Replicator Platform Heater PCB
by garyacrowellsr, published
The Replicator started out with the serious problem of failing HBP power connectors. In the process of failing and replacing the connector, the heater board is often damaged, mandating it's replacement as well. Plus, we've recently seen examples of the replacement 'red connector' also failing. As the Replicator fades into non-support by MBI there will be a need for a replacement HBP heater board that eliminates these problems.
This thing is the Replicator equivalent of what Dan Newman published for the TOM. (See ancestors.)
In addition, I recently made some extensive measurements of temperatures across the area of my stock HBP. What I basically found was that the corners and edges of the HBP were consistently 3-5 degC lower temperature than the central area of the platform.
I also discovered that one corner of my HBP is significantly cooler than the others. I'm suspecting unequal clamping force in that corner or something else that is resulting in poor contact.
This Thing, uses the rear center hole of this HBP in a 3-point leveling arrangement.
I sketched this up in a few hours following a short discussion on the Google Group. It was tested as a milled board that was worked pretty well, and now it has had a prototype sample run fabricated.
The trace dimensions and path have been placed nearly identically to the original Replicator board. I don't want go thru multiple rounds of prototypes, so I'm sticking pretty close to what is known to work. The changes I've made to the the board are as follows:
The surface mount connector is replaced by a thru-hole equivalent. In addition, solder pads/holes are provided so that the high-current pins can be wired to an off board connector. This is accomplished by putting the connector on a 'tab' that extends outside of the original board outline. I'm pretty sure there is ample room on the Replicator for this connector mounting.
No silkscreen on the heater surface. The objective is to keep the surface in close thermal contact with the platform plate.
Copper around the mounting holes. Changing the board thickness at the clamping points doesn't sound like a good idea; the clamping force might bow the board.
Copper fill on the heater side. This extra copper should make for a flatter surface and the additional contact area should aid in transferring heat into the plate. This can be done with a 12/12 mil trace/space, which is not a premium at most board houses.
Nearly all of the trace corners on the original board are around the edges of the board. A square corner increases the effective trace width at the corner, which reduces the heat generated by that bit of trace, which may be contributing to the temperature gradient I observed. I've rounded the corners with a constant trace width. (Yes, it would be a tiny difference, if noticeable at all. But rounding was easy to do.)
The heater trace width on the original board is a constant 1mm (except the trace corners, as noted). The resulting even heat generation across the board guarantees that the resulting heat flow will make the center of the board hotter than the edges. A potential solution to this is to use different trace widths in 'zones' of the board. The attached .jpg highlights the zones. It would be easy to get carried away with this, so I've only varied the widths by a minimal amount as a start. The traces in the yellow area are 1mm; in the center zone they are 1.05mm, and at the corners/edges 0.95mm. If I want greater variation, it will be easy to go back and select and alter the trace widths in the zones. (What appears in the .jpg to be uneven horizontal trace space/fill is an artifact of rasterization; everything is nice and even on the board.)
The surface mount LED on the original board is pretty useless; I've never seen mine (not even sure if its populated). So with a location available on the connector tab, I used a thru-hole LED that should be visible when the printer is in operation.
Extra area on the connector tab is copper filled in a small attempt to add cooling to the current-carrying pins.
- The connector tab is offset from center - this might allow a tiny price break in panelizing the board. (Unlikely, but possible.)
UPDATES: 1/19 =========================================
A bottom edge center hole has been added, in case someone wants to implement a 3-point leveling system by drilling their aluminum plate.
The connector tab has been enlarged slightly and a hole added so the user can zip tie soldered wires in place.
The LED has been moved, better to accommodate a printed cover placed over the connector area. I've added notches for a cover to latch onto.
I've dimensioned the board, but this still isn't a complete fab package.
Tightened up the heater traces around the mounting holes and added some trace in areas that were previously blank. This should help to reduce cold spots around the holes.
- The large pads near the thermistor have been retained 'cause I don't wanna be the new monkey that gets beat up trying for the banana.
UPDATES: 1/20 =========================================
Just cleaned up a few things, and added the rest of the Gerber files. There might be some more touch up, but the only thing incomplete is fabrication Gerber file; it still needs some fabrication notes.
UPDATES: 1/23 =========================================
Final touch-ups; should be good to go.
UPDATES: 1/30 =========================================
[see pic] Milled a test board that came out very well. It was done on 1oz. copper and came out to about twice the resistance of an MBI board, so it's pretty certain their board is in 2oz copper. I'll be milling a board in 2oz next, for assembly and a practical test.
UPDATES: 2/15 =========================================
I got some 2oz copper and milled a new board. The resistance came out to 5.1 ohms. Installed in my replicator it appears to work normally but I still have some measurements to take. Stay tuned.
UPDATE: 3/21/14 =======================================
I've been using the fabricated version of this board for about nine months now, and it works very well. The three-point leveling that it allows is perfect. I did some temperature measurements, and there is still a lot of temperature gradient towards the board edges, so I wasn't aggressive enough in varying the trace widths. I might work up a new version if someone wants to try it, but I doubt that I'll have a new board fabricated, as this one is working well. One point of caution, 24V is exposed on the connector pins at the back of the board, and it would be easy to short them with a metal tool. I strongly suggest that they be covered, at least with tape; a printed cover would be better.
Heh, I had to add a random .scad file, 'cause Thingiverse doesn't recognize my Gerber files as being 'real'.