1. Home
  2. Electronics Design
  3. Rewinding an SMPS (again) More Questions

Rewinding an SMPS (again) More Questions

I made the decision to disassemble an SMPS transformer from a power supply (which I do not have a schematic) for rewinding. My pre-disassembly electrical tests showed that the primary had an internal short. During my disassembly process I soaked the transformer in a strong paint stripper to attempt to remove the varnish/shellac that the unit was dipped in. This worked well, but not work well enough to allow the E-cores to be removed. Thus, they ended up breaking from too much force. I then managed to start the unwinding process, keeping track of the turns and CW/CCW directions of the two secondaries. When I removed the primary, I ended up not documenting the whether they were wound CW or CCW relative to the secondaries.

Finally, the questions:

  1. Not knowing whether this particular supply was designed as a flyback or feed forward leaves the question of which direction to wind the primary relative to the secondary. My limited knowledge of smps transformers says that with a flyback xmfr, the primary is out of phase with the secondaries. Is this correct? Is there a way to tell if the design was intended as a flyback topology?

  1. Since I severely damaged the E-cores, I need to find a replacement. The cores shattered into many pieces and cannot be glues back. I do have large piece that I can use for testing and getting dimensions from. My thought is that I can wind a test inductance and determine it's Al (inductance per turn squared) . I do know that the switching speed is 40 KHz, so with these parameters, I should (I hope) find a replacement that will be compatible with the original. Does this sound reasonable? I think that I should be able to use the original bobbin.

  2. The last question, is where is a good source of ferrite E-core material where I could purchased just a few pieces? Maybe bobbins, too..

I studied the construction of the transformer and see where the mistake was mad in the primary that caused a short between the windings. Thus I plan on taking extra precautions when I get to finally winding my version of the unit.

Your thoughts and tips are appreciated. Thanks Jim WB5KYE

Reply to
Jim Flanagan
Loading thread data ...

This sounds like more work and responsibility than you anticipate. It makes more sense to replace the supply.

That said; what was the make and model of the power supply?

If there are largish output inductors (or one inductor with multiple windings), then it's a forward converter. If the inductors are all small 'hash' types, it's a flyback. If the unit is rated above about

75W, it's unlikely to be a flyback. It may be possible to tell more, based on the winding info. A flyback would introduce the additional problem of determining gap size originally in the destroyed cores.

Measure the bobbin pin orientations (breadth and spacing) and identify which windings (turn count) terminated on any/each particular pin. Determine pin numbers, if molded/stamped on the body/board, or assign your own numbers for reference. Pin number one is often identified by a non-symmetry of the shape (cut corner or missing/staggered pin).

Measure the bobbin's center leg dims/diameter, and the length of the center hole (within ~ 1mm or 0.05 inches). This alone should allow identification of the core size/shape.

Do you have access to appropriately sized polyester film insulating tape? Did you record the winding and insulating tape layering/count and margins? These would have to be slavishly reproduced, without more general knowledge on your part.

If you returned this supply under warranty to the vendor or mfr, with the same fault correctly diagnosed, they would not likely repair it - they would replace it, where possible. This should indicate the most sensible course of action for you, as well.

RL

Reply to
legg

Hi.. Thanks for the response. The power supply is a 'dead' supply made by a company, Micro-Energy, which I believe is no longer in business. They sold their supply to HP/Agilent back in the early '90s for use in a line of HP Spectrum Analyzers. Thus, HP no longer supports this supply and with the OEM out of business, I'm stuck with either finding a junked spectrum analyzer that I might be able to steal the supply from or repair this one.

I'm attempting to make this a learning experience, as I've never really been involved with switching supply design before. The rebuilding of the transformer can be (with some luck) a fun experience.

Here is what I know at this point. The exact bobbin used is the COSMO

61531-0-515 and have some parts on order. The ferrite material size is a standard EE core and Ferroxcube Outline number E34/14/9 would be a match. Unfortanely, the ferrite material itself, I'm not sure of. I wrapped 20 turns around the dissembled transformer (the ferrite was still running through the center of the bobbin) and came up with an Al of 80nH/N^2. Don't know how accurate this is, but I will choose some ferrite material which is adequate for a 44KHz switcher (this I know).

The insulation tape that I have access to is Scotch #92 which is 2.5 thou Kapton, I believe. During the disassembly I also took notes as to the stackup of the windings. The primary winding was two layers (innermost) of 25 turns each. However, the mfg didn't insulate between these two layers, they wrapped the two on top of one another. This is where the shorting occurred. They did appear to use some sort of 'dope' on the windings, but I'm not sure what that substance was (Red 'Q' dope?).

How critical is the actual ferrite material to this process. If I obtain material which is specified for this

40KHz frequency and is the same size and shape, will I probably be OK, so long as my reconstruction technique is correct? Also, the supply is a 100W unit with large 1" toroid inductors on the output line, so I am going to go with forward converter when I wire the transformer.

Thanks again and if you have any further thoughts, I would appreciate them. Jim

Reply to
Jim Flanagan

So that was before the ferrite had been damaged?

80nH/T^2 might do well for a small flyback ( Also, the supply is a 100W unit with large 1"

Please don't assume anything. Read the circuit and prove it. At a glance: does it have one or two switching transistors? Does it use a single diode or two, and how? Flyback uses just one, half wave forward uses two (one to GND), full wave forward uses two (and a CT'd winding).

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
Reply to
Tim Williams

Model number of the HP SA?

HP assembly part number?

All this is good, except for the HP part. Not so easy to get good service docs, which would allow a better understanding of the broken part.

SAs and other analog measuring equipment often use non-standard or modified topologies to match the supply to the application. A common concern is noise - so hard switching tends to be avoided, where possible, by resonating the conversion current or voltage waveforms.

Without a complete ferrite loop with minimal physical gap, the material permeability is not derivable, as air dominates. If the structure running through the center of the bobbin is still intact, a physical gap, if present, may be measurable. Look more closely to see if the ferrite parts actually meet at the centre.

Listed 3M insulating tape should work if wide enough to cover the winding breadth. Note that the tape is actually composed of 1 thou tape and 1.5thou adhesive. Safety layers are built up using three layers min.

I'm surprised that it got by an HP design review. A layer of film tape is hardly a budget breaking item. They are/were usually more perceptive and beat on suppliers for demonstrated PPM defect rates.

You should be able to map out the secondary fairly easily, even measuring discrete inductor values, but the large chokes do indicate the likelihood of a basic forward's functioning. As I've said, though, this may be altered by a resonant or quasi-resonant primary, and just be an indication of massive filtering after direct rectification in the secondary.

RL

Reply to
legg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.