17mm buck-boost driver (TPS63020 and LT6106)

[foulmist]

I would suggest caps are changed to 10V.

Include vias points the the bottom layer for a plain ground - both beneath the IC thermal pad, and close the caps negative connections.

Also those inductance traces seem awfully thin

EDIT: already done

 

[laserluke]

I have added the eagle files for the board in the first post so anyone with goodwill can modify the layout and/or the schematic.

EDIT:
foulmist, please let us know the manufacturer and part number for 22uF, 10V with package 0603 capacitors.
The schematic is inconsistent without pcb vias... there are always been... check the "drill" and "drill.info" excellon files included in the gerber archive

 

[laserluke]

I'll order a few and let you know how they work :beer: that's why I was asking a parts list

I have updated the board.
Please be sure to send to manufacturer the updated gerber files.
I will follow you during the testing.

Edit: FINAL VERSION, happy testing

 

[AnthoT]

I have updated the board.
Please be sure to send to manufacturer the updated gerber files.
I will follow you during the testing.

Sure i'll send those

 

[laserluke]

Sure i'll send those

I have posted the final version ready for testing.
In this post, the user "rfhv" by using a similar schematic but with a less flexible (fixed) gain in the current sensing (zxct1009), has obtained an exceptional ripple of less than 10mA at 2A output (final accuracy of 0.5% and efficiency of 82% ).
With this open source project, by adjusting the values of R5, C7 and the optional C8, i think you can obtain the same, if not better, result.
You have also more flexibility (variable gain with lt6106 and constant output impedance of the current sensing circuit) by using the multi turn trimmer to set the output current.

 

[AnthoT]

Alright so ill be ordering 25 of these ill only be making 2-3 though so if anyone wants PCB's ill have them

 

[laserluke]

Alright so ill be ordering 25 of these ill only be making 2-3 though so if anyone wants PCB's ill have them

Very good
You can ask to pcb manufacturer to not apply the solder resist to bottom side of the pcb, which is the ground plane, and/or to manufacture it with a thin dielectric (less than the standard 1.6mm... i think 0.8mm is easily available) so to obtain a very good thermal contact when you attach the pcb to the heat sink.
Let me know...

 

[AnthoT]

Very good
You can ask to pcb manufacturer to not apply the solder resist to bottom side of the pcb, which is the ground plane, and/or to manufacture it with a thin dielectric (less than the standard 1.6mm... i think 0.8mm is easily available) so to obtain a very good thermal contact.
Let me know...

I get all my PCB's 0.8mm

 

[laserluke]

be sure to test them on scope first with a test load. And sharing the results would be nice

Alright so ill be ordering 25 of these ill only be making 2-3 though so if anyone wants PCB's ill have them

I think that would be nice and useful the foulmist's contribution to testing and design review. After all foulmist is a good man and could be useful to have two "test sites" in Canada and Bulgaria at about 7700 Km of distance with no risks in case of explosion :crackup:
What do you think? :yh:

 

[AnthoT]

I think that would be nice and useful the foulmist's contribution to testing and design review. After all foulmist is a good man and could be useful to have two "test sites" in Canada and Bulgaria at about 7700 Km of distance with no risks in case of explosion.
What do you think? :yh:

Explosion!?!??????? I highly doubt any driver explodes..... I will be sending him one as he has a scope and it would really help show how stable this driver is :beer:

 

[laserluke]

Explosion!?!??????? I highly doubt any driver explodes..... I will be sending him one as he has a scope and it would really help show how stable this driver is :beer:

should have been a joke but i forgot the emoticon after the word "explosion" :beer:

 

[AnthoT]

should have been a joke but i forgot the emoticon after the word "explosion" :beer:

Ah ok

 

[laserluke]

To find the right values, i would suggest to start testing with no C8 capacitor (you can add later to minimize the feedback ripple) and smaller values than those showed in the schematic for R5 and C7.
You can start with 220ohm and 10nF and starts to increase the capacitor.
Don't forget to start with the trimmer turned completely ccw and tests all the output current range for the actual values of R5, C7

 

[foulmist]

I will sure be glad to help :beer:

foulmist, please let us know the manufacturer and part number for 22uF, 10V with package 0603 capacitors.

hah that's a trick question isn't it I was thinking of replacing the caps with 22uF 10V but at least 0805. As there isn't a 0603 cap rated for that. Or at least I coudn't find on Digikey and farnell. . I think there is enough space for the output caps to fit the 2x0805 ?

 

[laserluke]

I will sure be glad to help :beer:



hah that's a trick question isn't it I was thinking of replacing the caps with 22uF 10V but at least 0805. As there isn't a 0603 cap rated for that. Or at least I coudn't find on Digikey and farnell. . I think there is enough space for the output caps to fit the 2x0805 ?

There is no space on the board for 2x0805 (or it's very difficult...) but i think it's not necessary to use 10V capacitors because the maximum rated output voltage of TPS63020 is 5.5V (from datasheet here) and the TI itself mounts 6.3V capacitors in their evaluation board here. The evaluation board is usually used as a starting point for design and for obtainable performances.

 

[foulmist]

Perhaps it's not necessary but I always like to put caps with higher voltage ratings. Just to be safe (although you can't say you will be safe even then but still... )

so.. 0603 6.3V it is