Power supply trouble with Button/LED matrix SEQ

[polpat]

Hi all,

this is my first post here.

My name is Patrizio and I have almost terminated the build of my midibox stepseq with full Button/LED matrix option (soon I will post some picture of it).

I am driving with a wall power supply 9V AC, 1A all the circuits consisting of 1 core, 3 + 1/4 DIN, 3 DOUT, 1 bankstick, 2 LCD 2x40 (but at the time only 1 connected), 64 duoled.

Tonight I have finished the cabling of the system and I attempted a first run to test the enitre duoled matrix to be ok without shortcircuits. Before I tested only half matrix (32 leds) each time during the build

I lighted up all the 64 duoleds in the matrix, I played a little with buttons and after 2-3 minutes working well with all led lighted...then the blackout!

I've checked the power supply outs: dead!

I took another wall PSU (this time 12V AC, 1A) to verify if anything worse happened to my system and all ok, and the midiseq ran smoothly.

I played again a little, I played a cycle with all led lighted on and....again the same purpose, the seq went off.

Again I've measured with my tester the PSU out..nothing, blown this also.  ???

Result: I never did an approximate calculation for total consumption of my system and before tonight for me 1A current for power consumption was enough but now I'm going to revise this idea.

I think to drive all the circuits in the pessimistic case of 64 led lighted on, more than 1A is required.

Have someone had the same problem as me?

Should I build an external power supply with at least 1.5A (or more) or better to leave a 1A PSU for the circuits and add a 2nd PSU only for leds and LCD?

Thank you for your help & tips.

Greetings from Italy,

Polpat

[stryd_one]

The answer is more Amps not more volts, but IIRC, the 7805 will only handle 1A - you may need to get a special 2A version. I'm no expert on this stuff, so don't blame me if you take my advice and blow it up ;)

[ris8_allo_zen0]

Ciao Patrizio!

(ti rispondo in inglese così altri possono beneficiare di queste info... sempre che siano giuste!)

Try to measure the current with your multimeter, if you get much more than 1A maybe there's something wrong with your hardware. I didn't find anyone in the forum with power issues with Duo-Led matrix (can you do a search too?), neither read anything about empowering the voltage regulator.

However it's better 9V 1A then 12V 1A, so the regulator has less work to do = less heat.

let me know, also I'm building a SEQ ;)

[polpat]

...sempre in inglese, è utile condividere con tutti!

Hi all!

Thank you for your prompt reply and your effort to help me!

Well... eh eh...the reason why I used a 12V PSU after the first blown supply it's because it was the only one left! :-)

I've downloaded the datasheet of the bicolor leds I purchased from ebay (here is the link, may be for someone is useful http://cgi.ebay.it/ws/eBayISAPI.dll?ViewItem&item=170099552887&sspagename=ADME:L:RTQ:IT:1).

Every led drive 20mA, so the calculation (if I'm not wrong) is pretty easy: 2 * 10^-2 * 64 = 1.28 A (not including circuits and display).

That's why my PSU are dead, I suppose... :-(

Now I think I have 2 possibilities: the first is a 2A power supply (or to be more safe, 2.5A) and change the 7805 with an adequate current consumption type (then more heat to dissipate and greater dissipator to use); the second is to use a single trafo 9+9V (1.5A each line) and build a PSU around it.

This last solution would let me to drive separately the circuits: one line is for core, dins, dout and lcds, the other is for leds.

In that way we have have less total current running in each regulator consequently less heat to dissipate.

Of course, the bad part is that I will have to build 2 circuit with upgraded 1.5A version of 7805 power regulator (and change the values for resistors and cons in that part of the circuit) to regulate things correctly and bypass this part of circuit in the core.

What do you think about? Is this idea too much to reach a simple result?

Thank you for your comments.

Cheers,

Polpat

P.s. Are you building an expanded seq like mine? If you need help (for what can I do) I am here.

[ris8_allo_zen0]

The better answer will come from who's successfully built a Duo-Led SEQ... that's not me :(

So at the moment you don't have any PSU suitable to power the Seq?

Some time ago I used the 5V line (red-black) of my PC's ATX power supply, which is usually able to source 25-30A. You can try this way just the time to measure the amount of current while "idle", i.e. with no led matrix on (or even with the matrix DOUTs disconnected!), and see if there's some current leakage somewhere.

P.s. Are you building an expanded seq like mine? If you need help (for what can I do) I am here.

Currently I'm building a non-expanded Seq. The only problem I'm having now is to find the knobs. Tried at ALBS but my order is below their minimum :(

[polpat]

P.s. Are you building an expanded seq like mine? If you need help (for what can I do) I am here.

Currently I'm building a non-expanded Seq. The only problem I'm having now is to find the knobs. Tried at ALBS but my order is below their minimum :(

[TK.]

LEDs are driven time multiplexed, in addition the weak 74HC595 output drivers limit the current, therefore your calculation is wrong.

I just did some measurements. When all steps are activated (all green LEDs on), and the sequencer is running (vertical line of red leds), the current is about 110 mA

Best Regards, Thorsten.

[ris8_allo_zen0]

I have some help for you about this....ebay again:

http://cgi.ebay.it/PKG50-Knob-for-Bose-Heavy-Metal-Hm-2-Pedal-Effects-BR14_W0QQitemZ300129341476QQihZ020QQcategoryZ10171QQcmdZViewItem

Lot of stuff there....

They're perfect! Thanks!!

As TK said, more than 1A is too much for a Seq. Check carefully (both visually and with a multimeter) that each pair of solderings in the PCBs are isolated from each other - sorry, it's quite late now to remember other troubleshooting ideas...

[polpat]

Hi Thorsten, hi ris8_allo_zen0,

thank you so much for your comments!

Thank you Thorsten for your exact current measurement, I tought more than 1A was too much for a system like that, 1A is a lot of current when we speak of digital....

My next move will be to resume my power supply bench (actually I don't use too much, but in that cases is really useful), my DMM and give a check for current driven in the system searching for some short circuit, also if I don't believe the problem is a short circuit (or...I won't to believe, after you will see pictures I will understand why.... :-) ).

The strange thing is that the seq is definitely ok: circuits works properly, bankstick loads, the only display connected for test respond well with a good backlighting, all leds toggle properly...the only trouble is that after some minutes the seq goes down.

Really confused about this!

But ok, I will measure current and only after I will make suppositions about the problem.

Bye...

Pat

[doc]

The most current is used by the backlights of the LCDs.

So please try to measure the current with and without backlight.

It's very strange that your PSUs broke while you're testing.

The normal behaviour of the 7805 would be to drop the voltage down if it's getting to hot or if the current rises over 1A. If the voltage drop goes down to 4V your CORE won't work again.

If you decide to take a 2A type regulator (78S05) a very good heatsink is a MUST!

The 78S05 is getting much more hot than the 7805. If you drive 1.5 A through a 78S05 a normal "finger" heatsink is not enough! In this case you'll have to mount the regualator (off-board) an a good heatsink and connect it via 3 cables to the core.

greets

Doc

[polpat]

Hi all, thank you doc also for your comments about my problem.

I have some news about my problem....a good news..probably I have solved! Or better... I suppose to have solved it, I go to explain why ;)

Today I did some measurement: I used my bank power supply (this time a serious one!), a handy tester to measure current and of course my seq.

I fixed the voltage to 9VDC (yes not AC as before), I passed through the tester and voilà: 380mA for 1 core, 3 and 1/4 DIN, 2 and 1/2 DOUT, 8 chips bankstick, 64 duoleds for the 4 tracks matrix, 16 leds for steps and 16 leds for other and only 1 2x40 LCD backlighted.

As TK said the duoled are multiplexed, so the total for 370-380 mA is absolutely reasonable (with one lcd connected).

I played for 20/30 minutes and it was all right, no problems about current nor system off or what else.

This time the only difference was the power supply: it furnished 9V DC not AC.

I am seriously thinking there is some problem with the bridge (may be defective or broken) in the core or the two wall adapters I used before were not so good as I thought.

I looked at the bridge and the code (if I remember correctly) is W06G, I suppose it's a 600mA type, I don't have the datasheet to verify.

The important thing however is the seq is now working!

Next step is to buy a DC wall adapter 9VDC, 1A and make the final try.

I could also remove the bridge but the core is fixed on the panel and digging on wires, so I don't want to disturb it :-)

Tomorrow I will connect the 2nd LCD and I will check again the current.

Thank you again to all for your help!

Greetings,

Pat

[midiboxxer]

Hi,

the W06G is a 1,5A type.

Greetz ::)