44V AC CT is OK. Make R16, R18 = 15k
Read carefully post #1, especially 4th paragraph (measure your JFETs' Id at 10V Vds with 10R degeneration - the value of R4, R6 will depend on that).
Edit: Actually, if you plan to use the PSU from post #1 (cap multiplier), make R16, R18 = 12k
Read carefully post #1, especially 4th paragraph (measure your JFETs' Id at 10V Vds with 10R degeneration - the value of R4, R6 will depend on that).
Edit: Actually, if you plan to use the PSU from post #1 (cap multiplier), make R16, R18 = 12k
Last edited:
Take BD13916 and BD14016 - it's the same thing in TO-126 package and both are made by Fairchild (suffix 16 means higher hfe).
That way the current through them can be raised to 10 ma and two pairs of output Lateral MOSFETs can be easily driven (if you wish that. But in that case your PSU should be up to it - you'll need 400VA transformer and double the filter capacitance).
That way the current through them can be raised to 10 ma and two pairs of output Lateral MOSFETs can be easily driven (if you wish that. But in that case your PSU should be up to it - you'll need 400VA transformer and double the filter capacitance).
Perhaps a little misunderstanding, my first question was not about how to calculate biases, it was about your personal reason behind the choice of topology change, between regular current mirror with hawksford cascode and wilson mirror. But anyway, more important to me is the answer you gave for the second question. Thank you.
No. First post, second paragraph says it roughly - you are lazy to read as much as I'm lazy to type...
Did you do the boards photoresist etch or iron transfer?Just finished making the Cubie 3 PCB's.
The reason I ask is the laser printout seems correct for transfer etching, but your board looks the same as the laser print, and it should be a mirror image. Just to avoid weirdness, look at it and make sure it is correct. If it is mirrored, you will have to pay very close attention to all polarities during build.
Last edited:
Thanks Juma, that explains it for me. I don't think I get it yet, but I think I'll understand better in time after working a bit more on it.
Another option is to build it inserting the parts on the copper trace side. The only problem is electro caps, since they block the solder pads if inserted fully. If one left just enough space to carefully solder them, and then use silicone to fill in the space, or let them stand on their leads (which could vibrate), one could populate a backwards board. I re-etched the Slewmaster board I originally etched backwards. Ps, the type (cubie3 and juma) should read correctly on the copper side.