http://www.diyaudio.com/forums/anal...otrodded-blue-dcb1-build-366.html#post3753478
made some changes:
- change IRP240 & IRFP9240 with match pairs
- change 7812 with different brand
- relay Rdopper only use 47R
results:
- Vout: (+10.16V) & (-10.25V)
- Vdrop ccs: +rail (1.45V) & -rail (1.33V)
- Vgs: IRFP9240 ccs (4.31V), reg (4.21V) & IRFP240 ccs (4.36V), reg (4.28V)
- relay voltage: after 7812 (12.03V), after Rdopper 47R (10.9V), in4002 (4.2V)
- dc offset : R (2.1mV), L (0.8mV)
it's so fun to complete this, some kind of fun that can never be replaced by buying finished product + great support from this forum
http://www.diyaudio.com/forums/anal...otrodded-blue-dcb1-build-366.html#post3753597
After several additional parallel resistors from 100R to 10R, I get 2 possible choices of 47R parallel with:
- 20R (14R in total) resulting 4.93V on relay
- 10R (8.3R in total) resulting 5.15V on relay
Then I install 10R only (0.26V Vdrop, 26mA coming to relay) and get 5.02V measured on relay re-measure dc offset, negative probe on ground and measured L channel -1.2mV and R channel -1.7mV. previously i never noticed polarity on dc offset
http://www.diyaudio.com/forums/anal...otrodded-blue-dcb1-build-367.html#post3753848
We see two JFETs electrically chained vertically in any DCB1 channel, one at the top for signal and one at the bottom with gate and source strapped together for CCS. When they are not 100% same IDSS in all temperatures and the supplies are not 100% same DCV in all temperatures there will be some little offset. Since in everything made on Earth by humans and nature there are tolerances and drifts, we inevitably measure very few mV DC on every DCB1's output at a point. Most builders report in the +/-2mV or less.
If we choose the top JFET to be the one with the slightly higher IDSS in a pair then the offset will result negative given the supplies are very near also. Theoretically we would like the bottom JFET not able to overdrive the top one. Still there is no practical difference that I have seen in THD after differences so tiny. Some builders had mutually swapped JFETs within each pair after the matching and PSUs tolerances conspired somehow to an X result in a finished board, so to try finally measure it tiniest possible and negative in polarity. For aesthetic purposes mainly since nobody reported any plausible benefit. Experimental errors and ambient temperature range can render all the above futile also. So back to square one, i.e. better kick back a beer.