[ppl]

Well in my endless prsuit of Headphone Nervina, I spent the time i should have alloted to Finishing the Board for The All BUF-634P Version of My Pocket Amp. The Folks at Intersil informed me the EL-2001's and 2002's are not recomended for new designs. However is and will be available for some time. depressing as this may be i thought the Good old BUF-634P operated at a moderate bandwidth on the output stages and low bandwidth on the Virtual ground Buffer. This ment making sure that several tolerences of BUF-634's and TLE-2426's track each other reasonably well or Additional Battery drain will result from the parts Fighting each other. I then made an adaptor socket so as to drop the BUF-634 in the Virtualground Driver's socket of my Presently constructed pocket Amp so as to Evaluate The Interaction between several BUF-634's and the TLE-2426 on the Opamp rails. Well if the date Codes on the BUF-634 and TLE-2426 are the same thay all worked if one year apart 6 out of 12 would not cause more than 1 Ma of additional Current over an Ideal match. I feared not wanting to make the pocket amp into a project if a Long Term issue of unavailability of critical Parts are a problem. Well the BUF-634 is going to be around for quite a long while so it is wise to base the Amp of the Future around The BUF-634P's rather than the Elantecs.

After conferming that at the worst case the Builder could Expect about 1 ma if additional Icq out of a random selection of TLE's and BUF's vs a Matched set of devices, that most folks would not complain and if it was an issue thay would buy several additional Devices each so as to have a Better than 1 in 6 chance of getting the Lowest possible Current Drain. The BUF-634's in the low bandwidt mode draw about the same to 0.8 Ma each more than the EL-2001's still Quite reasonable since Just replacing the EL-2001 with the Selected BUF-634P in the afore mentioned Adaptor socket resulted in the Icq of the Amp going from 10.28 Ma to 11.63 Ma. Still 9 Volt battery frendly. and a worst case of 12.96 Ma with the largest mismatch of Devices. The Outpput stage is going to add about 6 ma (3Ma Ea. Ch) If biased to the 70 MHz. bandwidth of the EL-2001's presently on board however that issue will have to wait untill i get the Dedicated BUF-634 board done and the Point i would like to talk about is the Differences in sound between various Buffers used as a Virtual ground driver while leaving everything else the same including the EL-2001's at one ea. per Channel. Driving my Sony MDR-V6 Cans.

Having Droped the BUF-634 in somthing unexpected happened the sound changed and the change was for the upper mids and Hig's to take on the sonic signature of the BUF-634 and I know what the BUF-634 sounds like. I changed back to the EL-2001 Device and the sound quickly reverted to it's old self. well this led me on q quest of mixing and matching output stage and Virtual ground drivers. the output stage Buffers was changed to several EL-2001's in parr with a similar amount on the Virtual ground Drivers in all combonations. Too manny to list so i will just keep it to the BUF-634,El-2001 x1 and EL2002 X1 Virtual ground Drivers and EL-2001 X1 and EL-2002 X1 Output stages.

Going from the EL-2001 to the BUF-634 resulted in sonic changes so i thought of trying an EL-2002 as the Vir Gnd driver with EL-2001 still in Output stage. Again the sound took on the charictor of an EL-2002. this had me wondering what would happen if the Output stage was also upgraded to the 2002's. Wow what a Synergy of sound resultred from this combo Quick fast Lighter and more Air than the Origenal 2001's. Changing from all 2001's to all 2002's made the bass seem thinner and alot faster maby? the improvement in the uppermids and the 7KHz untill you can't hear no more range now has the Imaging normaly asseated with the Midrange on the 2001. with the 2001' the upper octave is heared but lacks that 3D image it so well Projects in the Midrange and Evean the bass. The addition of just the 2002 vir gnd driver only with the 2001's still in the output gave Dimention to this upper midrange and Ultra Highs that were not present with the 2001 as the VGD. upgrading the output stage to the 2002's gave a Holistic 3D image of this octave rather than just a Dimention or two. The effect is quite noticeable and if Fast articulate transients are importent to you and you can give up a slightly fat bass for a thinner but better controled bass go for the EL-2002's all the way around. what elce will it cost you well the Amps Icq jumps from just over 10Ma to just over 21Ma. not 9 volt battery frendley but AAA frendley.

 

[Audio-Me]

I like your findings of going all EL2002, I would like to have best of both.

Could you describe how it sounds with EL2002 as buffers and EL2001 as VGD? I'm thinking of going stacked on all (mine is a bulky travel amp w/ external AC psu, hehe).

 

[ppl]

Audio and Me> The difference between using a 2002 as a VGD vs the 2001 with a set of 2002's in the Output stages is slight but it dose give an added deminsion to the 7-Can't hear Octave's. Regarding stacking if your Phones are 100 ohms and above not mutch to gain sonicly by parralelling Buffers. My MDR-7506 are 63 ohms and farly reactive so the More Output Current an Amp has the better the Sony's sound. This is evean at my rather tame listening Volumes that typicaly Put only 5 Ma into the Phones. using a Resistive load playing Music from my Pan. SL-SW860's Line out I still only measure 30 or so Ma Avg. I could not listen to the MDR-7506/MDR-V6 at that volume for long. I at times when using the Bass Boost and Dynamic recordings like the Telarc 1812 and Telarc's bachbusters managed to Bottom out the drivers and not be into Clipping.

Hearing a difference between VGD is somthing one would not expect Given that two Tonex 1000uF/10V Electros and 0.68uF/160 volt Wima's are bypassing the VGD and would think these would swamp the Impedance of the VGD within the Audio range.

Since my Amp is powered by two 150 MAh NI-MH 9 Volt's i am at the Present using a set of 2002's on the Output One Per Channel and a 2001 as the VGD this consumes 18Ma of ICq about as High as i would like to go from that kind of battery source. I may go back to Two EL-200'1 per channel and one EL-2001 VGD if battery life becomes a problem. This combo draws 13.26 Ma.

 

[morsel]

Great posts and reseach, PPL! Thanks for doing that. One point, you mention that there is not much to gain from paralleling buffers for headphones above 100 Ohms. It was my undestanding that even though there is no need for additional current, the lower output impedance of paralleled buffers gives tighter control of the buffers over the voice coil of the headphones, particularly for bass, thus theoretically providing some improvement for higher impedance phones, too. What do you think?

 

[aos]

I parallelled my output buffer last week (just soldered one on top of another) and I think that indeed bass is better but it might well be psychological. I don't have patience any more for A to B tests, I had enough of that last year. So these days I just take suggestions I read for granted (as long as they make sense), I try them and if nothing smokes and there is no audible decrease of quality, I say that mod worked and move on to the next one

.

 

[aos]

ppl, if I'm getting you, you say that if BUF634 is used as virtual ground driver, even if you use EL2001 in output stage, the sound will be as if BUF634 was in output stage?!

 

[geom_tol]

ppl. Cool that you are testing all these combinations. Is it possible for you to make two battery packs and listen to the sound with no ground driver?

I know of the safety and battery drain advantages when using a ground driver. But it would be nice to know the result for reference.

 

[ppl]

aos. With the BUF-634 as the VGD the sound takes on the Sound of The BUF-634 only slightly and in the upper octave only it is like the sound is a Blend of the Elantecs with about 10% of the BUF-634. I tried a similar test on One of my larger amps using the EL-2008 and Three BUF-634's as the VGD this amp uses alot more Rail capacitence. By Replacing the BUF-634 X3 VGD with an EL-2008 the sonic change was not diectly Noticeable i think the Bass may have got slightly Better, 2008 as the VGD, However it might not have. Regarding Paralell EL-2001's with High Impedance Cans. With my 220 ohm Koss one or 4 Buffers sounded the same. with my Newly obtained Sen 535 that jr music world was discounting at an unreal Price the sound did improve the Bass defintly got Tighter and these are 120 ohm cans so i gess the Improvement is Headphone dependent and not perhapps directly related to it's Impedance, so i would sugest Exparamentation to be sure. I have no doubt that the Improved Bass happened in your situation as this is somthing that really stands out and is not Subtle like other Mods.

Regarding using the Centertap on the batteries rather than the VGD this sounded the best but one battery went dead prior to the others and put 10 Volts DC on the Amps output and fried my then used SR-60 so i dont Recomend this as the improvement is not all that great and if operated in a Virtual battery supply a slight HUM was noticed by using the Batt cent tap rather than the VGD. this was using Ni-MH AAA batt pack if Alkalines are used the VGD sounded alot Better with Battery operation. This was years ago when i First started designing battery operated Headphone Amps as my Prior stuff was ALL AC powered and Class A or as Close to it as Possible to Class A the battery powering was somthing i did for other People in Analog instermentation in these cases the VGD is almost always used so i went that way after being unable to find rechageable batteries that were matched and if at first matched that later became unmatched and like i sed the improvement is not all that great and not worth a pair of cans as NI-MH batteries when Discharged go fast no by the time you notice the battery is going ded it is two late.

 

[aos]

Thanks for the info ppl. I originally designed PCB with BUF634 in mind as EL2001 was (and now isn't only because of tangent) exotic, very rare, buffer so I wanted to use my precious supply only as output buffers. Plus BUF634 has higher current capacity. However, at the moment price of EL2001 is available and much better priced and if it also sounds better I should probably make the change. Too bad I already ordered some PCBs but hey, exacto knife and few pieces of wire will sort that out nicely if it starts disturbing my sleep.

 

[ppl]

aos> you can do as I did and make an adaptor socket if you are using Dip's. Then your exsisting board is Ok. I am now back to my orig Configuation on my pocket amp Two EL-2001's on each channels output. one EL-2001 as the VGD. and it sounds great. The change with using the BUF-634 as the VGD is not bad and it dose not make the Sound worse just different. it is not like using the BUF-634 as an output stage. While i think the EL-2001's sound Better than the BUF-634 it is possible to change that with other Circuit tweeks like More feedback by a Higher open-loop gain this works great with the BUF-634 but This weekend i wil hopfully finish the All BUF-634 version and then start the long process of tweeking the design.

One note on parralelling Output stage Buffers be thay Elantec Bur brown or anyother OPEN-LOOP Buffer i like to have the devices that are parralelled matched for DC offset as if thay are raticaly different the Icq will go up as the Buffers Fight each other to get the DC offset of the lowest one. This is however only a small increce , But i also select all devices be thay opamps Buffers or whatever i select for the lowest current drain ones like the EL-2001 can drain from 0.33 to 1.5 Ma. at the same supply voltage so it is possible to select eather for more Class A operation Higher Icq ones to saving a few Ma of current drain in the amp and on 9 volt Ni-MH's a mil or two can result in quite a noticable difference in runtime before needing Recharge. Your Mileage may vary. The 100 ohm limit befor having to parralell buffers is based upon resistive loads and headphones are reactive loads and the more available output current the Better that is why i think the 1 Amp EL-2008 & 2009 is the Best darn Monolithic buffer recently available at under $100 ea. The Apex WB-05 is better sounding but is not only unavailable has no output Protection and is in a 8 lead TO-3 case. the EL-2008 & EL-2009 are darn close however.

So I recomend you try the Parralell thing yourself and see if you like it better or not. the last time I recomended parralelling BUF-634's I was acused of designing Complex and expensive amps, however that was when every one thought the Cmoy or seakers amps were state of the art and the concept of using Buffers was an alternate universe to most See http://headwize.powerpill.org/ubb/sh...&srch=buf-634;