Fully open source drivers! Both buck and boost!

[Giannis_TDM]

Well since I stopped selling atm due to issues with school and time, Here you go, Open source cheap and proven drivers fully documented!

Giannis_TDM's Laser shack

Laser/Led drivers

sites.google.com

 

[farbe2]

As you mention everywhere that one should short the output before connecting a diode:
The "solution" is easy, just add a resistor across the output. (as low as you can tolerate for power dissipation / efficiency reasons)

Lets say you got 10V max for a boost driver and add a 2,2k resistor across its output. Thats around 50mW losses with full 10V of output voltage and around 25mW with a high Vf diode.
I think thats acceptable. While still being enough to discharge the output caps from 10V to <2V in less than 150ms.

I know, i know, thats robbing you a few tenths of a percent of efficiency and space is always tight on these boards.

However you could just as easily use a lower value voltage divider for the FB, that eliminates the extra resistor completely while still providing good protection!

Your Boost driver uses 140K combined OVP feedback resistance, that discharges the 40u output caps in around 9s to <2V.
If you just use say 1,4k (by changing the 120k to 1,2k and the 20k to 200R) you will get a discharge time of roughly 90ms. Way better!

Lower feedback resistances also make the circuit more robust against noise and "fingering" it e.g. touching the circuit while its powered up. With high feedback resistances, there is always a change of changing the divider ratio while touching the board and therefore potentially killing something.
Thats not an issue with the IC you choose for the boost driver as the FB divider is just for OVP but other designs might benefit from this tip.

 

[Giannis_TDM]

As you mention everywhere that one should short the output before connecting a diode:
The "solution" is easy, just add a resistor across the output. (as low as you can tolerate for power dissipation / efficiency reasons)

Lets say you got 10V max for a boost driver and add a 2,2k resistor across its output. Thats around 50mW losses with full 10V of output voltage and around 25mW with a high Vf diode.
I think thats acceptable. While still being enough to discharge the output caps from 10V to <2V in less than 150ms.

I know, i know, thats robbing you a few tenths of a percent of efficiency and space is always tight on these boards.

However you could just as easily use a lower value voltage divider for the FB, that eliminates the extra resistor completely while still providing good protection!

Your Boost driver uses 140K combined OVP feedback resistance, that discharges the 40u output caps in around 9s to <2V.
If you just use say 1,4k (by changing the 120k to 1,2k and the 20k to 200R) you will get a discharge time of roughly 90ms. Way better!

Lower feedback resistances also make the circuit more robust against noise and "fingering" it e.g. touching the circuit while its powered up. With high feedback resistances, there is always a change of changing the divider ratio while touching the board and therefore potentially killing something.
Thats not an issue with the IC you choose for the boost driver as the FB divider is just for OVP but other designs might benefit from this tip.

The "solution" here is not really needed. By the time they de-solder their driver off their test jig and solder to the diode more than a few min would have past and with all the MLCCs being in parallel their leakages add up, a 2.2k resistor would just be another thing adding to the bom, More like 50 things due to the MOQ for resistors. These open source drivers are designed to be easy to understand and give wiggle room to a beginner, with beginner soldering skills as obvious by the generous space behind the diode pins and the overall layout, making them solder another resistor when not required is sub par.

I do not understand your point about the feedback divider, I will assume that you're talking about the OVP divider and labeling feedback as it directly influences the feedback loop under set conditions. Also I can't see why someone would be "fingering" the driver during operation as this style of driver is a set and forget and put in the pen host kind of deal that directly solders to the diode. Besides that the point you mentioned about divider resistances is valid but as you also mentioned, not really relevant for the OVP function it is used for.

 

[farbe2]

Leakage current should be in the order of >10M per C so way more than the already soldered divider.

I meant that you dont need the extra bleeder resistor but could rather use a existing feedback network (in case of your boost driver, the OVP network). Its already across the output and can just be changed to lower values to act as a bleeder and the intended feedback network.
That would not add any BOM item and still have the benefit of a lower discharge time.


These drivers dont have voltage control so there is no real voltage feedback network as a normal dc/dc would have.
I was just talking more general than your drivers. "fingering" can be a problem on these small chinese dcdc modules.
Some people like to try this stuff on there bench first. Imaging the driver lying on a ESD antistatic mat which could make a difference on high impedance divider networks.
Stuff like this can ruin the day

 

[Giannis_TDM]

Leakage current should be in the order of >10M per C so way more than the already soldered divider.

I meant that you dont need the extra bleeder resistor but could rather use a existing feedback network (in case of your boost driver, the OVP network). Its already across the output and can just be changed to lower values to act as a bleeder and the intended feedback network.
That would not add any BOM item and still have the benefit of a lower discharge time.


These drivers dont have voltage control so there is no real voltage feedback network as a normal dc/dc would have.
I was just talking more general than your drivers. "fingering" can be a problem on these small chinese dcdc modules.
Some people like to try this stuff on there bench first. Imaging the driver lying on a ESD antistatic mat which could make a difference on high impedance divider networks.
Stuff like this can ruin the day

Ok now I see, First off, I own an esd mat and that is a common confusion actually, the top material is dissipitive(not sure I spelled that right) aka not really conductive, the bottom is. Since I have an electrometer I just measured the resistivity. ~262Mohms/cm, furthermore the resistors are not the highest component on the board and such such it is not likely they will ever touch the mat to begin with although I do see your concern.

>These drivers don't have voltage control so there is no real voltage feedback network as a normal dc/dc would have.
Well not really, these drivers do assume the typical configuration inside the IC just with a way lower voltage internal secondary reference. Their feedback network is of course based on the voltage 'generated' across the shunt.

Your point about the values is valid but as said it will kind of be unnecessary and we should think about this from this POV too, Say its the first time you're assembling such a driver, you're a beginner, you solder the resistors in the wrong place, higher value resistors have a higher chance of not causing anything for the good few seconds that it would take you to realize your error.

 

[james1095]

This thread is a bit old at this point but it seems the boost driver is missing the schematic in KiCAD format, is this intentional? Also the only BOM I could find is a PDF I can't copy and paste from so I have to manually type in every one of those long part numbers. Looks like a nice little driver though.

 

[Giannis_TDM]

This thread is a bit old at this point but it seems the boost driver is missing the schematic in KiCAD format, is this intentional? Also the only BOM I could find is a PDF I can't copy and paste from so I have to manually type in every one of those long part numbers. Looks like a nice little driver though.

Not really, I think I lost the original schematic when my hard drive went kaput. Also you should be able to copy paste the numbers in the PDF,
Open it with the default windows edge browser.

 

[james1095]

I suppose if I needed the schematic I could create it easily enough, it just seemed odd that it was missing but that makes sense. I don't have Windows but I thought pdf was pretty standardized, strange, I ended up typing the numbers in since there aren't all that many parts, not sure why copying didn't work though. I'll build up one of these and play around with it.

 

[Giannis_TDM]

I suppose if I needed the schematic I could create it easily enough, it just seemed odd that it was missing but that makes sense. I don't have Windows but I thought pdf was pretty standardized, strange, I ended up typing the numbers in since there aren't all that many parts, not sure why copying didn't work though. I'll build up one of these and play around with it.

Have fun! any questions just send my way, will be more than happy to assist!