[Jupiter]

I was thinking about getting rid of the opamp(s) (or other components) in the signal path that's not needed when I don't need to amplify the signal.

Has anybody tried using a buffer with a volume control instead of an amp for low impedance headphones?

I have a few EL2009, would those work? The datasheet specify 1% THD, but doesn't say at which frequencies.

 

[Glassman]

DACT makes almost-passive preamp using their stepped attenuator and output buffer (Analog's BUF03).. this is the very same story and I think it's generally very good idea..

 

[jeffreyj]

By themselves, buffers will still deliver high speed and high output current, but distortion and DC offsets will be high; this is why an op-amp is pretty much always included even if no voltage gain is required. Trying to strip unnecessary components from the signal path is always a worthwhile endeavor, but don't throw the proverbial baby out with the bathwater.

edit: mispelled "throw"?!?!

 

[rickcr42]

I use the old LH0063 buffers on the output of my preamp on line out #2 .
The preamp gain control controls the headphone volume and only the main output has a mute switch (main power amp mute/buffer active)


I use the LH0033 stacked on the output of my passive recorder dubbing panel for monitoring but that box includes a volume attenuator

so the answer is yes .

If your buffers have a high enough input impedance and gain is not needed you can hang them on just about any line level output for current gain and cable driving

feed them well though (power supply is critical for best performance)

 

[tangent]

You could try building a discrete op-amp with a gain=1. This would let you choose output transistors for high output current, while still having feedback for low distortion. I believe the Szekeres design is along these sort of lines.

 

[Jupiter]

Quote:

Originally posted by Glassman DACT makes almost-passive preamp using their stepped attenuator and output buffer (Analog's BUF03).. this is the very same story and I think it's generally very good idea..

Thanks for the tip. I had a look at the CT101, but DACT recommends high impedance headphones (>600ohms) for it. I've emailed them asking if they have something that can be used for low impedance headphones.

Quote:

Originally posted by jeffreyj By themselves, buffers will still deliver high speed and high output current, but distortion and DC offsets will be high; this is why an op-amp is pretty much always included even if no voltage gain is required. Trying to strip unnecessary components from the signal path is always a worthwhile endeavor, but don't throw the proverbial baby out with the bathwater. edit: mispelled "throw"?!?!

I looked at the Intersil HA3-5002 datasheet. The distortion is low (0.005% at 10kHz) and the DC offset is usually OK (typical 5mv, max 20mv). It looks like this buffer would work, but maybe I'm missing something.


rickcr42,

I'll take a look at the datasheet for the LH0033 and LH0066. Thanks.

tangent,

The Szekeres looks like one solution to my "problem". Should be cheap, too.

 

[rickcr42]

good luck finding the LH0033/63 parts . They have been obsolete for some time.

funny thing is that while they do not spec out as well as some of the newer buffers they sure do sound damn nice

the "66" looks like a TO-3 can but it is multipin and goes into a custom factory heat sink/socket assembly.This baby is huge going 4X5X1.5 per buffer !

(the heat sinks)

the "33's" are also in a non standard wide dip ceramic package and heat sinking these babies was a real chore.

since I built the devices modular i pop the buffers out occasionally and try various compound layouts and still end up using them straight out.

The Szekeres is a wonderfully simple and at the same time frustrating design

to look at it it looks like nothing . Nothing until you get into how important the power supply and parts quality are mainly due to the simple circuit

straight from the project page and used with all radio shack parts plus a radio shack 12 volt regulated 3 amp supply and it sounds pretty darn nice.Not the ultimate but pleasing.

but use higher quality parts , isolate the front with a dual jfet , throw in a low noise wide band power supply and maybe an active current source (I like the resistor

) and it becomes a whole new animal

but for me an animal i can't leave alone !

I am always tearing it apart to "improve" it

oh well , there are worse things .................................


rickasaurus

 

[rickcr42]

main product page

http://www.national.com/pf/LH/LH0033.html

data sheet

http://www.national.com/ds.cgi/LH/LH0033.pdf

enjoy

rickeraptor

 

[Jupiter]

Thanks for the link.

Looks good, except for the output impedance. Typical 6 ohms, max 10 ohms. Isn't that too high for Grados?

 

[rickcr42]

I always put a 10 ohm resistor on the output of all buffer circuts for protection and even higher for line level , 51 ohm

sometimes it comes down to listening and not just what is supposed to be good

for instance : my line level buffers are mostly used for the tape record output where more often than not the actual recorder is some distance away ( actually it feeds a line level mixer near the deck) and on the output of line level tube gear where the high output impedance is not a good match for the solid state gear inputs.

so I use a 1meg ohm input impedance .

now this makes the stage susceptable to RF interference and is considered a bad thing but just throw a small rf cap across the resistor and the problem is solved !

same with output resistors . I feel much better having some protection than worrying about the Nth degree of resolution which i am not sure i can hear anyway .

I can convince myself that i hear certain things but when i have assistance and i do a blind test I am wrong 50 % of the time which means diddly.

now one thing i hear all the time is power supply differnces , but again different does not always mean better , just different

and btw (off topic but time i weighed in somewhere) ---the devil smileys are mine , were created for me (do a search) and are here to stay !

rickasaurus rex

 

[Steve Eddy]

Quote:

Originally posted by Jupiter Thanks for the link. Looks good, except for the output impedance. Typical 6 ohms, max 10 ohms. Isn't that too high for Grados?

Looking at the impedance plot of the Grado 'phones, it's virtually flat across the audio band. So the output impedance will pretty much just result in a bit of attenuation.

se

 

[Jupiter]

Great, then I don't see any reason why the LH0033 shouldn't work.

So far I only found the LH0033 at Future-Active, but they require you to order 100 pieces.

 

[Steve Eddy]

Quote:

Originally posted by Jupiter Great, then I don't see any reason why the LH0033 shouldn't work. So far I only found the LH0033 at Future-Active, but they require you to order 100 pieces.

Why not just go with one of these designs:

A Class A MOSFET Headphone Driver

A JFET-MOSFET Headphone Driver

se

 

[rickcr42]

Quote:

Why not just go with one of these designs:

perfectly legitimate designs and the "class A MOSFET Headphone Driver" or "Szekeres Amp" is one of my favorites if done right

both open loop buffers and voltage followers are fine for hanging anywhere you have a need for current gain but will require an additional stage for voltage gain-something i rarely have a need for.

The szekeres being extremely flexable if you know what you want , being infinately variable in operating points (VA)

I have had it as high as +30 volts @ 500 ma per channel !!!!!!!!!!!!!!!

at this time i run it at 24 volts/200 ma per ch and have a dual jfet front end which replaced a 5687 tube front end-too damn much heat was generated in the 8X6X3 chassis.the heat sinks are rated at 20 watts and that is just for the mosfets !

The current source resistors are bolted to the chassis bottom

monolithic buffers while not giving you the freedom to choose most of the parameters DO offer fast build time and easily forseable results bad

supply current to the device at spec "A" and get current out at spec "B" with voltage "x"

And I must say the National LH buffers are not the easiest to use nor are they the ultimate in final current output

The 33 is only +/- 100 ma while the 63 is +/- 250 ma and that in a package well larger than say the BUF634 for the same output

Like i said-the LH0063 requires a custom socket and heat sink and where you would come up with one I do not know though it does look very high on the cool scale , kinda like a small power amp

and the LH0033 ceramic version is non standard in all ways so heat sinking it for me was a bit*h

I ended up using a DIP IC heat sink for each one and cranking them down to the perf board with copper wire , after laying on the paste

the metal can package looks like the winner in ease of use

TO-5/39 heat sinks are easy

but again , the ceramic looks cool

oh yeah

you could also "clone" the buffers using discrete parts

only four parts in the 33 and not many more in most good buffers

 

[Jupiter]

I was thinking of trying the LH0033 (if I can find it) and some variation of the Szekeres.

I'm not just looking for the best solution, I also want try a few alternatives. I'm not going to try and clone the LH0033, I would need a schematic with values for that.