Mk2a PCB heatbed Woes

Hey folks,

I'm having some trouble getting my mk2a heated bed to reach 110C (or even 95C, for that matter). The resistance across the PCB terminals (no wire) is 1.7 to 1.8 ohms. The multimeter leads measure zero ohms resistance so I'm guessing my multimeter calibrates them out (or they are very low loss). This resistance seems a bit high. Including the wires, I measure 1.8 ohms, so at least there isn't much resistance in the wires running to the 12V PSU (which is rated to 20A @ 12V, and the 12V rail doesn't sag under load).

I've timed the warm-up to various temperatures, where are the numbers measured from the thermistor. An IR thermometer reads about 2-5C warmer, but I'm not sure if the glass interferes with the reading.
To 65C: 5 min
To 80C: 9 min
To 85C: 12 min
To 90C: 16.5 min
Max temp is about 92.5C
Ambient temperature is about 25C

Currently for insulation, I'm using a thin (1/16") thick piece of cardboard wrapped around with aluminum foil. It did help a bit over the bare heated bed, particularly in terms of warmup time to 65C. In hindsight, I realized that I probably should have not wrapped both sides with aluminum foil. I picked up some 4.8mm thick cork tile squares from Lowes and was thinking of using these for better insulation. Before I go ahead, though I'd like to ask some advice. Should I create a small air grap? Should I back the cork with aluminum foil (not the side that comes into contact with the bed)? Is 5mm of cork enough, or should I double stack for 10mm?

I want to do it right this time. Thanks for any advice!

At 1.8 Ohms and 12V, that's only 80W going into your heated bed. It's no surprise you can't get to 110C with that. I would contact the supplier you bought it from and see if you can get a replacement. No matter how much insulation you add you'll have difficulty getting above 100C.

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

NewPerfection Wrote:
-------------------------------------------------------
> At 1.8 Ohms and 12V, that's only 80W going into
> your heated bed. It's no surprise you can't get
> to 110C with that. I would contact the supplier
> you bought it from and see if you can get a
> replacement. No matter how much insulation you
> add you'll have difficulty getting above 100C.

The problem is I bought it secondhand about 6 months ago (long story). Changes are I wouldn't get a replacement :\.

Zangetsu57 Wrote:
-------------------------------------------------------
> The problem is I bought it secondhand about 6
> months ago (long story). Changes are I wouldn't
> get a replacement :\.

Ah, that sucks. The heated bed really should have a resistance of 1.0-1.2 Ohms, giving a total power of 120 to 140W. Are you located in the States?

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

Zangetsu57, unless you have a pretty good idea what your doing, I wouldn't try this right away. I'm posting it to get peoples thoughts first.

What if he picked a couple of traces that were right next to each other, scraped some paint off, and soldered a bridge between them to reduce the total resistance? Perhaps in a couple of places on the board?

(Something you could look for now though, my board has solid unetched copper on the top under the paint. One small place on the top had a bubble under the paint. It looks like paint does on wood when water penetrates below the paint. Scraping proved that there was corrosion under the paint. Check your board to make sure it doesn't have any such places on the bottom that would narrow the traces resulting in more resitance.)

appdev007 Wrote:
-------------------------------------------------------
> What if he picked a couple of traces that were
> right next to each other, scraped some paint off,
> and soldered a bridge between them to reduce the
> total resistance? Perhaps in a couple of places on
> the board?

I would worry about getting uneven heating that way, and possibly burning up a trace.

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

My thinking was the bridge would offer an alternative path for the current, so hardly any if not none would be flowing across the trace. This would cut that trace out of the loop, reducing the resistance of the board. Wrong thinking? Would adding more path result in over heating of the trace?

NewPerfection Wrote:
-------------------------------------------------------
> appdev007 Wrote:
> --------------------------------------------------
> -----
> > What if he picked a couple of traces that were
> > right next to each other, scraped some paint
> off,
> > and soldered a bridge between them to reduce
> the
> > total resistance? Perhaps in a couple of places
> on
> > the board?
>
> I would worry about getting uneven heating that
> way, and possibly burning up a trace.

There are twenty zigzags on each of the four blocks on the PCB, so cutting one out and replacing it with a bridge would reduce resistance by 5%?

If four zigzags per block (every fifth one) was bridged, the PCB would have about 1.4Ω. Is that right?

That would still require 16 bridges....

---

Yvan

Singularity Machine

NewPerfection Wrote:
-------------------------------------------------------
> Zangetsu57 Wrote:
> --------------------------------------------------
> -----
> > The problem is I bought it secondhand about 6
> > months ago (long story). Changes are I
> wouldn't
> > get a replacement :\.
>
> Ah, that sucks. The heated bed really should have
> a resistance of 1.0-1.2 Ohms, giving a total power
> of 120 to 140W. Are you located in the States?

I am located in the USA (southern California). As for the trace cutting ideas, I'm going to opt out since it is pretty risky. Even if I could short some of the paths to reduce resistance, it would be difficult to maintain a flat surface to put the glass over.

I'd still like to try cork insulation and see how that goes. There's always the chance that my multimeter is inaccurate since it was $30. Does anyone have a suggestion of how to attach the cork sheet to the bottom side of my mk2a heated bed? The bed is suspended over an aluminum plate via tensioned springs, so the easiest way is to attach the cork directly to the bottom of the heated bed. Kapton tape, unfortunately, doesn't stick to cork that well.

Edited 1 time(s). Last edit at 09/15/2013 05:50PM by Zangetsu57.

I ended up just clipping the glass / pcb / cork sandwich together with bulldog clips. Works reasonably well, however eventually you'll make a divot in the cork.

A $30 multimeter should be accurate enough to measure the resistance in most cases. Although improving the insulation would help, in the end I think 80W simply isn't enough power to heat up the mass you're trying to heat up (and keep it hot when everything is moving). Using a higher voltage would work, but that would require another power supply which would probably cost more than just buying a new printbed from a reputable source.

Yes, ignore all the sillyness about bridging the PCB!

As a temporary work around, you can enclose your printer,(cardboard box, anything really, as long as it is as compact as possible) that will give you a few extra degrees. Also if you have an extra heat source that can be placed nearby, that can help.

A new PCB from a known good supplier is the best most practical solution long term.

---

Yvan

Singularity Machine

@Zangetsu - To measure the resistance of the beds, you have to account the resistance of the leads since the resistance you are measuring is pretty low for a regular voltmeter. Short your leads and record the resistance, then subtract it to the reading you are getting from the bed. Thats your approximate resistance.

As for your slow heating problem, you could bump up your PS output to around 14V, assuming all your other components like controllers and hotends can handle the increase. This would increase the current drawn by the bed and heat up more. Also, make sure that you have thick wires going to the bed. #14 would be good.

Resistance in these beds are so difficult to get right, I should know. I recently rejected close to 100 units of beds due to out of spec resistances. Best way to accurately measure the resistance is with an adjustable power supply with volt and amp meters.

royco Wrote:
-------------------------------------------------------
> @Zangetsu - To measure the resistance of the beds,
> you have to account the resistance of the leads
> since the resistance you are measuring is pretty
> low for a regular voltmeter. Short your leads and
> record the resistance, then subtract it to the
> reading you are getting from the bed. Thats your
> approximate resistance.
>
> As for your slow heating problem, you could bump
> up your PS output to around 14V, assuming all your
> other components like controllers and hotends can
> handle the increase. This would increase the
> current drawn by the bed and heat up more. Also,
> make sure that you have thick wires going to the
> bed. #14 would be good.
>
> Resistance in these beds are so difficult to get
> right, I should know. I recently rejected close to
> 100 units of beds due to out of spec resistances.
> Best way to accurately measure the resistance is
> with an adjustable power supply with volt and amp
> meters.

Hey Royco, thanks for the reply. When measuring the resistance, of 1.7 ohms at the bed, I actually did account for the resistance of the leads. The resistance of the leads was measured to be zero as long as I used modest pressure when touching the leads to each other. Good idea on upping the voltage on the PSU though. I'll check the tolerance of my Arduino board and RAMPS.

Fortunately after improving the insulation with aluminum-foil covered cork and using a cover over the heatbed when warming up, I'm able to reach 65C in 6 minutes and 100C in 12.5 minutes.

I'm having an issue where the power input to the RAMPS is heating up and melting the plug, but not the whole plug, just the bed positive lead. I've checked the bed resistance, and it seems to be fluctuating between 0.6 and 1.2 ohms, but this could just be the multimeter. I can't see why the bed would exhibit variable resistance unless the LED and 1K resistor have capacitance. The power input for the bed was making intermittent contact, so I thought probably this melting problem has more to do with a poor power connection to this pin. But the weird resistance on the PCB has me wondering if it isn't really the bed after all. I would like to try soldering the power input to the RAMPS to make sure it has excellent contact, but I'm sure I'd make a mess. I'd also like to have a look at the traces on the RAMPS board to make sure they're not heating up too much, but I can't get in-between the Arduino and the shield to examine it in action.

Any tips on troubleshooting this further?

---

|
| Lead Developer of Marlin Firmware
| Help support my work at Patreon and Elsewhere.
|

Just double checking - you are measuring the resistance of the bed with the multimeter probes, without it attached to anything, and from the trace solder points, right? If you apply ample pressure with the probes to ensure good contact, there should be no fluctuation. If there is, it might mean intermittent contact or a short somewhere, especially since 0.6 ohms is very low for this bed.

If the screw on the plug isn't down *tight* and the wire *clean* at the plug, you can melt any of these connectors.

A good way to check this stuff is with an amp meter. A clamp on one that will measure DC (and has a low range scale) is ideal. Another option is to make a current shunt and install it in series with the power lead. A measured length of wire with a pair of taps on it is a perfectly adequate shunt in this case. A wire table will give you the ohms / meter of your wire. Ohms law will give you the amps to volts conversion. You want something that's > 50 mv if you are using a normal DVM with a 200mv scale. You probably want < 200 mv simply to not waste power.

I took another measurement at the bed terminals with the wires disconnected. At that point the meter shows a good 1.3 or 1.4 ohms. So that points back to the plug contact point as the main problem. The MOSFET has a large sink and gets pretty hot, but it never fails.

So, I turned on the bed heater with the bed disconnected, and while the power input got warm, there was no burning. Next I reconnected the bed, tightening down the screw terminals. Again with the bed on, the plug smoked. So I poked at the plug to make sure the contact was centered. Then I adjusted the socket pin to ensure it was centered. Finally I plugged it in, but I noticed it was hard to get just right; the clips get in the way. Once I got the plug in correctly I started the bed heater, and nothing started to burn.

It looks like the plug was the culprit, just having been bent and plugged in too loosely. Which makes sense, I guess. I wonder if there isn't a more foolproof plug design that could prevent this completely. Anyway, it went for months before this issue appeared, so how bad could it be?

Thanks everyone for your thoughts on the issue. I learned a bit about resistance and watts, and they are nothing but trouble. Superconducting wire had better be coming soon.

Edited 1 time(s). Last edit at 11/14/2013 06:23PM by Thinkyhead.

The old time answer (old = older than I am and I have gray hair) is to use barrier terminal strips and screw everything down. They aren't pretty and they aren't small. They also are not as easy to pull stuff out of. The other solution is a true high current connector with crimp contacts on each side. They work quite well. You need a (usually expensive) custom crimp tool for the contacts.