1 Watt or Less Efficient Led Light Project

[SkyWatcher123]

Hi all, I made some tweaks to the project.
I picked up a diode assortment at frys, happened to have 50 1n4007's in the set also.
So I made a new full wave bridge with 4 - 1n4007's in parallel per branch, so 16 total diodes for the full wave bridge.
Again, the brightness noticeably increased and the input dropped 30 milliwatts this time, with previous bridge,  it only dropped a few millwatts.
Here is a short video that gives an honest representation of the lighting level, input is 932 milliwatts now.
https://youtu.be/Lk00vIxcxLs

And here is the new full wave bridge.
Comments or questions welcome.
peace love light

Edit: Actually, I notice after watching the uploaded video, it degraded the quality a little and reduced the lighting that is seen with the eyes.

[gyulasun]

Hi SkyWatcher,

Nice results and good job with paralleling the diodes. I think that only germanium type diodes can beat your diode assembly if you are thinking of further improvement for the diodes. Let me show you this link:  https://picclick.ca/USSR-Rectifier-Germanium-Diode-D7J-D7Z-%D0%947%D0%96-If-300ma-382453550737.html where a Russian type Germanium diode, D7Z (400V, 300mA) is compared (by a diode checker DMM) to an also Russian military type diode (200V, 10A) probably Schottky type (but said to be Silicon but for silicon material the forward voltage is around minimum 500-600 mV or higher).
The meters show forward voltage drops,  and the forward current the meter drives through them is usually 1mA but for a single 1N4007 such 1mA may display anything between say 550-670: this display means the forward voltage drop in mV across the diode and not the resistance of the diode as stated in the picture.

IF you have a digital multimeter that also has a Diode symbol in its Ohms measuring range (or next to it), you can easily check this, also for any of the 4 parallel diodes in the bridge.
So the good thing for the D7Z type is that it has 130 mV forward voltage drop at probably 1 mA and it is sure that say at 100-150 mA forward current the voltage drop must be around 200-250 mV maximum.  And you can easily confront this with your 4 diodes presently in parallel:  what your DMM shows for them?

Then you can decide whether probably the higher than 200% reducement in overall diode loss would be worth the cost for the Ge diodes.  Here I assume your parallel 4 diodes have roughly 400-450 mV forward voltage drop at say 50-100 mA forward current.

If you do not have a DMM with Diode test feature, you can drive a given DC current in forward direction via any of the 4 paralleled diodes from the bridge and measure DC voltage across them with your DMM.
Say you have a DC source of a 12.7V battery (or any other type), then you calculate a resistor value and connect that resistor in series with the diode (or with the 4 parallel diodes of course) and then you connect this series R and Diode assembly across the battery.  For a 50 mA forward current the R series resistor would be 12 V / 0.05 A = 240 Ohm. For 100 mA forward current use a 120 Ohm resistor (these should be at least 1-2 W rated resistors or measure quicly not to let the R heat up too much).

Here is another ebay offer with free shipping for 10 pieces of D7Z Ge diodes:
https://www.ebay.com/itm/202340101831
or here is an offer for 4 pieces (but shipping is not free):  https://www.ebay.ca/itm/253341986352 
The technical specifications for the 200V 10A diode is here at the bottom of this link if you need them:
https://www.ebay.com/itm/323306203220

The 3.33 kOhm emitter resistor possibly dissipates 8-9 mW DC power (my rough estimation) if you consider that to reduce (but it would be hard... by selecting the transistors I suppose).

Gyula

[SkyWatcher123]

Hi gyulasun, thanks for the helpful information.
I checked some of the diodes i have with my Centech 61593 meter, it does have diode test setting.
Here is some forward voltage drop data of diodes i have.
1n4007 = 535 mV
1n5408 = 474 mV
1n4148 = 565 mV
1n5822 = 117 mV
6A6    = 453 mV
---------------------------
4-1n4007 in parallel = 465 mV
8-1n4148 in parallel = 455 mV

The 1n4007's in parallel seem the best so far, for this coil/core setup and the higher voltage is good safety margin, as without a load off bridge to bring down the voltage, the 1n4148's can die.
Still going to try and optimize the project and have some comparison tests to make and see how the light output compares to off the shelf led bulbs.
peace love light