Author Topic: Power Supply Line Filter with Trifilar Windings  (Read 3115 times)

0 Members and 1 Guest are viewing this topic.

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
Power Supply Line Filter with Trifilar Windings
« on: March 08, 2020, 12:16:56 pm »
Below are 2 photos of AC line filter "LF2" from two different Apple Macintosh 128k~Plus Analog/PSU Boards. The newest board is stamped 4192 (1986 PCB), which I assume means 42nd week of 1992, and uses the LF2 without the green Earth wire. The other PCB is a 1983-84 version stamped 0792, which I assume means 7th week of 1992, and you can see it has the green earth wire wound inside LF2. Indeed, most every Analog Board I've seen has the Earth wire wound in LF2 (i.e., trifilar windings). Search this PDF for "LF2" to see the schematic, which also shows it to be trifilar.
 
Why do you think the green Earth ground wire was not wound into LF2 on the 4192 board?
 
945486-0   945490-1

The 3-pin power connector mounts on the PCB, but the Earth pin has no PCB trace. Regardless of which board version we examine, the green Earth ground wire coming from the power cord screws down to the metal chassis of the computer so the chassis is Earth grounded. My question is centers on why that green Earth wire does or does not go through LF2.
« Last Edit: March 08, 2020, 12:18:31 pm by JDW »
 

Offline jonpaul

  • Super Contributor
  • ***
  • Posts: 3656
  • Country: fr
  • Analog, magnetics, Power, HV, Audio, Cinema
    • IEEE Spectrum
Re: Power Supply Line Filter with Trifilar Windings
« Reply #1 on: March 08, 2020, 12:33:11 pm »
Hello: This is a common mode EMI choke, and cancells the currents flowing common to both line and neutral (balck/white)

The green ground isnot needs.

Jon
An Internet Dinosaur...
 

Offline bsdphk

  • Regular Contributor
  • *
  • Posts: 212
  • Country: dk
Re: Power Supply Line Filter with Trifilar Windings
« Reply #2 on: March 08, 2020, 12:45:56 pm »
The ferrite is a formidable impedance in high dV/dt scenarios so maybe the idea is to cause incoming surge energy to arc from the live/neutral to the pe-wire at the toroid, rather than further into the PSU, without expending a MOV or spark-gap ?

That could make a difference when a drunk driver hits the utilitypole and the HV falls on the LV and when utility poles get hit by lightning strikes.

 

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
Re: Power Supply Line Filter with Trifilar Windings
« Reply #3 on: March 08, 2020, 02:09:01 pm »
So basically jonpaul thinks the trifilar Earth is unnecessary, while bsdphk is arguing it is necessary for drunk drivers and lightning strikes.

The schematic shows the trifilar design, which implies the original designers thought it was necessary:

945518-0

Also note the presence of the second LF1 filter in that same schematic, which means LF2 is not the only filter in this power supply.

Even so, we have a case where only a bifilar version (no green Earth ground wire wound inside) is used.  So was that use of a bifilar merely a cost savings measure, or could it really have been later decided that the triple winding was unnecessary?
 

Offline fcb

  • Super Contributor
  • ***
  • Posts: 2135
  • Country: gb
  • Test instrument designer/G1YWC
    • Electron Plus
Re: Power Supply Line Filter with Trifilar Windings
« Reply #4 on: March 08, 2020, 02:19:06 pm »
I would argue that LF2 isn't a common-mode choke. Just an RF choke.

The next choke LF1 is likely the common-mode.

The engineers deemed it necessary (probably after testing at EMC) - not seen it before in that configuartion.
https://electron.plus Power Analysers, VI Signature Testers, Voltage References, Picoammeters, Curve Tracers.
 

Offline Mr Evil

  • Regular Contributor
  • *
  • Posts: 76
  • Country: gb
    • mrevil.asvachin.com
Re: Power Supply Line Filter with Trifilar Windings
« Reply #5 on: March 08, 2020, 02:24:29 pm »
It seems to me that including earth on the core will render the filter ineffective. A mains common-mode choke works when current is flowing in the same direction on both live and neutral, so the magnetic fields generated by both currents add together, yielding a high inductance. That current doesn't just flow into the device and disappear though - it has to flow out again somewhere, and as long as there isn't a lot of leakage capacitance or a ground loop somewhere, it will all flow out through the earth wire. Since the current in the earth wire is equal to the sum of currents in live+neutral, but in the opposite direction, it will generate an equal and opposite magnetic field in the core, cancelling it out.

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Power Supply Line Filter with Trifilar Windings
« Reply #6 on: March 08, 2020, 05:03:53 pm »
I agree with fcb in that FL2 is there for additional EMI suppression.  Since FL2 is wound trifilar it will act as a common mode choke. I have some line filters of various types.  Some are single stage common mode, some are two stage CM and some are a combination of differential and CM.

Most common mode chokes have inductances on the order of 1 to 10 mH.  This inductance gives an idea of the frequencies that the filter was designed for.  JDW, since the green wire is free are you able to measure the inductance?   
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Power Supply Line Filter with Trifilar Windings
« Reply #7 on: March 08, 2020, 05:54:59 pm »
Perhaps their (RF) grounding wasn't quite good enough, and they needed a CMC on the line cord as a whole?  Perhaps it was an errant guess and including ground wasn't necessary, so they cost-optimized it in the next revision?

Ground filtering isn't usually necessary because ground can be used as a reference plane.  Grounded metal enclosure, or metallized plastic, for example.

A case where it might be necessary, is where that reference plane sucks.  PCBs with plastic enclosures for example, may have unsubstantial grounding, lengthy tracks near noise sources, etc., thus putting CM noise into all the line wires, ground included.  The only solution then is to pray you can put enough impedance in series with the whole cord to get the margin required.

The equivalent circuit, for all wires emitting common mode, is where the bulk of the circuit (chassis, PCB grounds/planes, other connected devices -- keyboard, peripherals, ports) has one [average] potential, and the line cord has a different potential, thus emitting common mode through all wires.  A two-line CMC will not help in this case.  The usual solution is to CMC the other ports, or to use a shielded enclosure so that all the ports can use the shield as a reference plane.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline Mr Evil

  • Regular Contributor
  • *
  • Posts: 76
  • Country: gb
    • mrevil.asvachin.com
Re: Power Supply Line Filter with Trifilar Windings
« Reply #8 on: March 08, 2020, 06:20:47 pm »
...or a ground loop somewhere...
Just quoting myself. Thinking about that a bit more, maybe that is the reason for it; including the earth wire ensures that the lowest impedance path for common-mode currents entering on live+neutral is back through that same earth wire, rather than out through any earthed peripherals that are connected.

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
Re: Power Supply Line Filter with Trifilar Windings
« Reply #9 on: March 09, 2020, 12:10:56 am »
Most common mode chokes have inductances on the order of 1 to 10 mH.  This inductance gives an idea of the frequencies that the filter was designed for.  JDW, since the green wire is free are you able to measure the inductance?

I cannot make time to desolder and measure the inductance of LF2 until later tonight.  For now, here is the official schematic from Apple:

945902-0

Another consideration that no one has yet thought about is this.  Not everyone has a grounded AC socket in their home, and some users would use a 3-to-2 prong adapter to be able to plug in the computer to a wall socket.  Furthermore, here in Japan, pretty much no home has a 3 prong socket.  In fact, the only grounds you see in Japanese homes are a screw terminal found near any electronic device that carries water, such as a washing machine, or a dish washer/dryer, or a bidet toilet which requires wall socket power.  So in cases where the Mac was not plugged into a wall socket with an Earth Ground, could there have been some negative implication to warrant the removal of the Green Earth wire from LF2?
 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Power Supply Line Filter with Trifilar Windings
« Reply #10 on: March 09, 2020, 05:00:21 am »
JDW, about your last question.  If the chassis is not connected to earth, it will tend to float towards half the AC voltage between line and neutral by virtue of C33 & C37 that form a capacitive voltage divider.

Also, regarding measuring LF2's value.  You shouldn't have to unsolder it if all the connections on its left are not connected.
 

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
Re: Power Supply Line Filter with Trifilar Windings
« Reply #11 on: March 09, 2020, 09:02:18 am »
You shouldn't have to unsolder it if all the connections on its left are not connected.

I measured LF2 on 4 different Analog Boards (boards used inside the Macintosh 128k through the Plus) using my DE-5000 LCR Meter and TL-21 alligator clip adapter, with meter set to 1kHz (and I confirmed the same values for 100Hz and 120Hz frequency settings too):

1. 946096-0 82µH (bifilar, no Green Earth GND, 1986 board)
2. 946092-1 57µH (trifilar, 1983 board)
3. 946088-2 47µH (trifilar, 1986 board)
4. 946084-3 35µH (trifilar, 1983 board)

I measured the same value above for each winding on the same coil.

Note that the word at the very bottom of (3) & (4) above is COILCRAFT.  Also note the very confusing labels on (1) & (2) above, confusing because the labels are the same even though (1) is a trifilar and (2) is a bifilar!!
« Last Edit: March 09, 2020, 10:07:47 am by JDW »
 

Offline jonpaul

  • Super Contributor
  • ***
  • Posts: 3656
  • Country: fr
  • Analog, magnetics, Power, HV, Audio, Cinema
    • IEEE Spectrum
Re: Power Supply Line Filter with Trifilar Windings
« Reply #12 on: March 09, 2020, 11:51:53 am »
Hello: Unusual to see two CM chokes in series.

Certainly the Japanese lack of any ground affects EMI compliance.

But in 1986, CISPR and other EMI compliance was not as strict.

I cannot understand the designers motivation of the Apple ii PSU, but I can say that all modern SMPS I have designed or seen use a 2 wire CM noise choke and that the ground (Earth) green is always direct to chassis. Just looping thru the trifilar means risk of interruption of ground path, e.g. bad solder on PCB.

Thus for safety compliance the mains green (or Green/yellow) wire  is directly to chassis.

Just the ramblings of an old retired EE

Jean-Paul
An Internet Dinosaur...
 
The following users thanked this post: JDW

Offline jonpaul

  • Super Contributor
  • ***
  • Posts: 3656
  • Country: fr
  • Analog, magnetics, Power, HV, Audio, Cinema
    • IEEE Spectrum
Re: Power Supply Line Filter with Trifilar Windings
« Reply #13 on: March 09, 2020, 11:53:44 am »
All the inductances are about right for a CM choke: The exact values depend on line Zo, Y capacitors to earth and required compliance for conducted emissions, vs the PSU CM noise spectrum.

JP
An Internet Dinosaur...
 

Offline poorchava

  • Super Contributor
  • ***
  • Posts: 1673
  • Country: pl
  • Troll Cave Electronics!
Re: Power Supply Line Filter with Trifilar Windings
« Reply #14 on: March 09, 2020, 12:18:33 pm »
This is sometimes needed. I recall one project in our company, where passing all 3 mains conductors (L, N, PE) through a CMC finally got us through conducted emissions. As is often the case with EMC: wo don't actually have a solid theory why this was so, but it worked :)
I love the smell of FR4 in the morning!
 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Power Supply Line Filter with Trifilar Windings
« Reply #15 on: March 09, 2020, 05:27:32 pm »
Hmmm, the measured values are lower than I would expect.  It would then make sense that Apple had to use two stages to meet requirements.  Having participated in EMI testing, it would be interesting to see the test spectra.

I checked some of the parts in my collection and see they're all in the mH range.  I understand that because the line/mains currents cancelling, the core can have greater permeability giving a greater inductance without worrying about saturation.   I see that the later assembly has a greater inductance so perhaps the design was simplified with a component change.
 

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
Re: Power Supply Line Filter with Trifilar Windings
« Reply #16 on: March 09, 2020, 09:47:00 pm »
But as I said in my previous post, the part number labels on (1) & (2) are exactly the same (look at the photos), despite one being a trifilar and the other a bifilar. How would you explain that?
 

Offline Phoenix

  • Frequent Contributor
  • **
  • Posts: 442
  • Country: au
Re: Power Supply Line Filter with Trifilar Windings
« Reply #17 on: March 09, 2020, 10:38:29 pm »
How would you explain that?

Same potting box, different custom wound choke.
 

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
Re: Power Supply Line Filter with Trifilar Windings
« Reply #18 on: March 10, 2020, 12:15:21 am »
How would you explain that?
Same potting box, different custom wound choke.

I could have guessed that but didn't because it seems too odd to me that all those numbers on that label merely refer to a "potting box" instead of referring to what that component really is.
« Last Edit: March 10, 2020, 12:42:38 am by JDW »
 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Power Supply Line Filter with Trifilar Windings
« Reply #19 on: March 10, 2020, 12:37:41 am »
Welcome to production land.  Bills of Material (BOMs) and part numbers.  Buy parts, make parts, programmed parts, assemblies, subassemblies, ECRs and ECOs, Deviation Permits, running changes, versions and revisions, etc., etc...  I hear that in aviation, the paperwork can be heavier than the part.

Sometimes the part number you see isn't for the whole thing.  EPROMs for instance, have the original manufacturers' numbers on them but it's the contents that make them special.  The auto industry has no end of parts that can't be properly identified by the obvious part number.  A part may have a raw forging number but may be machined and heat treated differently.
 

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
Re: Power Supply Line Filter with Trifilar Windings
« Reply #20 on: March 10, 2020, 12:48:24 am »
Welcome to production land.

I follow you about paperwork, labeling and part numbers.  But you must admit that is still speculation in terms of what we are talking about.  Indeed, I too can speculate that the part number on the label could in fact refer to the windings, yet I could speculate the windings differ due to an error at the coil manufacturer.  An "error" is not outside the realm of the impossible since I have never seen a bifilar LF2 on any Macintosh 128k to Plus Analog Board except this one board.  Assuming the bifilar was a mistake, I wondered if it could have negative repercussions if not swapped out for a trifilar.  But the consensus in this thread seems to be "a bifilar would do just fine."
 

Offline Mulligan

  • Newbie
  • Posts: 1
  • Country: us
Re: Power Supply Line Filter with Trifilar Windings
« Reply #21 on: May 01, 2020, 02:00:25 am »
I have a Macintosh Plus that I purchased new in December 1988. (It still works perfectly, having recently been awakened from a 25-year slumber in my closet.)

LF2 is bifilar on my analog board, so that suggests yours wasn't a mistake.

 
The following users thanked this post: JDW


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf