MAX7219 Current draw with digits not showing

[premelec]

I have an 8 digit LED unit from Ebay and the Mysterious Orient which has what seems very strange current draw... e. g. if I blank each digit and then insert 1 2 3 4 5 6 7 8 the current is about 200 ma not changing much with digits added [with low set PWM - 56K R current set] - then the mystery is that as I blank each digit in sequence the current goes very low with the first blanked digit. Is this normal or is this a bum chip? I couldn't find any actual description of the circuitry inside the MAX7219 chip... just block diagrams... Thanks... [PS the unit seems to obey commands correctly].

PPS I tried pulling both [socketed] LED clusters off board and the current is the same without them so it's something in the chip...

ITEM: http://www.ebay.com/itm/171241714572?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

 

[AllyCat]

Hi,

PPS I tried pulling both [socketed] LED clusters off board and the current is the same without them so it's something in the chip...

So still 200mA with a 5 volt supply? That suggests 1 watt dissipation in the chip. Is it getting (very) hot (finger tips are probably an adequate sensor)?

If hot, it's probably faulty (or there's a PCB/assembly fault). If not, you need to try to find where the power is going, or check the I/V measurements.

Cheers, Alan.

 

[westaust55]

There have been a number of past threads about counterfeit MAX7219 chips sourced from the same region you mention.
For example, See here:
http://www.picaxeforum.co.uk/showthread.php?16812

 

[hippy]

if I blank each digit and then insert 1 2 3 4 5 6 7 8 the current is about 200 ma not changing much with digits added [with low set PWM - 56K R current set] - then the mystery is that as I blank each digit in sequence the current goes very low with the first blanked digit. Is this normal or is this a bum chip?

It does seem odd that you get such different current measurements when adding digits over blanks and when turning digits into blanks. It is possibly an issue of how you are measuring the current and/or could be an issue of what data is being displayed and what is being blanked. I would suggest displaying all 8's with DP for testing.

The 7219 seems a little odd / clever in its multiplexing; the datasheet says with an RSET of ~10K, 40mA per segment but total typical chip draw of 330mA. That suggests it's more than one segment on at a time, probably eight, plus 10mA draw for the logic. Depending on which combination of segments you could see 40mA+10mA at any time or 8x40mA+10mA.

Looking at the "SEGMENT DRIVER OUTPUT CURRENT vs. OUTPUT VOLTAGE" graph in the datasheet with an RSET of 56K a draw of 200mA could be in the right ballpark.

Perhaps the only way of knowing if it's a dodgy chip or not is to compare with one which is known not to be.

 

[John West]

Premelec, I think I have some 7219's somewhere. I'll try to drop by with one for you to try. I'm sure the chip is the problem, as there's no logical reason for the behavior you described. You pay too much attention to detail to screw up wiring like I do.

 

[premelec]

Thanks for the replies... The data sheet says that the displays are turned to V+ when off implying that under certain conditions the chip might rely on the current regulation to turn off the drive by shorting the LED - and turn off the actual drive at other times. From the data sheet for Dig0-Dig7:

"Eight-Digit Drive Lines that sink current from the display common cathode. The MAX7219 pulls the digit outputs to V+ when turned off."

Anyhow clearly something like that is happening in some instances and not in others - It's particularly perplexing that when - after setting blanking on all digits - adding a single digit display raises the current which stays up as all digits are loaded - AND the current drops immediately when ONE digit is blanked though all the others are still showing.... there's a clue there in a certain symmetry of behavior - current up first digit loaded, current down when first digit of all show is blanked... [OK I haven't tested all possibilities just LOAD digit 0 and BLANK digit 7 in my test runs].

PS Westaust55 - the chip on this unit looks good - etched, top full pins etc... haven't got my x-ray microscope going however Thanks for the ref to Martin's comments.

 

[Hydroid]

I have an 8 digit LED unit from Ebay and the Mysterious Orient which has what seems very strange current draw... e. g. if I blank each digit and then insert 1 2 3 4 5 6 7 8 the current is about 200 ma not changing much with digits added [with low set PWM - 56K R current set] - then the mystery is that as I blank each digit in sequence the current goes very low with the first blanked digit. Is this normal or is this a bum chip? I couldn't find any actual description of the circuitry inside the MAX7219 chip... just block diagrams... Thanks... [PS the unit seems to obey commands correctly].

PPS I tried pulling both [socketed] LED clusters off board and the current is the same without them so it's something in the chip...

ITEM: http://www.ebay.com/itm/171241714572?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

I made myself a MAX7219 demo board - consisting of a MAX7219, two 4X7 CC LED displays, and fixed support components (10uf. 0.1uf caps and a fixed 12K resistor), so I can't match your exact setup.

But, for comparison, and extra data, this is what I measured by putting my Fluke in mA current mode on the supply line to the MAX chip:

Note: All measurements are made with the chip's intensity register ($A) set to '8' and 'Shutdown register ($C) set to normal (1).

Digits Displayed and Current drawn:

None = 6.1 mA
1 = 10.8 mA
1 & 2 = 21.4 mA
1 thru 3 = 32 mA
1 thru 4 = 40.6 mA
1 thru 5 = 52.7 mA
1 thru 6 = 66.8 mA
1 thru 7 = 74.1 mA
1 thru 8 = 89.7 mA

and with all digits displaying 8 and the decimal point (all segments lit), current was 135.4 mA

Hope this was useful.

Regards, John.

Edit: Just realized I never actually answered what you were asking, so...

If I start with digits 1 thru 8 displayed, current is again 89.7 mA. If I then blank digit 1, current dropped slightly to 85.4 mA - not a big drop like you were seeing....

 

[premelec]

@Hydroid... thanks for that info - no where near what I've got - and my intensity is set at "2" - I'll look into the Shutdown register being specifically set - see if that makes a difference... still not clear if bum chip or inept controls settings...

To wrap up this matter - Hydroid tested my code, 11 Apr 2014, on his chip and it does not show any unexplained high current as my unit does - my unit is suitable for non-battery supply use but definitely would kill batteries fast... evidently a circuit error in counterfeit MAX7219 part supplied...

22 Jun 2014 - 2 units tested:
I found a partial fix is to issue a digit off command to the position before it is written with the correct value - however issuing a standby command makes the current go up [200ma] with the digital info held ok when standby terminated and I've not found a way to prevent this - these chips are definitely defective! They seem to be the majority of 8 digit LED units listed on ebay... China, Hong Kong.

7 July 2014 - I got a larger digit unit with plug in MAX7219 kit unit from China for a few bucks more and it shows NO current draw problems - however... It addresses the digits in reverse order! 01234567 rather than 76543210. Tis can be fixed in software easily, but it was a surprise... I don't know if this is some Chinese convention or just PCB designer decided 12345678 was right [right digit is high in addressing for this unit].