T&L Publications

Well, there's a hole in one of my ICs.

My brother and I built a dual motor controller. We're trying to add a heater to the vw. He's rigged up tubing that goes around the exhaust pipes into a device that fits behind the back seat. It runs fluid through the tubing and then through an old oil cooler. A fan blows through the oil cooler and should (in theory) heat the passenger compartment.

I used the 555 based PWM controller from the Nov Nuts & Volts. I modified it to use a 556 so that we could control two motors separately (the fan and the pump).

My brother hooked it up today. Here is what he said:

Hey, it was all working for about a min, then it started smoking, then i turned it off. I took it apart to look, and the timing IC had melted partially, and there was a little chunk taken off of it

So, here is the original schematic:
<img src="http://img363.imageshack.us/img363/7945 ... 6ug.th.jpg" alt=" - " />

Here is how the pins on the 555 (left column) corelate to pins on the 556:
<img src="http://img363.imageshack.us/img363/3186 ... 9vk.th.jpg" alt=" - " />

And here is my finished diagram:
<img src="http://img363.imageshack.us/img363/8003 ... 5cj.th.jpg" alt=" - " />

The transistors are T1P42 pnps. The diodes are 1N4148s. The potentiometers are 100K linears. The resistor connected to the base of the pnp is a 1k ohm. The resistors connected in series with the potentiometer are 6.8k ohm. The capacitors are 1uF non-polarized electrolytics.

The pump is rated at 3amp, the fan doesn't have a rating but it's not too high powered. I estimate 2 amps or less (likely far less).

The circuit is fused at 10amp and didn't blow. There's a relay connected to a switch that turns the whole circuit on and off. The circuit is built on a perf board and crammed into a hobby box.

Any idea why the smoke?

<small>[ November 22, 2005, 08:10 PM: Message edited by: Matt Nuzum ]</small>

Absolute max supply voltage for 555/556 is 15V. Car battery voltage while charging is 14.5V, or a bit higher. Any spike at all on supply line will go over 15V, and blow the IC.

You've got the diodes in parallel with transistors, instead of the motors. This is okay for protecting the transistors, but dumps spikes from motors right into the supply.

I suggest a 500-1000 Ohm resistor from +12 to a 10V or 12V zener, paralleled with 10uF to common. Power the 556 from that filtered power. The transistor emitters remain connected to "raw" +12V. The reduced 556 supply voltage may prevent the transistors from turning-off completely. In that case, a low power NPN between 556 and PNP power transistor will be needed.

Cheers,

added: Change that 500-1000 Ohm resistor to maybe 150 to 220 Ohms. I just realized the 555 spec sheet I have is for CMOS version. Check the spec sheet for the 556 you are using and calculate correct values.

<small>[ November 23, 2005, 03:00 AM: Message edited by: Dale Y ]</small>

The schematic was a little hard for me to read, but I think the pot wiper does not have a resistor in series. Pin 7 of the 555 should have a minimum of 1K to Vcc. I would put a 1K between the pot wiper and the connection to Vcc.
I concur with Dale Y on the maximum voltage for the 555. But if you lower the voltage to the 556 it may not drive the TIP42 transistors. I think you are connecting the motor from collector to ground and connecting the emitters to + battery.
I suggest you use a NPN to drive the motors, connect the motor between the collector and + battery, and connect the emitter to ground. Then use a 7805 for power for the 556. Another important item is a 10uf electrolytic connected from pin Vcc pin of the 556 to ground, physically close to the IC.

It seems to me the 555 was handling too much current. I don't think brief supply spikes would do that.

I agree that the diagram has a possibility of connecting the discharge pin to +12V with a very small amount of resistance. Could be the problem.

I think the PNP as a high side driver is OK. I tend to use NPNs as low side drivers but that's just my bias (pun semi intended). I'd put the diode across the motor. by the way, did the diode burn out? What does your motor pull? the 4148 may be a little too lightweight. It's If max is around 200 mA, pulsed is somewhat higher, check the datasheet. a good emf kickback could take it out.

Well the heavyweights are on your problem, but I just wanted to ask - did you test the circuit before installing it in the car? Maybe hook up some dummy loads, and possibly scope the PWM output ?

Or are you like me and said nah, that takes all the fun out of it.

We tested with LEDs and it worked great before we put it in the car. We were fortunate because we made one connection improperly, plus we had the heat sinks hooked up together which is apparently a no-no on PNPs.

We're doing a custom board for this now. The perfboard was pretty ugly looking and we had to do a lot of disassembly to get the broken (literally) 556 out.

The advise of K7 regarding pot. minimum resisistance is a valid point. I doubt that a small supply overvoltage killed your timer, As the 15V max. spec. is where the manufacturer will no longer guarantee proper operation --this is not the point where meltdown occurs. I had mistakenly received an order of one hundred 74 HCxx gates last year (only ordered 10 and only was charged for 10). This gave me an opportunity to play around with this family of chips. One of the things I did was grab 10 at random and push their supply voltage to the max. None of them blew out below 20V .This is three times their max. supply voltage. However performance specs suffered long before this. My point being that its doubtful a minor increase in supply voltage would have caused your problem. Transients -yes , especially negative going. The advice you got on changing the transistors over to NPN is well worthwhile and would eliminate cutoff problems when you drop your IC supply voltage. Maybe also adding a small LC filter at the supply input.

P.S. If you build your circuits the same way you draw them --I can see why it failed.

[quote]Originally posted by ROBERT REED:
P.S. If you build your circuits the same way you draw them --I can see why it failed.

Now you know why I complained about my difficulties with Eagle/circuit software. I can draw it by hand very fast (maybe I should slow down?), but it takes me 45 min to draw it in a very simple, closed program such as Express PCB. In Eagle I don't have the patience to complete it yet. I'm working on it though...

I'm putting resistors in line with the wiper on the pots.

Also, it looks like one of the diodes was split in half. It's hard to tell if that was because of the disassembly or if it happened because of the overload. Does it sound reasonable for a diode to be split right in half?

Also, I did have a 0.1uF capacitor connected from + to Gnd right next to the chip. Do you suggest higher? What is the disadvantage of using a higher value cap there?

Every capacitor value/type has a series resonant frequency at which they will perform best at bypassing /decoupling. This of course assumes short leads and the area of interest is quite broad. As a general rule using 0.1, 0.01 and 10 MF all in parrelel should cover all bases (up to 100 mhz). In your circuits 0.1 mf should be sufficient for the rise times encountered.Its also good practice to install a 10 MF capacitor where B+ enters the circuit. The higher the value, the more efficient it becomes for removing lower and lower frequencies on the feed line. Ive seen diodes break apart from overload and heat stress but its usually accompanied by other signs such as scorching, etc. not to mention smoking. I use the Express PCB program and have grown to love it for its simplicity, but that wasn't always so. Early on it took me forever to do even the simplest schematics, mainly due to my mistakes and illiteracy of the program. Now, after many weeks of frustration and learning and several thousand cusswords I can breeze thru complex prints in a short time. As I mentioned simplicity--this is important when you aren't on it day in and day out so that in occasional use you can come up to speed quickly.

<small>[ November 23, 2005, 09:08 PM: Message edited by: ROBERT REED ]</small>

Originally posted by Matt Nuzum:
...
Now you know why I complained about my difficulties with Eagle/circuit software. I can draw it by hand very fast (maybe I should slow down?), but it takes me 45 min to draw it in a very simple, closed program such as Express PCB. In Eagle I don't have the patience to complete it yet. I'm working on it though...

wow. it doesn't take me very long to do a schematic in eagle. though I find the 555 to be one of the harder chips to draw a reasonable schematic for.

I'm putting resistors in line with the wiper on the pots.

Also, it looks like one of the diodes was split in half. It's hard to tell if that was because of the disassembly or if it happened because of the overload. Does it sound reasonable for a diode to be split right in half?

sure, if it's been subjected to significant over voltage or over current...

Also, I did have a 0.1uF capacitor connected from + to Gnd right next to the chip. Do you suggest higher? What is the disadvantage of using a higher value cap there?

I doubt that you need that much bypassing for this application - its just not that sensitive. In general some false triggering won't make much of a difference anyway. It won't hurt to add more caps but I doubt you will see much of a difference.

An undocumented feature of the 555 is its output stage. It is basically a totem pole arrangement like normal TTL except that it is designed to deliver several times as much current. It may also use slower transistors than TTL and there is no clamp diode to keep them out of saturation.

The result (like TTL) is that there is a short period of time during output transistions when both transistors are turned on. This results in a brief spike of current passsing through the device from the positive supply to ground. This can cause a negative spike on the positive supply. If other circuits are using the same supply, this can cause a fair amount of false triggering and other mayhem.

I would normally use a .1 uF directly at the package itself and something aroud a 10 uF electrolytic in the immediate vicinity.

If you are operating this off of the car battery, you will need to do a fair amount of filtering and limiting of the power supply to protect the circuit. A vehicle power source can have surges in the 50 or 60 volt range and transients of several hundred volts. These hve to be guarded against in your circuit. Motorola (now ON Semiconductor) used to make a form of heavy duty zener diode for this purpose.

A 10V spike ABOVE the 14.5V car supply could cause an emitter-collector short in a transistor inside the 556, allowing almost unlimited current to flow until melt down and/or package blowout.

With TIP42s the heatsink tab is also the collector. With two TIP42s on a shared heatsink, either transistor turned on, both motors should run. If heatsink is grounded... poof supply to ground short when either transistor turns on. To fix this use transistor mounting kits (film or mica insulator w/heatsink grease and non-conductive shoulder washer on screw).

If the motors use the motor frame for ground return connection, you've got no choise but PNP for main driver.

Cheers,

<small>[ November 28, 2005, 03:49 AM: Message edited by: Dale Y ]</small>

You know. Eagle can be overkill if you just want to draw a quick schematic.

Might I recommend SPLAN from ABACOM?


http://www.abacom-online.de/uk/html/body_splan.html

Creating your own shapes is a breeze...