Olin Lathrop embedinc.com> writes: > From: "Peter P." > > your regulators could be reduced > > to just one LDO regulator I think, cutting the parts > > count by 5 (5 parts to 1). > > You can get this opinion by looking at the regulator by itself. You need to > look at the context in the circuit and other bigger issues to realize why it > was done this way. Do you honestly think I'm unaware of LDOs and that using > them here never occurred to me? I do not make assumptions. All I can say is that if this project would be 'dropped' in my lap so to speak I'd probably shave off one third to one half of the parts in various ways. I'm not saying it can be done, I'm saying that's what I'd try to do. I think that your distributed regulator scheme tries to use delayed power-on to the MCUs which work as switchers such that they start working when the power rail supplied by external USB power is high enough, and presumably after USB power negotiation took place (I have no way to guess this it's a firmware thing), but in any case after IC3 will have reset properly and initialized (and they also isolate the programming Vdd when the 10Fs are ICSPd). I think that I would have used a LDO with built in voltage monitor from Seiko, Microchip etc. instead of IC1 and used the power good sense to release reset to IC3 and allowed the switchers to be off (idle) at power on, and turned on by IC3 when needed. I'd also have used a 1-way smpsu controller with 2 outputs instead of the two 10F's and used the 8.2V to make 17V. Also an active filter with gain for the PWM filters (saves parts for the same slope). Also no -2.2V supply (324 outputs pull down to 0 if loaded with 1k or active jfet or even bipolar current mirror half). But I'm just writing this email and it's your circuit, you tested it, and built it, and are marketing it. That's a lot more than any email can do. So this is just a guy's opinion. Peter P. -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist