PSU shuts down

[DeltaCon]

Recently I have been adding a 24V PSU to my Rmax V2. I choose this Delta Electronics 600W 25A PSU above the Meanwell, because of the better price and better shipment conditions (digikey sends shipment free to Europe):

http://www.digikey.nl/product-search/en ... 45-1068-ND

Installation went fine. SSR (crydom) and stepdown and also the PSU itself do not get warm at all. I noticed however a change in how the bed gets heated. On the Viotek PSU I could see the bedheat-led blink very fast when it came to temp. Fast switching on and off due to PID pwm. That behaviour changed. Now the bed heater switches on and off in seconds, or max 2x a second or so. It gets to temp very quickly, and keeps temp nicely. A bit irritating is that the PSU fan follows exactly the same pattern, switching on and off a lot. I would not mention this if there was no problem though...

Problem is that the PSU shuts down after about half an hour, even with unheated hotend and no motion. After a power reset it comes on immediately. You would think it can not handle the high load and has a temperature switch, but it does not even get warm to the touch. The high load when heating continuously seems to be no problem. I see no drop in voltage (trimmed down to 20,5V). I am thinking that the PSU might not handle the very fast PID switching very well. Does that make sense?

Maybe I should turn to bang-bang pwm for the heatbed. Advise is welcomed!

[Xenocrates]

I do suspect you are correct, and that the switching frequency is the problem with the setup. I think that the correct answer may be to set the control to either dead time (heat manager 3), or to bang-bang with a set period, and move the period to the lowest reliable value.

I recommend dead-time control primarily. If you were moving over to duet firmware, I know that it has different control rates for different heaters, such as chamber heaters only running at 10Hz. I don't know enough about the repetier firmware to be able to tell you how to change the PWM rate, which would honestly be the ideal solution, as you could keep PID control and get rid of the random shut-downs.

[DeltaCon]

I changed repieter to remedy the above problem. I failed to check the setting before I made changed, so in the end I am not sure anything has changed at all. Here's what I did:

First I changed to slow bang-bang with a delay of 1 second. When I heat to 50, I see an overshoot of at least 5 degrees. Of course then it drops to 50, the heat goes on again for a few seconds, and then again an overshoot of at least 5 degrees. This repeats. So this method wil not survive for heating my bed. Perhaps it works for an actively cooled peek because the intervals will be much shorter because of the lower thermal mass. But it does not seem fitted for beds.

Then I changed to dead time and forgot to set the dead time to a reasonable value. I saw exact the same behaviour as seen before I wrote the OP. So now I am wondering if my heat manager was set at PID at all before the exercise started... Should have checked...

Then I changed deadtime to 10 seconds instead of the 63 something it was before (a PID value). Temp rises to set temp minus 20 and then starts switching. I noticed the switching intervals are a bit longer then previously, but I doubt the difference is big enough. I find it strange that sometimes the heat goes on three spikes in a second, and then not at all for 5 or 6 seconds. Why not keep it burning a bit longer so there would not be so many spikes? The PSU has not shut down yet in over an our and a half, but I was testing on 70 degrees instead of 90 as did before. Tonight I wil test further on 90 degrees and see what happens.

Can something be said on the usual max switching frequency a good PSU can handle? I used to think that PID switches 100s or 1000s times a second, but I learned now that 15Hz is the most commonly used setting. Could such a frequency kill or safe-trigger a PSU indeed? Are people running their SSR's on such a frequency? And does it make sens at all to use PID on a SSR? I ask this because in my believe the Rambo also changes the voltage according to needs while heating, and that of course will be eliminated by a SSR (everything or nothing...).

I kindly ask my community friends here to share their heatbed settings on their 24V SSR setups. Thanks!

[DeltaCon]

I have been dry-running the bed for about 40 minutes now. I do not experience a lot of change in behaviour with this and my starting point, except for the fact that it has not shut down on me, yet I think I can be satisfied wit the temp curve as it is now.

Some spike do not show up on the graph though (not even when I zoom in). A deviation of 1 degree C is not bad I guess. Except for some small incidents that is. I expect it to be even better, and the spikes getting less frequent, when I add thermal mass (aka heatspreader). But if possible, would a few gents or ladies please share their bed setup with me? Heated print bed that is...

[DeltaCon]

Hmm, thought that my issue was over... I have been testing and calibrating a lot lately, and only today I found the time to do a test print. But I just ran halfway through a 20mm testcube when my PSU shut down on me again... I am still under the impression that this PSU does not like being loaded and unloaded so often by the heated bed PWM. I am running deadtime now with a deadtime of 21 seconds (found some guide on the big net). That works really well in terms of no overshoot and quick heat up. Is there a better way for PWM on a SSR'd heated bed? A way that switches less often? I see spikes sometimes 3 or for times a second. Looks like my PSU does not handle that well. Or are there other thoughts about this?

[DeltaCon]

Seemingly there is not a lot of knowledge and experience about the problem I am having and have described above. Or do people want to keep secret how they heat their beds

I've sent a mail to Delta Electronics (manufacturer of my PSU) asking if something can be said about the switching frequency. So I hope to get some info soon...

An other thought is that maybe EMI is triggering a safety feature of the PSU. I have my Crydom SSR strapped to the PSU housing just behind the fan, hoping to get some cooling. Neither SSR nor PSU get warm during operation, but maybe SSR's give much EMI? Maybe I'll open up the base again and give the SSR an other location...

[lightninjay]

Everything that I have always read has pointed toward 24v heating for the bed, leading to either dead-time heating or bangbang, just because of what you mentioned, as well as not wanting to overheat the SSR from constant switching.

[DeltaCon]

I agree. It just looks like my PSU thinks even dead-time switches to quickly. bangbang is useless because of the large overshoot. Dead time is spot on. I'll try to find out if EMI plays a role in this too.

[Jrjones]

Have you tried setting
#define PDM_FOR_EXTRUDER 0
to 0?
I was having the same issue with my ATX PSU before.

[Noircogi]

Are you sure you're not pulling too much current for your power supply and causing it to overheat?

The Onyx in my Max V3 is 0.92 ohms.To get reasonable current after doing my "24v, 600w" conversion, I turned the output voltage of my power supply down to 20.5 volts. I'm using a single supply with a regulator for the 12V section, so I limited the bed to around 450W, leaving some for the rest of the printer.

If your Onyx is the same as mine, running it at 24V would use 26 amps for the bed alone. More than the rated output of your PS, not even counting the steppers, hot end etc.

[DeltaCon]

Have you tried setting
#define PDM_FOR_EXTRUDER 0
to 0?
I was having the same issue with my ATX PSU before.

No... But I think my problem is in the bed, not the extruder. But likely there is a likewise setting for the bed? Can you explain the difference of PDM to PWM to me?

Are you sure you're not pulling too much current for your power supply and causing it to overheat? The Onyx in my Max V3 is 0.92 ohms.To get reasonable current after doing my "24v, 600w" conversion, I turned the output voltage of my power supply down to 20.5 volts. I'm using a single supply with a regulator for the 12V section, so I limited the bed to around 450W, leaving some for the rest of the printer. If your Onyx is the same as mine, running it at 24V would use 26 amps for the bed alone. More than the rated output of your PS, not even counting the steppers, hot end etc.

Yes, my PSU is turned down to 20,5V too, so I don't expect a problem there. Also the casing of the PSU, and the airflow out of it, do not even get warm to the touch. So overheating cannot be the issue in my perspective.

I will look into the PDM setting first, before opening the base again. Thanks for the direction.

[Jrjones]

This is the thread on the Repetier forum addressing the issue: http://forum.repetier.com/discussion/50 ... expectedly
PDM is Pulse Density Modulation, and PWM is Pulse Width Modulation.
With PDM at 50%, you would see lots of little on pulses, whereas with PWM at 50%, you would have one long on pulse for the switching frequency. So PDM has a higher switching frequency

I'm not sure if there's a setting for the bed PDM or PWM. I think it might have to do with the heater cartridge's induction, and switching it at a fast frequency causes some PSUs' safeties to trip.

[DeltaCon]

The same seems to happen to my PSU, but there is no pdm setting for the bed, so I can not turn that off (if it is on at all). The link you mentioned states that bang-bang does not hurt for the bed. But it does: when I set a temp of 50C the actual temp varies between 50 and 55. That can be seen in the outer walls of an object easily as a banding structure. I don't want that! Besides it overshoots as much as 12C after the first heat-up. dead-time is much better but gives too many too short bursts. It seems it cannot be controlled any further. Thanks for the link.

Edit:
#define PDM_FOR_EXTRUDER 0
was already off, but as mentioned the problem is with the bed.

[DeltaCon]

I took apart the base once again and dismounted the SSR. I hung it about 20cm above the PSU, as far as the cabling allows. In a testrun the PSU fails again. So I think I can rule out EMI being the cause. Also I tried PID tuning the bed, just to see what happens. The quick power switching causes the PSU to fail within a few seconds after PID kicks in. So I guess my problem really is related to the switching frequency... I received no reply from Delta Electronics. I am beginning to think I better would have thrown out a 100 bucks extra (partly on the higher price, but mostly on shipping) to get a Meanwell afterall...

[Jrjones]

Have you tried changing this value? #define HEATER_PWM_SPEED
That changes the PWM frequency.

Are you using dead-time for both the bed and extruder? I prefer dead time control for both.
And does your PSU shut off if you are only heating the bed?

[DeltaCon]

Yes I have changed that value from 0 to 4 and everything in between. But I think it only applies to PID control. I have the hotend at PID control, but the bed not (if I do that the PSU kills in a few seconds after the initial heating). For the bed it looks like bang-bang works reliably, but it gives me too much overshoot after each pulse of around 6 degrees C and that is very visible as banding in the printed objects. Dead time does a much greater job and a stable temperatuur, but in my opinion it gives too many too short burst too quickly and the PSU kills after some time. As far as I know there is no way to influence that behavior, is there? Yes, it shuts down also when only heating the bed.

We have the same taste in printer, but very different taste in motorbikes

[DeltaCon]

Yesterday I have been running the heated bed for about 2,5 hours non-stop on a different mains circuit. At first I was astonished that it seemed to work reliably, but in the end it still shut down on me. So probably still nothing to do with the mains. I do not have a lot of faith in this PS unit anymore. Delta Electronics is not responding to my inquiry about the switching frequency. So I am asking for an RMA with Digi-Key. I hope they will do a refund, so I can get a Meanwell to replace it afterall...

In the meantime I am considering taking the bed off the 24V and put it back on the rambo. But I guess the 12V stepdown 20A / 240 Watts will not like that... I really don't want to build in the Viotek anymore, because I would have to redo the whole cabling again.

[DeltaCon]

Digikey is sending me a replacement PSU of the same kind (Delta Electronics PMC-24V600W1BA). I think they offer great service, because I bought it somewhere in April already, and I explained the situation but did not have to beg. It took me one mail for them to ship a new one, and they don 't even charge shipment. So, after the weekend I will know whether my PSU problems are just a defect in the one I have, or whether it is a problem with this type of PSU overall.

Tuesday night I had an old friend visiting. He is a R&D tech at Océ Canon. We came to talk about my printer and I showed him the problem of my PSU shutting down. In his job he works both with Meanwell and Delta Electronics PSU's and I was surprised to hear that Meanwell generally is the cheaper one. Of course they buy in quantities of full ship-containers The 750 Watt type they don't use, but he can probably "spare" me a 600 Watt PSU. So if the shutdown is a problem for Delta PSU's in general, I hope to have a Meanwell at my disposal soon

Also my friend was surprised that there are not many parameters configurable for deadtime control in the firmware. According to him it would not be hard to implement parameters for "if triggered ON keep ON for at least X ms" and "if triggered OFF keep OFF for at least Y ms". That is really beyond me, but it might be an idea for DC42 or other people here working on firmwares. These values would be valuable to prevent some modern PSU's to go into "hiccup-mode"

So we'll see, progress is slow. Havent been printing anything for several months now, but I do everything in very little sparetime.

[DeltaCon]

I received the replacement PSU yesterday. First 2,5 hours have been heating with no shutdown. I had to break off to get some sleep
Nothing conclusive, but it gives good hope that that the problem was exemplary to my first PSU. Think I will close up my printer with the new PSU and start a real print this weekend.

[DeltaCon]

To conclude this thread, here the latest findings:

I received the new Delta PSU in just a few days. I have been printing and heating for about 12 hours in total, and the shutting down did not happen anymore. I think I can conclude that there was a fault in my first PSU. Cost me some time...

On to the next problem: tower lean!