very similar configurations;
Non Inverted,
21Avv (20K/1K/47u)
4u7 MKS type DC decoupling capacitor,
20K input parallel and 1K series resistor,
10K Tandem A curve ALPS potentiometer,
Bass boosted (8dB @ 20Hz) linkwitz NFB preamp with 4 Avv (LM833)
One (LM4780) feeding by a pair of MW 2x24 V SMPS and the other by 2x18v 250VA EI type transformer plus 8xBY80-200 + 10,000uF Philips caps.
Both aöps has 2200u+10u+100n on board snubbers.
I am testing them with my Wharfedale Diamon 9.6 floorstanding speakers.
I can hear some differences between them but not in bass. All I can say is LM3886 is more noisy because of transformer PSU, have more detailed mids and trebles but not exactly lees bass then the other.
There should be a design differences between two amplifiers if there are differences.

[AndrewT]

Hi Dx,
can you measure the noise of both chipamps with their inputs shorted?

[tiltedhalo]

What kind of sound card? and interface?? I had some good sound cards measure very poorly in frequency respone and noise.
Also I had one person that had a lot of ground loop problems between there PC and amp that used a transformer for isolation that lopped of almost eveything below 70Hz........ just a thought.

[Dxvideo]

Hi Andrew,

In fact I dont know the noise measurement methode. But as I heard from somebody, I need a pink noise source for it. However I dont have any.
But I made a test in last week. I work for a environment laboratory. And we have a calibrated Sound Pressure Level Meter which we use for industrial noise meusurements. Anyway, I coupled it to my speakers tweeter with its special adapter and connect my speaker to all my amplifiers by order,
LM4780 has 28dB
LM3886 has 36dB
RAS100 (Nico's MOSAMP) has 24dB
Noise in a band while all inputs shorted to signal ground. But I have to warn you one of them (LM4780) was feeding by a SMPS and the others were with transformer.
I know that; its not a scientific measuring but at least I have a data about noise of my amps.

[AndrewT]

Hi,
DMM set to 200mVac on the power amp outouts with the inputs shorted should give readings in the range 0.0mVac to 1mVac.
I hope not worse, because 1mV is pretty poor, at -86db below max output.
These DMM readings always look better than real audioband rms measurements due to lack of sensitivity to high frequency signals in the DMM AC measuring circuits.