Heat Distribution on Heated Bed

What started off a week ago as as "simple" change to my Z offset (as noted in my Senior Moments thread) has taken me down the path of upgrading to Marlin RC-8, a short bout with repetier and back to Marlin.

Now the firmware is all in place, and most of the tuning is completed, I took a look at some of the hardware. My heated bed was slow to heat anywhere warmer than about 85 degrees. A little measuring showed it was only getting 10.8v at the bed. I was losing 1.2v (only had a 12v Supply, which did pump out 12 v even with the bed on) going through the Solid State relay, which dropped about 1v and another .2v was lost in about 4ft of 12ga wire. I had a 25 Amp 13.6 v power supply gathering dust, so that was pressed into use for the bed heater. In supplies 12.4 v at the heater. Heating was noticeably quicker now.

The problem I now have is uneven heating, not 3 or 4 degrees difference across the bed but more like 30 degrees from back to front! (back is where the power is connected) I looked at the back of the bed (the original Makerfarm 12x12 bed) and it looks to be in pristine condition. From what I can see the power "rails" run across the top in each direction from the input and down each side with "elements" joining them going across the board. I suspect there is a significant voltage drop along these power rails so the elements at the bottom pf the bed are getting way less the those at the top.

My thinking is to run a wire (12-14ga), from the input to the board, to a p(o)int at the end of the rails by scraping off a little of the red covering and soldering it. This should bring the voltage up a bit.

Good, Bad or Stupid idea?

Thoughts please

How do you know the temperature varies that much? Are you printing on glass? Is the glass making good contact with the entire heat bed? If it is not, the heat bed will be warmer in those locations of poor contact, while the glass will be cooler.

My guess is you are not fighting poor conductivity. But you can scrape off some of the red silk screen in different locations and verify each rail (and line) is getting the same voltage. I would do that before soldering big wires on to give more even current flow.

Hi Roxy, initially I could feel by hand that there was a difference in heating across the board, so I got out the IR thermometer and there was a gradual decrease in temperature from back to front. The centre of the board (where the thermistor is) is close to the set temperature, the back about 10 degrees higher and the front about 20 lower. The glass is in good contact with the bed, and the same problem occurs with just the board with no glass on it.

I had planned to do some measurements, as you suggested, prior to doing any soldering.

Did some voltage measurements, not quite what I expected as there seems to be a greater loss towards the back of the board, i would have expected a more linear loss.

Click on image to make it larger.

BACK
Attachment 10038
Front

Just finished up running some 12 gauge wire around the outer traces, soldering to the bed at the positions of the red X's in the diagram. Vast improvement, temperature difference on heating is only 3 degrees back to front and the voltage drop from back to front is now 0.15v vs 2.0v previously. Heat-up time has improved as well.

The voltage loss is less towards the front since the traces there are only carrying the current for the forward section of the heater. Traces towards the rear are carrying more current for more of the heat bed, and likely explains why there's more voltage loss there.

Whether beefing up the distribution traces should be needed depends on whether the PCB layout person knew what they were doing. Trace resistance is a factor of trace width, trace thickness, and trace length. It would have been possible in the PCB design to tweak the trace widths or trace lengths in the different heater sections as required to achieve equal heat dissipation throughout the board. The reprap heat bed Wiki says the original MK1 design by Prusa should be within +/- 3 degrees C across the bed. Of course, the MK1 used by MakerFarm is custom, so what Prusa did may mean diddly-squat to the MakerFarm variants.

Note that you can't reliably use an IR thermometer on glass unless the thermometer has a setting for the emissivity of the surface being measured. Search on IR thermometers and emissivity for an explanation. Here's one reference.

How do you have the hot end fan oriented? A lot of people, including myself, flip the hot end fan around from the MakerFarm recommendation so that fan exhaust goes to the front of the printer. The bad side of this is that it can create a major draft across the heat bed in front of the hot end. I used to add a piece of painter's tape across the bottom of the hot end fan, shaped in a bit of a U to channel the airflow upwards. Now my fan mount angles the hot end fan upwards 10 degrees to help with this. A bit more of an angle would be even better.

IMO the forward part of the heat bed is naturally more prone to drafts in the room environment than the rear part. Most of my printing is done on a cold bed, but when I run the heat bed I *always* cover the heat bed with something like a piece of cardboard and let the heat bed soak at temperature about 10 minutes before starting the print. Regardless of how even the heat distribution is across the heat bed, I think that's important.

Hi printbus, thanks for the response. I understand about the lack of accuracy when using an IR thermometer to determine actual temperature, but for the measurements I took I clamped the IR above the bed and moved the bed under it, that way at least the angle and distance to the bed stayed the same. I still have the hot end fan exhausting to the back of the printer. In the early stages of warming the irregular heat distribution was noticeable just by touch, so I believe there was an issue with power distribution on the bed. At ABS type temperatures (bed ~110 degrees) where the heater is running quite a bit, ;leaving the bed for 10 minutes or so only slightly improved the heat distribution.

With the additional wiring in place, both the voltage and heat distribution have improved dramatically. I like the idea of covering the bed during warm up, will give that a try.

I realize now you have a 12x12 bed. With more current flowing around for the large heat bed, any issues with the current distribution would be more pronounced, perhaps as you are seeing. Maybe the design *is* bad, or maybe there's a problem in manufacturing inconsistency.

I'm going to be revamping my bed and print surface within the next few weeks - now I know to put a bit of scrutiny on how the traces were routed on my 8x8 bed. I've been holding off on my bed revamp for a long time, hoping that Panucatt would eventually restock their Helios bed, which sounds like a good design. I really liked the idea of being able to eliminate the screw heads sticking up in each corner. After waiting a year with no response from Panucatt I'm giving up on it and going in another direction reusing the MakerFarm MK1 heat bed.

It's hard to fathom that the original concept behind the Prusa bed was to just print directly on the back side of the circuit board. At least I remember reading that somewhere. The idea to use glass came later when people were having issues with the circuit board flexing. Duh.

Hi

With a proper insulator on the back side of the heated bed (I use Nomex), and a good supply (I use MeanWell), and a good SSR (with < 0.25V drop) .... I get temperatures that are uniform to < 5C measured with an IR thermometer. This is with the stock heated bed (which is set up to compensate for the drops in the feeds). It also is true with a 12V or with a 15V supply.

If you try to measure the temperature of a moving bed, you need to be careful about cooling the bed with the moving air.....

Bob

Hi Bob,
The power supply is good, I have two cork tiles on the back as insulation and a Fotek 60A SSR. At least I thought it was a Fotek, but just recently the relay started to show signs of melting! Last night the hot bet heated and heated ans heated, although the relay controller has commanded the relay off, it continues to heat. A little research shows I have a "knock-off" relay, so that may explain its demise. What relay do you have?

I am building up a MOSFET switch for temp/permanent use depending on performance. That's probably my last work for a while, we're off to our RV on the Alabama Gulf Coast for 3 months, thanks for your help in getting my system up an running over the last few weeks.

Hi

If your SSR has ~ 1V drop at 20A, at 60A the drop will be higher. That puts the dissipation up in the 80 to 100W region pretty fast (at 60A). The rated heatsink you are supposed to use with them is good to *maybe* 20W. I'm running a 100A Fotek off of Amazon. Price was < $20 so who knows what it really is.

Bob

Hi, I am only running 20A through the relay, not 60A (it would probably have vapourised on first contact!)

If you look at the label on the relay and it has the UL and CE markings one above the other AND "Made in Taiwan" on the label it is probably genuine. If UL and CE are side by side and NO "Made in Taiwan", probably a copy.

Quote:

---

Originally Posted by ex-egll

Hi, I am only running 20A through the relay, not 60A (it would probably have vapourised on first contact!)

If you look at the label on the relay and it has the UL and CE markings one above the other AND "Made in Taiwan" on the label it is probably genuine. If UL and CE are side by side and NO "Made in Taiwan", probably a copy.

---

Hi

If the SSR is *rated* for 60A, then you *should* be able to put 60A through it. Whatever voltage drop you measured at 20A, will just get higher at 60A. If the drop goes to 1.5V at 60A, then it would dissipate 60 * 1.5 = 90W. The "large" heatsinks they sell you to go with these relays are only good to about 20W. It is reasonably certain that your SSR either was defective or a fake. Virtually 100% of all this stuff is made in mainland China these days. It is unlikely that even a legit part would have "Made in Taiwan" on it.

If you got it off of one of the auction sites, and bought it from a random low cost seller - it's most likely a fake. Blowing up at 20A makes it a bad fake (there are perfectly good ones ...). Even buying from distribution, it can be tough to sort out who has what. I got my latest batch on Amazon and got lucky. It's barely a step above the auction sites, but you have a little better ability to get your money back when the part is totally bogus. That of course assumes it's shipped / sold by Amazon and not "Uncle Bob's Fine Pile of Junk".

Bob

Got fed up with the Solid State Relay approach! grabbed a 10k resistor and a Power MOSFET, soldered the MOSFET casing to a strip of brass, put thermal compound on the back of the brass strip and bolted it to the SSR heat sink. Connected it up to the RUMBA, Power Supply and heated bed, powered on and it works like a charm. 19.2 amp draw gives a voltage drop of 0.17volts across the MOSFET, with no appreciable heating! Back to work on more designs!