Input resistor

[Stabist]

Hi!

I'm building up 2 "Pateks" - one based on LM3875 and the other with OPA541 ...
Because I want to reduce costs I've decided to make a third "Patek" look alike chassis with passiv preamp in it - using fake Alps stepped attenuator 50k ...
So - when using a Patek - I will have a pot in the circuitry - it will only be phisicaly located in separate chassis - so I gues I can left out the input resistor in Patek's circuitry or not because of safety?!?
And if I have to use it - which value to choose considering the pot has 50k??

Thanks!

[Mad_K]

I'm not shure what resistor you are talking about. Do you have a schematic? Input to gnd resistor should not be left out, especially if you are using a separate passive preamp. 100K is a good value.

I prefer to use a 100K LIN pot and a 22K to gnd resistor on the input (keep the pot close to the input though)

[Peter Daniel]

The reason I'm using 220R resistor (in series with input, after the pot) is that I observed in certain setups (when pot is at full volume) DC offset increase. Placing that resistor, adds some "permamnent" resistance at the input, and this prevents DC voltage jump, regardless of volume setting. The value of that resistor is not critical and you might use anything between 200-1K values.

I'd say it's best to test it in your system, if no abnormal behaviour observed, you might leave that resistor out.

I was doing some comparisons over the weekend, and with carbon pot my Patek actually sounds better when additional 220R Vishay resistor was inserted after the pot (the sound is smoother, but mostly because of the sonic properties of Vishay S102 resistor).

[Stabist]

Hehe - Peter - you have always good advices for me
So in my case - I use this "schematic" (btw it will be non-inverted GC):
RCA plug --> a resistor in series with input --> resistor to ground on the chip + input; and then in preamp the 50k "Alps" ...

Btw - those fake Alps have 4 conectors - one is I guess a kind of "loudness" effect ...

Thanks guys for the help!

Primoz

[jam]

Peter,

Could you explain DC voltage jump............it could mean something like RF on the input. You might be confusing what is actually happening here.

Regards,
Jam

[Peter Daniel]

Originally posted by jam
You might be confusing what is actually happening here.

I know you like to confuse people, I'm only trying to isolate the input a bit.

After discussing attenuators (with you), I'm not eager to discuss input resistor, as I might get you lost again.

[MalichiConstant]

It is a little surprising that after all the discussion about input buffers, high frequency power supply decoupling, and (dare I say it......) zobel networks; that filtering the RF out of the input of the amp could be a mystery to anyone who builds amplifiers. Some of the other chip amp builders seem well aware of these details and can even recommend specific vales for the filter and further tuning tricks like zoble networks at the loudspeaker cable interface as well as the amp output. Of course they are offering something besides just attitude, like advice based on years of experience and an understanding of circuit design.

For the typical input bias current of 0.2 uA, an additional 220 ohms gives an input referred offset change of about 44 uV. For a gain of 10 this is 440 uV or less than 1/2 of a millivolt change at the output. The change in the offset voltage will much more from the bias current through the different setting of the pot than from the addition of 220 ohms resistance. It seems to me that you are seeing the results of RF on the input and that a series input resistor is not optional. More measurements and information and less attitude might be very beneficial.

Don't make me sick Carlos on you again.

http://www.analog.com/UploadedFiles/...277AN671_0.pdf

[jam]

Peter,

I rather think that you are getting me lost....... but seriously the effect you are talking about might be more more the supression of RF than.........DC voltage jumps (whatever they are).

You might also consider the base currents of the input assuming a bipolar input pair in the device in question.

Regards,
Jam

Your Zobel network advocate.

[Peter Daniel]

I'm not really sure what was causing the "DC jump", but it certainly was clearly observable.

When turning the pot fully up, in AMP-1 (with TDA 1543 as a source, in passive I/V mode with 1k resistors) the offset suddenly increased to 300mV or so. I decided to use a series resistor, small value though (in order to have least sonic influence) and it seems to cure the problem.

Having the input isolated seems like a good idea too, and by choosing different types of resistors one can almost adjust the amp's sound to personal preference.

Although Brian's GT kit is sold with 220 Caddocks, I personally prefer using 220ohm Vishay S102, as it's mild character tones down a bit of brightness observable with some pots. OTOH, a friend of mine prefers Caddock, as it adds more "space" and somewhat "fuller" sound

Just recently, I was experimenting with carbon pot, and although the pot sounded fine by itself, inserting additional series resistor after the pot, made the sound more to my preference.

[jam]

Peter,

I wonder if replacing the resistor with a capacitor would solve the problem as well. Sonically it might not be the best solution but would potentially point to cause of the problem.

Regards,
Jam

Your Zobel network advocate.