 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Microcontroller and PC projects : Overheating Pic32
Page 1 of 2   |
|||||
| Author | Message | ||||
| shoebuckle Senior Member  Joined: 21/01/2012 Location: AustraliaPosts: 189 |
Hi, The Pic32 in my B&W Maximite is overheating when running MMBasic 4.0. Is anyone else having this problem? This doesn't happen with previous versions of MMBasic. The symptoms are that after a short time running a program (a matter of seconds) the VGA text starts to shimmer (moves by about 1 pixel). It is still readable but very fuzzy. When sitting at the prompt, not running a program, the shimmer goes away. It makes no difference if power is supplied from a plug-pack with more than adequate capacity or via USB. There are no external connections other than PS2 keyboard, VGA monitor and plug-pack and it doesn't matter if I have it mounted in the Altronics-supplied case or not. I used a freezing can to show that the problem is definitely in the Pic processor and not any other component. The shimmer starts when the Pic32 case reaches 40 deg C measured with a non-contact IR thermometer. Below that temp, there is no shimmer. The Pic32MX7xx spec gives an ambient temp range -40 to +105C and junction temp -40 to +140C so it appears to be running within spec. The kit was supplied with the Pic already mounted on the board. Any suggestions would be very helpful. Cheers, Hugh |
||||
| isochronic Guru  Joined: 21/01/2012 Location: AustraliaPosts: 689 |
I would check some other things as well. Although the prompt seems idle the controller is still running a large throughput, so the heating is similar I think. 1) what is the temperature of the actual controller ? 2) Is the Vcap capacitor a factor ? Some ceramics have a high variation with temperature, voltage, phase of the moon, .. 3) How was it soldered ? Some of the surface mount literature suggests pcb's must be heated while soldering eg wave or batch soldering, the idea being that differential expansion causes problems, usually ok in small packages but this is larger. Try heating the pcb a little ? 4) Is anything else heating things ? [In general I have found the Microchip controllers astonishingly robust.] |
||||
donmck Guru Joined: 09/06/2011 Location: AustraliaPosts: 1313 |
Have a look at the thread starting at around: http://tinyurl.com/8duyova Don... https://www.dontronics.com |
||||
| ajkw Senior Member Joined: 29/06/2011 Location: AustraliaPosts: 290 |
Shoebuckle, My Altronics MM with 4.0 suffers the same problem. It is straight out a heat issue in the Pic. I have tried different caps without any change/improvement however a spray with freezer spray direct to the Pic once the shimmering starts fixes it very quickly (well for a little while anyway!). Cheers, Anthony. |
||||
| paceman Guru Joined: 07/10/2011 Location: AustraliaPosts: 1329 |
Hugh, Did you see Mick's post yesterday: http://www.thebackshed.com/forum/forum_posts.asp?TID=5240&PN =1 Your problem seems to be PIC temperature related but maybe if the capacitor's a problem it'd make the PIC more sensitive to it. Greg |
||||
| shoebuckle Senior Member Joined: 21/01/2012 Location: AustraliaPosts: 189 |
Thanks Don. Looks like the only solution found there was to add a heat sink to the Pic32 and as someone pointed out, that's just a band-aid. I suspect the problem will only get worse with later versions of MMBasic as Geoff works the processor harder. Cheers, Hugh |
||||
| shoebuckle Senior Member Joined: 21/01/2012 Location: AustraliaPosts: 189 |
Thanks Greg, I just had a look at that post and I have already tried an extra 10µF across both C1 and C5 without success. Cheers, Hugh |
||||
Gizmo Admin Group  Joined: 05/06/2004 Location: AustraliaPosts: 5078 |
I have a maximite running all day every day, its doing a fair bit of work, and it does get warm, but not too hot that I cant put my finger on the Pic32. The regulators have heatsinks fitted, and these get a bit to hot for comfort, I might look at making them bigger. I'm running 3.2. I agree we may need to look at ways to fit a little heatsink to the chips. The problem is mounting it. I dont think we need a big heatsink, something the size of a thumb nail with fins would be enough, like you would fit to a TO-220 regulator thats not working to hard. I'm wondering about the heat transfer capabilities of superglue? Is there a heat transfer compound thats also a glue? Glenn The best time to plant a tree was twenty years ago, the second best time is right now. JAQ |
||||
djuqa Guru Joined: 23/11/2011 Location: AustraliaPosts: 447 |
small heatsink and fan often there is a suitable one on old vga cards VK4MU MicroController Units |
||||
| shoebuckle Senior Member Joined: 21/01/2012 Location: AustraliaPosts: 189 |
Thanks for all your suggestions. Geoff has been very patient with me over the last couple of months trying to solve this problem but even he has given up. An interesting thing emerged when I was testing for over heating with a freezer can and a thermometer. The Pic32 actually runs hotter by about 2 or 3 degrees with MMBasic 3.x than with 4.0, yet it is 4.0 that seems heat sensitive. This makes me think that a) there is nothing wrong with my Pic32 b) the soldering of the Pic32 is ok c) something else is causing the problem. Geoff suggested replacing the two regulators even though they don't appear to be heat sensitive. I haven't tried that yet but I did go over them both with a soldering iron to check that there isn't a dry joint. I will try replacing them next. I am also beginning to suspect the type of decoupling capacitors supplied with the kit as I had trouble with the SD card resetting the processor whenever an SD card was plugged in. The solution was to add a 22µF across C9. The decoupling capacitors are all a grey retangular shape with 104J100 printid on the top. C13 is the same type with 334J100 on the top. In the Silicon Chip article, all these capacitors are round yellow ones except for C13, which is the same as mine. Cheers, Hugh |
||||
| donmck Guru Joined: 09/06/2011 Location: AustraliaPosts: 1313 |
I know Don Recardo had this problem with V3.2, and his immediate solution was: 
However he was ordering some self adhesive heat sinks. Just wondering how you got on Don? Sure since V2.7, each new version of MaxiMite has it drawing more current, as Geoff has indicated. This appears to be the major culprit, so a simple cheap ebay solution would be nice. Super glue? Don't know. Anyone experiencing these problems game enough to try? I believe it must be a small percentage of PIC32s that become operationally marginal at the higher current loads. Don... EDIT: Hugh, just noticed your mention of V3.x drawing more current than V4.x, so my blanket statement of later versions drawing more current is possibly incorrect. https://www.dontronics.com |
||||
| shoebuckle Senior Member Joined: 21/01/2012 Location: AustraliaPosts: 189 |
Yes, this is strange as Geoff believes that 4.0 should draw more current than 3.x and should therefore run hotter. I wonder what is going on in my MM. I haven't measured current; only Pic case temperature. Shimmering starts at 40ºC consistently. Hugh |
||||
| isochronic Guru Joined: 21/01/2012 Location: AustraliaPosts: 689 |
Just a hunch.. The observation is that the jitter occurs in MMB 4 a short time after extra/different processing starts. The SC article mentions the video and its sync is set up by code in an interrupt routine that is setup/triggered by a timer, and the interrupt code had to be expanded for MM4 colour. So it points to a jittery sync or video delay, and thus a jittery interrupt completion (??). If the prompt never jitters when idling (even when hot) it means the timer/interrupt is completing ok - I wondered if the timer was running off a RC oscillator which might change timing when hot (?), presumably it is crystal-locked though. So, that implies there is an interference in interrupt processing from the additional tasks - maybe the extra task is using more interrupt processing which clashes with the video timing ? Maybe it is both ? Does the jitter happen when there is no program screen output ? Pretty complex to sort out. (edit) The SC article says the start of the horizontal sync, triggers the video line loading into ram by interrupt code. Then at the end of the horizontal sync pulse the SPI's send the data out. A 31.xx kHZ timer is used. If a RC-based timer on the chip has been used to end the sync, possibly it is drifting a little with temperature/voltage, enough to end the pulse before the interrupt(s) have finished? |
||||
| paceman Guru Joined: 07/10/2011 Location: AustraliaPosts: 1329 |
I've used the silver paste heat transfer compound on my Notebook processor's heatsink and it worked well - much better than the silicone variety. Not sure if it's supposed to be a glue, but it dries out and holds on well - certainly well enough for a little TO220 heatsink which may be all that's needed (for the band-aid approach). Rockby have the silver paste for $5.40 a tube (stock no. 34274) and similarly the TO220 heatsink, less than a dollar (stock no. 36255). My Maximite is an Altronics one (one of the second delivery ones that took months to arrive!) and I've never had this problem either with any of the MMBasics including V3.2 & V4. Also the PIC32 chip only ever gets just warm - the "hottest" component is the 5v reg and even that is only "quite warm" but I did put a TO220 heatsink on that right from the beginning. Greg |
||||
bigmik Guru Joined: 20/06/2011 Location: AustraliaPosts: 2914 |
Hi All, I am wondering if there is something else the problem here.. Maybe we are looking at the Symptom and not the cause. Possibly, the code that is used to generate the video has a minor bug.. After all every screen has to be right shifted about 2 characters to get it into the `field of vision'. Has someone who has TV knowledge actually looked at the video output signal to check the timings, especially on a bad sync MM? It might be that the timing is `just' acceptable for a Display to lock on to and that the temperature (or other environmental variable) causes the slight drift to fall outside the range of `lock on' stability. This is an area outside my knowledge but I am certain there are a few TV Techies out there. Regards, Mick Mick's uMite Stuff can be found >>> HERE (Kindly hosted by Dontronics) <<< |
||||
| aargee Senior Member Joined: 21/08/2008 Location: AustraliaPosts: 255 |
Forty degrees is not the problem, that is well within specs for the chip. I agree with Mick, you're looking at a symptom - of sorts. The PIC may just be faulty or has suffered heat/electro-static stress whilst being put on the board during construction. It could even be a close margin or bug in the silicon - not unheard of. - Rob For crying out loud, all I wanted to do was flash this blasted LED. |
||||
| isochronic Guru Joined: 21/01/2012 Location: AustraliaPosts: 689 |
I suggest, trying heating the pic artificially slightly (while the idle prompt is the only thing running) and seeing the result. If the prompt then starts to jitter then it is a temperature effect. If it doesn't jitter then it looks like the fault is caused by something from the extra processing. A nearby soldering iron or radiator should heat things a little, the non-contact IR thermometers can be a bit imprecise at low temperatures though. (It sounds like it only needs a few degrees, watch it !!! I take no responsibility !!!). If it derives from the extra processing then, I'd try something without any screen updates and see if if starts the jitter, then screen outputs, etc, .. |
||||
| shoebuckle Senior Member Joined: 21/01/2012 Location: AustraliaPosts: 189 |
The saga continues, but with a strange twist. I loaded Geoff's MMBasic v4.1 yesterday and the VGA shimmer has gone completely... steady as a rock. So I reverted to v4.0 which was the first version that shimmered and that too is now steady. So what is happening? I haven't used the MM for a while and have made no changes so it has to be something external. The only difference I can see is the warmer weather. I will have to get out the freezer can again and see if there is a component which doesn't like winter. The only other possibility is that there was some other external electro-magnetic interference which has since gone away, but I don't know of any under my control. I am just thankful that I can read the text again without my eyes going all funny. Cheers, Hugh |
||||
| ajkw Senior Member Joined: 29/06/2011 Location: AustraliaPosts: 290 |
Hugh, I have had an ongoing issue with this also. I have also found that different versions behave differently, in particular 4.0B being rock solid. Geoff has been unable to shed any light on why the different versions shimmer or not but I am wondering now about these NOP's he says may move code to different areas of the pic's memory, refer ColorMax Pre-order thread. If the NOP's change the memory optimization does other code changes do so also?? Other things I have found/done in relation to this. Reinstating the components in the audio out circuits to original helps. I resoldered the PIC to eliminate any potential dry solder joints. This seemed to help in the short term but now I have lost my USB probably due to the mess I made, it is shimming now again anyway. Mucking with the capacitors did nothing. Freezer spray helps but it is short term until the PIC heats again. I have not yet tried 4.1 as I have lost USB and cannot update it now. :-( It seems some firmware versions are effected by/susceptible to a external source but why?? Is it the memory allocation?? Now that I have a version bricked MM I am tossing the coin on replacing it with another MM or CMM. Cheers, Anthony. |
||||
| bigmik Guru Joined: 20/06/2011 Location: AustraliaPosts: 2914 |
Gday Anthony, Where are you in Aus? If near Melb Nth West subs I am happy to look at it for you (or if you wish to pay postage both ways) .. I have a partially built MM I can complete and offer for a cheapish price but personally I think the money might be better invested in a Colour MM.. but I suppose that depends on your needs. Mick Mick's uMite Stuff can be found >>> HERE (Kindly hosted by Dontronics) <<< |
||||
| Page 1 of 2 |
|||||
 Print this page |
