1

I have searched the web thoroughly and find very little info on smps transformers. How to take them apart ... bare bones operations in Power supplies etc. I have a smps that still doesn't work after replacing ALL burnt components. So I'm thinking the transformer bit it as well. AFter blue ring testing seems the thing is toast. Took it apart to rebuild. Noted as much as I could and remembered the rest. This transformer is pretty simple ... 2 primaries and two secondaries ... primary one (pins 1 & 2 ) is from the mains measuring approx 165V and is part of the main switching transistor to produce +18V and -18V on it's secondary which of course has a center tap on secondary for the split supply. The 2nd primary is for the 5V supply for the digital supply. It's pins 8 & 9 on the schematic ... it's primary input voltage is the same as the 1st primary on pins 3 & 4. So when I took it apart the winding ratio's didn't make any sense for the respective ins and out voltages.

Primary 2 ... pins 3 & 4 was approx 44 winds @ 28 awg. It's secondary was 4 winds @ 23/24 awg. This one sorta makes sense. This primary was the innermost wind ... it's secondary was the outermost wind.

The 2nd primary was only 2 winds (possibly 1 wind) and it's secondary winds were not counted but was at least say 20-25 winds. Insulating tape was used between all winds but not sure how thick each layer was so subsequent winds after the 1st weren't known exactly.

So problem is I need secondary voltages of +18 and -18 on the 1st primary and secondary then need 5v from 2nd primary and secondary ... according to what I know about turn ratios they don't line up.

SMPS power supply that is in question

After rebuild 1 I had 5v supply working but found out that priamry one should be attached to pins 3 & 4. Got that fixed ... haven't finished the rebuild at that point. BUT the windings for second primary and secondary don't make sense for +-18V One or two winds for primary and say 20-30 winds dual for split secondary. +-18V

Any help would be appreciated

Kevin Smith
  • 11
  • 4
  • One odd thing that has happened using the blue ring tester tesing the winds as I do them. Wind the 1st primary and it rings ok the wind 2nd primary and it rings ok but then 1st wind is bad again ? WTF ? – Kevin Smith Aug 11 '18 at 12:20
  • Of course with the ring tester the primaries without the secondaries don't ring like they should but I do see the ringers pulses. – Kevin Smith Aug 11 '18 at 12:21
  • Also note ... when taking apart the unit I broke the lower and upper portions of the E part of the ferrite core ... I am hoping crazy glue can reattach these arms and still function. I have done this with ferrite cores for magnet motors without issues. – Kevin Smith Aug 11 '18 at 12:24
  • Since this is an audio piece of equipment should I just build a linear supply for it ? Problem is a linear won't fit inside the old box ... will have to have a separate box for it. – Kevin Smith Aug 11 '18 at 12:26
  • It's an SPX90 effects unit for studio or live effects – Kevin Smith Aug 11 '18 at 12:33
  • The 2nd primary (pins 1 & 2) was wound on top of inner most primary (pins 3&4) hen it's secondary wound on top of it and lastly the 5v secondary was the last outer most wind ... so two primaries then two secondaries in that order from inner to outer. – Kevin Smith Aug 11 '18 at 12:39
  • Advice: Unless you know exactly what you are doing then get a new one. You are unlikely to appreciate the insulation requirements between primary and secondary and you could end up destroying valuable equipment and putting users of that equipment at risk. That transformer sounds like it's been mangled enough. Search on line for something compatible and don't take risks with this one. – Andy aka Aug 12 '18 at 17:03

1 Answers1

0

It seems to me that you are confusing some things. The two primary windings are on the LEFT in the schematic, pins 1-2 and 3-4. The two on the RIGHT, are secondary windings, pins 5-6-7 and 8-(9).

The 2nd primary is for the 5V supply for the digital supply. It's pins 3 & 4 on the schematic

It doens't work that way. In a transformer, all the coils influence each other. If I understand the schematic correctly, the primary coil that 'does all the work' (power goes IN) is the Pin 1-2 winding. The 3-4 primary winding is for feedback and power for the primary switching transistor/circuit. So although it's on the primary side, power doesn't go IN, but OUT. I'd call it a primary winding, because it's on the live side of things. But maybe to call it that is wrong, and it should be called a secondary winding? I'd be interested to learn how that works actually. I guess it depends on how you look at it.

primary one (pins 1 & 2 ) is from the mains

Primary 1 ... pins 3 & 4 was approx 44 winds

Let's stop the confusion, and call Pins 1-2 Primary 1 (P1), and pins 3-4 Primary 2 (P2). Pins 5-6-7 Secondary 1 (S1) and Pins 8-(9) Secondary 2 (S2). Maybe you should edit your post to match...

So count the windings of P1. Then count the windings of S1. This should be something like 9:1:1 (for 50% duty cycle at 165V). So for every 9 winding on P1, S1 Pin5-7 has 1 winding and Pin 6-7 has 1 winding (2 windings in total). Of course there are guestimates, the actual amount depends on the desing of the SMPS. Also take note that the direction of the turns matters!! The turn direction is marked by a dot in the schematic.

Then for S2, you still compare with P1. This should be something like 33:1 (again, for 50% duty cycle). Update: Apparently this SMPS has quite a low duty cycle at 165VDC, see corrected values at UPDATE below.

I don't know about your broken E-coil, but I'd say to glue it is a long shot, and you'd be better off with a new transformer... But I might be wrong.

Also note that not all broken parts will seem 'burned'. Usually the main switching transistor blows, taking more parts along with it. You can't always find the faulty parts by eyesight.

I hope you'll be able to fix it. Just make sure you insulate the primary windings from the secondary ones properly. Use the right tape for it, as your safety depends on it!

Hope this helps. Good luck!

UPDATE: So the correct number of winds should be:

  • P1 (pin1,2) has 44 winds
  • P2 (pin3,4) has 2 winds
  • S1 (pin5,6,7) has unknown winds
  • S2 (pin8,9) has 4 winds

 

  • Up1:Us2 = 165:5 = 33:1
  • P1:S2 = 44:4 = 11:1
  • 33/11=factor 3

 

  • Up1:Us1 = 165:18 = 9:1
  • divided by factor 3 gives you:
  • P1:S1 = 3:1

Since P1 has 44 winds, 44/3=14.7 (approx. 15). That would mean that S1 should have 15 winds for 18V (so 2x 15winds for 2x 18V -or- 15 winds between pin6-7 and 15 winds between pin7-5)

You can read more on how it works here

About turn direction:

Look at the dot in the schematic. If you connect the start of the wire to the dot, start turning in one direction, say CLOCKWISE. Then for the other coils, also connect the start of the wire to the dot and start turning in the SAME direction - in this case I chose CLOCKWISE. This way you'll do the direction right. You can also choose to turn CCW, but then you should do the other coils also CCW.

IMPORTANT
The dot on P2 is in the wrong position. The dot in the schematic should be on pin4, and NOT on pin3!

About dot notation:

Yeah, the dots... there's so much confusion about them, and I've looked at many schematics online only to find so many dots in the wrong place.

  • P1 is the coil that creates the magnetic field. The + is connected to the dot, so current flows from the dot terminal to the non-dot terminal. Remember that.
  • S1 is a coil that does the opposite. It turn the magnetic field into a current. Here, the current flows in the SAME direction, so from dot to non-dot. To make that happen, the + must be at pin 6.
  • Same for S2. Current flows from pin9 to pin8, so the + must be at pin 8. That matches with D8 and D10, otherwise nothing would happen.
  • For the + to be at pin3 of P2, current must flow from pin4 to pin3. So the dot must be at pin4.
Community
  • 1
MartinF
  • 1,230
  • 7
  • 13
  • Hi Martin, thanks for pointing out my errors in the post. I corrected it. Your tips were interesting especially about that wind at pins 3 & 4 ... that one is throwing me for a loop. Feedback sounds like a good conclusion but who knows ? About the ratio you stated, shouldn't it be Hi:Lo instead of Lo:Hi ? To step down ratio should be for example 4:1 no ? I found another core same dimensions as the old one so can build a second one after I rebuild old one and see if I can get it going. My insulation tape is glass cloth class B insulation. It seems to be doing the job since I ran it once already – Kevin Smith Aug 12 '18 at 15:28
  • No sparks flying or magic smoke getting out. I couldn't find anything else at the electronics supplier in town and they really didn't have any idea what kind of tape would be used for transformers anyway. Looked online and didn't find out much either. Things like paper, oil and one other used in large power transformers ... nothing about small stuff. Although I know for flybacks that use high voltages one needs something like silicone or similar to insulate from arcing. – Kevin Smith Aug 12 '18 at 15:29
  • After the 1st rebuild, I found a note I made about the most inner wind ... seems it was connected to pin 3 & 4 so I had 1 & 2 and 3 & 4 reversed which would definitely cause a problem. Weird tho is that the secondary for the 5v supply worked, had just over 5v when powered up but nothing on pins 5 to 7. Currently correcting that error and rebuilding again will get back here with results. – Kevin Smith Aug 12 '18 at 15:44
  • Some extra note about number of winds I know for sure. Pins 8 & 9 had 4 winds. pins 5,6,7 were not counted ... forgot. Pins 3 & 4 had approx 44 winds and Pins 1 & 2 had 1 to 2 ... I'm leaning towards two due to the actual length of the wire then when wound it was 2. So I'm trying to number crunch your suggested ratios and they don't seem to jive. – Kevin Smith Aug 12 '18 at 15:59
  • Some extra note about number of winds I know for sure. Pins 8 & 9 had 4 winds. pins 5,6,7 were not counted ... forgot. Pins 3 & 4 had approx 44 winds and Pins 1 & 2 had 1 to 2 ... I'm leaning towards two due to the actual length of the wire then when wound it was 2. So I'm trying to number crunch your suggested ratios and they don't seem to jive. If pins 1 & 2 is the main/only primary at 2 turns and pins 8 & 9 was 44 winds then the ratio ends up being 1:22 not 1:33. – Kevin Smith Aug 12 '18 at 16:11
  • One question about the wind direction you mentioned to take note of ... I didn't make notes when disassembling but I'm thinking to wind all in the same direction ? keep it all same ? – Kevin Smith Aug 12 '18 at 16:15
  • Yeah, I'm sorry about the turns ratio, you're right. I got confused as well. I fixed it in the post. I find it hard to believe that on pins 1-2 was only two windings. Of course, I only have the schematic you've posted, not the physical unit. Are you sure you're not mixing things up? Because looking at the schematic, I see a path of 165V-P1(pin1,2)-Q1-R13(1Ω). Not good for just 2 winds. 44 windings sounds much more plausible, and 2 windings for P2(pin3,4) also sounds good to me. I'll post an update about the other numbers. – MartinF Aug 12 '18 at 16:52
  • As you can see in my suggested turns, I assume pins 1-2 have 44 winds and pins 3-4 two winds. Look at how the transformer fits into the PCB, and compare it with the schematic. Pin1 must be tied to R2 and *this* winding should have 44 turns. I refuse to believe otherwise ;-). Good luck! – MartinF Aug 12 '18 at 17:25
  • Just for fun, check out D1: The + and - sign are wrong... :-) – MartinF Aug 12 '18 at 17:59
  • Hi Martin, thanks for the update ... I think I see now how you came up with those numbers now. After winding the split wind for the +/-18v it looks like 15 is very close as after my wind I counted 11 using the length of the old wire. Interesting about the dot on the schematic re the starting point for the wind ... in all my teachings i got in school for electronics this was never talked about. – Kevin Smith Aug 13 '18 at 01:32
  • So I'll run with your suggestions on the second core I found. Just finished rebuild on the old one again and ready to ring test and install. Interestingly enough when I wired the 44 winds to pin1/2 I did get 5v on the corresponding secondary but nothing for the +/-18v end of things. I'll let ya know how things go. And thanks again for your input. – Kevin Smith Aug 13 '18 at 01:36
  • Hi Martin, ya bout D1 ... I noticed that early on when I was troubleshooting for blown components. So one mistake in the schematic might think there could be a few more like the dot for P2 might be reversed. SO the dot indicates polarity just like on other components with polarity ? Seems right to me ... – Kevin Smith Aug 13 '18 at 02:17
  • The dot at pin 3 seems right if the dot indicates pos polarity since pin4 is only 1 ohm away from ground. But interestingly enough on the secondary side the dots are reversed ... on the ground end of things. But then you also mentioned the dot is the starting point for the beginning of your winds so not sure myself how that affects any polarity. Never got much teaching on transformers in school. We were all about digital for the most part of the course. – Kevin Smith Aug 13 '18 at 02:21
  • One more note regarding those dots ... I vividly remember when I took the very 1st wind off at pins 8 & 9 that pin 9 was the outer most end of the wire where the dot is at 9. Also looking at pins 5 & 7 they have dots and when I took that one off the 2 wires were joined at pin 7 and were again the end of the wind as the 2 wires were bifilar wound side by side. At least I'm 90+ % sure it was that way. – Kevin Smith Aug 13 '18 at 02:34
  • I've updated my answer once again. So the dot doesn't indicate 'polarity', as that leads to confusion, because polarity for the coil *inducing* the magnetic field (P1) is reversed from the coils doing the opposite (P2,S1,S2), as you've already noticed. It makes more sense to use current. The current flows from dot to non-dot (or opposite, depending on who drew the schematic). That way it'll always work right. Then you'll also come to the conclusion that the dot of P2 is in the wrong place. I've studied the schematic some more, and I'm now certain of this. **The dot needs to be at pin4.** – MartinF Aug 13 '18 at 11:54
  • What you say about pin7 (center tap) being bifilar wound makes absolutely no sense to me. Start winding at pin5, make 15 winds and end up at pin7, wind another 15 in the SAME direction, and end up at 6. That's the only way it will ever work. The only way a bifilar winding could work, is if you start at 5, wind a bifilar (2 wires) coil of 15 winds to 7, and tie one of the wires to 7, but continue with the other (without attaching anything) to wind another 15 winds and end up at 6. But efficiency wise that doesn't make ANY sense to me. I'd say stick to the plan :-) – MartinF Aug 13 '18 at 12:02
  • It's quite possible you got 5V with improper windings, as the schematic has feedback. PC1 signals the primary side to shut off when the secondary side reaches 5V. But with improper windings, the duty cycle will be completely off, and the 18V part will not work properly. Also under load, things will collapse and/or destroy some parts in the process. So it's important to get the turns right. I'm pretty positive, that if you follow the numbers I gave you, and use the right directions, including the dot at pin4, it'll work (unless some other parts are broken as well, as I warned about.) Good luck! – MartinF Aug 13 '18 at 12:07
  • Hi M, here's a good post here about the dots. https://electronics.stackexchange.com/questions/66708/dots-in-a-transformer-symbol#66709 – Kevin Smith Aug 13 '18 at 12:59
  • Yeah I saw that one. There are actually a LOT of posts about the dots, even on EE.SE. I saw many of them, but I still think talking about 'polarity' can be confusing. It depends on the way your brain works I guess :-) – MartinF Aug 13 '18 at 13:03
  • It's phase, which does not necessarily mean polarity. Also saw a photo I took where I see pin 9 being inner wind or start of wind so I was wrong in previous comment. After looking at the schematic I'm confused again coming up on winding the split supply. I got 44 winds on pins 1 & 2 again and the 2 winds on 3 & 4 per your suggestion which I agree with ... makes much more sense. When I disassembled ... 95+ % sure pin7 was 2 wires and end of wind but dots on schema woudn't suggest that. I made note for that wind was that pin6 was most outer wire from a bifilar wind starting @ pin7 running CW. – Kevin Smith Aug 13 '18 at 14:42
  • Next time I'm gonna run a video cam for the whole disassemble !!! I made notes and took photos and still lost a lot of info. – Kevin Smith Aug 13 '18 at 14:57
  • I rest my case, as I've said what I needed to say. I won't change my mind about the winds of S1, unless the schematic doesn't match the PCB, which seems unlikely, but possible considering the other two mistakes in the schematic. Please consider accepting my answer. Thank you. – MartinF Aug 13 '18 at 15:23
  • Hi Martin, thanks again for all your help. I'm implementing all of your suggestions here and making headway. I just tested a build and got some life out of it, of course all the wrong voltages. I did wind that split wire S2 my way and it isn't working so will do as you said and wind it like that. I was thinking of doing that before I even saw your update. Gonna try one more time to get it going. If can't then I have to build linear supply for it. Maybe you could suggest pass along a schematic for linear +/-18v and +5v rails. – Kevin Smith Aug 14 '18 at 04:16
  • I more or less have an idea how to build a linear for this unit but was wondering about VA rating for the transformer. The unit is rated as using 20W according to label on back of unit. So I'm thinking what ? 25 to 30 VA for 20 watts ? – Kevin Smith Aug 14 '18 at 04:19
  • Well, if the -18V was wired improperly, it'll never work as the feedback circuit with PC1 needs the -18V to apply proper feedback. Because you've tried several times with wrong wiring, other components might have failed by now. These things are really not to play around with, and SMPS is usually not very forgiving. Be careful and don't have stuff blowing in your face please! I'm not even kidding. – MartinF Aug 14 '18 at 11:27
  • As for your question about a linear PSU. That's really a different question. I suggest you do some research, come up with a proposal and post that as a new question on EE.SE. Problem is, you know the total power consumption, but not how much current you need for the 5V, the 18V and the -18V individually. But if you post the SMPS schematic again, maybe someone will be able to figure it out, looking at the size of the inductors and the capacitors at the output? I don't know. – MartinF Aug 14 '18 at 11:34
  • Also consider buying a ready made PSU, as the voltages you have are VERY common. Maybe [something like this](https://www.ebay.nl/itm/TRACO-POWER-TOF15-Triple-Output-Embedded-SMPS-2A-5-V-dc-15-V-dc-H7GR-4666783/223096888254?hash=item33f19c5bbe:g:V7EAAOSwRLZaexsr). This one might not have the right power, but you get the idea. Besides +5V, it's got +/- 15V, but that might still work. – MartinF Aug 14 '18 at 11:36
  • Hi, Just finished another build exactly how you instructed. Waiting for glue to dry for the core and lid. The lid had some plastic spacers so I'm guessing an air gap on this unit, I hope it's close to spec. That ready made looks promising save that the +/-18v supply had to supply 2 more regs for +/-15. Stupid design if you ask me. Since this is audio gear, they should have used a linear from the get go but would have probably been far to big to fit in the chassis and more expensive. I'll let ya know how this last effort goes, if fail then it's linear in a separate box, as it will not fit box. – Kevin Smith Aug 14 '18 at 12:58
  • Hi M, ok just one more question, Final build is done, before I powered up I rechecked a bunch of parts. Low and behold Q2 popped again and the part in there wasn't the original but replaced with one that came close. Now the new one is even closer as the original is no longer available. Other than that all was still good to go. SO question ... can you tell from the schematic if this smps will run without a load ? I powered it up without load before but that may have caused a shut down. I'm worried about damaging the digital end of things if the 5v supply is wrong. No regulation on 5v supply. – Kevin Smith Aug 15 '18 at 06:21
  • If I want to power up without connecting to rest of circuit ... what would you suggest to use as a load for the two voltage rails ? Say a couple of high watt resistors but at what value ? I'm gonna say high value like maybe 5 to 10k ? – Kevin Smith Aug 15 '18 at 06:23
  • So Q2 broke. Was it open, or shorted? Because if it shorted, you might also want to check Q1 and PC1 and all the diodes on the primary side. You see, it's all connected to each other and if one thing goes, more goes. There's no reason why Q2 would break if there isn't something else broken as well. Unless the replacement part wasn't as close of a match as it should've been. **I would surely NOT connect the PSU with the rest until you know it has the right voltages!** If you want to load it, probably 15-100mA would be fine. You could try 220R/0.25W for 5V, 1k/0.5W for 18V and 1k/0.5W for -18V. – MartinF Aug 15 '18 at 11:13
  • If you get somewhat good voltages at both the 5V and the +/-18V, you can use VR1 to adjust the 5V. The 5V is critical, and the 18V will follow the 5V somewhat, but you can't adjust those separately. So adjust the 5V to measure 5V with VR1. If +/- 18V is too far out, you could add or lose one or a few windings on S1. _For now: Pray for a miracle ;-)_ I hope it works, but I have my doubts. Good luck! – MartinF Aug 15 '18 at 11:17
  • By the way: Yes, the SMPS should be fine without a load, but since the -18V wasn't ok, there was no feedback and that's bad. Also with Q2 broken, depending on how it broke (shorted or not), Q1 was on all the time, which will destroy it because without Q2 the overcurrent protection with R13 won't work. Also if the 5V is too low, feedback will continually tell the circuit to give more, which will continually trigger the overcurrent protection, causing lots of stress on Q1 and the transformer, ... So better check well if all is ok _before_ you turn anything on! It's really all connected. – MartinF Aug 15 '18 at 11:25