sziklai-darlington diamond buffer (SDDB) for power output stages

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

I've seen two interesting threads recently, one is
four transistor emitter follower (diamond buffer) as power output stage?
capslock said,

The good old four transistor unity-gain buffer which consists of a complementary emitter follower, another emitter follower pair that replaces the bias diodes and two current sources is very popular in complementary integrated op amp designs as well as some discrete designs. For a schemetic, take a look at Sonny's page:

http://www.geocities.com/sonnya5000/basicb1.jpg

He did some nice simulations about distortion and how to improve it for low impedance loads.

What I wonder about : why is this nice topology apparently not used for power output stages? Is it because it is too difficult to keep the bias constant?

Puzzled,

Eric


the other is The Ultimate Suplifier

and found Marc. Vi's design gave me an idea , why not use
sziklai-darlington diamond buffer (SDDB) for power output stages ?

So I changed a little circuit of my own DB-01 headphone Amp,
to test the SDDB power output stages, and find it works very
good with my AKG-501.

I think it can be changed with only few components to
be a small Class-A power Amp.

here are the two circuits' pdf.files and pictures about the head
phone amp, please note the first output stage with only 4 bjts
, and the second with 6 transistors.

M.T.



DB-01 headamp amp circuit

ULTIMA-SDDB CLASS-A POWER AMP

http://homepage20.seed.net.tw/web@3/asaas/images/DB%20with%20SDDB/DB-01-headamp-512x384.JPG

http://homepage20.seed.net.tw/web@3/asaas/images/DB%20with%20SDDB/SDDB-01-512x384.JPG
 
ULTIMA-SDDA amp can be seen as a classical op amp plus
a 250mA high-speed buffer just like Burr-Brown BUF634 with
a op amp on page 9 Fig 5 high performance headphone dirver.

http://www-s.ti.com/sc/ds/buf634.pdf


(A) 1st differential stage + 2nd p-JFET SE VAS as normal op amp.

(B) sziklai-darlington diamond buffer just works like BUF634 but
can get higher current gain and supply more current using
6- transistor like Marc Vi's output stage (BC639/640 +BD139/140
+ 2SA1943/2SC5200).


Fig. 5 OPA134 or 132 + BUF634 headphone driver
 
lumanauw said:
Nice design. How is the sound?
In your photograph, if it is a building, not a PCB, it will need an elevator from basic floor to 4th floor :D

According to several diyers' reports who bought my licenced bare
PCBs from DIYZONE , the
DB-01 sounds very well, because I choose Linear System's LS844
Dual-JFET as the first input stage and Silinoix or ON's 2N5461 P-ch
JFET as the 2nd stage, they both have very low noise and large
linear dynamic range with good sound quality characters.

So I can forcast the new pure DC coupled class-A ULTIMA-SDDB
will also sound very well ,too.

If I make a six-channel amp by using the same PCB structure ,
then it will be a 12 story tower with several parallel pipelines
on the backside. :eek:
 
Wow now this input stage you have used in theory must be perhaps the most Linear Differential input stages available in discrete Transistor Form. The DC qualities of the supper matched (I have tested and they are supper matched) JFETs LS840 available from Linear Integrated systems are IMHO the Finest JFETs available for ultra low noise DC stable Differential input stages. http://www.linearsystems.com/datasheets/LS840-2.pdf

I use The LS840 and the faster LS5912
http://www.linearsystems.com/datasheets/LS5911-2.pdf
that’s also available elsewhere or at least used to be in this location a lot and they sound better than any others I have used at about half the Input current one expects to operate such devices at. These things are superb at 1mA per transistor making use of these in portable equipment. These also make for the best sounding cascoded current sources I have used. Substantially better than the 2N5486/2N5486 pair as depicted hear http://tangentsoft.net/audio/bitmaps/opamp-bias4.png

A self-biased pair of LS840’s will typically maintain from 0.95-1.2 mA from about 2 volts to 20. Now that’s pretty good voltage compliance IMHO. Considering the low capacitance of these devices it just seems natural to use the previously mentioned cascoded LS840 pair to bias the LS840’s used in the Diff amp of the Circuit under consideration http://homepage7.seed.net.tw/web@3/asaas/PDF/ULTIMA-SDDB POWER AMP V1.0.pdf

I like how the gate is returned to the source so as to apply local feedback. Boy this works great but sometimes can be pone to instability, However one its working its great Ultra low DC drift and offset and low distortion naturally prior to the application of Global feedback. Loading this stage with a current mirror again a great idea for good DC performance and obtaining a higher slew rate at a given Bias current. The transistors I would use in this location are the LS 352 or MAT03 http://www.linearsystems.com/datasheets/LS350-2.pdf

MAT-03
http://www.analog.com/UploadedFiles...rm=Transistor&fterm=Dual PNP Transistor&la=en


The 60 volt VCE of the LS part is preferable to the lower by almost half Vce Max of 36 volts for the AD Device so as to minimize the possibility of the early effect. By virtue of the inherent matching heat in these supper matched devices DC drift is again minimized. So far it would appear we have an ideal input stage for some real detailed tunes without any distortion. Providing the remaining circuitry is of the same caliber or better. Something I would have liked to have seen in this stage is a casscoded current mirror, its not like you need to worry about dropout voltage and the lower miller effect of the casscode configuration will make for lower capacitance and hence again higher slew rate per mA of Bias current, This IMHO again would be the icing on the cake for this input topology. I must close for now so hopfully this is some food for thought for you.

BTW boy the guys over at DIY Zone sure do some nice work http://www.myclie.info/images/OP0330-12.jpg
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.