2
\$\begingroup\$

Do Thermoelectric Generators (TEG) in series require a power balancer, when storing the energy to a battery source? If not, what should I do for my design to distribute to a battery source more effectively.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Consider that any TEG module is already lots and lots of little TEG cells in series. \$\endgroup\$ Commented Feb 15, 2012 at 13:42

2 Answers 2

2
\$\begingroup\$

TEGs do or don't require some form of power balancing depending how serious you are.

  • Putting them all in series will work well enough in most cases.

  • Paralleling is liable to lead to weaker modules making even less power than they otherwise may.

  • The excessively keen may consider using MPPT converters so that each module or group of modules can be operated at optimum load point and also provide optimum output voltage. Gains in output from MPPT have to exceed the losses in the MPPT converter (about 5% if carefully done and depending on overall topology)

I'd be interested in hearing details of what you are doing with them as I have a dozen or so crying out to me from their storage shelf to be used or at least played with.

  • I bought them for use with a trial fan blown stove that was able to use natural hydrocarbon based fuels that flowed about as well as week old porridge if not heated. Lots of fun playing but it didn't happen.

If all TEGs are notionally identical and have identical hot and cold face conditions then they will have the same optimum current and voltage conditions.
If truly identical you could parallel them, but you'd want to be confident that they were truly well matched. If used in series there is less problems when they are matched reasonably well but not perfectly BUT the current provided will be approximately that of the lowest output module.

Note that each TEG module already has many junctions in series internally so series connection is "part of the territory" However, you can expect (or at least hope) that the junctions in a single module are reasonably well matched and all tend to have the same hot and cold temperatures.Thermal equality assumes that there is no notable temperature differential across the face plates - not something you'd expect except in extreme situations (such as eg a point flame impinging on the centre of the face plate.

If you are serious about optimising output you may wish to consider providing an MPPT (Maximum power point tracking) converter per module or per group of modules. The gain from MPPT has to exceed the converter losses. If boosting you should be able to get 90%-95% efficiency with synchronous rectification and if bucking (less likely) maybe 95% or a tad more with exquisite care.


I'm working on an ultra low power, medical device solely powered by body heat. The design is a watch that keeps the modules fastened against the skin. Temperature gradient is optimized through heat sinks / heat pipes / ambient temperature. The problem with my current design is having equal surface area skin contact with each module, which led me towards the power balancing direction. I have looked over using MPPT (Cymbet EnerCHIP) on a bundle of TEG, but came to realize that perhaps the MPPT is optimizing a 'non-optimal' sum of the TEG in series

OK - this is a VERY serious application where every possible gain is needed.

MPPT SHOULD optimise power if that's what it is measuring. BUT with unequal heating per module you "have problems". As I noted, power will tend to be limited by lowest output module. MPPT per module is desirable by your low voltages eliminate many devices. IF your budget allows then a per module MPPT is best.

Also, you MAY be able to select modules with more junctions per module to increase voltage(but, you know that). Your energy output will be very very low - you know that too. Carnot efficiency from body (say 38C or less) to ambient- say 25C is (38-25)/(273+38) = 4%. TEGs with Carnot efficiencies in the 20%-30% range get 5% actual if very good. You may get 1%? No doubt you know what you can achieve.

For maximum net output its not obvious that there is any alternative to optimising the energy output per module and then summing them.

Some energy harvesting ICs use a high turns ratio transformer plus analog oscillator to boost the voltage from tens of mV to Volts. You could implement per-TEG low cost low size oscillators with discrete components - possibly one transistor and a coupled inductor per TEG, and then deal with the resultant higher voltages. You can also get voltage boost just by sing a MOSFET plus inductor and PWMing it to get HV flyback (where "HV" may be a few volts. Having all TEGs flyback into a common load capacitor in possible gives you independent operation of each TEG.

One channel:
TEG has reservoir capacitor connected across it. TEG - Inductor - MOSFET - ground connection. Output diode connected from MOSFET drain to output cap.

MOSFET is on for less time than it takes for inductor to saturate - not hard with very small inductors at these voltage levels.
Turn off FET. Inductor rings to load-cap voltage. While FET is off TEG is charging reservoir cap.

FET on off % can be managed to optimise output.
FET could be turned on when TEG reservoir cap reached some preset voltage.


Questions:

What module area?
What power out?

\$\endgroup\$
5
  • \$\begingroup\$ I'm working on an ultra low power, medical device solely powered by body heat. The design is a watch that keeps the modules fastened against the skin. Temperature gradient is optimized through heat sinks / heat pipes / ambient temperature. The problem with my current design is having equal surface area skin contact with each module, which led me towards the power balancing direction. I have looked over using MPPT (Cymbet EnerCHIP) on a bundle of TEG, but came to realize that perhaps the MPPT is optimizing a 'non-optimal' sum of the TEG in series. \$\endgroup\$
    – TTT
    Commented Feb 15, 2012 at 9:07
  • \$\begingroup\$ IE: Testing 3 individual TEG: (0.25mV, 0.10mV, 0.15mV). The sum should be 0.50mV, but when measured over in series, the series sum may perhaps be 80% of the desired sum. So, which got me thinking that perhaps the MPPT is optimizing the 80% rather than the 100% (not sure if this is true, as of yet. Perhaps you can clarify this.) \$\endgroup\$
    – TTT
    Commented Feb 15, 2012 at 9:07
  • \$\begingroup\$ Therefore, it looks as if my only path now is towards power balancing with MPPT, but I was wondering if power balancing is a 'must' for my current project due to problems with equal module voltage production. Do you think there are alternative solutions for optimization or additional tools to be used for optimization? \$\endgroup\$
    – TTT
    Commented Feb 15, 2012 at 9:08
  • \$\begingroup\$ MPPT SHOULD optimise power if that's what it is measuring. BUT with unequal heating you "have problems". As I noted, power will tend to be limited by lowest output module. MPPT per module is desirable by your low voltages eliminate many devices. IF your budget allows then a per module MPPT is best. Also, you MAY be able to select modules with more junctions per module to increase voltage(but, you know that). \$\endgroup\$
    – Russell McMahon
    Commented Feb 15, 2012 at 13:06
  • 1
    \$\begingroup\$ @TTT: MPPT maximizes power output from varying levels of power input. Unlike voltage regulators, MPPT changes voltage levels via MPPT resistance to get max current & voltage combination from generator (increased efficiency) similar to what overhead transmission cables on roads attempt to do (by raising transmission voltage to minimize energy loss during transmission). Since your application is based on multiple devices with varying levels of power, you can MPPT to maximize power output. For Cymbet, EnerChip EP provides MPPT to charge their batteries eg CBC51100. \$\endgroup\$
    – Zimba
    Commented Sep 7, 2022 at 4:18
0
\$\begingroup\$

What you seem to be looking for is efficiency, and therefore MPPT.

MPPT maximizes power output from varying levels of power input. Unlike voltage regulators, MPPT changes voltage levels via MPPT resistance to get max current & voltage combination from generator (increased efficiency) similar to what overhead transmission cables on roads attempt to do (by raising transmission voltage to minimize energy loss during transmission).

Since your application is based on multiple devices with varying levels of power, you can MPPT to maximize power output. For Cymbet, EnerChip EP provides MPPT to charge their batteries eg CBC51100.

Batteries normally have a max. charging voltage, which is solved by voltage regulators. Both PWM & MPPT can provide a fixed output voltage. This can either be directly connected to battery (DC coupling) or via an inverter to battery inverter (AC coupling) with optional BMS.

The MPPT Difference

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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