mishka
Is it me?
- Messages
- 247
Following a prod from Weldequip I have opened up my Eutectic DC inverter (more details in this thread) and taken a couple of photos.
First up, the component side:
Starting at the bottom right, there is
The other side of the main PCB:
Mostly for completeness, but it also shows how simply laid out the board is. The lack of component pins also indicates the extensive use of surface-mount components for the control circuits, reducing cost and increasing reliability.
One aspect I really like about this design is the simplicity of the output circuit. The output winding on the transformer is two single-turn windings made from ordinary braided copper strip. The centre-tapped design means that the load is evenly split across two diodes, keeping the average current down. Also the output diode pack is OVER-RATED - this welder goes up to 140A, but the ouput diode pack is rated up to 200A. So the limitation is purely thermal and easily within spec. Finally, the connections in the output circuit are made using copper strip - no soldered joints, no PCB tracks, just bolts and copper.
So, overall I'm impressed.
Sure, there are things that could be done better (you knew I was going to say that, didn't you?!). I'd prefer to see more robust controls on the front panel. And changing the heatsinks to black anodised would increase the 100% duty cycle current (black heatsinks have slightly better thermal conductivity to air). And with a bit more thought they could have made the cooling air path completely separate from the rest of the box, minimising PCB contamination from airborne dust, etc (such as is done in the Inverter Fusion range). But all in all these are minor gripes.
First up, the component side:
Starting at the bottom right, there is
- The cooling fan;
- The main heatsink for the four switching transistors (ST GW30NC60WD 30A 600V IGBT, replacements available from Farnell, £2.94 each);
- On top of which is the main transformer (really - its a thin planar transformer, about 1cm thick);
- Then the heatsink for the output diode pack (ST BYV255V200, again, available from Farnell, £35.57 each);
- The little PCB sitting on top of the diode pack is most likely the voltage-sensing module, for monitoring what's happening at the electrode/tip;
- The silver metal strip is one of the output conductors, and in this case is also being used as a current shunt to measure the output current;
- At the top we have three large capacitors on the input side of the inverter;
- The stripey-yellow lead sticking out of the top is the earth strap for the metal case;
- A single-pole high-current relay for energising the main power circuits (I think the control circuits are on a separate low-current supply - at switch-on there is a 4sec delay before the main circuits+fan comes on);
- Finally, the mains input filter (double choke + PTC + capacitors).
The other side of the main PCB:
Mostly for completeness, but it also shows how simply laid out the board is. The lack of component pins also indicates the extensive use of surface-mount components for the control circuits, reducing cost and increasing reliability.
One aspect I really like about this design is the simplicity of the output circuit. The output winding on the transformer is two single-turn windings made from ordinary braided copper strip. The centre-tapped design means that the load is evenly split across two diodes, keeping the average current down. Also the output diode pack is OVER-RATED - this welder goes up to 140A, but the ouput diode pack is rated up to 200A. So the limitation is purely thermal and easily within spec. Finally, the connections in the output circuit are made using copper strip - no soldered joints, no PCB tracks, just bolts and copper.
So, overall I'm impressed.
Sure, there are things that could be done better (you knew I was going to say that, didn't you?!). I'd prefer to see more robust controls on the front panel. And changing the heatsinks to black anodised would increase the 100% duty cycle current (black heatsinks have slightly better thermal conductivity to air). And with a bit more thought they could have made the cooling air path completely separate from the rest of the box, minimising PCB contamination from airborne dust, etc (such as is done in the Inverter Fusion range). But all in all these are minor gripes.
Last edited: