Agilent E3648A Power Supply Transformer

[ddcc]

I recently picked up a broken E3648A power supply that doesn't power on. Since the schematics are available online, and I have a working E3648A unit, I figured this one shouldn't be too hard to repair.

Unfortunately, it seems that the fault is inside the transformer, which is a beast. The primary side has two windings, white-black and yellow-red-orange (red is center tap), to support 100V, 115V, or 230V operation. Since I'm in the US, which is 120V, the two primaries are connected in parallel, white-black and yellow-orange. In the schematic below, both switches S100 and S101 are in the up position for 120V. When I measure resistance, there is no connection between white-black and red-orange; only yellow-red is still connected. Of course, this means that unit is completely inoperable, because even for 100V or 230V operation, the first white-black winding is connected in series to the second winding.



My next thought was to measure the secondaries, to see if I could buy an equivalent transformer. It turns out that there are nine secondary windings, of which six have center taps. The voltages aren't labeled on the schematic, but since I have a working unit, I can just measure them myself. Not including the center taps, these voltages are 27, 27, 29.4, 29.4, 38.8, 22.4, 8.8, 8.8, and 5.8. Crap, this is a completely custom transformer.

Hmm, so perhaps I can get a replacement from Agilent. The part number is E3648-60013, and although all four models of the E3646A - E3649A series are very similar, the transformers appear to be different for each unit (part numbers E364x-60013), probably to save cost. The plastic bobbin for the primary does have E3648A/E3649A printed on it, so this appears to be reused between the two units, but the actual winding might be completely different. But, E3648-60013 doesn't show up in their online parts store, so I doubt it's something they'd sell directly. There's also not much printed on the transformer, so I don't know if these are from an outside vendor.

I briefly also considered rewinding the transformer myself, but the whole thing is completely coated in varnish, and given the complexity of the transformer, I doubt that I can disassemble and repair it without breaking something. The transformer also appears completely fine; there are no signs of charring, scorching, or any sort of damage. And, the leads appear fine, up until they enter the transformer assembly. For reference, here's a photo.



The whole situation is unfortunate, because cosmetically, the unit is in great condition. Plus, when I temporarily swapped the transformer from my working unit into this one, it works fine; passing self-test and regulating the output. So, I'm stumped, because I don't know how to just repair or replace the transformer. Does anyone know how much Keysight would charge to repair the unit? Or any other suggestions?

[Samogon]

Sorry to hear that.
I think it would be cheaper to find another broken unit with good tranformer, than gettint new from keysight especially taking into account that keysight sells only to businesses.

[skennedy]

This is a bid of a sorry story. Best of luck getting this power supply going again.

I wonder if you have checked the winding that are open close to the transformer itself? I would have suspected that the most likely point of damage would be the transition between the free hanging wires and those wrapped around the bobbin. Perhaps piercing the wire close to the bobbin to see if there is a break in the cable?

I know that it isn't shown on the schematic but possibly there might be a thermal fuse inside the transformer in series with the winding. I have repaired a Yamaha AV receiver a few years ago by cutting away the plastic bobbin and revealing the thermal fuse. It actually wasn't as terrible to fix as it might sound.

I have a couple of these similar Agilent E36 series power supplies and the transformer certainly is a work of art (and a lot of mass)

[duak]

Toroidal transformers are generally smaller and lighter than EI core transfomers and often have two primary windings and two secondary windings.  I'll bet there are only a few windings in the E3648A transformer that have to supply ampere level currents.  The rest could use smaller transformers and thus isolate their circuits from each other. 

If the toroidal transformer is unpotted, you can add extra windings by winding them on top of the existing ones. These windings could also be used to tune the voltages of particular windings by wiring them in series to either buck or boost its voltage. 

To maintain the input line voltage range and simplify the primary wiring you might also connect the primaries of the smaller transformers together and drive them in parallel with one primary winding of the biggest transformer.

I found Signal Transformer, Hammond and Stancor to be a good sources of various sizes of transformers.

I got a Wavetek 23 function generator with a burnt power transformer.  In the end I was able to find a Stancor transformer that was close enough to generate the required voltages with a few extra resistors.  Before then though, I worked out a scheme using a couple of toroidal core transformers to generate the exact voltages needed.

Hope my description makes sense.

Best o' luck

[HighVoltage]

After repairing many of these Agilent power supplies, I never encountered a broken transformer, because they are usually overkill.

If you apply primary voltage, you get any output?
I would test all primary and secondary windings under load.
You might find a simple problem.

[ddcc]

I wonder if you have checked the winding that are open close to the transformer itself? I would have suspected that the most likely point of damage would be the transition between the free hanging wires and those wrapped around the bobbin. Perhaps piercing the wire close to the bobbin to see if there is a break in the cable?

I know that it isn't shown on the schematic but possibly there might be a thermal fuse inside the transformer in series with the winding. I have repaired a Yamaha AV receiver a few years ago by cutting away the plastic bobbin and revealing the thermal fuse. It actually wasn't as terrible to fix as it might sound.

I'll double check, but the wire feels very solid, leading into the transformer. I'm also a little hesitant to cut off the outer insulation on the transformer to look at the windings. The primary windings also appear to be underneath the secondary windings.

Toroidal transformers are generally smaller and lighter than EI core transfomers and often have two primary windings and two secondary windings.  I'll bet there are only a few windings in the E3648A transformer that have to supply ampere level currents.  The rest could use smaller transformers and thus isolate their circuits from each other. 

If the toroidal transformer is unpotted, you can add extra windings by winding them on top of the existing ones. These windings could also be used to tune the voltages of particular windings by wiring them in series to either buck or boost its voltage. 

Hmm, so it is possible to retrofit an off-the-shelf transformer. Yeah, only two of the secondary windings are for the main output channels. The rest are for the various bias supplies on the bottom board, the bias supplies on the control board, the VFD filament supply, etc.

After repairing many of these Agilent power supplies, I never encountered a broken transformer, because they are usually overkill.

If you apply primary voltage, you get any output?
I would test all primary and secondary windings under load.
You might find a simple problem.

Yeah, this is not a failure I expected. There's no output at all when the primary is connected, and the transformer doesn't even hum. The unit also used to work fine before the transformer broke; I had to break an Agilent calibration seal to open it.

[ddcc]

I called Keysight, and they were surprisingly helpful. They confirmed that the transformer used on each of the E364xA series is different, gave me the internal part number for this transformer, and also gave me the vendor that supplies the transformer. But, since the only supported repair method is whole unit exchange, they won't sell me the transformer individually.

I subsequently contacted the transformer vendor, but they refuse to sell custom designs to anyone but the original customer.

So, I will either need to figure out a transformer repair, or find a parts unit.