Home built demo board problem

[Frei]

Hello!

I'm using a MC56F8365 unit for a motor control project. I decided to build my simple, self built demo board according to the 56F8367 evaluation module schematic. The first version had some serious problems because of high current consumption (200mA without any extra peripherals) and had a serious overheat on the voltage regulator . I decided to build a more simple test circuit, connecting only the microcontroller with these part: all the Vdd, Vss pins, OCR_DIS on ground, Vosc on Vdd, CLKMODE on ground, all Vcap pin connected by the proper capacitor, reset with the 74AC00 circuit, 8MHz Crystal with 1M resistor parallel to it...  To generate the 3.3V voltage, I'm using an LD1117-3.3 chip. When I power up the circuit, it consumes around 150mA and the voltage generator(LD1117) is getting seriously overheating( it's maximum current is 800mA)... so I must turn it off... Now I found out that when I disconnect the 8MHz crystal, or disconnect only the parallel 1M resistor, or keep pushing the reset button, the current consumption is falling down to 50mA and the voltage regulator is working properly... It is a very strange symptom.... The circuit is correct, there are no wrong pinout-pin connections, short circuits... The schematic is simple. I've tried another microcontroller, but it is from the same sample package... It seems to me when the clock circuit is working a short circuit comes off inside the microcontroller... 

So what may be the mistake? Did I forget something? Is it normal? Are the microcontrollers refused? I've checked the soldering by x-ray. The controller was never programmed. Please help me!

Thanks in advance:

                                              Laszlo 

[kef]

Datasheet mentions 150mA Ivdd_io @ 60MHz. I'm not sure what should be default speed with 8MHz oscilator and PLL not yet tuned up to 60MHz, but your current figures don't sound abnormal.

You didn't mention what is supply voltage to LT1117 regulator. It it was 5V, then heat to disspate would be (5V-3.3V) * 150mA = 0.255W. With junction to ambient termal resistance of about 50C/W (it depends on PCB construction, etc), LT1117 would be just +12C warmer than ambient. Are you supplying it from 12V? If that's the case, then no wonder it is hot. (12V-3.3V)*0.15A * 50C/W = +65C more hot than ambient! I think you should find linear regulator with lower thermal resistances and maybe one, you could attach some heatsing to. Also you may consider switched mode regulator.

[Frei]

Yes, but my circuit is free of any peripherals.. no rs232, can, led's... nothing. Datasheet mentions a maximum of 50mA for normal operation. But it consumes 120-150mA... Why? LT1117's supply voltage is 9V, so the junction thermal must be around +42C... If I 'm using 6V supply, the current on the controller is around 116mA.

[kef]

Maybe I don't understand your device datasheet, but see table 10-7 Current Consumption per power pin (should be pins I think). RUN1_MAC case. Total device currnet is Idd_io + Idd_adc + Iodd_osc_pll = 207.5mA. Conditions, RUN1_MAC was measured, are listed. Idd_io current includes Processor Core current, internal voltage regulator current and IO buffers current. IO current can be bigger, because Idd_io is specified when no pins are being switched or statically loaded. Of course you observe lower currents if PLL is not engaged and if device clock is low.

Using 9V regulator supply, temperature of LT1117 should be not around +42C, but around ambient plus 42C. That's well over 60C when operating in the "living room".

[Frei]

Yes you're right, it is interesting how much current it consumes officially... I also counted with the ambient + room temp, but it was even higher... So I decided to lower down the supply voltage to 6V and now it's temperature is fine, but the current consumption is still that high.

But there is another interesting thing: on the 56F8367 evaluation board, there is no heatsink  for the regulators even if it has a 12V power supply... How could it work for them?

http://icbank.com/data/ICBShop/product/mouser/images/freescale/lrg/MC56F8367EVMIMG1.jpg

Now I tested my circuit with parallel jtag programmer but "surprisingly" it's not working. On windows xp, ecp mode parallel port, using 56836x_flash.cfg config file, css connenction type, etc it tells me that Device is not powered up or any connection error... It will be a long week...

[kef]

If you look at sheet 13 of EVM schematic, you may notice that there is not single but 3 linear regulators. Common one that reduces 7-12V DC/AC to 5V, and two others that are reducing 5V to 3.3V. So the same heat to dissipate is radiated not from single but from 3 devices.

If you compare LT1117CM  to MC33269DT, you may notice that MC device has lower junction to case (metal tab) thermal resistance.

Maybe multilayer PCB has solid planes, which helps reducing junction to ambient thermal resistance. etc

I have no idea why your device isn't working, maybe it still consumes to much current. Try comparing your schematic to EVM.

. http://cache.freescale.com/files/dsp/doc/user_guide/MC56F8367EVMUM.pdf

[Frei]

Maybe... Even if it uses 3 devices the common one reducing the voltage must get overheated without heatsink... There must be some tricks...

I'm using the schematic you linked in... Is there any trick or procedure to test the parallel programmer functioning the right way? I connected the 74HC244 to 5V supply so the computer must be able to use it.

[Frei]

Some updates:

I reassembled the circuit with the schematic containing only the LVC244. Now I can download my code and works well but the current consumption is very high again. I constructed a simple serial port "hello world" demo code. It works fine but I must switch off the board after a minute because the DSC is getting overheated. Consumption is about 200mA. If i turn on the board without the DSC it consumes only 3-4mA so the problem is around the processor. If I break the clock signal by disconnecting the quartz or VddOSC, the consumption falls down to 80mA. It is also the same situation when programming the circuit or leaving in debugger mode (only 60-80mA). The high consumption is a normal mode peculiarity depending on the clock working or not.  There is a ~100mA fault consumption. It is the same situation when I change the DSC. The board is as simple as can.  Can anyone help me please?