Germanium BJT recommendation

Thread Starter

k1ng 1337

Joined Sep 11, 2020
539
Hi,

I'm working on a solar project with a super capacitor. Under test, a boost converter is able to consume energy until ~800mV at which it slowly cuts out. I would like to replace the silicon npn BJT with a germanium or other technology (FET?) that can reduce this value further to get the most juice out of the capacitor at night.

Design requirements are minimal:

- 100k switching frequency
- 500mA current

A recommendation for a part number and where to buy would be super.

Regards,
Mark

This topic should probably be moved to power electronics, oops.
 
Last edited:

Papabravo

Joined Feb 24, 2006
18,431
Hi,

I'm working on a solar project with a super capacitor. Under test, a boost converter is able to consume energy until ~800mV at which it slowly cuts out. I would like to replace the silicon npn BJT with a germanium or other technology (FET?) that can reduce this value further to get the most juice out of the capacitor at night.

Design requirements are minimal:

- 100k switching frequency
- 500mA current

A recommendation for a part number and where to buy would be super.

Regards,
Mark

This topic should probably be moved to power electronics, oops.
It will be difficult to find a germanium transistor. I'd suggest consideration of a FET, but that said, your expectations for the operation of a boost converter with a low input voltage may be overblown. There is just no available input power to get any meaningful output power.
 
Last edited:

ronsimpson

Joined Oct 7, 2019
2,074
As MrChips said there are "low voltage energy harvesting" ic on the marked. LTC______ for example. There is good reading in the data sheets. These parts need some voltage to get started, but when running they work down to a very low voltage.

Germanium transistors are slow (not 100khz), leaky (loose power), not in production.
 

MisterBill2

Joined Jan 23, 2018
11,897
Some varieties of mosfet devices do work down to lower voltages. But you have hit the hidden wall with capacitor storage systems, which is that the voltage is proportional to the amount of charge left in the capacitor. One possible work-around would be a scheme to put the charged devices in series for discharge. That could allow more charge to be recovered.
Because switching will be a serious pain, I suggest a re-draw of the circuit to keep the caps in series and charge them with separate sections of the solar cell array. Even a split into just two segments will double the low end voltage. Certainly this is more complex but it will not require additional components and it wil be very reliable as a result.
 

MrSalts

Joined Apr 2, 2020
1,762
There is just no available input power to get any meaningful output power.
"Meaningful power" is in the eye of user and their particular system If they are only looking to power an unlit LCD display or send a low-power Bluetooth message once a threshold voltage in a storage capacitor is achieved, then I think enough power is present. TI and Maxim and others make a voltage step-up modules exactly for this purpose.
 

BobTPH

Joined Jun 5, 2013
4,916
What voltage was the capacitor charged to? If it was, for example , 3.2V you have already used 96% if the energy when it gets ti 800mV.

Bob
 

Papabravo

Joined Feb 24, 2006
18,431
"Meaningful power" is in the eye of user and their particular system If they are only looking to power an unlit LCD display or send a low-power Bluetooth message once a threshold voltage in a storage capacitor is achieved, then I think enough power is present. TI and Maxim and others make a voltage step-up modules exactly for this purpose.
I was not thinking about output power, but input power available from a voltage source that is declining will make it decline faster by requiring more current. At some point it will be insufficient to run a boost converter. Where is that point? Wherever it is you will be getting there in a hurry.
 

Thread Starter

k1ng 1337

Joined Sep 11, 2020
539
Thanks everyone for the input. The cap is charged to 2.7V and I am aware that usable energy declines significantly as V approaches zero. My project will have three power modes: 1) in daytime a uC will operate (Pico), 2) at night between ~1.25-2.8V a LED array will be lit, the lower limit is due to oscillator stability 3) below 1.25V oscillator is unstable but still operates so a very low power mode to light a reduced led array. I am designing my own boost for 2/3 as 1 has one on board.

Tests reveal I can get LEDs to light to decent brightness at ~1mA so this topic is about using as much juice as possible in the third power mode. I'm impressed on how LTC chips are able to start at so low a voltage and I prefer to build everything myself for this project so this topic will turn to how those devices are able to do what they do. I suspected I could use a network of charge pumps prior though I have not found anything simple to wrap my head around. I studied the block diagrams of the LTC chips and they are too complex for me at this time. Perhaps someone could recommend a method to achieve a similar result with discrete components to push the 800mV limit a bit lower.

I'm making this post at work so I don't have time to go into detail. In the meantime, thank you :)

In retrospect I assumed the 800mV cutout point was due to the silicon in the power transitor but now I'm thinking it may be the limit of the TLC555 I am using for switching or both. I am aware I am below spec on the 555 although does anyone know the actual lower limit of a CMOS 555? As voltage drops below spec, the frequency declines which is acceptable.
 
Last edited:

MisterBill2

Joined Jan 23, 2018
11,897
Once again, if you have multiple solar arrays and multipl3 storage capacitors, arranging the capacitors in series with each charged by a separate solar cell array, you will multiply the voltage without losing anything except max current availability.
You will have taken the same amount of stored power and multiplied the available charge that can be recovered, all without any additional electronic components. Of course, if the entire system consists of only one solar cell array that can not be split, this option may not work. At that point switching would be required and it would no longer be simple.
 

MrChips

Joined Oct 2, 2009
26,118
Solar panels are tricky because the internal resistance varies with load conditions and solar input.

With no load, they output voltage but zero current. The power output is zero.

With zero resistance load, they output current but at zero voltage.

Hence you need to adjust the load resistance for MPP (maximum power point) in order to achieve maximum efficiency.

solar panel mppt.jpg
 

MisterBill2

Joined Jan 23, 2018
11,897
How does post #11 relate to charging supercaps connected across a solar cell??? WE are talking about a small system. See posts #8 and #9. The added burden of a max power circuit is not trivial.
 

MrChips

Joined Oct 2, 2009
26,118
I am making the assumption that TS is attempting to harvest solar energy. For maximum efficiency, solar power to electrical power, you would want to do MPPT.
 

MisterBill2

Joined Jan 23, 2018
11,897
I am making the assumption that TS is attempting to harvest solar energy. For maximum efficiency, solar power to electrical power, you would want to do MPPT.
In post #9 the TS talks about lighting individual LEDs and getting useful light. THAT sounds like a small system. For a "small system" any circuits to maximize power collection (MPPT) will be a large chunk of the package. Thus I question the value added versus the cost of adding it.
 
Top