The Progression of Kit278

May I know the metal case of crystal for CPU grounded ?

Quote
tru168
May I know the metal case of crystal for CPU grounded ?

Not too sure hard to see, but 2 of the pins are grounded, pins 2 and 4 are the ones that are grounded.
Pins 1 and 3 are the actual crystal pins 2 and 4 are the can I guess.

Paul

---

RS Ormerod No 436

Looked at the thread about freezing and the non- scientific test, I suspect poor earth or none at all but who knows without testing it.

Dieter

Ormerod #257

Its grounded then. I need to wait until I receive the Ormerod from RS for further investigation on the duet board. 1 month waiting , sigh.........

Edited 1 time(s). Last edit at 01/10/2014 09:42AM by tru168.

This weekend I had very limited time to work on the machine. Still need to check the large PSU (and mod the fan for voltage drop). So I am still busy with PLA prints.

1) I am still on the older firmware. For PLA the bed needs to be 70C and the extruder at 210C else the print becomes stringy and/or does not adhere to the table. I assume Version 53 will sort that out.
2) Still need to solder my USB socket. I can see some movement so it might be already have fatigued the 4x connections. I filed out the SDcard slot. Much better now.
3) My bed is so skew it is not even funny anymore. Before I started fiddling with it the left side was 5mm lower than the right. I took everything apart and tried different spacers. I got it good enough now so that I can at least print all over the middle 75mmx 75 mm part. But I need to either redo the bed frame in Alu or devise a better method with screws.
4) My IR sensor, for Z axis, calibrates all aver the place. My ligting stayed the same. Solution is to calib to zero, move the X,Y to the middle and then manually turn the Z axis to get it to the right position. A bit of a pain but at least I then get repeatable results.
5) I probably printed about 50meter of PLA filament over 3 days. My lab is Gauss shielded (my cellphone signal drops by 20dBm even with the front door open). I did not take any cellphones in there. My 3G internet runs from other unshielded room. All my power (each and every single wall socket) runs independently back to the distribution boxes in another room. So I don't have the problem with soldering irons and other switching devices. Good news is absolutely no crash at all. Yesterday I printed 8 parts nearly non-stop and the last 4x were part of a single gcode file.
6) I installed the new part to stop fan backwash. I got 1x part to lift slightly but I think that was because the bed was too cold. Heated it up to '70C' and then the next set worked fine.
7) The RRP supplied PSU is not up to scratch. With both the bed and extruder heaters on (no motors running) the voltage on the screw terminals drop to about 10.1V I forgot to check with all motors running, but it should be worse. (PS. South Africa, +-230VAC). As the heaters kick in you can hear all the fans slowing down. Thus when bed and extruder is on the cooling to the hot end drops. Maybe not a good thing. But with the bigger PSU my output voltage stays 13.8V right up to 30A.
8) Y belt still not perfect. When I did the 4x print I found visible errors in the Y axis. I need to make a part that gives me control on the tension, maybe with a screw.


Well, the 3D printer is now printing parts for my LDI (Laser Direct Imaging) system. It is an old salvaged engraver with 0.5Nm motors on 5mm pitch screws (with absolutely no detectable backlash). So basically only the X and Y axis with around 150mm by 150mm area. The Z plane seems nearly perfectly flat so it shout make a nice 3D printer although the max height will be limited to around 4cm. Over the weekend I printed pillow cases for the 6mm bearings and they came out very nicely (not professional looking but quite functional). The Z axis must contain receptacles for both the laser (which is focuses by going up and down, and a small NEMA8 motor to control the laser line angle as it is not a dot) and a Dremel drill (37 000RPM) for drilling pad, via and ref holes. So far the 3D printer is very helpful for the bearing pillow cases, rod ends, attaching section to X axis, end bearing holder and Dremel tool holder. The Z axis movement is via 400 step-0.5Nm stepper and 5mm V-thread, but I have machined Alu nuts (2x 10mm) that are tuned to provide zero backlash (as long as it is well oiled). With the 3D printer I can now design a part that takes these 2x nuts and allow it to be tuned for best movement and zero backlash. I will post some pictures when I am done. With this LDI machine I can "draw" 1mil/25um features directly in photo resist which should allow minimum PCB features of around 3mil tracks and spacing. No artwork and better resolution than what PCB fab companies can do at the moment! Initial calculations show that I should be able to expose a 10cm x 10cm PCB in raster format in around 70 minutes, thus roughly 2 1/2 hours with holes drilled as well. In vector mode (where backlash becomes very important) it should be a lot faster. Using LPKF's through hole metal-polymer solution I can do through hole double sided PCBs in another 30 minutes (but this stuff is seriously expensive). The photo image-able photomask exposes in the same way and is actually a lot faster as it only needs to expose the parts that needs to be removed. I have a 1Watt blue laser as well that I want to use to burn the "silk screen" on as well. It makes a very fine black mark (also around 25um) but what is nice about this is that I can add text that you cannot do with normal silk screen. This attachment requires a fan to cool the laser off, thus yet another application for the 3D printer.

So RRP, your Ormerod design still needs lots of work but I am beginning to enjoy working with it.

Thankyou for sharing your activities, your kit sounds great for prototype PCB making. Watch out for under-etching on such fine traces, leaving a track with almost no area adhering to the board. Having fine but readable silkscreen markings will certainly be a great help on tight & dense SMD boards.

dmould

For 3mil tracks the copper thickness is 18um. On 35um it will fall off if you rub against it. A more realistic value is 5mil. Then I am assured that I can send the PCBs out for mass produceable quantities at a PCB fab. Here in South Africa we have a place that can do downto 4 mils. But again it depends what you etch it with (and the speed of etching).

Silk: Just got a board back where we could not put any silk markings under the fineline BGA. The entire section was drawn in 1mil text that we get the CAM people to remove. With the laser I should be able to actually draw it. The poor PCB assembly guy now needs to sit and work out from an assembly drawing where everything goes ;-)

But this 3D printer concept is awesome. I think it, CAD it and then (in a short while) I have it ;-) And this then allows me to make other machines that run along the same line.

OK, big PSU time. Xytron 13.8V 30A Constant 32A Surge. Now that is a power supply.

I tried with the supplied power supply again before this. I only get to around 90C. I tried to heat up the bed with a heatgun. Get to 110C and then it falls down to 90C. Even the extruder only gets to 245C when I demand 250C. Then the PSU goes funny, electronics causes crap and all hell breaks lose on the USB bus requiring reboots.

Plugged the monster on using the same flex power cable. The cable itself heats up to 36C (Ambient = 27C About an 80mV drop!). Bed demand = 115C. It reports 115C. Fluke 62 IR reports 111C to 112C. Pronter reports extruder at 245C (demand is 250C) but I can see more overshoot (13.8V ??). Unfortunately I don't have a spot where I can put the current clamp in, so I have no idea what the consumption is.

But I am printing ABS and it looks much better than anything I printed in PLA! So my conclusion is that the PC power supply sucks! Big time! Sorry RRP, it does.

Another worrying thing. The electronics also heat up. The CPU is now sitting at 47C. Max temperature over entire board is 55C. I think you guys need to rethink the bed-heater bus plus electronics away from the Duet.

The print just finished. Definitely the best one so far. The only drawback is the bed is so hot you can cook an egg on it and takes a very long time to cool down. But if that is the price to pay then I am OK with it.

Quote
victors

Another worrying thing. The electronics also heat up. The CPU is now sitting at 47C. Max temperature over entire board is 55C. I think you guys need to rethink the bed-heater bus plus electronics away from the Duet.

something along these lines?

[forums.reprap.org]

Is ambient 27°C a good thing or a bad thing (ambient 8°C wher eim typing in my outhouse while smoking)

Ray

Ray

Ambient temperature adds to whatever you are generating as dissipation. So 8C is much better for the electronics, but takes more effort to get the bed up to 110C.

Many electronic components only work up to 55C but then others, like regulators, power MOSFETs and stepper drivers can operate over 100C. I measured the 55C on the stepper chips.

The majority of the power goes to the bed. So for this I would suggest proper 20A to 30A cable (for the bed), going to a MOSFET circuit that is switched by the existing output. Then it would be possible to make the ugly ribbon cable neater.

Quote
victors
So my conclusion is that the PC power supply sucks! Big time! Sorry RRP, it does.

RRP seem to have been shipping two different power supplies, Ace and Alpine. I have the Alpine, which is a 550W PSU rated at 32A on the 12V output. I got the bed up to 120C, read using the old firmware (it might read a little differently with the post-0.52 firmware) and it was still rising, although it took around 20 minutes to get there. I think this PSU is fine for PLA and probably just about OK for ABS. However, I have purchased a 12V 300W PSU that I will switch to when I want to print ABS, because it will allow me to heat up the bed faster.

Quote
victors
Another worrying thing. The electronics also heat up. The CPU is now sitting at 47C. Max temperature over entire board is 55C. I think you guys need to rethink the bed-heater bus plus electronics away from the Duet.

Whilst it would be nice to keep the high current bed heater circuitry off the Duet PCB, I don't see any real need to do it. The power dissipation in the bed heater mosfet is less than 0.5W. But if you want to use an external bed heater mosfet, there is an output (J14) provided for this purpose. As for the rest of the electronics, it is normal for stepper motor driver ICs to run warm. However, putting the electronics board in what is almost a sealed plastic box is not ideal, and I think it would be good to add a few ventilation holes at the top of the case or the back plate to allow cooling by convection.

Edited 2 time(s). Last edit at 01/14/2014 03:54AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Victors,

Nice! The main reason I purchased the 3d printer was to build other machines and hard to get parts and custom parts. Not really as a toy or nice to have at all...

Ok, so the psu is crap. I figured there must be some voltage sag with the bed current and the band cable, seems a bad idea for that kind of current to the bed. So note, new psu will solve a lot of issues.

The heater is rated at 40watt, so at higher 13+ volts you should get higher temp on the head. Probably no more than 4amps from the print head at the very max. Wires are thinnish for that sort of current and you should see a voltage difference at the head vs supply - BAD - upgrade wire guage to heater head.

Those on board regulators on the duet will be generating the heat. Then the stepper drivers too. Seen some designs put heat sinks on the stepper drivers. Maybe check all axis are smooth and nothing is grabbing or labouring the steppers.

I don't have the heat issue but that must be due to the higher voltage from the psu upgrade.

Interesting, a photo of that 'LAB' of yours would make an interesting post.

Dieter

Ormerod #257

Hi Victors

We have been supplying two different ATX power supplies. The Alpine is actually the higher output one, and should say on it that it supplies +12V with 30A. The Ace PSU is lower rated, at 22A. However, the machine should not draw even close to this much current, at 12V.

The PSU we supply with our Mendel kit (12V 20A PSU for LED lights) is capable of having it's voltage turned up a bit. The usual advice I give when doing this is to BE CAREFUL! I appreciate you probably know what you're doing, but here's some info...

Stepper drivers - won't heat up with extra voltage (up to their limit) because it's a current chopping device - turning up the current (via M906 in config.g) makes them run hotter. The motors will actually work better, because they energise quicker with higher voltage.
ARM chip - won't run hotter, because it's running through a number of regulators that cut the voltage from 5V to 3.3V to 1.7V (on the core). It should be able to cope with pretty high temps anyway. The 5V regulator on the board can cope with up to 35V.
Hot end heater - runs on PWM. It will heat up quicker, but ultimately the duty cycle will kick in. It should have a resistance of 3 ohms, which means it will draw more current.
The MOSFETS may run hotter, but they are well within specification.

So it's all about the bed heater. Check the resistance of your bed; it should be around 1.2 or 1.3 ohms.
At 12V, with a heatbed resistance of around 1.3ohms, the bed draws 9.23A, so there is 110.7W of power going to the bed.
At 13V, same resistance, the bed draws 10A, 130W.
At 14V, same resistance, the bed draws 10.77A, 150.7W
At 15V, same resistance, the bed draws 11.54A, 173W

If you're generating a lot of heat on the heated bed, it may be radiating through the Duet enclosure. Perhaps put some silver foil facing the bed. If it's generating heat on the board, check the following, which is the first advice I'm giving to people with power supply cutting out (very few emails on this, not sent out any PSUs or ATX power PCBs as warranties yet):

1. Check that the power cable is pushed fully into the PSU.
2. Check that the PSU plugs are pushed ALL THE WAY IN to the ATX power PCB. It is easy to have a partial connection, which will generate resistance, and so heat.
3. Check the power wire that goes from the ATX power PCB to the Duet, that the screw terminals are tight, and the wire is firmly held. Did you use the crimps, or 'tin' the ends of the wires with solder, or are they bare? Bare is bad - a good connection here is absolutely vital! Again, if it is not, it will generate resistance, and heat.
4. Check the resistance of the heated bed. Turn the power to the printer off, disconnect the two 2x6 IDC connectors on the bed PCB, and measure the resistance between the solder pads on the bed PCB with a good digital multimeter. It should be around 1.2 to 1.3 ohms. If it is lower, it will use more current, and could lead to the PSU being overloaded.
5. Unplug the hot end heater plug from the Duet board (brown and yellow), and measure the resistance of the hot end heater; it should be around 3 ohms. Again, check the hot end connector for good connections - it's possible the pins have been pushed out.

Hope that helps.

Ian
RepRapPro tech support

DC

I do a lot of design for military systems. Heat is the enemy. I agree that the current PSU is OK-ish for PLA. It does not overload and keeps my bed at 65C. My problem comes in when heating for ABS. I never reach 100C, the PSU gets hot and then causes funny stuff on the controller and USB. I also agree that these computer power supplies are cheap. So here is a suggestion (other than my very expensive transformer based PSU) for other users with the same problem as me:

1) Use 2x power supplies. This means buying another one for around $35. The second one is just for the bed.
2) Tap the signal from J14 and run it into a circuit using a common MOSFET like IRPZ44 (can do >50A). To drive it you will need xsomething like a NE555 circuit to switch a voltage multiplier to get it to switch the gate on at 10V to 15V higher than the PSU voltage. Or just use e.g. P channel IRF9Z24N because it is easier to switch them on. Will 2 to 3 units in parallel as I think they can only do 12A max. If you can handle the clicking noise then a 30A automotive relay should also work.
3) Print a little box to house this circuit inline with the "flow" of the ribbon cable and mount it on the black perspex. This will look a lot neater. Will still need to get the thermistor signal though a long with J14's enable signal.

I also want to print another cover for the electronics and mount a fan in it. 4x fans will surely make it look more impressive ;-) (2x PSU, 1x hot end and 1x electronics).


Dieter

Using this printer to make more printers and machines... exactly why I bought it. So I don't really worry about final print quality. I am already getting the parts together for a system with a 0.35mm nozzle and more rigid frame. I am waiting for an order for 96 enclosures that I need to print for a client. The Ormerod will probably be too small for that(but again maybe not). Currently I am busy with my LDI system, with which I can rapid prototype circuit boards. Then on to my 50cmx50cmx25cm CNC machine (for which I also need to make plastic parts and circuits). Then onto the Pick&Place machine. In parallel I am busy with a CNC plasma cutter that needs to be up and running by mid Feb.

The 4A for the hot end does not worry me too much. Seems like all the wires are doing their thing. The drivers are actually well within spec. I know the DRV8825 chips cut out at around 150C so 55C is OK. I don't know what the spec is on the Atmel or other chips, so it won't hurt putting another fan on them. The drivers have power pads on the bottom so I assume the PCB provides enough heatsinking. I think I might still have a set of stick-on heatsinks. One can make little Alu cut-offs and stick it onto the chip with one half epoxy glue and the other half heat sink compound, but this is messy. I would just stick with a fan blowing over the board.

--------------------------------------
A photo of my lab :-) Then I first need to clean it up a bit. Like I said before, it is a real man cave. The entire room (+-50M2) has metal shielding around it and is well earthed (reduces EMI in and out). The brick walls have a thick layer of fiber glass on and then rounded off with 12mm drywall (sound dampening). The windows are all double glazed. 2x different types of thermal covers for the roof. 9mm ceiling right at the top and then a 9mm office ceiling at 2.4meter. Plus a safe door to the entrance. Each double wall socket (8x of them) runs on its own breaker switch. 12 sets of lights for all types of applications including lights with low EMI. BTW, this room also doubles as a recording studio. Then I have a walk-in safe room (with its own safe door and outer sound dampening door) that is approx 1m x 6m. This stores my one big computer (connected via long DVI and USB cables) as well as all the other expensive stuff. When all the doors are closed you cannot hear this monster PC (with 9 fans in to keep it cool). Then I have all the nice to have electronic equipment. Lots of multimeters, different type of oscilloscopes, spectrum analyzer, huge power supplies, several soldering stations, hot-air station, 6x microscopes (each one has a different feature/ strong point) etc... I even have compressed air in the room. Next door I have all the loud stuff (another +-40M2) with the compressor, welding equipment, plasma cutter, drill/XYZ mill, metal lathe, various cutters etc... This is where the 3 meter plasma cutter and pick&place/oven are going to be housed. Since we bought the house in 2006 we just kept on expanding. My kids have 25M2 rooms each with a queen size bed, study section and lounge area with 32 inch LCD TV and stand. They love me ;-) I am currently busy with a 50M2 kitchen (keeping the wife happy;-). Later this year I am planning to add another double storey 200M2. I will probably stop when the town council says no to more building plans.
---------------------------------------

Ian

With the bed at 110C a lot of the heat radiated to the controller. Will do some more tests later. I will check the resistance on the bed. The hot-end seems fine. Used the connectors as supplied.
But the difference between the Xytron PSU and the PC PSU is quite noticeable. With the PC PSU the power drops out after trying to get it to >90C and eventually everything goes hay-wire. With the Xytron the bed heats up with ease (13.8V) and all the electronics remains happy.

Hmmm... Strange that the Ormerod not tested on ABS print before selling them on RS ? When I first heard that ormerod came with PC power I knew it will not working properly as a normal 3D printer.
My Ormerod still on its way , but I need to buy proper power supply for my ABS printing.
Why not Reprap pro stop using PC PSU as standard power supply and include an industrial type together with kit set? I don't mind if new products having bugs / minor problem but if there's known problem and keep supplying the same old PC PSU, I'm not really happy with that. (just my personal point of view )

Most customers, who are printing ABS, are doing it with our supplied PSU - see the other links in this forum.

Quote
tru168
Why not Reprap pro stop using PC PSU as standard power supply and include an industrial type together with kit set?

It's not within the budget range of this printer. The ATX PSU is cheap, ubiquitous, has standard connections, passes CE tests, has a power switch, and is easy for people to use. There is no other PSU that encompasses all of these. Compare with what customers have to do with our Mendel kit: [www.reprappro.com]
Many customers are uncomfortable doing this, and we wanted there to be minimal wiring.

Ian
RepRapPro

Quote
reprappro
We have been supplying two different ATX power supplies. The Alpine is actually the higher output one, and should say on it that it supplies +12V with 30A. The Ace PSU is lower rated, at 22A. However, the machine should not draw even close to this much current, at 12V.

My Ace PSU is 550W with 12V rated at 24A (maybe Max value).

Quote
reprappro
So it's all about the bed heater. Check the resistance of your bed; it should be around 1.2 or 1.3 ohms.
At 12V, with a heatbed resistance of around 1.3ohms, the bed draws 9.23A, so there is 110.7W of power going to the bed.
...

4. Check the resistance of the heated bed. Turn the power to the printer off, disconnect the two 2x6 IDC connectors on the bed PCB, and measure the resistance between the solder pads on the bed PCB with a good digital multimeter. It should be around 1.2 to 1.3 ohms. If it is lower, it will use more current, and could lead to the PSU being overloaded.

My Bed Heater measures at 1.0 Ω so its drawing around 12 Amps right there.

Quote
reprappro
5. Unplug the hot end heater plug from the Duet board (brown and yellow), and measure the resistance of the hot end heater; it should be around 3 ohms. Again, check the hot end connector for good connections - it's possible the pins have been pushed out.

Well my heater is measuring about 2.45 Ω so that's about 4.9 Amps there.

Paul

---

RS Ormerod No 436

I do not trust either the bed or extruder temperature reports from the supplied sensors - I suspect mine under-read by lots. After finding my contact digital thermometer was broken, I thought of getting a IR non-contact thermometer, but found that the best minimum spot size (min size of area to be measured) of any that I can find that are in stock is 13mm - too big to measure the extruder, and I cannot quite justify the cost of a thermal imaging camera. Plus it means making the measurement area black to get an accurate reading (bare aluminium has too low emissivity for accurate IR readings), and measuring the temperature of glass with IR is apparently not so good either, so I've ordered a normal thermocouple device instead, which besides being cheaper will give a more trustworthy reading without having to wonder whether the emissivity factor is set correctly on an IR thermometer or camera.

Major problem. Sagging head!!!

Tonight I started printing some real ABS parts for my LDI machine. Bed 110C, Hot-end 240C.

At first I did not understand what was happening. The layers were "stepping" away (from the Y motor). I thought it was the belt again but it is fairly secure now. Then it became progressively worse. By the time I realized what was going on the head was swinging like it was held together by bubblegum. My fan is actually blowing harder now running at 13.8V.

So major stripping now on the head to see how big the damage is.

;-(

I had a problem with head sagging over time when printing with PLA. I solved it using a new nozzle mount as detailed in this thread. You might want to make it the first part you print in ABS.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

DC

OK, damage not that bad. Basically what you experienced. So for now I have reassembled the screws with isolator washers (just to finish the new mount). So far everything looks straight. Thanks for the tip.

Also my fan had 3x 1N4007 diodes in series to drop the voltage to it. Removed it and the fan is much faster now. Hope I don't burn it out. But so far it is behaving.

Victors,

Get yourself a piece of 0.5m solid PTFE 20mm round bar from Maizey Plastics in Pretoria. You have the equipment it seems and make a new mount that replaces the part that mounts the aluminium block to the printer from PTFE. It will not fail even over 300'C. The PTFE bar cost me R103 ex vat and worth every cent and 100% fix for this problem. The screws transfer heat on the original setup and melt the thing. Failure will occur if you are running ABS.

The guys at Maizey Plastics Cape Town really know their stuff when it comes to various plastics. Hopes this helps you.

Dieter

Ormerod #257

I checked the specs of a few 12V 40mm fans at Farnell, and the ones that gave an operating voltage range specified either 13.2V or 13.8V as the limit. So you might want to reinstate one of those diodes.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Victors, I'm not sure what temperature that bracket reaches, but when mine melted (during a glitch that blackened the white PLA I was running), I made a replacement by milling and drilling some delrin, I've used it for a few hours a day printing with ABS for ten days or so. It might be quicker for you to look around the man cave for some delrin (or acetal copolymer if you don't have homopolymer) and milling it up. I slotted through for the bowden tube, but left the screw holes as holes with no slot or countersink (you need to counterbore the horizontals of course to prevent the verticals interfering).

The heat deflection temperature for delrin is MUCH better than PLA, and with your high ambient temperature it might be worth skipping PLA though ABS may cope unless your fan totally fails, which would also stress the delrin (heat deflection temps @ 66psi: 170°C for delrin, 105°C for ABS and 65°C for PLA - presuming they're solid not printed ). www.makeitfrom.com has quite a lot of comparative data if you have some other plastic lying around. Using metal would just spread the problem to the carriage...

[edit] just saw deiter's recommendation for ptfe - good temperature resistance, but pretty soft and flexible at best, PEEK would be a better option if your buying new stock, or Torlon (both available from RS too), the ultimate high temp plastic I found at my local stockist is celazole (intermittent use at 500°C, continuous use at 350°C, and it's harder), but costly [/edit]

Ray

Edited 1 time(s). Last edit at 01/14/2014 03:45PM by rayhicks.

The fans are so cheap here in SA from local electronics suppliers, R35.00 converted probably around $3.50 that if it blows, just replace it with a better one local supply.

Dieter

Ormerod #257

Dieter

I will go to Maizeys in Pretoria.I also have a small sheet of PTFE somewhere. Another option is to use heatsink bushes. The screws will thus not touch any plastic. The idea of making a solid Alu head mount and isolating it with the PTFE sheet also seems tempting.


DC

Will put back 1x diode. This seems like a double effort fix. I have just installed the 1st new mount. It printed ugly because my head started to sag again. Now for printing a nicer looking one ;-)

Ian and Victors

Just going back to the PSU posts of earlier, I have the 550W Alpine PC PSU that came with the kit. I ran a test now, voltage around 12.10V with bed off. Switch bed on to 110'C, voltage drops at the bed to 10.8V and overall system voltage down to 11.33V. Despite this, my heater bed gets to 110'C within 3 minutes or so and then switches in and out. The bed read 26'C before switching on which is about the room temp here in Cape Town at the moment inside. Bed measured 1.2ohm

I didn't bother reading the current as clearly the PSU is already struggling despite the 12V 32Amp rating which in my opinion is vastly optimistic on the manufacturers part. If I then add in the load from the hot end, stepper motors, fan (not much) it all starts to look like it's going south quickly.

Dieter

Ormerod #257

As I understand it the ATX PSU is effectively running "open loop" as the opto-isolated feedback path to the transformer chopper driver is taken from the 5V/3.3V rail. Not having had the time to trace out the PSU I'm not sure, but reprap have had to artificially load those rails to get the PSU to play ball. Maybe an active dummy load could be substituted that takes its cue from the saggy 12V and adjusts the loading (power MOSFET on heatsink) to maintain a constant 12V supply?

Edit: I have started a new topic to explore this possibility.

Edited 1 time(s). Last edit at 01/14/2014 05:35PM by Radian.

---

RS Components Reprap Ormerod No. 481

Radian,

You have a very good point on the opto feedback and it just further illustrates that these power supplies are just not up to the required standard in this application. If I were to run this power supply on a PC (mine will kill it) with a high end ATI or Geforce graphics card in it, it would be a disaster!!! Those cards all require 12V ATX plugged directly on to them as some are rated at 500W+ and really drag on the 12V line, the very reason why people (me being one of them) put well reviewed and tested 750W and larger power supplies into their machines that cost $150+, voltage stability under heavy load.

Dieter

Ormerod #257