Jump to content
Electronics-Lab.com Community

indulis

Members
  • Posts

    488
  • Joined

  • Last visited

Everything posted by indulis

  1. A correction is needed to my explanation above (thanks to Zeppelin who caught it), heat sinks are rated in degrees C per watt, not he other way around. Heat sink and component surfaces are not perfectly flat or smooth. To maximize surface area contact, you want to fill all those imperfection with some thermally conductive material. It can be thermal grease, sil pads, gap fillers... etc. The point is to lower the thermal resistance as much as possible. In some cases (in the SMT world), ceramic substrates are used, or even copper clad aluminum instead of the "old" G10, FR4... etc.
  2. Yeah.......... OK 150 is a bit high. Our work de-rating guide lines call for max junction temp of 130
  3. A "google search" for heat sinks will show you all the different sizes and shapes. Heat sinks are rated in watts/degrees... in other words, a heat sink can dissipate a certain amount of watts and only increase in temperature so may degrees. Say you had a MOSFET with 10 mOhms on resistance. You push 100A thru it so the MOSFET dissipates (I^2*R) 100 watts. Now, say the MOSFET manufacturer says the the thermal resistance is 2 degrees/watt junction to tab (we'll say it a TO-220 case), so that means the the tab will rise 200 degrees (that's above ambient, and this hot is a bad thing). But they also tell you if you have the device on a heat sink rated xx degrees/watt the thermal resistance is now .5 degrees so the temperature now will rise only 50 degrees above ambient (this is better and a good thing). In general you don't want junction temperatures to go above 150
  4. Heat sinks is better lumped into Physics than electronics. Basicaly there are 3 ways to transfer heat, conduction, convection and radiation. The most effective of these is conduction. Metal/s, conducts heat a lot better than air. The magic is in in the amount of surface area exposed to the air (that's why you have all those fin's...
  5. If he in fact has "voltage regulation", then Q1 can not be on, as it would pull U2's output low, and with U2's output low, there is no output voltage.
  6. You might try looking at the data sheets and app's notes for the Linear Tech LT3486 and LT3474 for ideas. They seem to control brightness by voltage not duty cycle directly.
  7. Not true... For iron core transformers... The voltage equation is V=4*F*f*a*N*B*10^-8 Where: V is the voltage across the winding F is form factor (1.11 for AC) f is the frequency a is the cross sectional area (cm^2) N is the number of turns B is the flux density (gauss) The power equation is P=.707*J*f*W*a*B*10^-8 P is power (VA) J is current density of the transformer (amp/cm^2) W is area of core window (cm^2) What this tells you is that the size does matter, and the variables are the cross sectional area and the window area.
  8. Did you consider that the MOSFETS run for a long time in their "Ohmic Region" as the cause of them getting hot? Since the current, like the voltage, is a sinusoid, at "crossover", the MOSFET's shouldn't be conducting any current at all. As a general statment, that's a little misleading to those that aren't familar with op-amp circuits. Yes a reference is chosen, but it can also be ground just as easly as a DC... or for that matter, an AC signal.
  9. John You can get 100uF ceramic cap's nowadays if you can live with a low voltage. 2.2uF caps are a dime-a-dozen and nothing special... check TDK or Murata.
  10. Any number of resistors of the same value in parallel is equal to the resistor value divided by the number of resistors, it's not just limited too 2... i.e. 10, 100ohm resisistors in parallel is 100/10=10ohms... it's one of those things that works out that way given certain conditions. The equation for any values in parallel is Rval= 1/ (1/R1+1/R2...1/Rn). I suppose you could algebraically manipulate the equation to get Rval=R1/n when R1=R2=R3... if you wanted to really prove it to yourself.
  11. Those capacitors are not that hard to fine... the first catalog I took off my bookshelf has those values and voltage ratings (Cornell Dubilier), but most importantly, what is the case size?? CD# 4700
  12. Prior to the current limit threshold, U3 is "acting" like a comparator, sitting at the positive rail waiting for the voltage across R7 to increase. Unless someone wanted to use the supply as a "regulated current source" whiles it's in current limit, does current regulation matter?? Not that it isn't a good idea, but I can't think of a reason why a "typical hobbyist" would care.
  13. Actually, wire wound resistors do have inductance. As far a wire, it's about 20nH/inch. Yup... zero ohm jumpers can be/are used for configuring different options. They are also used a as "bridge". When you have to keep cost's down, manufacturers use them to avoid having to go to a double sided PCB. A few "jumpers" are a LOT cheaper than drilled plated via's on a multi-layer PCB (as in double sided and up).
  14. Yes, but not exactly... true, the U2-p3 node voltage doesn't drop quite as fast (.1
  15. If you have a triangular waveform already, why not just apply it directly to a resistor? That will give you your triangular current waveform. If you need more power, apply the signal to the resistor with a emitter follower. A triangular waveform is not a pulse. Which do you need? What is a EDMT machine?
  16. Yes it would. Will the circuit run without those components in place... yes, but under certain conditions, it might be possible to "flame" some components without it. R19?? Did you mean R9?
  17. When your output is linear, you are probably operating in the MOSFET's "ohmic region" where the FET looks just like a "pot". Once your gate voltage gets too high, the FET goes full on and there is no longer any correlation (linear) to the gate voltage vs. drain current. It would help if you post the full circuit.
  18. Where is there "shielded cable" connected to the output of U2?? QUANTIFY the amount of capacitance it would take to make U2 possibly oscillate, and why that would happen given C9 is in the circuit?? In this configuration it's some sort of gain limited integrator in combination with R15, Q2 b-e, R16, R12/C6 and R11... I don't suppose you have worked out the transfer function?? If Q1 turns on, for whatever reason and there had been a voltage at the output, D10 is now forward biased, and C1 is discharged via D10, R15 and Q1. How fast can you turn the knob? Something tells me that transient response of this supply isn't all that good!! besides, R16 would do the same thing.
  19. C4 has NO affect on U3 until you hit current limit, at which point U3 turns on just hard enough to reduce (clamp) the voltage at U2-pin3 such that the drop across R7 becomes small enough to "release" U3. Having R9 will slow down the discharge of C4, not speed it up. Until you hit current limit, U3's output is at the positive rail... so it will not oscillate. In the end R9 doesn't really do much... C1 has a large ESR (good for killing Q) but high ESL at line frequencies (60/120 Hz)?? C9 IS part of the feedback loop. U2 can't drive a transistor?? It's junction capacitance is so large that U2 will oscillate? R15 is there to limit the current through Q1 when it has to discharge (via D10) the ouput capacitor (and load cap, if any).
  20. That is correct, they do have resisistors inside of them that are trimmed with lasers when the wafer is being tested. BTW- there are capacitors in there too, and some even have inductors.
  21. The key is Vgs relative to the drain current. Remember "load lines"?? Vgs will dictate how much current can flow. Also... use caution when looking a simulator results!! They are only as good as the model!!
  22. Even running the simulation within the recomended frequency range, I couldn't get much more than 12V. Now "tweaking" the LC, that will do a lot for the simulation (36Vout), that doesn't mean much unless it's backed-up with real world data.
  23. The die will be the same, only the current limit point will be trimmed (laser) to a different point. The carrier size is for heat.
  24. Unless you know the core characteristics, you will not get too far!! Thing like saturation flux density, best frequency range for the material, etc.
  25. Neither... an unlimited current square wave generator at frequencies from 20KHz to 300KHz.
×
  • Create New...