[ppl]

Regarding Your Concerns

(1)The PPA prototype board I using has OPA637/627 combo and bias @ 16-20 Ma no stability issues. With the Triad Module the primary reason that stability issues arise were too low of a bias 10mA combined with having the Module socket and mounted what are stilts placing the output stage well away from the remaining circuitry. In addition since the PPA2.0 has the buffer on board and as an integral part of the board then the rest of the amp can be designed to accommodate this. So with the default @20 mA this will not be a problem.

(2)While not finalized yet alternate resistor values in the emitter stage for portable low bias use and high bias home use.

(3)AMB correctly pointed out that the Bias is also adjustable with a trim pot so one could set it with in reason to achieve there desired goal

(4)As for concerns regarding the sound everyone that has replaced the IC buffers with 5002’s have reported sonic improvement not one person has
reported a reduction in sound so these concerns are not warranted

(5)Regarding the LED I think that one led position to the side is plenty as was done on the M3

(6)With the Bass boost pot mounted in the center this places the volume control, Bass boost and Headphone Jack at even intervals and surly one would not use the same size Knob for bass boost as they would the Volume. There are exceptions Like xtreme 4099

 

[skyskraper]

morsel: i have deeply seated anxiety issues.

 

[ppl]

To Illistrate what a smaller Bass boost know will look like i present a photo of my GR Headphone Amp/DAC http://ipodstudio.com/photopost/data...Front_view.gif

Also note that the 1 meg resistors have been changed to 470K to reduce excessive feedback in the gain channels and make sure High speed opamps are stable. Moreover if i find room on the PCB adding a Zobel network is a good idea as this type of network is most often required to maintain stability in all Audio Amps. with the HA5002 IC buffers we were able to skate around this however as things improve this may become needed with HP opamps however im not using one and the resistor changes resulted in OPA637/627 stability with a rather low bias of 13 mA

The 4k resistors in the ground channel should be changed to 10K this makes the ground channel more stable

 

[Voodoochile]

Is there any chance that the boards will be available before they are actually completed?

 

[tangent]

Quote:

With 300 ohm phones, the max peak output current just before clipping (assuming 24V power supply) is about 50mA

Can you give me a reference for how to do that calculation? Everything I've read so far doesn't seem to go deep enough, or I just haven't understood it yet.

Quote:

In the M³ [there is] no difference in distortion whether the load is 300 ohms or 32 ohms.

In my latest PPAv2 RMAA test, I found the same thing at 36mA bias. My buffers' resistors aren't optimal, though, so we may find a lower point where the distortions equal.

Also, it should be pointed out that these RMAA tests are a little unrealistic, because nobody listens to 32 ohm phones at the same voltages as 300 ohm phones, but we test at the same voltages for convenience. All else being equal, the 32 ohm phones need less biasing than our tests indicate.

 

[morsel]

Voodoochile: For you dear, anything.

bg4533 and PPL, about that bass boost pot:

Quote:

With 1.15" spacing between pot shafts, there is not enough room to use 1" knobs comfortably. The problem with moving the bass boost pot over is that the space is needed for the headphone jack(s), LED, etc. There is less than 1.5" between the bass boost pot and the edge of the board. If we move it .5" to allow for 1" knobs, that leaves less than 1", and also puts the bass boost knob off center.

If we move it over, it only affects those people who are actually planning to use the bass boost knob. If those people who are into bass boost are OK with this, and/or plan to use oversized cases, in which case (humor, arr arr) there will be plenty of room to panel mount everything, then it doesn't matter if we slide the bass boost pot over, and thus allow room for the use of 1" knobs. Also, if compact Switchcraft or CUI Stack jacks are used, 2 can be mounted vertically to save space. What do you think?

addendum - If you want knobs spaced farther apart and don't require battery operation, you can always build an M³ instead. Leaving the PPA knob spacing as is makes it more practical for Eurocard case use.

 

[amb]

Quote:

Originally Posted by tangent Can you give me a reference for how to do that calculation? Everything I've read so far doesn't seem to go deep enough, or I just haven't understood it yet.

It's not exact, but say we have around 10% output voltage swing loss due to the output transistors and emitter resistors (and the opamp not being rail-to-rail), with 24V rails we get about 21Vp-p, or (21 / 1.4142) = 14.8V peak (not RMS). 14.8V / 300 ohms ~= 50mA peak current. The 90% swing may be a bit optimistic, but real world loads are also reactive rather than purely resistive, which increases current draw under dynamic conditions, so it's good to estimate a bit high rather than low.

As for the quiescent current needing to be at least half of the peak output current in order to remain in class A, this is true for any push-pull type output stage. See Nelson Pass's article on his A40 power amp, which has a good description about this.

Quote:

Also, it should be pointed out that these RMAA tests are a little unrealistic, because nobody listens to 32 ohm phones at the same voltages as 300 ohm phones, but we test at the same voltages for convenience. All else being equal, the 32 ohm phones need less biasing than our tests indicate.

It's true that 32 ohm phones don't need as much voltage swing as 300 ohm phones to get to the same loudness, but RMAA also tests at a fairly modest volume level (nowhere near amplifier clipping). So I think it's an "easy" test on the amp. At any rate, how much bias is considered "optimal" is not only based on required voltage swing, load impedance, output device rating/power dissipation management, and power supply capacity. It is also a subjective choice of the designer such that the amp could be comfortably called "class A" under all operating conditions that the amp is supposed to encounter. Some designers take pride in a dose of overkill, and the Gilmore dynahi comes to mind. That amp has quiesces at over 300mA per channel.

 

[ppl]

i vote not to move the Bass boost pot its ideal where its at for the hammond case the PPA is intended for.

 

[morsel]

Let's leave the bass boost pot alone then. The right channel has to be jumpered anyway, it is not much more effort to wire both channels to the pot for those who are going to use a larger case.

Are there any more requests or suggestions?

 

[ppl]

Quote:

Originally Posted by morsel Let's leave the bass boost pot alone then. The right channel has to be jumpered anyway, it is not much more effort to wire both channels to the pot for those who are going to use a larger case. Are there any more requests or suggestions?

I agree also consider if a smallere pot is board mounted then the vertical hight of the bass boost will be lower than the headphone jack and volume control. woth the Alps blue the vertical centers of the Vol,Bass and Headphone jack are the same

 

[tangent]

Quote:

say we have around 10% output voltage swing loss due to the output transistors and emitter resistors (and the opamp not being rail-to-rail), with 24V rails we get about 21Vp-p, or (21 / 1.4142) = 14.8V peak (not RMS). 14.8V / 300 ohms ~= 50mA peak current.

That's clear enough, but doesn't it assume that you're trying to put a full-scale output signal into the load?

Put another way, if you only need to put, say, 2Vrms into the load, don't you only need about 10mA quiescent to claim class A into a 300 ohm load?

 

[amb]

Quote:

Originally Posted by tangent That's clear enough, but doesn't it assume that you're trying to put a full-scale output signal into the load? Put another way, if you only need to put, say, 2Vrms into the load, don't you only need about 10mA quiescent to claim class A into a 300 ohm load?

Sure. I come from a background of building speaker power amps before indulging in headphone amps, and in that world you don't claim an amp to be class A unless it could remain in class A at its full rated output power into the standard 8 ohm load. With headphone amps there is no standard impedance, so it all depends on how you qualify the claim. Hi-Z cans do require a large voltage output swing to get loud, so it's not unreasonable to base it on the full output.

 

[individual6891]

Will there be any possibility of transistor substitution/rolling for the buffer stage? Some of those transistors are very hard to source in europe.

-Andy

 

[ppl]

Yes like the BC,BD and BF series parts as you all ready know the BD 139 and BD140 are available all over the EU the problem with the to-92 parts is the BC and BF parts have different pinouts most often the BC parts can be installed by rotating but this will not be good for the transistors that should face each other exactly what transistors are hard to soure ill see abut subs
also most EU semiconductor makers like Philips have almost discontinued throught hole parts in favor of SMT Semins might still have some however
the PN4392 can be subed with a J111 i think check pinout of the manufacture uou are using as Jfet pinouts can be different depending upon supplyer fairchilds data sheet is http://www.fairchildsemi.com/ds/J1/J111.pdf

digikey UK had the 2N5088,2N5087,MJE243,MJE253
http://dkc1.digikey.com/uk/digihome.html

 

[Nisbeth]

Quote:

Originally Posted by individual6891 Will there be any possibility of transistor substitution/rolling for the buffer stage? Some of those transistors are very hard to source in europe. -Andy

We could do a european group buy from Digi-Key when the boards become available

/U.