Zen -> Cen -> Sen, evolution of a minimalistic IV Converter

C2 capacitor choice

just checking as i'm getting the straggling couple of pieces that i don't already have to populate these kits, MUST C2 be a bipolar electrolytic, or is that just a suggestion?

thats the great thing about this build, i have just about everything i need already. the psu will require a few bits, but not too much.

if anyone would like to throw in for an A123 or olevinpower LiFePO4 bulk order let me know. we may be able to organise them to be dropshipped, as shipping any type of lithium cell, even the safe ones like this is often very expensive depending on the carrier; one assumes Hobbyking has this worked out. we may be able to organise a multiple cell balance charger GB as well, although its too late for me as i already have that sorted
 
one more thing, since we need to double the jfets for stereo balanced on the 9018 vref version do we get a couple extra heatsinks in our kits? if so excellent, if not how do i make that happen?

I think i only need it for the SEN, i will be trying both, but i'm pretty convinced i'll prefer the balanced single ended sound sig and i certainly have a larger bevy of well matched quads of 2sk170
 
Last edited:
The LM317 supplies around 60mA and some tens of mA are shunted to ground. For one CEN-SEN i would go for 300mA per winding. Some may prefer a bigger transformer but i can not see an advantage besides that big transformers have even a bigger stray field.
I was quite shocked to find a really tiny transformer in the most expensive version of the Tom Evans "The Goove" so there is a school that prefers the smallest transformer possible.
I am somewhere in between. Solve the job well but nothing more.
 
The current consumption for 4-JFET CEN/SEN per single ended channel is about 16~22mA. SO for one channel balanced about 45mA max, and for 18V Vout a 10VA transformer should be plenty. Even 5VA would do. I would suggest changing Joachim's design value of to 25mA for the LM317 per each single ended circuit (by replacing the 1R22 with a 5R resistor). The shunt device still has enough current to sink.


Patrick
 
Last edited:
did anyone end up building the untested current conveyer bipolar (would this be an apt description of the topology?) stage posted by Patrick for 9018 a while ago? i wonder why it slipped by in the night without a wimper? is it the lack of jfets? i have a bunch of bc550/60 hfe matched lying around and i could easily bend lt3092 to the task at least temporarily for testing, anything stopping me from breadboarding it as posted in post # 238 ? anyone do it and find any issues?