Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Szekeres's Class A MOSFET amp with high imp. cans...
New Posts  All Forums:Forum Nav:

Szekeres's Class A MOSFET amp with high imp. cans...

post #1 of 12
Thread Starter 
I'm looking to build a version Szekeres's Class A MOSFET amp using an opamp gain stage similar to the CMOY amp as well as Tomo's constant current source. I'm just wondering how well it will work with higher impedance cans. I have some AGK K-501s (120 ohms nominal) and I'm hoping to get some Sennheiser AD-600s soon (300 ohms nominal). Since this amp acts like a current source instead of a voltage source, will it be able to swing the voltage high/low enough to drive high impedance cans like the AD-600s? (My PS will be 15V regulated)

Also, will the amp be able to drive both sets of cans without any design changes, or will I need to build a version that can change some things with a switch, such as the gain resistors?

One last question: Has anyone experimented with the output capacitor? It suggests a 470uF cap, but recommends a 1000uF replacement if one can be obtained. Can someone comment on the sound difference? (Obviously, a bigger cap here will give better low frequency responce, but how does it sound? Will it muddy up some of the higher frequencies? etc.)
post #2 of 12
As far as the impedence of the phones, you shouldn't have much trouble between 120 and 300 ohms. I think the biggest difference would be between, say, some 600 ohm HD-545's and some 32 ohm Grados.

As for the output capacitor, its primary purpose (I would imagine) is to lower the overall output impedence so it could nicely drive low impedence cans. With HD-600's, a nice, fast 470uF output cap should be more than enough.

As for using more capacitance on the output, the main criticisms I hear is that electrolytics (which you're pretty much forced to use) will give a "wetter" sound. Granted, this often isn't necessarily bad, since SS circuits in my opinion sound harsher. It just depends on your taste... of course, if you haven't had the chance to develop the taste, I'd recommend less electrolytics on the signal path.

My two cents,
post #3 of 12
Depending on which version of the Szekeres amp you used, the output capacitor is probably most importantly used to remove the DC voltage from the signal going to your headphones. Only the schematic labelled "Direct-coupled" shouldn't need an output capacitor if you adjust the output voltage to 0 as per the circuit description.

Try searching on HeadWize for discussions on the output capacitor. I know one thread discussed the calculations for finding the high-pass corner frequency (using the impedence of your headphones) while at the same time discussing actual sound characteristics of trying different size capacitors, from 20+ uF to the thousands of microFarads.
post #4 of 12

I can tell you from my experience that Szkeres' amp will work nicely with high impedance headphones as well as low impedance headphones as long as the signal source can output large enough voltage swing. This is true for most of modern home CD players. (Not true for portable CDP)

The output coupling capacitors must be chosen according to your headphone impedances. This is because the capacitor and the headphones form low pass filter. I have done calculations before using usual equations; at the time, I refered to them as Apheared equations ...

1 / (2*Pi*R*C) = Cutoff freq
C = output coupling capacitance
R = headphone impedance

Cutoff freq is where low pass filter kick in and attenuate signals with freq lower than that strongly.

From my calculations, I concluded that you can do very well with caps as low as 220uF if you want around 50Hz cut off freq. 470uF caps are recommended for more reasonable cut off freq around 20Hz covering all audible freq. Making the output coupling cap bigger than that is unnecessary. Instead you should concentrate on the quality of caps used here.

Personally, I used 24uF Solen film caps. This is ONLY possible if you use 600ohm headphones. I can testify that sound quality actually improved while keeping equal bass response. In fact, the sound quality is actually better. Electrolytics have inferior performances. (However, unavoidable for low impedance headphones.)

Keep in mind of the equations while changing coupling capacitance. You will get optimal performances for low capacitance accordingly with the Apheared equations.

Note that paralleling caps are popular to achieve high capacitances. However, this is not optimal. You will get parasitic qualities which cause phase shifts. Bad for sonic quality.


post #5 of 12
I wouldn't say in the 470uF range that your only option is electrolytic. Your only _sane_ option is electrolytic, but there are still alternatives. Solen makes very high values of their fast caps. Even http://www.partsexpress.com holds values up to 200uF/450V. By running a couple in parallel, you've got a desirable cutoff frequency for your 32 ohm Grados. Granted, that's four caps, and each is $50

Of course, you could just go out and buy Blackgates or Cerafines which are, in some cases, even better than their film counterparts and would be a cheaper option than polypropylene.

post #6 of 12

Stay away from large capacitance film caps. Of course, if it really is a film cap you might want to consider using it. However, some of them are film caps but they are immersed in oil or some other liquid. This is not necessarily good. Most of these will sound pretty loose.

For examples, Vitamin Q caps are one of them and these distort sound a lot. I stopped using these in signal paths long ago. Currently, I only use these in PSU.

I think getting film caps at this size is rather not insane but unreasonable. If you do the calculations, for 120 ohm and 300 ohms, you should be able to have lower capacitance. Do them. It's really easy.

Electrolytics are simply not good. Of course, there are some film caps (not in oil) which sound worse. However, this is not the point since you will choose decent film cap which will sound better anyways. Electrolytics are slow and their impedance increase with freq. This is very bad. You can improve this by bypass the cap with a film cap.

Nonetheless, non-electrolytics are better at doing this theoretically and experimentally. (I did use Cerafine before.) I can testify to this.

However, the "best" capacitors are the "no" capacitors. DC coupling solution is the best idea if it is possible. It is possible with Szkeres' amp, but doing so will comprecate the procedures.

Yari Yara,

post #7 of 12

Standard Szkeres' amp can be easily built and you will get pretty good quality. Of couses, it will not sound like 1000 USD amps. But it definitely will surprise you considering the ease and the cost of the Szkeres' amp.

In addendums, you will see multiple versions. If you are up to a bit more difficult version, you can try out the version with stages DC coupled. This is pretty excellent yet it still only require single power supply. (truely DC coupled version will require dual PSU which makes it bit more comprecated ...)

A 470uF should be bypassed with a 4.7uF~1uF polypropylene capacitor. This will improve your sound quality. For now this is your best bet without dramatically redesigning your amp.

Cheers and Good luck,

post #8 of 12
I don't think oil should be condemned completely. Jensen makes a bunch of paper-in-oil/copper-in-oil and other capacitors that are considered top of the line (they should be considering the cost of some..)

Granted, I'm sure it's up to one's preference. I don't want to make this a separate thread on capacitor coloration

post #9 of 12

Jensen Oil cap and equivalents


Yes these sound pretty good. But it cannot be used in placed where signal critical. We use these caps not because it has less distortions or high speed etc. We add these to "craft" sound.

Oil caps tend to make the sound warmer. Unfortunately, this comes from distortions that oil cap addes to the sound. So excessive use will add considerably distortions. You will find the sound gets muddy and loose.

I usually do not recommend oil caps since it require a lot of experiments to make them sound right. I found this pretty hard and time consuming to do. (Then I grew more fond of accuracy/sweetness more than warmth in sound. So currently, I only use film caps or DC coupling for maximum transfer.)

post #10 of 12

How about transformer coupling?

Has anyone ever tried to tranformer couple the Szekeres amp? I would think that this would be a viable alternative. Heres a link to an interesting transformer:

post #11 of 12

Transformer coupling and capacitor coupling, both of them will distort sound. You can use them to "craft" sound. Which tends to be very hard and time consuming as I have mentioned. So bewarned.

Now here is how to use output trannies.

You have two ways of dealing with output trannies in Szkeres amp.

1. You remove the resistor, R4, and replace it with output transformer.

2. You take the output and skip the cap and feed that to the output transformer.

First is simply out of question since the current needs of the MOSFET is way too high for any output transformers.

Second is more reasonable. However, the output transformer must be connected in very specific ways. Simply replacing the output cap with a output tranny will cause big problem. There can't be DC current thru the tranny. So you will have to connect the other wire of the tranny to ground via a capacitor. Note, however, that you have both a cap and a tranny in the signal path.


post #12 of 12

But the cap can be a high quality film.

Yes you will still have a cap and a tranny in the signal path but you won't have a large electrolytic cap in the path merely a high quality film cap. I think that this would make a large difference.

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Szekeres's Class A MOSFET amp with high imp. cans...