Jump to content
Electronics-Lab.com Community


  • Posts

  • Joined

  • Last visited


Everything posted by Thomas

  1. Hello everybody I finished building my modified project maybe a year ago and promised back then to post some pictures as soon as i could get my hands on a digicam :) Well, here they are: Outside: The output can be shortened and switched off by two flick switches. This allows adjusting of voltage and current limiter before feeding it to the circuit. In order to prevent heating problems, I used a sandwich design for the bottom of the case. The 92mm fan sits directly behind the filter on the bottom. It blows up to the mounting plate, which has several holes in it to allow fresh air to get to the board, the trafo, and some other places. The rest of the airflow goes to the back and is forced through the heatsink. Inside: Looks shocking, but at least there's no space wasting ;) Both BD 241 and the 0,47R resistor are mounted on the sheet of aluminium bolted to the black heatsink. The 10.000
  2. The stupid thing will only show the current flowing trough it, but this hasn't always to be the current flowing trough the load :D In plain: You can connect a switch across the output terminals of the project and produce a shortcut by pushing it. Then you can adjust the current limit, flick the switch back and use the project the normal way with the ammeter showing the current flowing trough the load.
  3. Did you make sure that every diode in the negative 5,6V supply is soldered in the right way round?
  4. I think the terminals of D7 are connected in the wrong way. Just try to rotate the thing by 180
  5. I put an additional 8.5VAC winding on my toroidal core transformer in order to increase the voltage so 30VDC could be delivered. Because the heat generated at worst case, which is a shorted output and full current, was far to heavy to dissipate with my little heat sink, the additional winding is switched off when the temperature gets too high. This is done by a simple PTC-based comparative circuit with a relais output. I just changed a few things to achieve that the comparator can't reset itself, but requests a push button. I wanted to go all the way and tried to switch off the "bigger" original winding instead, but apparentely the voltage was too low so the circuit for the negative voltage stopped working. Without the negative supply, the project did act quite unstable and could not be adjusted. Of course, your smallest voltage is higher than mine, but I just want you to know that there's a limit. Can't say for sure how low you can go, though. There is also a moment of output instability when the switching back to full voltage occurs. I think it has something to do with C1. Do you intend to switch the windings automatically? Your 31.9V could be just a little to less. Depending on your transformer, you could perhaps put windings in series. The current is a little low, too.
  6. Well, I used a heatsink and thermal grease. But I soon discovered that the heat can't get away only by thermal convection, so now there's a 92mm @8V fan on it. I'm positive that this should work quite reliable. Maybe I can post some fotos of the case I made for it as soon as it's done. @MP: Then let's just say that I interpreted your post wrong, okay? What I read from it was that the changes had been unnecessary, which clearly not the case. Sorry for causing any troubles ;) @Mix: Well done!
  7. I think I should add something here... I made a few tests by myself while building my own (modified) project. I had to use a 29VAC trafo and a 10.000µF cap in order to get 30V/3Amp with (nearly) no ripple. There is no way that the original 24V trafo and the 3300 uF cap could have been sufficient to achieve this. I can't say much about the other changes audioguru and the others made, because I built the revised version. But, for instance, after making some tests with 1V/3A, my pair of 2N3055, which were mounted on a heatsink, burned down. I don't think the single one in the original project can be sufficient. So I'm quite glad that the project had been improved before I built it. Anyway, I don't think that this fight between you has only technical reasons, but I just wanted to put in my two cents ;)
  8. Hi anyone I've got a suggestion to make the project cheaper and perhaps less temperature-dependent. TI's TL431 makes a temperature-compensated Z-Diode, which output voltage can be adjusted by two (temperature-stable) resitors. That can replace the expensive U1. There is a problem, though, because the thing can only stand 37VDC. That could be solved by putting a resistor in series with it, and a Z-Diode of, let's say 15V, parallel. This way, we would get our 11.2VDC reference cheaper, with less space needed and with better stability over the temperature range. What do you think? The datasheet: http://www-s.ti.com/sc/ds/tl431.pdf
  9. Using a fan blowing over the heatsink could be a solution, too, if you don't want to buy a new heat sink. You need a very big heat sink to get 123W away without active cooling and even than the thing will get hot as hell. That is not good for the caps and will spoil the accuracy of voltage and current regulation, because the heat will spread inside the case.
  10. Well... my datasheet (SGS-Thomson) says something different: When the reg is lying before you with its legs towards you and the writing on top, it goes from left to right: GND, Input, Output. A 78xx, on the other hand, has Input, GND, Output. Annother chip manufactory says the same: http://www.fairchildsemi.com/ds/LM/LM7905.pdf But anyway, powering the dvms from different sources is a much better idea. Remember to use separate sec. windings for each dvm, because they will burn if a AM and a VM is connected to the same source.
  11. Hi Divljo, The 7805/7905 circuit seems strange to me... You know their pinout is different from each other? Pin 2 is ground with 78xx, but input with 79xx. Why are they connected with each other? Plus, filter caps are missing near those voltage regulators. Maybe they will get unstable under load.
  12. Hm... Well, I didn't measure that 3VAC drop. But 4,24A is still much smaller than the rated 5,83A. But annother thing puzzles me: The trafo is rated for sec. 20.5V, 5,83A and 140W. 20,5V times 5,83A equals only 119,52W. Something doesn't fit there ??? 23,5V times 5,83A would equal nearly 140W, but why should anyone want to know the "unloaded wattage" ??? :D
  13. Ehm? Maybe there's a bit of misunderstanding... With no load, I have 23,5V. When 5,83A are drawn, the voltage would go down to 20,5V. Never tried that, but it is printed to the trafo and sounds plausible. I don't draw that much current, therefore the voltage won't go down that far.
  14. Hi audioguru, Well, before that, I used 4 paralleled resistors of 39Ohm/2Watt. They became glowing red after 5 seconds, and smelly, too :D A burning 2N3055 doesn't smell that well, too, I've found out ::) The next problem with these resistors is their temperature drift meaning the current will go down after heating a several seconds. I don't understand that 33%? Do you mean the 20,5/23.5VAC of the original trafo? I wrote something about that in my last post... Right now, after adding 3 more windings and the line voltage being 235VAC, the ripple is gone. If I had a higher voltage rated C1, I could go even higher with the windings, but I think I will let it be as it is now. 30V/3A is a very extreme thing and I'm quite sure I won't need it anyway. 12,000µF/63V sounds quite well, but begs the question what happened to good ol' law "1,000µF/Amp and you are fine!" Well, you never stop learning ;)
  15. Hi audioguru, The trafo is rated for 140W and seems to stay quite cool @30V/3A output. But I could test that high load only half a minute before my load resistors (12 paralleled 120Ohm/approx.5W) are getting bloody hot. The secondary winding will have its 20.5VAC when loaded with the rated 5,83A. The whole output voltage from the trafo is at the moment 32VAC, and 30VAC when the output (DC) is fully loaded. That doesn't look that bad, imo? Maybe my C1 isn't as good as I hoped... it is a used one and I don't know how old it is. By adding a second (used) 10,000µF cap, the ripple is gone. But the trafo certainly wouldn't like that :( R3 and D7 do get warm, but are still touchable. R2 can't be touched anymore, but has still room before burning down. A 2W one should be a better solution, anyway.
  16. Hi audioguru, The original winding has only 20.5V rated voltage, but gives 23.5V when unloaded. Right now I've finished some further tests with the additional voltage being smaller. 8VAC unloaded works nearly fine, with little ripple at 30V/3A. If the voltage gets smaller, the ripple increases. There's not much room left for saving C1 :( The spring terminals (WAGO 236, rated at 16 A) do really look good. They are often used in serious industrial devices, too. You can visit the homepage of wago and look for the CAGE-CLAMP system for further informations. There are different types of spring terminals, so you really can't say they are all rubbish :D I use them to connect the rectifier, the cap, the 2N3055s, R7 and the output to the board. Screw terminals can get loose after a while (especially when the wire is tinned), and after ~ 25 times of openings and tightening again the soldering to the board will break. Plus, they cost about the same as the spring ones. Soldering the wire to the board is fine and cheap until you don't move it much. But I don't like the look of it ;)
  17. I' ve just soldered it together and made a few measurements. It seems to be quite fine :) At 30V and 3A Output I couldn't measure any ripple at all. My 10.000µF cap does its job well ;). Because C1 is off the board and connected with wires to it, I soldered a 220µF on the board in order to compensate the inductivity of the wires and of C1 itself. That should help when heavy load changes occure. R2 is getting quite hot. The part list should be updated for a 2W one, there's just enough space on the board for it. C2 and C3 get at least lukewarm. I don't like that, but I don't think there can be done something about it. Q1 and Q2 are mounted on a heatsink (3mm of aluminium sheet) and connected to the board with short wires. That solves two problems at one time, the cooling of Q2 and the leg twisting, which looks awful. R7 is mounted on the same sheet and the sheet itself will get bolted to the heatsink later. I did a few changes to the layout with a CAD program in order to make space for spring-type terminals. I like them better than soldering the wires directly to the board. If anyone is interested, say something ;) Three secondary coils had been added to the transformer. One for the additional 8,5V, and two for digital LED-Panelmeters. Equals 30m of wire to the toroidal core! ::) Maybe i can decrease the 8,5V coil a little, because C1 has 46V across it when unloaded. That's a little heavy for a 50V one ;) The voltage of the transformer doesn't go down when loaded as much as I expected and my rectifier bridge doesn't have that much voltage drop, so I think additional ~7 V would do the job. I thank you, audioguru, for answering my questions that well :D
  18. I use Target 3001. You can get a free demo at http://ibfriedrich.de. The demo can handle up to 100 pins (enough for small projects). A 50€ Version gives you 400 pins.
  19. Hi jpp, Did you probably forget the filter caps of 100n and 330n near the regulator?
  20. Well... What did mean by "higher wattage"? Do you want to have more current or more voltage? More voltage can be done quite easily by just connecting them in series, but I assume that's not what you wanted... More Current, on the other hand, is very difficult to achieve because you have to connect them parallel. So both psu try to regulate the same voltage, causing their output to get very instable. Maybe a diode in each rail of each output could help, but you also get much heat and less output voltage. I don't think you get happy with this one...
  21. Hi Blacque, We are talking about these, right? I think you are right, they shouldn't get that hot when they are bolted to the case... I think I'll try one of these. I just now discovered that they have also only 50ppm. That should work. Annother advantage of "my" current measuring is that you can always see how much current flows to the load. You just need a switch that can stand 3 Amps. Hi audioguru, Okay, now I'm informed about Q1. Thank you :) When both values can be presetted, the usage of multi-turn pots should be no problem anymore, right? Saves space, can be well adjusted and a accidential touch of the knobs would not produce that much smoke ;)
  22. Hi Blacque Jacque, Annother way to preset the current is to shorten the output with a suitable switch and measure the voltage drop across R7. This voltage gets to a 100k/27k potential divider. Now we have 0,3V @ 3Amp across the 27k resistor which can be fed to a panelmeter. A 1k pot can be added in series to the 27k for fine tuning. A second switch can be used to disconnect the output from the psu. That gives us the ability to preset the voltage. I intend to do it that way. Maybe there could appear a problem when R7 dissipates 4.23 Watts and gets mighty hot which will change its value. Those expensive Resistors in a TO-247 case could be a solution because their temperature coeffizient is less than 50ppm, but I can't find any with 0.47 Ohm. Hi Audioguru, 1) Can a 0.56R or a 0.39R Resistor be used for R7? 2) I've read both threads about this psu and am a little confused about the emitter leg of Q1. Is the original schematic correct or not? ???
  23. Hi audioguru, Well, there a not that much AC in my neighbourhood and the line here is quite stable. But after all this power supply should be very reliable, so I'll take the 445APs. Little expensive, but so be it. I can get also more output voltage this way and probably even less ripple because the higher voltage should push more electrons in my filter caps. Hi MP, That leads me to an interesting question: Where are you people located? Quite strange time in this forum, ACs are a normal thing in every house... Probably Greece?
  24. Hi audioguru, Thank you for the new list. I wondered whether the OPA604 could be used instead of OPA445. This is an audio-tuned opamp with +/-24V. I did some measuring with my toroidal trafos and the unloaded voltage is 14,634% higher than the rated one. And after rectifying and filtering, this unloaded voltage gets multiplied by the root of 2. 26VAC*1,14634*1,414 equals 42,15VAC. Together with the -5,6V supply we are pretty close to the rated voltage of the opamps. The absolute max. voltage is +/-25V. So can I tune my trafo up to 26VAC rated voltage and use this opamps, which are about 5 times cheaper? Thanks.
  25. Hi audioguru So, when those (expensive :o) opamps are used, the only thing to change in the project would be the layout, probably more 2N3055 with each one serial resistor and a big rectifier bridge, right? Getting 5A is impossible with my transformers but I'm sure I don't need that much current. P.S.: I intend to go with TI's friendly sample policy. Did anyone manage to register by using netscape browser 7.02? I tried it several times but without success. Nothing happens when "register" is clicked after filling out the form. ???
  • Create New...