darrins

Unfortunately, all the really good books are also big books. Try "The Art of Electronics" by Paul Horowitz and Winfield Hill. Or "Practical Electronics for Inventors" by Paul Scherz.

Ante, Just to clarify, I know I'm not controlling the phase angle of the triac. I'm simply turning the triac on and off for short periods of time. More heat needed -- longer "ON" time (and vice versa). I suppose the low frequency is needed to ensure you get some zero crossing with the AC. As I recall, the reason PWM was recommended in the first place is that the phase control technique has EMI problems because of large current spikes. Right?

I did that once when I was really tired. I put my oscilloscope probes across the AC mains on a power supply and POOF, one of my probes got shortened by 2mm. Also scortched a nearby cap. At least the fuse did what it was supposed to do. My scope was unharmed.

audioguru, This is how I got confused in the first place. I want to control the phase angle to the triac so I can controll how much power is delivered to my heater. The first circuit I built uses a heat sensor, a comparator, and a standard light dimmer switch. When my temp sensor is below the setpoint, the inverting comparator turns on a mechanical relay to my dimmer switch. This bang-bang controller has poor characteristics, e.g. overshoot, ocsillations, etc. I'd like the amount of power to the heater to be proportional to the difference b/n my heat sensor and set point. This could be done by rotating the dimmer switch knob, but obviously I'd like an automated way of doing this. This is where I thought the PWM and optoisolator triac driver came into play. No? If there is some other way to control the phase angle (other than hand-turning a pot), I'm open to suggestions. Thanks. Darrin

Thanks Ante and audioguru. I will try experimenting with a few different designs, including the one Ante simulated. I think I will also try stuffing a few of my temperature sensors (LM34) in a copper tube and use a summing amplifier to calculate an average temperature. A lot of the temp control circuits I see use only one sensor. A body of water can have a large temperature gradient (as can large pieces of metal, etc.). Well, I better hurry and post this. We are having a big storm here in Texas and the power keeps going out. Too bad I already dismantled my UPS. :-\ Have a nice weekend. :) Darrin

Ante, Looks like that site has been shut down. :( Darrin

Ante and audioguru, Just found another interesting circuit. This one is from http://home.cogeco.ca/~rpaisley4/LM555.html#7 This circuit looks like it might be a little easier to predict and the pulse width can be between 5% and 95% of nominal.

Ante and audioguru, The voltage difference between my heat sensor and reference will be, at most, around 300 mA. I've read that, as a rule-of-thumb, that the control voltage at pin 5 should be between 45% and 90% of Vcc. Obviously, I'll have to amplify my signal if I'm going to use a Vcc of say 5V. Given the limitations that audioguru described (with respect to the 555's PWM capability), will this still be adequate to drive my optoisolated tric driver? ---- I have another dumb question. The triac that I bought came with a mica washer. It is my understanding that the metal part that connects to the heat sink is LIVE :o. I've read that the mica washer should be installed between the metal part and the heat sink, along with some "thermal grease". What is this thermal grease and can I use silicon compound instead (or is silicone compound the same thing)? Looking at the inside of a light dimmer, I noticed that its triac had only a non-metalic washer (no grease). I've heard that there are silicone rubber washers that can replace the mica/grease. This seems a lot less messy. Can these washers be found at most electronics stores? There is supposed to be a full moon tonight, so maybe I should hold off on building this circuit. ;) As always, thanks for the help. Darrin

Ante, Cool. It worked!!! 8) This will help me greatly. :D Modeling and simulation are great time savers (not to mention money). Thanks again. Darrin

Ante, I downloaded the shareware version of Proteus (Proteus 6 Lite). Would it allow me to simulate this circuit? I don't know how to use the program, but I'm sure I could learn. Heck, I used SPICE in college (pre-Windows days). Is it possible for Proteus lite to load the model you have created? If so, is there a way for you to upload the model to this forum (if you're willing)? Thanks. Darrin

I found a circuit that is very close to what I am trying to do: This is from http://www.saburchill.com/tech/electronics/elect039.html This circuit uses a sound sensor to vary the brightness of a lightbulb. I also figured out some timing formulas based on threshold voltage (Vth) and triggering voltage (Vtr). As you know, normally, Vth = 2/3Vcc and Vtr = 1/3 Vcc. These values result in the "standard" timing formulas. Applying the control voltage, Vc, to pin 5 directly affects these values. It is applied directly to the 2/3 point on the voltage divider. 3 equal-valued resistors make up this voltage divider. The reference voltage for the threshold comparator is b/n the first and second resistor (hence 2/3Vcc) and the ref voltage for the trigger comparater is b/n the 2nd and 3rd resistor (hence 1/3Vcc). If I can figure out how the control voltage affects these reference voltages, the following formulas should work. In monostable mode: (pulse width) T = [ln(Vcc) - ln(Vcc-Vth)] * R * C In astable mode: T1 = -ln[1 - (Vth - Vtr)/(Vcc - Vtr)] * (R1 + R2) * C T2 = [ln(Vth) - ln(Vtr)] * R2 * C (pulse width) T = T1 + T2 (frequency) f = 1 / (T1 + T2) (duty cycle) % = T1 / (T1 + T2) ------- I plugged in values Vth = 2/3Vcc and Vtr = 1/3Vcc and got the same values that would result from using the "standard" formulas (discounting some roundoff errors). Thanks again. :) Darrin

I found a website that describes using the 555 as a voltage controlled oscillator. http://www.analog.com/library/analogDialogue/archives/29-3/single.html At the bottom of the page, they have a formula for output frequency. Charge time, Tc = (R5 + R6)*C5 * ln [ (1-Vin/6.8V) / (1-Vin/3.4V) ] Discharge time, Td = 0.69 * R6*C5 Frequency, f = 1 / (Tc + Td) --------------- In the formula, I'm not exactly sure where the 6.8V and 3.4V come from. These values may be specific to their circuit. Anyway, the formula is nonlinear function of control voltage. I'm very sceptical of this formula, (at least the charge time) because when the control voltage is 0, the formula should reduce to Tc = Tc = (R5 + R6)*C5 ln(2). Any thoughts? Darrin

Thanks for the advice guys. I read in one of my textbooks that the control voltage affects the reference voltages of the IC's internal comparators (normally 2/3 and 1/3 of Vcc), but it doesn't say exactly what this effect is. Here's what is said about pin 5 on Tony van Roon's tutorial page: "Pin 5 (Control Voltage): This pin allows direct access to the 2/3 V+ voltage-divider point, the reference level for the upper comparator. It also allows indirect access to the lower comparator, as there is a 2:1 divider (R8 - R9) from this point to the lower-comparator reference input, Q13. Use of this terminal is the option of the user, but it does allow extreme flexibility by permitting modification of the timing period, resetting of the comparator, etc. When the 555 timer is used in a voltage-controlled mode, its voltage-controlled operation ranges from about 1 volt less than V+ down to within 2 volts of ground (although this is not guaranteed). Voltages can be safely applied outside these limits, but they should be confined within the limits of V+ and ground for reliability. By applying a voltage to this pin, it is possible to vary the timing of the device independently of the RC network. The control voltage may be varied from 45 to 90% of the Vcc in the monostable mode, making it possible to control the width of the output pulse independently of RC. When it is used in the astable mode, the control voltage can be varied from 1.7V to the full Vcc. Varying the voltage in the astable mode will produce a frequency modulated (FM) output. In the event the control-voltage pin is not used, it is recommended that it be bypassed, to ground, with a capacitor of about 0.01uF (10nF) for immunity to noise, since it is a comparator input. This fact is not obvious in many 555 circuits since I have seen many circuits with 'no-pin-5' connected to anything, but this is the proper procedure. The small ceramic cap may eliminate false triggering." I may take a look at the 555 internals and see if I can derive timing formulas that include the control voltage. If I figure it out, I'll post it. It seems like it would be useful information. After all, the "standard formulas" are what I use to select "reasonable" values of resistors and caps. Otherwise, I'm selecting them by trial and error. I want to use PWM to send pulses to an optoisolated triac driver to control a heater (you guys have already answered many of my questions on this topic). I have a reference voltage set by a potentiometer and another voltage from a heat sensor. If there is a large differnence b/n these 2 voltages, I want to sent longer pulses (or more frequent pulses). If there is a small difference, I want to send shorter pulses (or less frequent pulses). If my heat sensor has a higher voltage than reference, I don't want to send any pulses (or at least only short, infrequent pulses). I don't necessarily have to use a 555 if there is a better option. Thanks for the feedback. This group is quickly becoming my best source of information. :) Darrin

I read several datasheets and application notes on 555 timers, but I can't seem to find the information I'm looking for. Specifically, I'd like to use the 555 as a pulse width modulator by using the control voltage (pin 5). The datasheets indicate that in monostable mode, the output pulse width can be varied by modulating the voltage at pin 5 (V5). They also indicate that in astable mode, modulating V5 will modulate pulse position. What I'm wondering is this: What is the relationship between the control voltage, V5, and the values of the timing resistors and caps? There are standard formulas for timing based on these resistors and caps. How are those formulas affected by the introduction of V5? Thanks. Darrin

I think those batteries are toast. I'll send them someplace to be recycled. There's still a lot of other useful parts in the UPS, so it wasn't a total waste. It has a gigantic transformer in it. I don't know what I'll use that for -- maybe a boat anchor. Cheers. Darrin

Ante, The batteries are connected in series. The voltage across both batteries measured about 9V. The "bubbly" one measured 7V and the other one measured 2V. Something's definitely wrong there. I would expect them to be more balanced. I'll hook everything back up and see if one of them gets hot -- I expect maybe the one carrying more load. Thanks Darrin

At work the other day, we smelled something burning in our server room. The culprit -- an old uninterruptible power supply (UPS). It's more cost effective for us to just throw these things away and buy new ones than try to repair the faulty unit. One man's trash is another man's treasure, so I decided to bring this UPS home and see what I could do with it. This thing weighs about 50 lbs and now I know why. It has a gigantic transformer and a couple of large 12 volt lead acid batteries. I was thinking of differnent things I could do with the batteries and I thought about making my 3 year old daughter a little electric car, kind of like a golf cart. I was wondering if these batteries would provide enough power and what kind of life they would have. Any thoughts from the group? I figured I could salvage the rest of the UPS as a battery charger. I haven't yet figured out what part of the UPS was getting overheated, but one of the battieries has a couple of bulges (about 1 inch in diameter and upraised about ¼ inch). Is there any danger in continuing to use this battery? Thanks. Darrin

I don't mean to beat this into the ground, but I'm very curious. What would happen if the pulses were sent to the triac driver such that they never crossed the zero intercept of the AC signal? In other words, what would be the output of the driver if the pulses were entirely contained within the AC waveform? Excuse the crude drawing (below). . . . . . . . .......... . . | | . . | | . ===================== zero volts ========= Thanks. Darrin

audioguru, Do you have a schematic for your LM317 regulated battery charger? I built my variable power supply using a LM317 (1.5 to 30V). Very similar in design to this one: http://electronickits.com/kit/complete/powe/ck402.pdf Perhaps my design could be modified to accomodate the battery charger? Thanks. Darrin

A few questions about this supply. It has 2 primaries and 6 secondaries. Where do you find transformers like this? And why the need for all the secondaries? A theoretical question: If you have a single set of AC leads split between two transformers, T1 and T2, will a load attached to T1 affect a circuit attached to T2? Assuming the same scenario (T1 and T2), would a load connected directly to the AC affect T1 and T2 circuits? Thanks for posting this. I eventually want to make a custom power supply for my project. For now, I'm just using some 9V batteries as I discovered I was spending more time designing a power supply than designing my project. Cheers. :) Darrin

I think I may have found my "poor man's" solution. I'll just use my temperature sensor and comparator circuit to control ON/OFF of the dimmer switch and adjust the pot on the dimmer until I'm happy with the response characteristics. I may try my hand at controlling it with PWM to the phototriac driver when I learn more. I'm sure it's somewhat similar to motor speed control using a 555 timer, but I'm still not clear on how the pulses get synchronized with the AC frequency. Thanks again for all the tips. Darrin

I guess what I'm really building is a solid state relay for my heater, like this one from Tony van Roon's tutorial on relays: From: http://www.uoguelph.ca/~antoon/gadgets/relays/relays.html In my application, the heating element would be the load, instead of the light bulb. Correct me if I'm wrong, but the components C1 and R5 are a snubber circuit to prevent false triggering of the triac, correct? It is also my understanding that the phase angle of the triac output is controlled by the resistors R3 and R4. Is that correct? (the R4 in the figure is labeled 2k2 -- 22K?) Or is the phase controlled by the MOC? I'm confused? ??? I can certainly build this circuit as is and experiement with it. It's my guess that using a pot for R4 would allow me to vary the phase, but it's just a guess. I certainly have appreciated your input. Hope I'm not flooding this board with my dumb questions. Darrin

Ok. I'm starting to feel really stupid. I just want to use the triac so I can limit the power to my heater (via phase control). In other words, I probably won't be varying the phase angle while my heater is ON. I'm building a VERY crude bang-bang controller. When my voltage from my temperature sensor is too low (comparator will tell me this), I just want to turn on the heater until the voltage reaches the temperature set point. I'm not sure I understand how to use the NAND gate, since I'm dealing with an analog signal. Can I just run pin 2 of the MOC to ground? Thanks again. Darrin

Oops. That last post didn't enter the entire url correctly. Here's the image.

I'm can't post an image of my circuit (I don't have a website), but I can post a web page of something very similar. Instead of using a thermisor to measure temperature, I'm using an IC (LM34), which outputs 10 mV/deg F and is much more linear than the thermistor. I don't know how this type of circuit would interface with the NAND gate on the triac driver IC. Any suggestions? Thanks again. Darrin