Jump to content

Headache


esaj

Recommended Posts

  • 2 weeks later...

Not much progress with the welder for now, although the controller & software shouldn't need much work, just haven't gotten around to design & mill the board. In the meantime, I did take some boards apart from a broken old tablet (just a dead battery, but it was a Motorola Xoom from somewhere like 2011, so really not worth trying to find a replacement) and an old laptop (circa 2008, abandoned to me after I saved the data from the hard drive, already taken the battery apart for the cells a long time ago ;)). I expected that more modern computers (like... made in this century) have next to no usable parts in their boards, especially laptops, ditto for tablets in general.

Turns out I was wrong, actually I got a good amount of 1206 10µF caps, tantalums and more heavy duty-SMD -mosfets out from those, plus a couple of thermal sensors, bunch of ceramic SMD power inductors and even a couple of ultra LDO (150mV dropout) voltage regs. Not that 1206 caps are expensive (something like 1 +- some fractions cent/piece at 100 pieces or more), but since they were there... Rest was mostly either too small to be useful, BGAs (Ball Grid Array) or too special purpose to be of use for me.

AlOvmOF.jpg

Close up on some part of the boards. The large metallic looking (it's actually ceramic, I found out) part on the left portion is an SMD power inductor, part with marking "76" upside down is a diode. The black square is a BGA-chip, rest are SMD-caps (brownish) in various sizes from 1206 down to 0201 and likely some ferrites/small inductors (black ones), although one next to the BGA might be a resistor too (they don't put markings even on the resistors at that small size).

I've mentioned the sizes that shrink towards dust speckles many times before, but once again, to give you some perspective:

3qTjAfv.jpg

That's five SMD-caps side-by-side, starting from 1206 (3.2mm x 1.6mm or 0.12" x 0.06"), then going down to 0805, 0603, 0402 and finally 0201 (0.6mm x 0.3mm or  0.02" x 0.01"). That's where the names come from 1206 = 0.12" x 0.06", 0201 = 0.02" x 0.01". Similarly, in metric-sizes (it seems usually imperial sizes are used for components, even in metric countries), 1206 is known as 3216 and 0201 is known as 0603. What's confusing is that the naming partially overlaps (metric: 3216, 2012, 1608, 1005, 0603 etc, imperial: 1206, 0805, 0603, 0402, 0201) which could sometimes lead to problems (imagine thinking you're ordering 0603 in imperial-size and get it in metric... :P). Probably that's why most people stick to the imperial size codes. Anyway, as you can see, it gets really small really fast. There's even one step smaller size (01005), but didn't have any of those in these boards (in general, they're not common, as even many automatic pick'n'place -machines can't handle that small). Here's a 0201 stuck on my sharp tweezers with the left over flux over the component:

LNczWTf.png

Now imagine that 01005 is quarter of that size (half the width, half the depth). And the set next to an XT60 to give more scale:

UBE64kW.png

I think it'll be a good while before I'm going down to even 0603's, not to mention the smaller sizes :D I think I could still mill boards for 0603's without the CNC tolerances / cutting quality becoming a problem, probably even 0402's (they're 1.0 x 0.5mm, and I often use 0.5mm traces when I need to), but lower than that, it might get tricky if there are burrs on the cuts and not to get shorts while soldering (no solder masks, remember ;)).

I still have a broken older desktop mainboard lying around in the closet, might pick that apart some evening, lately I've been mostly just reading through Art of Electronics (a forever project, the damn thing's something like 1100+ pages, and slow to read as you really need to concentrate to grasp everything, although a very good book...).

Edited by esaj
  • Upvote 3
Link to comment
Share on other sites

2 hours ago, esaj said:

I've mentioned the sizes that shrink towards dust speckles many times before, but once again, to give you some perspective:

Soo small! Great perspective!

2 hours ago, esaj said:

as you really need to concentrate to grasp everything, although a very good book...).

Ugh! :sleep1:I have a hard enough time keeping up with journals and reading about Trump!:)

  • Upvote 2
Link to comment
Share on other sites

  • 2 weeks later...

@esaj Where the hell are you? 

My first thought was you are riding the hell out of your KS16S but then it came to mind that you may be preoccupied working on the third HIE-Isolde cryomodule or involved in the connection of the GBAR experiment to ELENA’s new antiproton deceleration ring or maybe just adding the final touches to the ICARUS modules at the CERN Neutrino Platform.  :) :smartass:

Miss your dialog buddy!

  • Upvote 3
Link to comment
Share on other sites

17 hours ago, Rehab1 said:

@esaj Where the hell are you? 

My first thought was you are riding the hell out of your KS16S but then it came to mind that you may be preoccupied working on the third HIE-Isolde cryomodule or involved in the connection of the GBAR experiment to ELENA’s new antiproton deceleration ring or maybe just adding the final touches to the ICARUS modules at the CERN Neutrino Platform.  :) :smartass:

Miss your dialog buddy!

Not that much riding (shitty weather, like most of the summer, or just too tired) or that much projects... For the past few weeks, I've been on a pretty perpetual work-eat-sleep -cycle, and mostly just laying on the couch on my spare time and at the weekends. Low on energy and interest on doing pretty much anything bigger, although I've got a few tidbits done here and there, just haven't gotten around to write anything about them. Off the top of my head:

  •  I finished up the software for a simple frequency counter/-meter + pulse counter, it's not attached to the wooden rack face plate yet, as I don't want to put the displays there until I got more stuff and cutting more holes requires full disassembly each time
    • Should be good from some millihertz up to and above 40MHz,which really isn't that much for a counter, but I didn't even have anything that would give large enough amplitude to try it above 40MHz
  • Made a simple shift-register breakout for breadboarding (not that it's hard to wire up one or two on breadboard, but it's faster to have a ready-made board you can just connect and go, plus it's got its own led-bar display for seeing the states on-the-fly ;))
  • Made a new ad-hoc continuity meter straight to a dot-matrix board, after I managed to somehow destroy my old one
    • I've got multimeters, but this one's "always on" and I don't have to deal with the multimeters' running out of batteries or turning off automatically after a while when soldering
  • Looked a bit into the "Blue Pill" STM32F103C8T6 -boards (same chips as what drive at least the King Songs and Gotways, incidentally ;)) and programming those
    • I turned one of the boards into a "Black Magic Probe", a kind of a chip debugger / programmer that works with JTAG and SWD (poor man's STLink?), and used that to program a couple of others (testing STM32Duino-core and such).
    • If you want to do the same with the STM32F103C8T6-Blue Pills, do note that the current Black Magic Probe doesn't fit into C8T6 -chips, you either need CBT6 -chip or modify the BMP code, or use the disabled upper 64K on the C8 & risk that it hits some faulty memory locations (C8T6 = CBT6 with the upper 64K of Flash disabled, because one or more bits are faulty in the upper region, but it's possible to write into there)
  • Last Sunday I was supposed to be doing something useful but somehow ended up building a simple wooden frame for a robot car with belt-drive on the back and simple steering mechanism in front, didn't come out that good as I pretty much planned it as I went, but I guess it could be used for trying stupid ideas
    • No controller or servo attached for steering yet, I just slapped a stepper-motor for the belt-drive (only motor I had in suitable size that isn't designated for something else already)

Might sound like I'd actually gotten a lot done, but most of these were something like a 40 minutes to some hours of work here and there, over the span of 3+ weeks or so, which to me seems very little compared to how much time I've usually spent on this stuff. Still no controller for the welder :P I've gotten some second thoughts whether I really want to use it with the lithium-cells (some horror stories about burning/exploding cells when welding them with home-made equipment :D), but I've got some other uses in mind for that, once I manage to actually get started on finishing it.

Oh, also enrolled to an embedded systems design & development -course at the applied sciences uni, supposed to start next week on Friday... hopefully I'm less lethargic by then, and got my sleep rhythm in check, as it's early in the morning ;)

Edited by esaj
  • Upvote 3
Link to comment
Share on other sites

5 hours ago, esaj said:

Oh, also enrolled to an embedded systems design & development -course at the applied sciences uni, supposed to start next week on Friday... hopefully I'm less lethargic by then, and got my sleep rhythm in check, as it's early in the morning ;)

I've been a bit tired lately too.  I don't know what it is.  This summer for us has been sweltering hot and pretty nice (except for the forest fires next province to us smoking the skies up).  I find though that going for a ride on the wheel really helps get the energy going again.   There's something about moving and balancing around at medium speeds which revitalizes the spirit.  :w00t2:  That plus getting enough sleep, light exposure, and drinking enough water (not coffee!) helps.  Actually I should start drinking coffee as I'm so sleepy most of the time at work.  :efef1285c5:  I think I need to get a new wheel or do something different as lately it's been blahsville for some reason.

Glad to hear that you signed up for a course.  I wish I had some time to learn some new things at university.  After working for a while and doing the same old, same old stuff it gets a little boring.  :sleep1:

  • Upvote 2
Link to comment
Share on other sites

6 hours ago, esaj said:

Not that much

Are you kidding! I just read your long list of accomplishments!

6 hours ago, esaj said:

Low on energy and interest on doing pretty much anything bigger, although I've got a few tidbits done here and there, just haven't gotten around to write anything about them. 

We all have those low energy days and weeks!   Winter weather with zero sunlight really affects me and most others I work with. 

6 hours ago, esaj said:

Oh, also enrolled to an embedded systems design & development -course at the applied sciences uni, supposed to start next week on Friday... hopefully I'm less lethargic by then, and got my sleep rhythm in check, as it's early in the morning ;)

Awesome!

Yes sleep is so important! I never could fathom how you functioned on such little sleep. 

  • Upvote 2
Link to comment
Share on other sites

54 minutes ago, Hunka Hunka Burning Love said:

Actually I should start drinking coffee as I'm so sleepy most of the time at work.  :efef1285c5:  I think I need to get a new wheel or do something different as lately it's been blahsville for some reason.

Have you tried 8 hour energy? My god that little drink not only increases your energy but also your concentration. A new wheel will also give you a boost buddy! 

  • Upvote 1
Link to comment
Share on other sites

  • 4 weeks later...

Looking to buy a 100V+ power supply... If I want to get more into building things that work with the wheel (and power themselves from the batteries or measure the high voltages), I "need" (read: want) a good power supply that can go up high enough in voltage to match 20S batteries (84V) and above (to "simulate" possible overvoltage conditions during regenerative braking) and limit the current/voltage if things go wrong. The idea of using actual batteries for testing doesn't really appeal to me, as testing with different voltages would mean charging / discharging the packs in-between and I don't want to find out that whatever contraption I've made shorts/burns up and possibly takes the battery BMS, cells and the house with it ;) 

So far, the model I've had in mind has been HP6634[A or B] (100V / 1A, 100W linear supply), but haven't found a suitably cheap unit:

SgLabs_HP_6634A_1.JPG

 

Most of the cheaper units (around 200+€ + shipping) in eBay seem to be either in the USA or Israel, which would mean having to pay import taxes + VAT, which pushes the price of unit + shipping up a good ~30%. And the shipping from US is usually >100€ alone. There were two auctions in the German eBay with the rack-mounting brackets from the same seller, which ended about an hour ago, last I checked the prices were something like 71€ and 81€ (+ shipping), but the seller had marked that he won't ship internationally, so I didn't bid. Maybe I should have :unsure: Not sure if the seller actually sells them at that price, compared to the "usual" prices, those seemed really cheap. 

Another option I've found is TDK-Lambda GEN100-15 (100V / 15A, 1500W switching mode supply stuffed into a 1U case), at least the output power wouldn't be an issue though, and a new one would be around 450€ + shipping and used for about 340€ + shipping, but it would be a bit of a compromise, as it's already a bit pricy (well, not really when talking quality rack power supplies) and doesn't have numeric keypad on the front panel for programming the values, that I like so much on the HP's:

SgLabs_Lambda_GEN100-15_1.JPG

Also I don't think switchers can also act also as electronic loads in general, the linear HP663X's can do that, but it's not a strict requirement.

If anyone knows a place selling second-hand rack power supplies with reasonable prices, please point me there :P So far I've found at least places like Rosenkranz ( http://www.rosenkranz-elektronik.de ) and http://www.used-line.com (through which I've found even more European companies selling used/refurbished units, but nothing really cheap so far). The prices for companies selling second-hand lab equipment are of course usually slightly higher than internet auctions, because they make their money by buying used equipment and selling it forwards, but then again, they also usually test (and maybe even calibrate) the devices, maybe even give a short warranty. Not really in a rush, as I've got my hands full with other stuff right now, but I'll keep looking. Also if anyone knows other good robust options from other reputable manufacturers (like HP/Agilent/Keysight, Lambda, Keithley, Tektronix etc.), please let me know. Basically my requirements are:

  • 100V or more max output voltage (DC), linear or switching, adjustable
  • 100W or more
  • Must work with 230V / 50Hz mains (single phase)
  • Constant current and constant voltage modes, automatic changing
  • 19" rack mounting
  • Must be adjustable from the front panel, preferably by a keypad, but knobs are ok too
  • USB/RS-xxx/GPIB/whatever control is a plus, but not necessary
  • One output is enough, more is a plus but not necessary

I'd like all that in fully working condition for about 200-250€ max, maybe 300-350 with shipping, preferably less and inside EU (so I don't have to pay import costs / VAT ;)). Finding high quality power supplies for cheap might be a bit much, but one can dream... :P  I did buy that fully working HP6632A (20V / 5A, 100W) for 130€ including shipping though, so it's not entirely impossible. But like said, I'm not in a rush either, maybe I'll find one in some auction later on...

Edited by esaj
  • Like 2
Link to comment
Share on other sites

  • 1 month later...

:blink:  Your work reminds me of being stuck in an Escape Room trying to work the puzzles out and not having a "Hint Button" to rely on.  Glad to see you delving into the deep doodoo stuff,  but I am guessing you will be pulling down the average vacation time off for Finlanders in general.  ;)

  • Like 1
  • Upvote 1
Link to comment
Share on other sites

Fine-tuned the motor control a bit, it's still not too good, tends to overshoot and brakes too late when getting close... but enough to showcase it a bit tomorrow. Youtube video editor has been retired, I'd suggest turning down the volume (or mute it), as the sounds are just scraping and motors whining, and it's annoying  ;)

The range of a single IR-LED on the breadboard I'm holding isn't very much, the battery's pretty empty... also the big tires block the middle sensors to certain directions, but this is just a prototype.

I ordered the fabricated boards yesterday, hopefully they'll arrive soonish... this thing's got a deadline in about 4 weeks, including full report on everything (calculations, schematics, test results of everything...). I still made a few tweaks on the boards (mostly just added more connectors, moved around existing connectors a bit), since 10 boards costs as much as 5. Figured I could use the rest as more or less general-purpose -sensor platforms. If I had more time and patience, I could have made them much more compact and sensible using the second layer, now it's mostly the same layout as in the single-board version.

 

ZJYznvG.png

a2foujr.png

 

 

 

 

 

Edited by esaj
  • Like 2
  • Upvote 1
Link to comment
Share on other sites

15 hours ago, Hunka Hunka Burning Love said:

Looking good!  When's the self-balancing tech going in?  :popcorn:  And when can I bounce the IR off my frig so it will bring me a cola?  :w00t2:

No self-balancing in these, just dragging itself along with two motors, although technically, it would be possible to make it self-balancing by putting the tires in front, adding a gyro/accelerometer and controlling the motors more precisely ;)

The general idea of the project is to show the concept of swarm robotics / swarm AI behavior, that is, a group of robots acting together. Or as Wikipedia ( https://en.wikipedia.org/wiki/Swarm_robotics ) puts it:

Swarm robotics is an approach to the coordination of multirobot systems which consist of large numbers of mostly simple physical robots. It is supposed that a desired collective behavior emerges from the interactions between the robots and interactions of robots with the environment. This approach emerged on the field of artificial swarm intelligence, as well as the biological studies of insects, ants and other fields in nature, where swarm behaviour occurs.

We chose a very simple goal, to make the robots move to the one in the group that has highest light reading from an LDR (light-dependent resistor). Sounds simple, but the one that has the highest reading might be behind other robots (or simple obstacles, like a corner, except our bots have no other "vision" than IR, so they wouldn't see a wall for example and might try to drive through it  :D) or too far away for the faint IR-signal to reach to the farthest ones, so all the robots do not know where it is located physically. Instead, they must rely on the ones that do see the target robot, and use their positions to track a path to the "master", and follow that path to reach their goal (so the other robots in the way act as "beacons" to guide the ones further away). I guess this simple concept has pretty much zero real-world use though :P

There have been many similar projects before, some of which show very complex behavior patterns through (relatively) simple actions of single robots, like Swarmanoids. I guess these could fetch you a cola from the fridge:  ;)

 

These are autonomous robots, no human control involved. The footbots forming the linked path for the others to follow is pretty much what our swarm bots should be doing.

Edited by esaj
  • Like 3
Link to comment
Share on other sites

  • 2 weeks later...

A momentary digression, if I might...

This has got to be the longest running headache ever! Have you considered that it might be a tumour?  

Edited by The Fat Unicyclist
  • Like 2
  • Upvote 1
Link to comment
Share on other sites

  • 2 weeks later...

Aaand... done. I don't actually expect anyone, even the teacher of the course, to read through this in its entirety. I'm still not 100% happy with the report, as I had to leave out things due to time restraints, and would have had to make more measurements, but it's starting to rival my thesis in size and I'm out of time :P

If you're really interested in the subject (66 pages, about 14MB because of all the pictures, although they had to be compressed down a lot and suffer in quality):

https://github.com/esaj/SwarmRobot/raw/master/Documentation/FinalReport.pdf

Probably not  ;)

Edited by esaj
  • Like 2
Link to comment
Share on other sites

  • 5 weeks later...

Not much has happened since returning the final report. While changing from remote working to actually having to go to an office every day has been a more welcome change, the downside is that it leaves me with lot less free time, and after the final squeeze on the course, I didn't want to do much anything for a couple of weeks. I did work on a couple of pedal designs (and actually do have some simulations made for a simple pedal a friend wanted me to make) around the holidays and last night finally managed to work out the fault and fix the CNC motor speed controller. Not that it was that hard, as expected, the mosfet controlling the motor speed had failed, but getting the controller out, ensuring that the problem was with the mosfet and replacing it, adding protections to prevent it in the future and then putting the entire machine back together seemed more like a chore, although it didn't take but a few hours  ;) 

I already mentioned the bulk of the issue in another topic, so I'll just copy-paste it here:  

A couple of months back, I destroyed my DIY motor driver in my CNC by static electricity. I milled, drilled & cut more than 50 different PCB designs, some more than one piece, on that same controller without any problems.

When it failed, I was changing the bit, which I do a minimum of 2 times (if there are no holes to drill in the PCB) per board, usually more times (at least 1 for milling, 1 for cut out, and the amount of different sized drill holes) and always in a similar way (two spanners to release & tighten the ER-11 collet). When I touched a spanner onto the collet of the motor, suddenly the motor started running at full speed. The viable explanations are that it was a high voltage static discharge causing either an avalanche breakdown or high dV/dT (voltage change over time) -spike, do note that the part was connected to the drain of an N-channel mosfet, not the gate (which can only withstand +-25V difference to source, usually +-20V, but this mosfet is a bit "unusual" in that sense) or source. Both the gate and the drain-source -channel were destroyed (20 ohm resistance from drain to source, 36 ohms from gate to source), leaving the mosfet open at all times. Just last night I replaced the broken mosfet and the controller is running happily again. Also added some transient voltage suppressor diodes to prevent this from ever happening again.

 

The original circuitry already had some protections in place:

33CGv9H.png

The mosfet is a Fairchild FQP30N06, rated for 60V VDS , 21.3-30A continuous and 120A pulse current, and +-25V gate voltage. There's a TVS-diode across the motor connector (don't remember the voltage though and it's not marked in the schematic, 30V?), as well as 10A -rated freewheeling diode for negative spikes, but still the fet was destroyed by a static discharge, likely as the voltage against ground went to very high value. The 15V zener on the gate-side has a forward resistor so it wouldn't get destroyed, but that's a design fault. There shouldn't be a resistor in front of the zener, as it will allow the gate voltage to stay higher than the zener reverse drop, essentially NOT protecting the gate from transients. In this case the voltage spike occurred at the drain, coming from the motor connector (P_SP_PWR_OUT1) and reading through the mosfet failures, the likely candidate is a high dV/dt -spike that destroyed the gate insulation, leaving the mosfet open, regardless of whether there was any voltage at the gate or not:

The cause of this failure is a very high voltage, very fast transient spike (which may be positive or negative going). If such a spike gets onto the drain of a MOSFET, it gets coupled through the MOSFETs internal capacitance to the gate. If enough energy gets coupled, the voltage on the gate rises above the maximum allowable level – and the MOSFET dies instantaneously. The process less than a nano-second! The initial spike destroys the gate-body insulation, so that the gate is connected to the body. Once that has happened, the MOSFET explodes in a cloud of flame and black smoke. We have one documented case where the battery wire worked loose, causing a spark. It must have been this that caused the gate breakdown for the explosion of flame and smoke did not happen until the battery wire was re-connected some time later! Which demonstrates how very difficult cause and effect can be to connect!

In my case, the mosfet didn't explode (there were no outside signs of damage), but was definitely destroyed.

To make the controller more robust against such incident in the future, I've added a 30V TVS-diode (unidirectional, as I didn't have any bi-directionals above 20V, but it won't matter much there) in parallel with the fet, and another 15V bidirectional TVS going from the gate to source (well, in this case, ground). I should now be able to get back on track with making boards, but I still haven't designed any since the robot. Currently, I'm looking into making some ATX-breakouts (for using computer ATX-power supplies in more projects), a DC load (mostly just an issue of finding a suitably large heat sink, the circuitry itself is very straightforward) and prototyping the pedal for a friend.  In other unfinished projects, there's (still) the charging circuitry and controller for the spot-welder, and the robot software... I'm getting sort of cold feet with the idea of spot-welding lithium-cells with homemade equipment, and right now writing software outside of work doesn't seem that tempting  ;)  And there's always the time vs. energy -stuff getting in the way, the downside on working with electronics- and software-hobby projects is that (at least for me) it's hard to get anything done in so small parts that I could only do an hour here and there, instead I'd need at least 4-6 hours straight,  preferably an entire day without distractions, and only after a good night sleep (either of which I haven't had lately). Trying to do any designing or writing code after a full work day is usually impossible, although software engineering is just "sitting around and typing", it's pretty exhausting mentally, especially with tight deadlines.

 

 

Edited by esaj
  • Like 3
Link to comment
Share on other sites

3 hours ago, esaj said:

We get to pick our phones once every two years, so I waited over so I could order the Cat S60, a rugged mil-spec phone with dual processors, 3800mAh battery (which for EUC-people sounds ridiculously small, but is actually about 2-3 times larger than most phones), IP68-rating and...  FLIR thermal camera. The camera is the same spec as the Flir One (not Flir One Pro) -module sold sep

Your new phone has an appealing industrial appearance along with many positive reviews. I like the idea of a built in FLIR camera. FLIR does have apps available for smart phones but the ratings are 3 stars at most. 

 

24788519407_6639c776e9_b.jpg

 

3 hours ago, esaj said:

What the pictures show is our "genius" heating system, a true marvel of the 20th century: ceiling heating. I pushed the temperatures up to see them (although it might seem more extreme looking at the colors, the hottest spots are about 28 degrees Celcius / 82 Fahrenheit, if memory serves). Actually, in the above picture, there should be one more heating panel near the windows. Broken. There should be one in the first picture way at the back. Gone. Only way to replace them is to tear the entire ceiling open, but I'll think I just put on normal electric radiators on the walls when enough of them have come to the end of their lifespan.

I’ve never heard of installing heat panels in the ceiling. They are colorful. :) Not very intelligent unless that was the only option. We have forced air ducts in the ceiling at work but they serve for both heat and A/C. I can’t imagine having to remove the ceiling to perform routine maintenance on the system. 

  • Upvote 2
Link to comment
Share on other sites

10 hours ago, Rehab1 said:

Your new phone has an appealing industrial appearance along with many positive reviews. I like the idea of a built in FLIR camera. FLIR does have apps available for smart phones but the ratings are 3 stars at most. 

I haven't played around much with it, but other than the FLIR-camera, it seems plenty fast and the battery should last a good while between recharges, but I can't really tell yet, because I've only had it for a day. The industrial look is nice, and clearly it's directed to people like construction workers etc, as it should withstand a 6-foot drop to concrete, and 2m/7 feet submersion for a while (5m/16 feet up to one hour when the small switches in front are turned to the "5m"-position, which seal speaker and microphone etc. holes) as well as the military specification demands for dust/shock/etc. -proofness, not that I have tested dropping or submersing it, nor do I plan to. Figured it should last better than most phones at least ;)

 

Quote

I’ve never heard of installing heat panels in the ceiling. They are colorful. :) Not very intelligent unless that was the only option. We have forced air ducts in the ceiling at work but they serve for both heat and A/C. I can’t imagine having to remove the ceiling to perform routine maintenance on the system. 

They're not very common really. The upsides are the low cost vs. other types of heating (at least when installed during construction ;)) and they actually do work, most people think that heat only goes up, but these work by radiation, not convection (warm air does go "up", or actually above cooler/denser air/gas, radiated heat doesn't care about gravity or buoyancy or such). On a cold day you can actually feel that it's (somewhat) colder under tables and such, as the radiation comes from above. Illogically,  normal "radiators" actually work more through convection (warming air that then rises up and out from the radiator) rather than radiation, and that's why they're placed low on the walls.

They're completely maintenance free, until they break down. The element itself is just a "film" or a prebuilt-panel with heat resistor wire, but over time as the wires contract and expands with temperature, they'll break.

ceiling2.jpg

ecofilm_ceiling_floor_constr.jpg

Since the wood paneling or sheets or whatever is used on the ceiling surface material is usually nailed shut, the only ways to get to the panel are either by tearing down the ceiling or trying to reach it from above, but that's not that good here either really, as there's so much insulation in-between (2 x 50mm insulation sheets + plastic sheet to prevent humid air from rising out and condensing, and 400mm of rockwool on top of that) and the attic flooring joists run perpendicularly a good 40cm above them... Like said before, I doubt I'll even try to replace them, instead just switch to radiators once the rest of the elements start to die.

The panels that are broken might have been for who knows how long without anyone really noticing, they've been in place for over 30 years. The only place that's clearly colder (around 18...19 degrees C when it's cold) is the bedroom, as both heating panels there are gone and there's nothing else heating it up, except warm air coming through the forced ventilation system and through open door.

I was expecting to find at least a few dead panels, so no surprise here. I'd be a lot more worried if the heating wires inside the floor concrete slab would have been broken, as replacing those is hard (tear off all floor tiling, jackhammer the slab, remove old cables, install new ones...)

Edited by esaj
  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...