 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 5 of 8   |
|||||
| Author | Message | ||||
sPuDd Senior Member  Joined: 10/07/2007 Location: AustraliaPosts: 251 |
Thankyou Oztules & Dinges, I had my suspicions about the error amps, I though it had something to do with the 300uA internal constant current supply etc etc. It makes sense to me now, that the error amps should be looking at a DC level (the output) and adjusting it up/down. For some silly reason I thought about the 494 as a current mode system, where removing the AC on the feedback would kill the pulse by pulse control system. Tsk tsk – it just a 494 
I will try all those things tomorrow, test with CRO in all situations and report back. I was worried it was my rewound transformer, but I fitted a factory one and got the same effect. I'm using a 330uF 200V 50Hz cap from another ATX front end, and I suspect there is more AC across the O/P than the 240Vac coming in! I'm sure if I filter the feed to the error amp, I can trick it into thinking the output is clean. Yes, I know I should have more O/P filtering, but it in testing phase
sPuDd.. It should work ...in theory |
||||
oztules Guru  Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Spud, The squeal is in the audio range. If it were a reflection of the ripple in the output, it would be at twice the running frequency of the 494.... so I wouldn't lose sleep on that part. It is <15000hz or you wouldn't hear it. (Dinges found it in the 1-2khz range) I assume it is more to do with feed back (like microphone feedback) or motor boating in an audio amp.... positive feedback loop somewhere. The fact that Dinges could pull it on with touching the input pins supports this theory. My solution was just to damp it out. I'm not ... repeat not an electronic engineer, so some (all) of my seat of the pants solutions are not based in firm theory... rather observation and explosive reactions   
I am however, still firm favorite for the title of village idiot when the job comes up. 
..........oztules Village idiot...or... just another hack out of his depth |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
Oztules, tis OK mate. I wear my safety glasses when working on anything that can go boom. FrankenSwitcher blew a MOSFET into my eye, so I resolved to protect my eyes from then on. Oh, also I use a pair of matching 1.5VA, 15VA & 60VA transformers wired back to back which double as cheap isolation transformers and power limiters. They fold back hard when excess power is drawn. I start on the little one and work my way up once each safety test is complete. Cheap and easy. I move to the 500VA ISO then mains when done. I've heard microwave oven transformers are good for ISO too, but remember to drill off the secondary earthing lug first. And use 2x matched transformers. sPuDd.. It should work ...in theory |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
Well, I just spent several hours on it. The error amps needed to be setup for correct gain, filtered the error amp inputs, filtered the output, mucked about with the drive system. All for naught. Cleaned up the error amps somewhat, which cured the squeal and made it somewhat controllable. Still getting a 1Vpp ringing spike on the error amps inputs. Output drive when controlling voltage is terrible. I can only say that this particular unit was built to a price, and would be worthless in its current design. It may be useful if all you do is put in a pot to tweak the 12V rail up to 13.8Vdc. I'm quite convinced they cheated and used the DTC to hold the voltage correct. The parts may be useful donations for a properly designed dedicated PCB, but other than that... I should have used a higher quality unit, but at least it cost nothing - just some hair. sPuDd.. It should work ...in theory |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Must be a fault somewhere. At least you beat the squeal. The ringing should be damped at the transformer (resistor and capacitor on the input and output). The advantage /problem with emptying the board with the heat gun, and re-populating, is that all problems are of your own making..... If you leave their stuff in there, there is plenty of chance that it will interact at some point.... and problems arise.... I hooked up a huge ferrite transformer last night... just to see some pyro's. The ATX did 30A@26v. Yes Dinges, that little rectifier was still in it, and survived, the little switching trannies didn't. They had trouble driving wide open.... time to rewind and try again. The tranny stayed cold. Will get some pics soon I hope. .........oztules Village idiot...or... just another hack out of his depth |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
Sweet sweet fireworks... 
Its possible that the ATX I picked had issues before I started. That may be why it got scrapped. I've selected a new victim from the pit, and have load tested it prior to doing anything. I'll plug at its pins tomorrow with the CRO and see what makes it tick. I added a homebrew resistor/cap snubber to the output windings, also no good. And the output diodes are fast soft recovery units from a Siemens 5kva UPS. The ringing is synchronised with one edge of the transistor switching. I guess I can't expect much from a 2nd hand ATX that was $25 new... Oztules, did I see what looked like a homemade ATX style PSU in your Fieldlines files? sPuDd.. It should work ...in theory |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
[quote=Oztules] I hooked up a huge ferrite transformer last night... just to see some pyro's. The ATX did 30A@26v. Yes Dinges, that little rectifier was still in it, and survived, the little switching trannies didn't. They had trouble driving wide open.... time to rewind and try again. The tranny stayed cold. Will get some pics soon I hope.[/quote] Impressive. You never know what is 'enough' until you know what is 'too much'.  Looking forward to pictures of it.
Pushing nearly 900W through those transistors I'm not surprized they tossed in the towel. Easy fix though. The diodes surviving such abuse is a near miracle. [quote=Spudd] I've selected a new victim from the pit, and have load tested it prior to doing anything.[/quote] Spudd, some of the PSUs I tackled were a breeze, others were a never-ending nightmare (the first one was one of those) that ended blowing themselves up. Luck of the draw I suppose. [quote]did I see what looked like a homemade ATX style PSU in your Fieldlines files?[/quote] You mean the one with the two transformers? Saw that too. It even looked good. Spudd, I've got two identical transformers from 800W UPSes lying about. Could connect them back-to-back as an isolation transformer. I'm sure you shouldn't have much trouble finding two similar UPSes for parts. I'd prefer to use such transformers over microwave transformers with their high-voltage (2000-ish Vac) secondary, where one end of the 2ndary is grounded too. Not really floating. Microwave transformers are nice and big, but the high voltage is very unforgiving of mistakes. You could drill the groundlug out and disconnect the secondary, but the insulation of the secondary (w.r.t. earth/transformer iron) is not designed to withstand that high voltage. Under normal conditions it shouldn't matter... but still. An isolation transformer is a good thing to use when experimenting with things like PC PSUs. Electrocution is always a bit embarrassing. I've got a 200W isolation transformer that I rarely use, but it *does* get pulled out when playing with switchmode PSUs. (hark, the sound of Oztules making chicken noises in the background!) Been saving those UPS transformers (240-24Vac) to make either a 48V batterycharger or a 1kW isolation transformer. Those UPSes get tossed in large numbers as their batteries wear out. Probably the same computershop that you suggested for the PC PSUs would also be able to supply you with UPSes.... Oztules once suggested another way (than your small transformers that fold-back) to more safely experiment with those PSUs, by inserting an incandescent lightbulb in the line. Any problems that would normally instantaneously destroy the PSU would now only cause the lamp to light brightly. You may have to experiment a little with bulbs of different wattage to find the suitable value. It's an old trick of the trade, so I figure I'm not telling you anything new here :) Peter. |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
G'day Dinges, yeah the light bulb trick is all good for limiting, but I also need isolation so I can connect a CRO or USB Scope to the switcher without killing the CRO, Scope or PSU (or me). The small transformers help by providing both limiting & earth isolation in one go. Also I find that a PSU can blow easily even with a tiny light bulb, as the main cap has enough stored joules to smash any transistor. So these days I also remove the main cap and fit say a 10uF one. Especially useful in low load testing. sPuDd.. It should work ...in theory |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
I was about to start poking a Yaha (Yeha more like it) ATX, It powered up and seemed OK with my ATX load tester on it, but I thought I'd just measure the rails & ripple before starting. 5V rail - 3.5V 12V rail - 8.5V Not good. My ATX load tester just pulls the PWR_ON lead low and applies about 100W or so of resistive load to the 3.3, 5 & 12V rails. It does not measure the voltage or reject it if it is too low/high like a mobo would through its PWR_OK lead. Looks like I'll have to fit a few LCD volt meters to it. Or build V2.0 with PWR_OK LED as well. So, upon testing several other ATX units I find the same thing. THUS why they were rejected! That'll teach me to check first. I've found a different brand working 600W victim, with nice big TO-247 package transistors and diodes! I'll poke around in that for a while
My ATX Tester V1.0: 
sPuDd.. It should work ...in theory |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
OK, new victim uses a custom made ATX IC WT7520 by Weltrend. It incorporates the comparator IC into itself. Unit claims to be a 600W PSU, Captain Spuddly will decide that after a good long soak test. Its transformer is twice the size of the "400W" units. Has nice big TO-247 switching transistors. I think I'll make this one a 13.8Vdc unit. I added a 5k multi turn pot where the diagram shows in the example on the error amp input, and I could get about 13.3Vdc before the over volt shut it down. At least it works. I've started pulling all the O/P stuff out and tomorrow I'll set it up and correct the over and under volt inputs so they trigger at the new range. I think there is a spare input I can use for current limit/shutdown. Should make a nice 40amp PSU
sPuDd.. It should work ...in theory |
||||
| Janne Senior Member Joined: 20/06/2008 Location: FinlandPosts: 121 |
Spudd, looking back at the last few pages you have some really neat stuff going on with these PSU's.. nice work. I assume that 600W box is rather new..? Looking at the pictures makes the new psu's much more desireable than what i thought before. If at first you don't succeed, try again. My projects |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
Janne, the 600W PSU is about a 2006 vintage. The IC spec sheet shows about a 2005 release date. Compared to the AT units, the ATX are better in all respects. More power for starters. Shake down your local computer repair shop. One note - avoid any ATX that says "Power Factor Correction" on the side of it. All of those I've thus far found have tried to "Correct" a surge or dip and gone up in smoke. Plus it’s a lot of extra SMD rubbish on the front end you don't want to deal with. sPuDd.. It should work ...in theory |
||||
| Dinges Senior Member Joined: 04/01/2008 Location: AlbaniaPosts: 510 |
[quote=Spudd]One note - avoid any ATX that says "Power Factor Correction" on the side of it.[/quote] Spudd, I've only seen one PC powersupply that had powerfactor correction (PFC); it was a completely normal ATX supply, with the difference that it had a large choke in series with the input rectifier. There were no extra parts inside that PSU for PF correction apart from that large choke. I removed it (was just taking up space and made working on the board harder) and ended up with a perfectly normal, easy to convert supply. The choke had the size of a small (50-75 W) transformer and accounted for much of the extra weight of the PSU. It wasn't mounted on the PCB but bolted to the metal casing. The extra weight (which I initially thought indicated oversized transformers, cooling fins, etc.) was what drew me to that PSU in the first place, but it turned out to be only caused by the useless PFC choke. Maybe powerfactor correction has become smarter over time? Your ATX dummyload looks pretty professional, btw. I just used a few H4 bulbs and a 12V/150W electric car heater as load.... 
(the upside of the lamps is that it also gives quite a bit of instantaneous feedback on what is happening to the power rails) 13.8V @ 40A... what else is there to say but WOW! Peter. |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
G'day Dinges, They seem to use an inductor (makes sense as the ATX is capacitive in nature) and seem to pulse it somehow with a MOSFET and some kind of correction IC, DIP or SMT. You'll see two well smoked inductors in these pics. I could stuff about and bypass it, but it's probably damaged the rest of the circuit. Edit: I've seen the ballast you speak of in another unit. I'd say that is Passive V's Active power factor correction. End edit. 
Funny thing is the second one was in a DVR, and I played it back to see the moment it died, hoping maybe to see something. Nope, just stopped in the middle of a regular night. Was less than 12 months old in an air-conditioned office on a UPS. PS Gizmo, I'm finding that half the pics I try to upload error when done, and I have to find them manually through my forum member info. Any idea what’s doing it? Also it leads to extra pics in the member folder, which is wasting drive space. sPuDd.. Edit: Spelling - hey I'm fussy! It should work ...in theory |
||||
vasi Guru Joined: 23/03/2007 Location: RomaniaPosts: 1697 |
I can confirm. Same problem, uploading from linux and win98. Hobbit name: Togo Toadfoot of Frogmorton Elvish name: Mablung Miriel Beyound Arduino Lang |
||||
GWatPE Senior Member Joined: 01/09/2006 Location: AustraliaPosts: 2127 |
The UPS probably killed it. Most of these UPS have modified square wave outputs. If the UPS supplied power, it would be likely to be poor quality. Gordon. become more energy aware |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
G'day Gordon, yep the UPS is a 600VA square wave job. It was passing regular mains when it died though. A cheap UPS is still better than a DVR that turns off with every power dip. The PSU I re-fitted just has the passive ballast style correction choke. It should last longer. I'll be avoiding the "Active PFC" PSU's in future though, more trouble than they are worth. Got the 600W ATX working, after feeding the comparators with fixed voltages. That part kind of makes the TL494 systems look simpler. Need a bit more cleaning up on output & back to error amp, but so far looks good. Will need a transformer rewind for full power and better diode packs. sPuDd.. It should work ...in theory |
||||
| sPuDd Senior Member Joined: 10/07/2007 Location: AustraliaPosts: 251 |
OK, finished playing with the 600W unit. The Weltrend ATX IC has its own drawbacks for modders. There is no simple and gentle way of adding current limiting. Having to preset all the over/under voltage comparators is a pain. Does not seem to like running with less than ~50W load. It turns the whole IC on and off in a chattering mode when it can't minimise the duty cycle. Never mind - it will donate its organs to another unit sometime. Pic of the mods: 
EDIT: Will have to try later, upload function not working atm... OK, tis working again. sPuDd.. It should work ...in theory |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Sorry folks been very busy lately. A few things I have noticed. The power factor problem is not of the typical type where we are using Vars or making Vars. I think that anything that distorts the power from being utilized 100% gets lumped in as power factor.... however, I think Rossw correctly pointed out that it was really crest factor that we are talking about here. In this instance, we rectify the AC and after that charging up the caps or driving an inductor is a DC problem... it is not the classic rising and falling of a magnetic field in an AC environment (lagging), or the charging of a capacitor and discharging it... (leading)... so we can't cancel out the poor power factor with an inverse arrangement. The poor power factor in this case is caused from only using the crest of the waveform. None of the wave is used until the caps voltage is less than the incoming DC voltage... so there is an incredible surge of current from zero to maximum it can get at the top of the wave. This distorts the wave,(try's to make it square in fact) and harmonics and hash get sent back up the line. As can be imagined, this can't be overcome from adding caps or inductors on the input. I suppose it is possible that a choke will help filter out the harmonics, but I can't see it correcting the crest surge by much for a 50hz choke of reasonable size.... I may be wrong, but it feels right.... comments welcome. So the only real way I know of correcting the power factor in this instance is by using a boost converter stage before the real switcher. (the charred bits by the looks of it Spud) Even the inverter air conditioners use this in the front end... (huge reactor torroids.. yum) It depends on the output voltage of this converter as to whether you can just bypass it.... probably can... if it is a 400-450v cap or there abouts it is a fair bet you can just full wave rectify and feed that to the 400v cap toss out the fet, diode and reactor coil, and let it run from there. Sad that you couldn't get the 600watter running well. What chip is the PWM chip? I can't see how you tied up the error amps, and still had some load control.... better man than me. I will endeavor to get some photo's up... (trying to find the USB cord) Sorry Janne, this thread has life of it's own from the looks of it. ..........oztules Village idiot...or... just another hack out of his depth |
||||
| Janne Senior Member Joined: 20/06/2008 Location: FinlandPosts: 121 |
It's fine for me. Besides, it's still pretty close to the original subject... me thinks. I've been rather busy with other stuff, but have scrounged a few more atx boxes with TL494's inside.. Both have "normal" layouts. Will start playing with them again as the worst spring rush of work passes. If at first you don't succeed, try again. My projects |
||||
| Page 5 of 8 |
|||||
 Print this page |
