Some questions about the Szekeres headphone amp:

Bricolo · 23rd November 2002, 11:53 AM

Hi

I've got some questions about this amp:

What is R1 used for?
I made some simulations with spice, and with all default values, the output voltage is linear from 40 to 10kHz, from 10kHz to 30kHz, it drops from 1.8V to 1.6V

Bypassing it, or reducing R1's value from 4.7k to 470 totally cancels this.

Are there any risks to change it to a 470Ohm one? (except the risk of having a better sound

)

Another question: why do we need to use another output cap value when using the ccs modif?

peranders · 23rd November 2002, 02:35 PM

Quote:

Originally posted by Bricolo
Hi

I've got some questions about this amp:

What is R1 used for?
I made some simulations with spice, and with all default values, the output voltage is linear from 40 to 10kHz, from 10kHz to 30kHz, it drops from 1.8V to 1.6V

Bypassing it, or reducing R1's value from 4.7k to 470 totally cancels this.

Are there any risks to change it to a 470Ohm one? (except the risk of having a better sound )

I can think of two reasons:

1 current limitation in input overvoltage conditions

2 Create some roll off at high frequenies.

If you have control over your signals I think you can omit this resistor without any trouble.

Bricolo · 23rd November 2002, 04:38 PM

what does "roll off" mean? (sorry, I'm french...)

dhaen · 23rd November 2002, 05:12 PM

HF Roll off:

Increased attenuation with increase of frequency.

LF Roll off:

Increased attenuation with fall of frequency.

Bricolo · 23rd November 2002, 09:40 PM

but we don't want roll off from 10kHz to 30kHz, do we?

halojoy · 24th November 2002, 06:07 AM

we don't want roll off from 10kHz to 30kHz, do we?

Normally not in a buffert amp.
But say you want a more "round, soft" sound
and not too "crispy, overbright".
Then you can use a low pass filter.

Lowpass filter you usually do with one resistor connected to one capacitor. The resistor is like Rg in series with signal
and the capacitor is connected to ground (0v).
With 1 R and 1 C, you get a first order filter. It rolls off with -6dB/octave. Octave is to double the frequency.

If MOSFETs was without input capacitance we shouldn't get a filter
with Rg+ Gate of MOS. But the gate is like a capacitor, it is charged with current and disscharged. Current is stored and dis-stored.
We can say that the Gate sucks!
a small current, at AC-signals.
And spits it back out.

This is one reson to not have too high resistors around
the input of a bigger mosfet.
higher resistors values creates higher voltage drop
at higher freq.

If I remeber correctly, Nelson, who is well studied in this subject,
recommended 15 kOhm and 22 kOhm, to set the working point for the MOSFET.
This can also be expressed as:
Having a lower input impedance to the circuit.
Impedance is something like "resistance at AC-signals"

That current starts to create a voltage over Rg,
and that voltage drop make the out signal lower.
The higher the freq the higher the drop.

Any R+C filter have a rolloff point depending on resistor and capacitor value.
Lower resistor, higher f (freq)
Lower capacitor, higher f (freq)

There is a formula!
I do not know what US-patent number that formula has????

So with same resistor and another MOSFET you probably get another roll off freq.
Also if there is other resistors before in series with the signal, this is added for calculating the roll off.
That can be output resistance of the CD-player

Bricolo · 24th November 2002, 06:51 AM

thanks, I've understood the concept

I thought that it was a RC filter, now you confirm it.

But you talk about Rg, Rg is a low value, to prevent some oscillations

I talked abour R1, that is 4.7K
I must check my electronic lessons, but a low pass RC filter is, as you said, a resistor in serie with the signal, and a cap to the ground
here, we have both R1 and C1 in serie

Another thing I'll have to check: the hitachi MOSFET has a lower capacitance that the IRF510, I'll maybe heve to recalculate R1 according to this
But I'll need some help for this, I don' know the formula

halojoy · 24th November 2002, 07:00 AM

Also if there is other resistors before in series with the signal, this is added for calculating the roll off.

So in this case R= resistens of potentiometer + R1 + Rg and so.

That should mean that the RxC is dependent of the setting of the potentiometer.
Instead of doing the math,
you can feed the circuit with different freq signals
at different potentiometer settings.
And then take actual messurments.

This is a much more accurate (easier) way
then the math!!
With math we can try to estimate the behavior
of a circuit, but due to the imperfectness of components, layout
especially when handling AC-signals,
messurments gives us the real figures.

Bricolo · 24th November 2002, 07:04 AM

OK, i'll run spice right now

another thing I've seen with spice: someone told me to pur C1 before the pot, to have a better sound

this gives me other frequancy responce with spice, i'll post the results

Bricolo · 24th November 2002, 07:44 AM

This is the results I've got

Original diagram, with R1=470R

Original diagram, with the 100k pot to 100%

Original diagram, with the 100k pot to 50%

Original diagram, with the 100k pot to 90%

C1 before the pot, with the 100k pot to 100%

C1 before the pot, with the 100k pot to 90%

C1 before the pot, with the 100k pot to 50%

23rd November 2002, 11:53 AM	#1
Bricolo diyAudio Member Join Date: Nov 2002 Location: Grenoble, FR	Some questions about the Szekeres headphone amp: Hi I've got some questions about this amp: What is R1 used for? I made some simulations with spice, and with all default values, the output voltage is linear from 40 to 10kHz, from 10kHz to 30kHz, it drops from 1.8V to 1.6V Bypassing it, or reducing R1's value from 4.7k to 470 totally cancels this. Are there any risks to change it to a 470Ohm one? (except the risk of having a better sound ) Another question: why do we need to use another output cap value when using the ccs modif?

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
NE5532 as a pre amp for a Szekeres headphone amp	zebra100	Headphone Systems	2	1st July 2009 09:15 AM
Some questions on the Gilmore headphone amp.	Artie	Headphone Systems	2	22nd April 2009 12:21 AM
Headphone amp questions???	sunrise	Headphone Systems	13	22nd August 2005 04:27 AM
Stax headphone questions	amt	Headphone Systems	2	26th May 2004 08:10 PM
DIY Power supply for a Szekeres headphone amp	Bricolo	Headphone Systems	93	30th November 2002 05:58 PM

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read


	Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving

23rd November 2002, 04:38 PM	#3
Bricolo diyAudio Member Join Date: Nov 2002 Location: Grenoble, FR	what does "roll off" mean? (sorry, I'm french...)

23rd November 2002, 05:12 PM	#4
dhaen diyAudio Moderator Emeritus Join Date: Jun 2002 Location: U.K.	HF Roll off: Increased attenuation with increase of frequency. LF Roll off: Increased attenuation with fall of frequency.

23rd November 2002, 09:40 PM	#5
Bricolo diyAudio Member Join Date: Nov 2002 Location: Grenoble, FR	but we don't want roll off from 10kHz to 30kHz, do we?

24th November 2002, 06:07 AM	#6
halojoy On Hiatus Join Date: Nov 2002	we don't want roll off from 10kHz to 30kHz, do we? Normally not in a buffert amp. But say you want a more "round, soft" sound and not too "crispy, overbright". Then you can use a low pass filter. Lowpass filter you usually do with one resistor connected to one capacitor. The resistor is like Rg in series with signal and the capacitor is connected to ground (0v). With 1 R and 1 C, you get a first order filter. It rolls off with -6dB/octave. Octave is to double the frequency. If MOSFETs was without input capacitance we shouldn't get a filter with Rg+ Gate of MOS. But the gate is like a capacitor, it is charged with current and disscharged. Current is stored and dis-stored. We can say that the Gate sucks! a small current, at AC-signals. And spits it back out. This is one reson to not have too high resistors around the input of a bigger mosfet. higher resistors values creates higher voltage drop at higher freq. If I remeber correctly, Nelson, who is well studied in this subject, recommended 15 kOhm and 22 kOhm, to set the working point for the MOSFET. This can also be expressed as: Having a lower input impedance to the circuit. Impedance is something like "resistance at AC-signals" That current starts to create a voltage over Rg, and that voltage drop make the out signal lower. The higher the freq the higher the drop. Any R+C filter have a rolloff point depending on resistor and capacitor value. Lower resistor, higher f (freq) Lower capacitor, higher f (freq) There is a formula! I do not know what US-patent number that formula has???? So with same resistor and another MOSFET you probably get another roll off freq. Also if there is other resistors before in series with the signal, this is added for calculating the roll off. That can be output resistance of the CD-player

24th November 2002, 06:51 AM	#7
Bricolo diyAudio Member Join Date: Nov 2002 Location: Grenoble, FR	thanks, I've understood the concept I thought that it was a RC filter, now you confirm it. But you talk about Rg, Rg is a low value, to prevent some oscillations I talked abour R1, that is 4.7K I must check my electronic lessons, but a low pass RC filter is, as you said, a resistor in serie with the signal, and a cap to the ground here, we have both R1 and C1 in serie Another thing I'll have to check: the hitachi MOSFET has a lower capacitance that the IRF510, I'll maybe heve to recalculate R1 according to this But I'll need some help for this, I don' know the formula

24th November 2002, 07:00 AM	#8
halojoy On Hiatus Join Date: Nov 2002	Also if there is other resistors before in series with the signal, this is added for calculating the roll off. So in this case R= resistens of potentiometer + R1 + Rg and so. That should mean that the RxC is dependent of the setting of the potentiometer. Instead of doing the math, you can feed the circuit with different freq signals at different potentiometer settings. And then take actual messurments. This is a much more accurate (easier) way then the math!! With math we can try to estimate the behavior of a circuit, but due to the imperfectness of components, layout especially when handling AC-signals, messurments gives us the real figures.

24th November 2002, 07:04 AM	#9
Bricolo diyAudio Member Join Date: Nov 2002 Location: Grenoble, FR	OK, i'll run spice right now another thing I've seen with spice: someone told me to pur C1 before the pot, to have a better sound this gives me other frequancy responce with spice, i'll post the results

24th November 2002, 07:44 AM	#10
Bricolo diyAudio Member Join Date: Nov 2002 Location: Grenoble, FR	This is the results I've got Original diagram, with R1=470R Original diagram, with the 100k pot to 100% Original diagram, with the 100k pot to 50% Original diagram, with the 100k pot to 90% C1 before the pot, with the 100k pot to 100% C1 before the pot, with the 100k pot to 90% C1 before the pot, with the 100k pot to 50%