buf634t supply voltages question?

[crippledchicken]

hi, i've replaced the lt1010 buffers in my adcom gfp 565 preamp with the buf634t. the voltage supplys to the buffers is -17.34 and +17.48 are these values safe enough, for the +/- 18v. max of the 634t? or should they be reduced slightly? these sound much better in this preamp running full band width. there bypassed well, on original heat sinks, and don't seem to run hot. has anyone ran them that close to max v. without any problems? thanks in advance crippledchicken

[Alcaid]

The buffers should be fine. Other forum users who uses the BUF634T at +/-18V can be found here:

High-end preamp for my GC

[peranders]

crippledchicken, you could lower the idle current a bit without no harm. Use 3.9 k at the BW control. Check how I have done.

http://home8.swipnet.se/~w-80046/hif...0schema_p1.pdf

[crippledchicken]

peranders, in that schematic is the jp-1 there to create two options of biasing, full or trimed? i'm not real good with schematics. also, by adding the 3.9k resistor will it change the sound to some degree? i assume, the resistor is mainly to lighten the current load, and keep the temp. down. i don't plan on using the phono, or tape out at all so i've removed one end of the small wire jumpers, feeding supply voltages to the tape out op amps, the phono op amps, and lt1010 buffers since there all running class a, and not being used any way. i figured this might lighten the total current draw since it all runs off the same +/- 18v source. if you think it would be better on the buf634t buffers to add the 3.9k resistor to reduce the bias current, i'll surely give it a try. thanks alot you guys are the best, crippledchicken

[peranders]

My experience about "full speed" and "medium speed" is that the difference is very small. You should test it yourself.

[Christer]

Originally posted by peranders
My experience about "full speed" and "medium speed" is that the difference is very small. You should test it yourself.

Unless I am mistaken it is not only a question of speed. What
you program is the bias current of the output stage which
determines how much output current you can deliver while
still staying within class A. In the widebind mode you still have
a bias current of only 15mA if I remember the figures correctly.
That may or may not be sufficient. I made som measurements
recently with my headphones and tried a bunch of CDs to figure
out the maximum volume setting I would reasonably imagine
using. I then turned up the volume one nothc higher to get
some margin, played a CD with a 100% modulated sine and
measured the current and found it to be on the order of
50mA. That is, the 15mA seems not quite sufficient ot stay
within class A for the loudest listening levels. Maybe it doesn't
matter so much to get into class AB for the loudest passages
though, but with a discrete diamond buffer it would be no
problem to do so.

[peranders]

I must add that I mean the buffer in a feedback loop together with a AD8610 opamp.

[Christer]

Originally posted by peranders
I must add that I mean the buffer in a feedback loop together with a AD8610 opamp.

Yes, but as long as we are talkning BUF634, does that matter
for my argument?

If it matters sonically to get into class AB for high levels is
a different question, though.