Super diode (active rectification) question

Thread Starter

molecool

Joined Oct 24, 2011
32
Well, the circuit I posted doesn't compensate for battery voltage droop either, but molecool would like it if it did.
Exactly - the main requirements are this:

- The circuit should only run the motor if the battery polarity is negative to input A and positive to input B. If inverse the motor should not run (basically as if I put a diode in series but without the forward voltage loss).
- There should be as little forward voltage drop as possible - preferably less than -0.05V.
- The AAA's initial 1.5V should be maintained as long as possible, which is why I thought the integrated joule thief was such a cool idea (if it could work).
- Finally, the components should be pretty cheap and commonly available.

Hey, nothing like a good challenge! ;-)
 

praondevou

Joined Jul 9, 2011
2,942

Ron H

Joined Apr 14, 2005
7,014
Hey guys - how about using Ron's joule thief in combination with something like this:

http://search.digikey.com/us/en/products/LP3992IMFX-1.5/NOPB/LP3992IMFX-1.5-ND/1872932

This way I squeeze most of the juice out of my AAA and enjoy constant voltage output at 1.5V!

Will this work?
That is a linear regulator. Linear regulators always require that the input voltage be higher than the output.
You need a type of switching regulator called a boost converter.
Switching regulators are generally more efficient than linear regulator.
 
Exactly - the main requirements are this:

- The circuit should only run the motor if the battery polarity is negative to input A and positive to input B. If inverse the motor should not run (basically as if I put a diode in series but without the forward voltage loss).
- There should be as little forward voltage drop as possible - preferably less than -0.05V.
- The AAA's initial 1.5V should be maintained as long as possible, which is why I thought the integrated joule thief was such a cool idea (if it could work).
- Finally, the components should be pretty cheap and commonly available.

Hey, nothing like a good challenge! ;-)
Here's a solution that works ok for your basic requirements.

It doesn't regulate at all; it only acts like a "super diode".

I wired it up with a little vibrator motor I have that came out of an old pager. I tested it with a single AAA alkaline cell. The motor runs when the battery is connected in the proper polarity and in the reverse direction, nothing happens.

Then I connected a variable supply in place of the battery. The motor runs until the input voltage drops a little below .8 volts. With 1.0 volt applied, the loss in the transistor is about .05 volts. With .8 volts applied, the loss in the transistor goes up to about .065 volts.

The Vbe reverse voltage rating is 5 volts, so if you don't apply anything more than a single AAA cell, you should be ok. I thought about putting a capacitor across the motor to absorb any voltage spikes resulting from commutator sparking. I had a look with a scope and I saw short microsecond spikes reaching a maximum of about 3 volts. A .1 μF cap held them down to under 1 volt. After letting the motor run for 5 minutes, I noticed that even without a cap, the spikes had pretty much gone away. I guess the rough edges on the commutator got rubbed down with the carbon brushes. You might be ok without any capacitor.

The key to this circuit is the use of this exact transistor:

http://search.digikey.com/scripts/DkSearch/dksus.dll?x=20&y=17&lang=en&site=us&KeyWords=ztx788b

This transistor is designed to have very low Vcesat. It looks like the price of the transistor will be about 50 cents in several thousand quantity from Digikey.
 

Attachments

Thread Starter

molecool

Joined Oct 24, 2011
32
That is a linear regulator. Linear regulators always require that the input voltage be higher than the output.
You need a type of switching regulator called a boost converter.
Switching regulators are generally more efficient than linear regulator.
Yes, Ron - which is why I suggested a joule thief in front of it.
 

Thread Starter

molecool

Joined Oct 24, 2011
32
Here's a solution that works ok for your basic requirements.

It doesn't regulate at all; it only acts like a "super diode".

I wired it up with a little vibrator motor I have that came out of an old pager. I tested it with a single AAA alkaline cell. The motor runs when the battery is connected in the proper polarity and in the reverse direction, nothing happens.

Then I connected a variable supply in place of the battery. The motor runs until the input voltage drops a little below .8 volts. With 1.0 volt applied, the loss in the transistor is about .05 volts. With .8 volts applied, the loss in the transistor goes up to about .065 volts.

The Vbe reverse voltage rating is 5 volts, so if you don't apply anything more than a single AAA cell, you should be ok. I thought about putting a capacitor across the motor to absorb any voltage spikes resulting from commutator sparking. I had a look with a scope and I saw short microsecond spikes reaching a maximum of about 3 volts. A .1 μF cap held them down to under 1 volt. After letting the motor run for 5 minutes, I noticed that even without a cap, the spikes had pretty much gone away. I guess the rough edges on the commutator got rubbed down with the carbon brushes. You might be ok without any capacitor.

The key to this circuit is the use of this exact transistor:

http://search.digikey.com/scripts/DkSearch/dksus.dll?x=20&y=17&lang=en&site=us&KeyWords=ztx788b

This transistor is designed to have very low Vcesat. It looks like the price of the transistor will be about 50 cents in several thousand quantity from Digikey.
Now THAT is awesome. Finally a hands-on schematic with a focus on the actual problem I was trying to solve! Many many thanks - I'm paying someone more experienced now to help me implement this and will share your insights, which I'm sure will expedite our effort.

THANK YOU! :D
 

thatoneguy

Joined Feb 19, 2009
6,359
That IS a Nifty solution!

Are there any other transistors that come sort of close in performance to be used as a substitute, albeit an inferior one?
 
The truly astounding Vcesat at low collector currents needs a forced beta of 10; with a forced beta of 100 it's not quite so good.

The ZTX788B has better guaranteed HFE at low currents.

The ON semi part looks to be fairly new whereas the ZETEX part has been around for nearly 20 years.

I wonder what else ON is working on.
 

thatoneguy

Joined Feb 19, 2009
6,359
Check this out. Unfortunately, it is not stocked by Digikey or Mouser. Mouser says 13 weeks lead time. It is in a TSOP-6 package.
PNP transistors have really improved since I've last looked at low power switching! Only downside is the package. The roughly 1206 package is still big enough to solder without magnification, but you need a board to put it on, unless you use the dead bug method with wire wrap wire. A LOT of the cool stuff is coming out in SMT only, so get practicing now!

<OffTopic>
I'm hoping Radio Shack will start carrying SMT proto boards, at least for a few standard pitches on one board. Something like This

I do like the lack of drilling when working with SMT, but clipping scope probes on isn't as easy as it used to be, and you need to be dead sure of your layout before etching.

For SMT: Flux Pen (preferably refillable w/liquid flux), temp adjust iron with 1-2mm chisel tip, GOOD tweezer set, steady hands. The magnifying glass on the "Helping Hands" most of us already have is adequate for prototypes, the fab shop can have the robots do the sub-millimeter package reflow soldering.

</OffTopic>
 
That IS a Nifty solution!

Are there any other transistors that come sort of close in performance to be used as a substitute, albeit an inferior one?
ZETEX has been the major supplier of high current BJTs in small plastic packages (especially TO-92 type) that I have been aware of. They have a part with higher voltage rating, the ZTX750 that Motorola once tried to copy; MOT called it MPS750. It's a 2 amp part and the higher voltage rating gives you a worse Vcesat, but it's not too bad at 100 mA collector current.

ZETEX has been bought by Diodes, Inc. There's an index of all their BJTs here:

http://www.diodes.com/products/catalog/list_extended.php?parent-id=138&page=0

Sorting on collector current and doing a quick glance through, I found this one:

http://www.diodes.com/products/catalog/detail.php?item-id=7201

It has a higher voltage rating than the ON semi part, and higher guaranteed HFE, but not quite as good Vcesat. There are undoubtedly other candidates for the dedicated searcher.

A possible advantage of the ZTX788B is its age; it may be stocked by more distributors. But it may also be scheduled for obsolescence.
 

Ron H

Joined Apr 14, 2005
7,014
The truly astounding Vcesat at low collector currents needs a forced beta of 10; with a forced beta of 100 it's not quite so good.

The ZTX788B has better guaranteed HFE at low currents.

The ON semi part looks to be fairly new whereas the ZETEX part has been around for nearly 20 years.

I wonder what else ON is working on.
I see what you mean. Still, the Vce(sat) at high forced betas is remarkable. With Ic=50mA and Ib=10mA (pretty weird, I know), the sim of the nss20300mr6t1g showed Vce(sat)=-1.6mV! I wonder if that is realistic.
Under the same conditions, the ZTX788B was at about -12mV.
 

Thread Starter

molecool

Joined Oct 24, 2011
32
That's a neat part, but it won't handle the reverse voltage that he wants to be able to apply.

I would really like to know why his supply voltage might be of either polarity.
Well, could I simply use two of them in separate branches of a polarity switching circuit?
 
Top