PCB for turbo pump

[CM800]

Just finished this a few days ago, ordered the board yesterday.

Looking back on it today... Did I overdo the thermal vias? 

[Gyro]

I suppose there must come a point where the combined area of 'hole' outweighs the combined area of copper but I think you'd be hard pushed to achieve it in practice. 

[MagicSmoker]

And I sure hope you specified that the vias within the boundary of the cream layer are filled or plugged, because otherwise all your solder paste is going to get sucked down into them.

[CM800]

And I sure hope you specified that the vias within the boundary of the cream layer are filled or plugged, because otherwise all your solder paste is going to get sucked down into them.

I will be handsoldering it thankfully. I have no reflow gear.

[max_torque]

Loads of Vias, but tracks that look quite thin?

With modern low low FETs, if you're moving enough current to need to really sink them hard thermally, you'd also want thicker tracking no??


BTW, i'm not sure what a "turbo pump" actually is (oil pump for  turbocharger??) but save yourself a lot of hassle and check out Infineons range of automotive rated power switches and drivers!  For example BTN89829  When you see what they can do and how few external components they need you might start to wonder why you roll your own these days!

[CM800]

Loads of Vias, but tracks that look quite thin?

With modern low low FETs, if you're moving enough current to need to really sink them hard thermally, you'd also want thicker tracking no??


BTW, i'm not sure what a "turbo pump" actually is (oil pump for  turbocharger??) but save yourself a lot of hassle and check out Infineons range of automotive rated power switches and drivers!  For example BTN89829  When you see what they can do and how few external components they need you might start to wonder why you roll your own these days!

It's a turbomolecular pump. (those fancy things that spin at 90,000rpm to help create a complete vacuum)

It's just a home project.

I don't think it will draw more then 2A or so during operation, I just got a little carried away. 

[calexanian]

Out of curiosity what brand and model of turbo pump are you using. I have some interest in them.

[Wolfram]

This looks similar to a published schematic for a Pfeiffer TPH055 driver, http://www.fusor.net/board/viewtopic.php?f=10&t=4036#p21444 , with an ICL7667 or pin-compatible gate driver added.

Some tips for the circuit:

Is X2 a stand-in for a regulator? If so, there should be some capacitance on the regulator input. Many voltage regulators oscillate without proper decoupling on the input. Even if there is no regulator, it is still beneficial to have some capacitance on the DC bus feeding the pump, to minimize ringing caused by the switched motor current.

I prefer to keep the bottom copper layer as a complete ground layer and route most traces on the top layer. This makes it much easier to verify that each signal has a proper ground return without excessive loops. Also I see some of the hall effect sensor wiring is routed between the power MOSFETs and over the switched winding traces, this is not ideal for signal integrity. Most likely it will not be a problem, but it takes very little effort to avoid this potential trap once you're aware of it.

The component designators are placed on top of pads in several places. This can make the components hard to identify without consultin the board layout, and it takes very little effort to fix before sending the board to production.

[CM800]

This looks similar to a published schematic for a Pfeiffer TPH055 driver, http://www.fusor.net/board/viewtopic.php?f=10&t=4036#p21444 , with an ICL7667 or pin-compatible gate driver added.

Some tips for the circuit:

Is X2 a stand-in for a regulator? If so, there should be some capacitance on the regulator input. Many voltage regulators oscillate without proper decoupling on the input. Even if there is no regulator, it is still beneficial to have some capacitance on the DC bus feeding the pump, to minimize ringing caused by the switched motor current.

I prefer to keep the bottom copper layer as a complete ground layer and route most traces on the top layer. This makes it much easier to verify that each signal has a proper ground return without excessive loops. Also I see some of the hall effect sensor wiring is routed between the power MOSFETs and over the switched winding traces, this is not ideal for signal integrity. Most likely it will not be a problem, but it takes very little effort to avoid this potential trap once you're aware of it.

The component designators are placed on top of pads in several places. This can make the components hard to identify without consultin the board layout, and it takes very little effort to fix before sending the board to production.

You are right on all accounts.... I really should have thought about signal integrety more and so on.

I didn't properly look over it before sending it over admittedly... Let's hope it works.

[calexanian]

I would like to see how this works out. I have always wanted to play with one of those pumps. I still primarily use diffusion pumps in my home lab.