enter image description here

enter image description here

I need advice if it is possible to replace switches ( see drawings ) with suitable transistors and other parts.


  1. Base leg voltage to "open" transistor ~0.2 - 0.3 Volts.
  2. Transistor remains "open" until base leg voltage drops to less then 0.1 Volt. OR duration for "open" position of transistors 5-10 seconds to allow maximum energy transfer from C1 and C2 to C3.
  3. Amperage levels on all stages are 0-2 Ampers.

Purpose of this device is to "upgrade" energy available from very low-level sources to more useful levels. I tried many different kinds of DC to DC up-converters. Some of them are really well made and do great job.

But all of them have very low efficiency levels (10 - 40 % ) and if more higher level of upgrade requested ( for example from 0.1 Volt to 5 volts ) lower is efficiency. Multitude of experiments show that supercapacitors ( directly connected to the source ) do much better job. Circuit on the drawing works, but need constant human intervention. I think automation of this process is possible.

My knowledge and practical experience in Electronics are very poor. Please, help with selection of components needed for this unit from existing marketplace.


Use electromenchanical relays with changeover contacts. They have life operations in the tens of millions of cycles and if you needed to switch one every ten seconds, then you get over 1000 days before you might need to service a breakage. Toggle the relays with a low frequency oscillator but you are then wasting energy in the relays (as you are when using human intervention to manually operate switches).

Or use SSRs (solid state relays). A bit better efficiency compared to normal electromechanical relays of course.

However, I note what you say about the minimum voltage being 0.2 volts and I would like to point out that there are offerings from both TI and Linear Technology that work down at this level to produce 5 volts on a super cap. You also mention efficiency and I think you are somewhat misled in thinking that dumping charge from one capacitor to another is an efficient process - it isn't because the energy lost is 50% for each cycle. And you are doing this twice per cycle - once when you parallel the capacitors across the primary energy source and then again when you try and raise the voltage across supercap 3.

I think you need to look at the losses and ask yourself big questions.

  • \$\begingroup\$ +1 for "need to ask yourself big questions". We really don't know what OP tried, and 40% efficiency, even for boost/inverter is a bad efficiency. \$\endgroup\$ – Marcus Müller May 6 '17 at 9:05
  • \$\begingroup\$ Thank you Andy aka! Matter of efficiency. Equipment in use: 18 of TEC 127 6A connected in groups of 3x6 ( 6 in series and then those three groups connected in parallel ) . Temperature difference hot end and cold end 4-5 Celcius ( ambient ~25 C for hot end and 20-21C for cold end ). \$\endgroup\$ – Nioko May 6 '17 at 10:51
  • \$\begingroup\$ Voltage from this configuration without load ~ 0.45 Volt. Load is small motor from the phone. Voltage under load is ~0.23 Volt, current ~0.04 Amper, so output of device ~0.01 Watt. Motor makes ~750 RPMs withsmall 5 cm two blade propeller. Next item is Maxwell supercapacitor 2.7 Volts, 350 Farad and a last one LT 3108 based DC to DC microharvester.Here is a link: ebay.com/itm/… \$\endgroup\$ – Nioko May 6 '17 at 10:51
  • \$\begingroup\$ @Nioko why are you telling me this? \$\endgroup\$ – Andy aka May 6 '17 at 10:52
  • \$\begingroup\$ I'm describing experimental data to address matter of efficiency difference between direct " energy dump " to S- capacitors and DC to DC upconverters. \$\endgroup\$ – Nioko May 6 '17 at 11:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.