 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Electronics : Variable input voltage bat charger
  Page 3 of 8   |
|||||
| Author | Message | ||||
| Janne Senior Member  Joined: 20/06/2008 Location: FinlandPosts: 121 |
Hi oztules, I tried making it into a jpeg file.. picture size was pretty big but it was still very messy.. not going to work that way I'm afraid. (though if the world is blurry anyway, does it make a difference  )
If at first you don't succeed, try again. My projects |
||||
oztules Guru  Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Ok Janne.... I have decided to play with the nearest ATX I could find.... just to give you some positive things to think of. Here is the complete ATX before being "interviewed".. 
And this is what that ATX supply looks like when you remove some one hundred and 5 components that you don't need. 
As you can see, it requires very few components around the 494 to make a perfectly stable voltage and current controlled device. It is better than the AT in that the supply voltage is supplied from the aux flyback supply, and so can go to very low voltage be fully controlled at max current. This one is now set to 20A max, can be shorted and still push 20A through the meter. It requires only a handful of bits to work. The oscillator requires 1 resistor and 1 capacitor only. (pin 5+6) Pin 4 grounded.... unless you want soft start then a cap (10uf?) to pin 13, and a 10k to ground to bleed it slowly... soft start. Pin 3 .. a cap (222)and resistor (67k **see edit#2 below actually 330r) to pin 2 and the same again as per Dinges circuit... 330r and 10n from pin 3 to pin 15. A voltage divider to place a ref voltage on pin 2 (from say pin 13 (5v) to ground) to give say 2v on pin 2..and a sample of output voltage on pin1 (with trim pot for control) Pin 16 to ground and pin 15 to current sense (piece of wire) and trimpot (and 47k to pin 13/14. for bias) NO other controls are necessary for perfect stability and control. The outputs stay the same .. ie pull down resistors (1.5k or so) and then off to the totem driver trannies as before.... etc. From the 494 onwards it stays the original..... all the controlling has been done by now. The 494's power comes from the flyback aux transformer via small inductor and diodes.... the proper control gets the tossed out, as it is an inhibit circuit. Get rid of it all and approx 30v will go straight to pin 12 on the 494 at switch on. (as it does originally anyway) So it looks like brutalizing the ATX is every bit as easy as the AT, but requires a lots of unsoldering (I used a heat gun in the finish as so much has to be removed) if you want to bare bones it so you KNOW what is going on..... rather than tracing miles of useless stuff you don't need. Dinges will be the first to tell you I'm a grubby animal with this stuff. Lots of good fun.... Dinges this should take you only 40 mins or so. ............oztules Edit... Janne, I replace the glass fuse with a 270R 3w resistor for the initial test to stop runaway before you establish the right range for V control.... otherwise it may drive the pulse width to max if you dont get the trim pot in a safe range ( so V max doesn't exceed say 28v (change the caps voltage on the output filter for this kind of voltage). Then check you can control A down to 0 and just over to prove control. The resistor should stay cool at this power level. When you have the range right, replace fuse... and enjoy. Ps. Dinges, swines with white hair seem to be unaffected thus far....(silver bullet gene)... but those squashed grapes make a pig out of me 
Edit #2 the 67k is wrong it is 330R . The technical term for this is a stuff up late at night  Village idiot...or... just another hack out of his depth |
||||
GWatPE Senior Member Joined: 01/09/2006 Location: AustraliaPosts: 2127 |
Hi Oztules, perhaps you could put finishing touches to this current limited unit and add some controlled variability to the current limiting, so that the current limiting was directly increased, proportional to the input voltage. This would give a unit that was suitable for direct connection to a windmill to give a loading that automatically increased with the wind energy. This would be similar to the way my own boost cct works, but with a transformer coupled system, rather than a single inductor boost. Would be OK in the first instance for power levels up to 250W or so. Just something to do in front of the fire on a cold winters night. Gordon. become more energy aware |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
I agree with Oztules, Janne's PSU looks very unfamiliar; strange layout, very small heatsinks and strangely positioned. Unusual, but of course doesn't mean it can't be (easily) modified. Oztules' PSU, on the other hand, looks exactly like the ones I've butchered. Did you have any specific reason for disabling the soft-start by grounding pin4/TL494? [quote=Oztules]Pin 3 .. a cap (222)and resistor (67k) to pin 2 and the same again [/quote] I suppose this is meant to be 47k. [quote=Oztules]A voltage divider to place a ref voltage on pin 2 (from say pin 13 (5v) to ground) to give say 2v on pin 2..and a sample of output voltage on pin1 (with trim pot for control)[/quote] The reference voltage you use is the minimum voltage the PSU will then be able to put out; most PSUs I've seen use 2.5V as reference to the voltage comparator, so output range becomes 2.5-xx V; I've had one that used the 5.0V as reference, straight from the TL494's reference output pin (i.e. without a divider); that PSU couldn't go lower than 5.0V output, so I modified it to a 2.5V reference with some resistors. I've never tried it, but 1.0V as reference may work fine too and would give an even larger voltage output range. May try that myself next time. [quote=Oztules]Then check you can control A down to 0 and just over to prove control. The resistor should stay cool at this power level.[/quote] See above remarks; the output voltage won't be able to go below the reference voltage. So all the more reason to aim for an as low (as possible) reference voltage to the comparators. [quote=Oztules]Lots of good fun.... Dinges this should take you only 40 mins or so[/quote] Oztules, it takes me that long just to drag my sorry bones into the shop and heat up the soldering iron. 
Seeing your bare-bones PCB was quite a shock; reminded me a bit of photos of WW-I battlefields: the vast emptiness... the craters... remnants of what it once used to be... But yes, it's amazing to actually see how few components are needed. I normally leave most things in place, apart from the bits around the output filters and capacitors, because I find my replacement capacitors need all the space they can get. After I've removed the 2 diodes between the LM339 and pin4/TL494, I don't bother with the circuit around the LM339 anymore; it would just be extra work removing all the components... unless you use a heat-gun, of course  . But in my experience, using a heatgun on those pertinax (phenol/paper) boards quickly leads to a ruined board and dirty smells in the shop. (though Oztules would probably consider that smell to be like perfume... the smell of succes!)
20A out of one those PSUs is, well, err, 'impressive'. Mostly because the diodes are probably only rated for 6A.  Short periods of 20A should be fine, but I strongly suspect a continuous rating of 20A is wishful thinking... let alone 20A at, say, 24V output; the transformer will probably start to sweat at those power levels. Most PSUs I've worked on used the same 3A diodes (2 pcs. of them in parallel), even when the PSU claimed to be able to give 10A. Being mounted on cooling fins and with a fan blowing at them will help a lot, but still. Not my philosophy; if it's rated at 6A, I'd use it at maximum 4A. (Dinges hears Oztules making chicken sounds whilst flapping his arms)
Peter. |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Dinges, Pin 4 .... grounded through laziness. 67k.... It was blue violet orange..... Your correct about the voltage divider and V min. Amps min can be zero or so close as to not matter. (my amp meter does not register the lowest A setting) Interestingly, the duty cycle on the diodes is 50% of whatever the pulse width is. I have used the original ones for 3hour periods (battery charging applications) starting at 15A and finishing at 3A or so. Remember they are 2 diodes in the pack as well. The ATX seem to use heavier stuff than the AT. These ones can support 10A each forward current(pushing to the limit in the data sheets) with 50A surge ratings @ 200v. (Oztules accounting says... 10 amps all of the time or 20 amps half of the time is close enough.... no wonder no one will employ me ) This ATX one now has d92 diodes (I bought a thousand of them by an ordering mistake years ago) 20A@200v twins. These can run 20A easily.... but still need heat sinking ... very much so... but will stay under 100 degrees @ 20A. Once the board has been emptied, it leaves lots of room on the heat sink and board for extra diodes in parallel if your keen. The ATX transformer is quite large on these ones, and so 20A at less then 14v will be sustainable for extended run times I suspect. I have rewound larger E cores and used them for other applications... much bigger. The aggressive fans on these things allow you to push them surprisingly hard. (not for the feint hearted) Funnily enough (love the smell of burning components in the morning), if you use the heat gun at the right range, it seems to not denature the board. All the bits rain down on the bench, and she is denuded very very quickly. Gordon, easily done, and plenty of room to do it. .. perhaps I should have left the 393 on the board.. be handy maybe?? ..............oztules Village idiot...or... just another hack out of his depth |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
[quote=Oztules]67k.... It was blue violet orange.....[/quote] Nice try. Now pull the other one. 68k, yes, I could believe that. That would even be in the E12 series (hell, even the E6 series; just checked). Not even the E192 shows 67.0k... and would require 4 colour bands: blue-violet-black-red. Unless the resistor factory did a special run of 67k resistors (with +/-20% tolerance? <evil grin>) just for that power supply, perhaps. I'm afraid I'll need photographical evidence to believe your above remark.
You sure that violet wasn't grey? Those colour bands can come out very funny when they're not lighted by natural light. And some FLs are much worse at colour-rendering than others... Peter (who is reminded of the joys of interpreting colours: 'that band, would it be brown or (faded) red?'. I'm sure I'm not the only one who had to get a second-opinion on the colour of a resistor band at times... 'no, no, you're wrong, it's not red, it's more bordeaux-ish with a hint of mauve and a dash of cyan') |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Was worth a try I suppose. On taking it outside in the sunlight, the blue turns out to be a bluey green, the violet seems to be a browny purple, the only one that appears to be a correct is the orange. (I'll blame the chine4se ink.... and the hot air from the heat gun..... and anything else I can clutch at.) So using the trusty ohm meter, it read 330R !!!!!?????...... This had me going for a second, until I realised it was academic really, I had put a 330R in parallel on the back of the board, and forgot to get rid of the "50k come 67k" resistor... So I clipped the itty bitty (4mm)resistor off the board... and it sailed of the desk and down the back of the desk wall.... where it will probably stay until I ever pull out the desk. Not to be denied, I had a second one of these supplies, and that resistor clearly appears to be green brown orange (I'm claiming heat now).... and reads 50.5k... close enough to 50k for me. Mystery solved. Sherlock oztules Village idiot...or... just another hack out of his depth |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
[quote]and reads 50.5k... close enough to 50k for me.[/quote] And just as close to 51k as well. Then again, brown can look an awful lot like black.... Dr. Watson. |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
AAaaaaaarrrggghhhh.....!#@@#!#@!@@!  Village idiot...or... just another hack out of his depth |
||||
| Janne Senior Member Joined: 20/06/2008 Location: FinlandPosts: 121 |
Hi, Yeah sure, I can pull 105 components out from a psu too in 40 minutes with a heat gun.. though it's 100% sure it will never work after that ..¤%@Q%.. "first atx box i had in hand" adding a little insult to injury eh?  now after 7 atx boxes opened one of them looks promising, with KA7500b sitting in the middle.. (also i got another with UC3843, this one actually works, i'll put it into the "mayby" bin.)
Anyways, I went ahead and removed the undervoltage lockout on my "TEMERO" psu.. Now it works down to 6V(with the 20k trimmer at it's maxium), and with current control it will work down to 0V when i short the output. The bad news is that the current control propably monitors the primary winding on the transformer, as when i short the output the current limit magically rises, with the output limited to 2A @ 6V, when shorted it jumps to 8A. This kind of monitoring might work with our battery charger case, but where true current limiting is required it's a no-no. As is, the box seems still useful, so i don't want to "ruin" it by making the wire modification on the output side.. What comes to the battery charger case, the use of the voltage doubling swich will be requided(or a rewind), it takes over 140VAC+ to reach 16V on the output. I can't quite get my head around it, if using all 3 of the phases is possible with the voltage doubling circuit? On quick thinking it looks to me like we're limited to using only 2 of the phases if we will go with doubler switch route. (We're also still researching other options). In the mean time, I've opened the other AT box for surgery. This one has the "IR3M02" ,sharp equivalent of the tl494. It also has the second error amp disabled, so it seems like a good candinate for the sensing wire current limit-mod. This one is quite odd also: 
The cover removed. Everything is boxed between the PCB's and heatsinks. 
The workings removed from the case and top PCB opened. The heatsinks (= the 2 aluminum plates) seem rather puny. If at first you don't succeed, try again. My projects |
||||
| Janne Senior Member Joined: 20/06/2008 Location: FinlandPosts: 121 |
removed the things i messed up If at first you don't succeed, try again. My projects |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
The ATX with the KA7500 looks like a good candidate. If you get the trusty heat gun out, just empty all components off the board from the 7500 back to the outputs. Just leave in the oscillator res and cap. All the output stuff from the 7500 needs to remain (pwm outputs to driver network. So 100 odd components later you just need to add the current limit system (4 components), the voltage limit system (another 4 perhaps) and your rectifier and filter (RF at least, and a cap of some size)... and soft start if you need/want it. This way you can't fail, as all the "logic" will be new from you. Even if it's blown up, it will be easy to bring to life as there will be almost no cunning circuitry left to fix. It really is much easier than it looks. If your only after 28v or so, maybe change to full wave rectification on the secondary, and 1 supply will be enough. I wasn't a fan at trying to get newbies to convert the ATX ones, as they are pretty messy for the novice to unravel, but after bare boning them, this seems to make it easier to get people new to them to see how simple they are. When they are fully populated, it is hard to convince people. Thats why I emptied the board for you.... just to prove how simple the control side is. After that, the driver is simple to unravel, and the HV side is self explanatory.... psu solved forever. They do offer a few good things, like more powerful transformers, stable input drive voltage for the switcher, and generally bigger switchers and output diodes. The AT's made up for this with simplicity and space...... although somehow you have managed to scrounge up the most complicated looking things I have seen in those boxes. One can't imagine what they were thinking when they designed them, when others were making them much simpler and cheaper on a single board. That said, nearly all the complex stuff can be done away with, and you can get down to business with the simple mods. It is odd you need 140v just to get 16v... 140v usually gets you near full power. Something wrong with your control I think. ............oztules Village idiot...or... just another hack out of his depth |
||||
| Janne Senior Member Joined: 20/06/2008 Location: FinlandPosts: 121 |
Hi oztules, Emptying an atx board and starting from almost scratct is indeed starting to sound intriguing.. There is one part that i don't get still If the tl494 gets it's power from the aux flyback supply, how is the initial power provided to the ic to start switching the totem pole drivers? In one atx schematic i saw the power to tl494 was provided directly from the HV side by a voltage divider circuit, and the workings of that are clear to me... I think i understand the tl494 based control circuits fairly well now, but the actual business part on the psu is still a bit odd to me. If at first you don't succeed, try again. My projects |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
Janne, An ATX powersupply contains two supplies: the 'normal', primary supply, just as an AT PSU. And a tiny, secondary +5V supply that powers various stand-by functions of the PC. This secondary supply is always on (as long as the powercord is plugged in), the primary supply only operates when the power-on line (green(ish) or lime greenish or even white(ish) ) is grounded.
In the schematics of the ATX designs I've seen, the TL494 operates off the (unregulated) power of this secondary power supply (see e.g. http://www.pavouk.org/hw/en_atxps.html) I figure that if you simply remove everything unneeded around the TL494, but leave the secondary power supply intact then you'll be fine: the TL494 will always be powered. Or you could remove the unneeded parts of the secondary power supply too, I suppose... But you never know what use a regulated +5V line (even if only a few mA capacity) can be for powering other parts of a DIY project (some logic circuits, a LED, or....?) If I'm not mistaken, the bootstrap method (using a resistive voltage divider, powered by the primary voltage), is used in AT supplies, which do not have a secondary (permanently on) power source as the ATX. If you decide to depopulate the board, I beg to differ with Oztules' remarks about the heatgun. I've never managed to get it to work with pertinax/phenolboard. It works very nicely on epoxy boards. Then again, maybe Oztules is just handier with a heatgun than I am? Also, often components in PC PSUs have their legs bent, which makes them impossible to bulk-remove using a heatgun; even using a soldering iron and removing them individually becomes tedious due to this.
And yes, the TL494-equivalent with the strange marking that I couldn't recall... it was a IR3M02 too! For the past few days I was trying to recall the number, in vain. I had a 'yes, that's it!' moment reading that part of your post. Now I can sleep soundly again.... The full-wave rectification bit is new to me; don't want to divert this thread much more (rather would let you and Oztules up to working things out so you get your PSU running), but if Oz could explicate this a little I wouldn't hold that against him. Or in a separate thread perhaps. Thanks Oztules, for your relentless efforts on the topic of killing PSUs. Have learned a few more tricks, again. Personally I think you could do a great work by writing an article on these conversions for an electronics magazine (paper). Peter. Edit: your best bet for finding out how *your* TL494 is powered is by simply tracing the traces connected to the Vcc (pin 12/TL494). You'll quickly find out how your specific PSU powers the IC in standby-mode and during normal use. |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Janne and Dinges, I don't have access to the "weird" supplies you seem to have over there. I have noted that compaq seem to use an array of weird stuff in their designs... but the clones seem to use the half bridge system, most compaqs I have looked at seem to favour the single switch flyback type for their AT units. These types are generally 150w or less. AT's. This discussion is for the dual switch halfbridge, as seen in most supplies for the AT clones of the 200-250w style. In these units, the HV side is completely isolated from the LV switching side. The power to bootstrap them is derived from a cunning positive feedback turn on the driver transformer. If you look at one of these supplies, you will notice that the driver transformer has a thick turn or two around the coils stack. This provides the drive voltage to the switchers. The net effect of being wrapped around the base drive windings is a cyclic oscillation of about 50hz or so of a spurious nature when the power is applied. This haphazard waveform is fed to the base of the switchers, which drive the output transformer. If you overload a standard supply, you will induce the scr to turn off the 494. In this state, the 494 stays turned off. Only by turning off the ac for a few seconds, will the supply start again... How? Well, with the 494 turned off by the scr, no power should be available to keep the scr latched, it should give up and automatically restart... but it doesn't. This is because the spurious waveform set up by the positive feedback loop, keeps enough power in the secondary, to keep the scr latched.... so it does work. Use a scope while it is in this state, and you will see the wave train which is doing this..... this is how it bootstraps. In normal startup, the 494 sees this voltage being generated in the secondaries, and starts to wake up. It then drives the proper switching pulses into the driver, and swamps the little feedback wave train and we run normally.... until an overload arrives and sends the 494 into stall mode again, requiring a complete power down to recover... thats why we need complete power down, otherwise the bootstrap pulses will keep it alive enough to keep the 494 turned off. In an ATX, a simple uncontrolled flyback (on the HV side) provides approx 30v onto Vcc of the 494, and it is alive. Other control voltages keep it sleeping until called upon, and it still uses this uncontrolled supply for it's vcc. In AT's with a resistive bootstrap, it can only be that your looking at a single switch flyback, or dual switch halfbridge... but the pwn chip will be on the HV side of the board, and will require opto isolators to ascertain what the secondaries are doing.....for control. Janne "In one atx schematic I saw the power to tl494 was provided directly from the HV side by a voltage divider circuit, and the workings of that are clear to me..." This must mean the pwm chip is on the HV side....(likely flyback topology) other wise the LV isolation is compromised. So I would be surprised to see the chip on the LV side for that reason. Dinges "If I'm not mistaken, the bootstrap method (using a resistive voltage divider, powered by the primary voltage), is used in AT supplies, which do not have a secondary (permanently on) power source as the ATX"...... see above. As for the heat gun... I use a makita heat gun with no nozzles... broad heat indeed. But if you hold the board with something, and face the draft towards the LV half of the board from a few inches away (but the exhaust heat leaving the board at an angle to keep the good bits out of the hot air, localised burning heat is avoided. In a minute or so, the whole solder regime under "inspection" is melted, and a tiny jewelers screwdriver can easily prise off the bent component lead devices, the rest fall out. These boards have oversize holes I have found. and few components need help. It must work, as I just did it??.. and the board is still fine. In the case of the board above I did, all components not used in directly powering the 494 from the aux got the chop. A reference voltage is still available from the 494 at a few ma if required. Will do more with the full wave in a day or so... have a wool shed to dismantle today... muscles are aching from yesterdays effort 
Edit Dinges, I just looked at the circuit you linked too. It seems the third "positive feedback" winding is still in the driver transformer. It may be possible to use a single diode and capacitor to boot the 494 up without the aux circuit on the atx. (boot pulses should still be there... legacy???..Interesting, although the aux is already in place in the atx and provides stability in the low volts region where the AT may fail. Another advantage with the ATX, is that you can fire up the 494 and disable the b+ to the switchers, and test your control circuits with a lower b+ and watch the pulse width change as you fiddle the inputs(on a scope). (inject your own voltage into the error comparators and see the results without smoking it up. May be a good learning tool. .........oztules Village idiot...or... just another hack out of his depth |
||||
| Janne Senior Member Joined: 20/06/2008 Location: FinlandPosts: 121 |
Once again I have to say thanks to you guys patiently explaining this stuff out. Much about the bootstapping was still new to me, especially how the AT power without auxiliary supply gets the initial power. Will have to check the tl494's output on the odd psu nr.1, if the high voltage requirement indeed has something to do with the control. At least i could try to disable the dead time all together(or to the 3% minium) to see if it makes any difference. If at first you don't succeed, try again. My projects |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
We all learn something every time we revisit this stuff. I notice on my AT boards there is provision on the circuit boards for the extra ATX components (Dinges can you look at your AT clones for me if you could and see if they have them as well?) As an exercise I will ATize an ATX supply. I note that the transformers on my ATX do have the few thick turns on the outside windings. They obviously don't need these to start, and they certainly dont need them to drive, so they must use the same transformers in both versions. I will try pulling off all the components for the flyback, and use a diode and cap to self start the ex-ATX supply.... interesting experiment.... hopefully in the next few days (wool shed is still ongoing... so are the aches and pains). .........oztules Village idiot...or... just another hack out of his depth |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
[quote=Oztules](Dinges can you look at your AT clones for me if you could and see if they have them as well?)[/quote] Blasphemous as it will sound to you... but all spare PC PSUs have been discarded (about 15-20 pieces  ). When time and circumstances allow, I'll dive into those PSUs again. Only have one spare ATX supply left now, meant as a replacement for my PC should the PSU of it fail.
However, I do recall seeing PSUs that had a lot of vacant place with print on them that indicated extra components could be installed; whether this would've turned it into an ATX I can't say. Looking forward to the results of your experiment. Not because of the practical benefits of this modification, but slowly yet surely more and more secrets of the PC PSU get revealed. [quote=Oztules]I don't have access to the "weird" supplies you seem to have over there. I have noted that compaq seem to use an array of weird stuff in their designs... but the clones seem to use the half bridge system, most compaqs I have looked at seem to favour the single switch flyback type for their AT units. These types are generally 150w or less.[/quote] In all fairness... I critically selected the PSUs I converted. Plenty were weird or looked strange, or had a transformer without flying lead. And yes, Compaq did some funny things with its supplies. I recall seeing a Unitrode UC-???? IC in one of the Compaq PSUs I opened. Not saying that these supplies can't be converted, but I simply closed them up again and tried another PSU from the heap; why bother to reinvent the wheel if you've got plenty of more suitable candidates to choose from. In fact, if the insides were too dusty... I just grabbed another one: that was quicker and easier than cleaning them...
Peter a.k.a. 'the lazy bastard' |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Well, it looks like they do use the same transformers. A quick unsolder of the driver transformer, and unwrap of the outer sheath.... shows us the 2 turns are there as suspected. If you disconnect the flyback driver and look for the pulse train.... it's not there..... hmmmmm. A quick check around the driver bits and pieces and we find that there is no base collector resistor. By slapping two 330k resistors from the base to the collector and we find the pulse train strong and clear. If you look at the circuit you pointed to, you will see C9 and C10. The 330k resistor goes from neg terminal of the caps to their respective collectors.... and we have an AT startup legacy device. Useless I guess but we now know a little more. However, if you find a dead ATX supply, that has a faulty flyback circuit, you can bypass the flyback stuff, add the resistors and go on to modify it as normal. .......oztules Edit: As an aside, I have noticed that if the flyback driver dies, everything in the whole area dies.... all the diodes, half the resistors, the transistor/s... the whole lot... and all hard to get to. There is a lot to be said for having the driver gear on the LV side of the transformer. I don't like HV logic switching for this reason.... when it goes wrong it goes really wrong. Village idiot...or... just another hack out of his depth |
||||
| Janne Senior Member Joined: 20/06/2008 Location: FinlandPosts: 121 |
/me spares a tear for all the psu's that went to the dump 
I've been a bit busy in the last few days, and haven't done any work with the computer psu's.. But we've given up on the though on modifying a psu for the battery charger. The guy this charger was going for got bored waiting with the mill up in the air doing squat, and he found a good charger that would work off the shelf. I said go for it.. The charger he's getting is ADC4370 from powernet. Datasheet in here, if anyone's interested. datasheet The charger even has an analog 0-5V input to control the power output, so if one would like to get fancy it could be made to follow the power curve of the mill (Gordon you reading this?) The plan is to use 3-phase rectifier to feed the charger, and if need be, add capasitors to smooth the input ripple down. I will post how it turns out. I will continue the work with the computer psu modifications... for example, I could use another lab supply power, (how have i survived so far with only 3??), and when the time comes my bigger heating mill is also going to need something to charge the battery that powers the control equipment.. previously it has been done with an auxiliary car alternator, but it would be fun to ditch the extra alternator. Another psu might also be handy for feeding the field current for the main generator. If at first you don't succeed, try again. My projects |
||||
| Page 3 of 8 |
|||||
 Print this page |
