mfj 4225 SWITCHING power supply demons

Discussion in 'Radio Circuits, Repair & Performance' started by WB8VGE, Nov 20, 2021.

ad: L-HROutlet
ad: l-rl
ad: abrind-2
ad: Left-3
ad: L-Geochron
ad: Left-2
ad: L-MFJ
  1. WB8VGE

    WB8VGE XML Subscriber QRZ Page

    Hi, gang,

    I've searched the data base looking for information on this supply. I got the schematic from MFJ, basically useless as you can't read it.
    I've had this for years, and used it mostly on the test bench. It never seemed to work correctly but after you cycled it on and off, a few times, it would settle down. What did it do you ask?

    The output voltage would be too high. There is a center detent that sets the supply for 13.8 V. If you set the control to the detent, the output voltage goes too high and trips the supply. The lowest setting gave you 12 V. When it was working, the lowest was 6V, and with the front panel voltage adjust pot in the detent position, the output would be 13.8V. The cheapo meter would indicate the same. Life is good!

    Now no matter what I do, cycle on and off, thump on it, talk to it nicely, threaten it...I can't set the voltage to the detent at 13.8V. I can adjust the output with the front panel control, but it's not correct. (i.e. centered with the detent)

    Nothing is cooked, no swollen capacitors, nothing visibly wrong. The output voltage is correct, I can set it to 13.8 V and while I haven't checked, I'm sure it will supply current.

    Being a MFJ product, I reflowed all the solder joints. Thinking it might be a heat issue, freeze spray did nothing. I used a heat gun to warm things up and----No Joy!

    I'm up for suggestions, hints or magic spells.

    I'm Mike, WB8VGE
  2. G4COE

    G4COE Ham Member QRZ Page

    How about trying a replacement 'pot', you could even use a preset resistor just to try it's got to be of the right value obviously... as for 'pot's I'd be using a linear taper one.

    Worth looking at any preset resistors that may be used to set the voltage 'to whatever' at the detent point.

    You say you've at it years..... is there not a cluye here I ask? I'd be certainly looking ath the electrolytic caps, replacing them with low ESR 105deg. C. types of good quality.

    As you probably know 'switchers', has no AC isolation transformer, so anything on the 'LIVE' side will be at mains potential, the smoothing capacitors after rectification can give very nasty shocks - so discharge them first say through a 1K resistor of at least 1 Watt. The isolation is on the DC side and is the switching transformer, I call it the 'chopper transformer' since the rectified mains is at DC and it is this that is switched OFF and ON.

    There are various ways regulation is done, one is to use a pulse from a low voltage winding back to the 'primary side' via a transformer, sometimes an opto-coupler is used, this feedback tells the control chip to turn on longer or shorter. Imagine a light switch turn it off and on slow and you get it light bright.... turn it off and on faster and it isn't as bright, mind you these switching power supplies do this at about 20,000 to 30,000 or so time a second.

    Switch mode supplies can self destruct.... imagine what would happen if it loses regulation feedback, it could switch harder and harder ON until the switching devices can't take it and 'pop', this usually damages the control chip as well.

    I would swap any electrolytics around the feedback loop for sure and any that are in series with the gate of the chopper FET or the base of the chopper transistor, swapping any around the control chip too. Opto-couplers can and do give this type of problem.

    Another cause is bad soldered connections so a resolder is a must in any case. Many a power supply may have two switching devices working in push pull.

  3. KA9JLM

    KA9JLM Ham Member QRZ Page

    The first thing that I would do, Is replace that 50 cent voltage pot with a better one. A good one might be $3.00 or more, But worth it.
  4. WB8VGE

    WB8VGE XML Subscriber QRZ Page

    Thanks, dave, for your insight.

    While the capacitors look fine, no bulges or goo coming out of any of them, it wasn't much work to replace the lot. So I did, and of course that didn't do a thing.

    I reflowed all the solder joints, thinking one could have a hair line crack or cold joint. No joy!

    The schematic I have (From MFJ) is barely readable. I found a resistor that's by the voltage control pot (on the front panel) and reflowed all the solder on those connections (it's a small pcb and not the main one) All the resistors in the voltage divider are right on the money value wise.

    I found the data sheet for the PWM chip, and the reference voltages are there, the adjust voltage moves around when you move the panel control, no smoke, burnt marks... Short the output, and the over current protection trips as it should. Other than the voltage being out of whack, it seems to work fine!

    I did notice something odd. I took the whole shebang out of the cabinet, and in doing so, disconnected the fan, it runs all the time, and that's when I hear sizzling, cracking, and a high frequency whine. Okay, the whine is fine, that's a switcher doing its thing. The sizzling.... could be windings on one of the transformers.

    Any, poking around in it, I managed to short one of the heat sinks to the chassis and POP! There went one of the TO-247 transistors. So, I'm dead in the water while the parts come in, (I ordered more than one!)

    Before I invented a few new words after I zapped the transistor, I disabled the over voltage protection. When the voltage adjust (panel mount) all the way, I get 22 ish volts on the output.

    I've noticed this since I owned this supply. IF I short or over current the supply, it shuts down in a milliseconds. However, you can't turn it back on until the volt meter drops to zero. When it trips (over current) the volt meter hoover around 4-6, then 10 secs or so, drops to zero. At that time, you can turn the supply back on.

    I'm mike, wb8vge
  5. G4COE

    G4COE Ham Member QRZ Page

    Oh dear, sorry to hear that, when it comes to shorting 'switchers' on the 'primary' side are not so forgiving... some power supplies can be zapped by attempting to measure the voltage around the chopper FET.

    If it's the chopper/switching transistor thats blown I really would be tempted to swap the control chip as well especially if they are DC coupled to the control chip also, if it is the chopper thats failed check for any low value resistors in the emitter/source of the devices.... they may be used for current sampling to protect the devices from over current - the voltage across the resistor being fed back to the control chip to reduce or cut the drive..... exact same values must be used here, usually about 0.5 Ohms or even less.

    The sizzling noise indeed could be the transformer or it windings, it also could be arcing as well, so check any damper/snubber diodes or capacitors strung between the HT side and the Collector/Drain side of the chopper device .... these clamp the switching spikes. Applying pressure to the actual transformer whilst working may alter or cure the noise, this indicates it's either a winding or a loose core, sometimes a cure can be had by hot melt glue or epoxy, don't overlook the winding connections on the actual transformer pins.

  6. W9GB

    W9GB Ham Member QRZ Page

    MFJ 4225 Switch-mode Power Supply
    Manual with schematic

    The MFJ 4225 uses the TL-494 Pulse Width Modulation (PWM) Control Circuit
    Numerous designs, videoes, and schematics on Internet.
    Sizzling is NOT a good sign with SMPS, unless transformer rewinding is desired.
    The TL-494 was historically the most popular Control IC for SMPS.
    Texas Instruments
    On Semi
    Tom Hammond, N0SS (sk) schematic for SEC-1223 far easier to read.

    Samlex originally used the Fairchild KA7500B IC, but now discontinued (ON Semi acquired)

    Digi-Key recommends substituting Texas Instruments SG 2524
    Last edited: Nov 24, 2021
  7. WB8VGE

    WB8VGE XML Subscriber QRZ Page

    I had some of the TL-494s here in the shop, and since it is a single sided pcb, and I have the tools to suck the solder up it was a simple task to replace the IC. I had no expectations that replacement of the IC would cure the problem.
    If you read the post about me having popped one of the switching transistors, I'll replace them and trying again. If I can't figure it out, I think I'll make a new faceplate and turn this supply into a speaker!

    Mike, wb8vge
  8. WB8VGE

    WB8VGE XML Subscriber QRZ Page

    I should have known better! I had a hunk of cardboard laying across the chassis to keep the pcb from touching metal, and I noticed the bare switch terminals on the primary voltage select switch. I was going to wrap some tape around those lugs, and well.... I'm more concerned about the sizzle sounds from the supply. When the new parts arrive, I'll poke around the transformers.

    Mike, wb8vge
  9. W9GB

    W9GB Ham Member QRZ Page

    Frying Bacon never a good sign, with an energized chassis.
    Yes. SMPS can be very frustrating — as any Computer Shop Tech since the 1980s could tell you.
    Apple Techs dealt with Sony SMPS with the Apple II, which was actually reliable.

    You can never have enough speakers. :)
  10. WB8VGE

    WB8VGE XML Subscriber QRZ Page

    Way back, way back, I did some computer repairs for the local Apple dealer. The #1 issue with the power supplies in the Apple II and the IIe was the power switch failing. Back in the day, that switch was constantly cycled, and it wore out. The fix was to use a switched power strip. Turn the power on and off from the strip and not the switch on the supply.

Share This Page