TV Power Supply MOSFET replacement

Thread Starter

wayneh

Joined Sep 9, 2010
17,496
I'm trying to repair my daughter's RCA television, and I need advice about choosing a replacement MOSFET. I've just ordered a 2SK2917 from Mouser to replace the 2SK2842 I pulled from the power supply. The current ratings are higher for the 2917, but my concern is the higher (but still small?) capacitance and inductance of this "larger" MOSFET. Will this be a problem?

FWIW, the 2842 I pulled from the board has higher current ratings than the 2SK2543 shown on the attached schematic. Probably means they knew this was under-specified?



I was lucky to find this schematic online, along with training materials, a troubleshooting guide and so on. It was soon obvious that the power supply is dead, and I removed TP020 (left side middle of schematic), the switching MOSFET for the transformer primary coil. It tested dead short. The info I found suggests replacing all the active components on the primary side if you find the MOSFET dead, so I've ordered all of the transistors and zeners. All cheap compared to the MOSFET.
 

Attachments

mcasale

Joined Jul 18, 2011
210
Interesting schematic. I'm not familiar with some of the component designators. Is LP020 an inductor or ferrite thing? I wonder if it's supposed to help diminish ringing.

Looking at the spec sheets, it seems these parts are pretty close, except for the gate capacitance. It seems like the swap is worth a try. You may want to drop the value of the gate resistor, RP021, to compensate for the higher capacitance. Are these passive parts through-hole or surface-mount?
 

Thread Starter

wayneh

Joined Sep 9, 2010
17,496
Thankfully, the parts are all thru-hole. I'd have probably given up already if they weren't.

Good point on the gate resistor. I could just piggyback another resistor in parallel. But without a scope, I'm not sure how I'd detect any change. :(

I'll look at LP020 to see if I can identify it. It must be an inductor of some kind.

So, you're not terribly concerned that the new MOSFET will not work?
 

mcasale

Joined Jul 18, 2011
210
The time constant on the gate is less than 80nSec typically. Hard to tell if it's significant.

Otherwise, I'd say full steam ahead. Let out the magic smoke!!!

Good luck.
 

Thread Starter

wayneh

Joined Sep 9, 2010
17,496
Let out the magic smoke!!!
Ha, thanks. That'd be more fun if it wasn't a $5 MOSFET :eek:, plus a comparable shipping fee.

Those waveforms show "T = 12µS". Does "T" refer to a single cycle (meaning ƒ = 83.3kHz) or is it the entire x-axis (ƒ = 250kHz)? Either way, it seems that 80ns is "small" by comparison?
 

mcasale

Joined Jul 18, 2011
210
My guess is that T=12uS is the period of the pulse. So, 80nS is small compared to that, but you still want the edges to be as fast as possible to minimize time in the linear region and self-heating.

Looking at the schematic again, I notice the gate resistor is NOT driven by a low impedance. During pull-up it is driven through RP023 and RP027, which is 1.1K (if I'm reading it correctly. That gives an overall time constant of about 3uSec. That's probably why the gate voltage looks squirrelly. So it looks like turn-on is the slowest time.

The turn-off current goes through the emitter of TO023, which should be a much lower impedance, I THINK.

I wonder why they didn't use a rail-to-rail driver, with all these discrete parts.

Still, I guess it still works, even if the FET gets a bit warmer than it should.
 

Thread Starter

wayneh

Joined Sep 9, 2010
17,496
Still, I guess it still works, even if the FET gets a bit warmer than it should.
It's held against a large piece of aluminum (with thermal paste) by a spring clip, so I guess they designed for that. Since the replacement is much larger (lower Rds) than the original design, I'm hopeful. We'll see.
 

crutschow

Joined Mar 14, 2008
34,281
After it runs 15 minutes with the new transistor, unplug the TV and see how warm it is. If not too hot to hold a finger on it, then it is probably OK
 

Thread Starter

wayneh

Joined Sep 9, 2010
17,496
I'm guessing this is a little late...
Yes, but thanks. I really should have looked at cross references before I ordered. Duh. I was even able to find the original part from other sources, but wanted to order everything from one place. So I just used Mouser's filters and ended up where I did.
 

debe

Joined Sep 21, 2010
1,389
With SMPS replace all the electrolytic capacitors. Un less you have am ESR meter to check them. Also make sure CP008, the mains filter cap is good. Also check RP006 & 007 120K resistors as they kick start the powersupply at turn on
 

Thread Starter

wayneh

Joined Sep 9, 2010
17,496
With SMPS replace all the electrolytic capacitors. Un less you have am ESR meter to check them.
Can they be checked this way on the PCB? I didn't think to order replacements.
Also make sure CP008, the mains filter cap is good.
I can confirm it holds a helluva charge, since I purposely shorted it before working on the MOSFET. Yikes! :eek: Won't do that again.
Also check RP006 & 007 120K resistors as they kick start the powersupply at turn on
Good point, and an easy test.
 

Thread Starter

wayneh

Joined Sep 9, 2010
17,496
I'm looking for advice on where to go next.

Spent yesterday afternoon replacing all primary side transistors and diodes (other than the bridge, which is working). TV set is still dead :confused: but the power supply is now buzzing :D.

The MOSFET replacement was the wrong size! Can't believe I didn't check for that but anyway it installed easily with just a little bending of the leads. Was surprised and very glad the larger pins still fit the PCB holes, and that there was room on the heat sink. Whew!

The bridge is supplying about 174V for "RAW B+", and the oscillator seems to be working as judged by the buzz that starts a second after plugging in the set. But there's essentially nothing going thru the transformer on either side. I'm seeing no more than 1V anywhere on the secondary side.

My to-do list. Please help add things to look at!
• I previously replaced RP020 (current sense power resistor) and will check it again next time.
• Will check RP006 and 7 as suggested by debe.
• Will look for voltages on top of CP024 and 25, and CP040. This doesn't really test them, but might be interesting. I'll first test caps for shorts with power off.
• Looking hard at opto-isolator IP001, which provides feedback from the secondary back to the primary. According to the pdf, it must light up to draw more power thru the secondary. If it burnt out, I think it produce my dead-set symptom?
View attachment THOMSON ITC008-Power.pdf
 

THE_RB

Joined Feb 11, 2008
5,438
It's common to have a secondary side short that originally killed the PSU. Check the horiz out transistor (big transistor that drives the HV flyback transformer). It is likely shorted which is bringing down the secondary rail.

In the old days it was easy enough to disconnect the secondary side after the diode rectifier and connect a 100w light bulb etc as a load, and run the PSU and check it is regulating and producing the right voltage 120v DC. These days the PSUs are smarter and have standby systems and feedback systems so it's not always possible to run the PSU by itself on more modern sets.

A test you can do on the secondary is check it for shorts, ie attach a DC PSU after the diode and see if the voltage comes up ok or drags your DC PSU down into current limiting. You don't need a 120v supply, even a 20v or 30v supply will be fine for that test.
 

Thread Starter

wayneh

Joined Sep 9, 2010
17,496
It is likely shorted which is bringing down the secondary rail.
I'll look into this. But for my own understanding, wouldn't this cause a lot of current to flow on the primary side? I'm not seeing any. There is about 174V on top of the main filter cap and the primary coil, and only maybe 3V on the bottom (the rest across the MOSFET I just replaced).

Hmmm... Maybe the voltage drop across the current sense resistor is turning it all off?

Anyway, thanks for the tip.
 

dataman19

Joined Dec 26, 2009
135
Have you eliminated the flyback?
Sounds like you could have a bad flyback (the buzzing)..
Incidentally, this is a switching regulator to drive the flyback transformer.
There are numerous issues and several suspects that need to be eliminated.
..
Question: Was the original MOSFET actually bad?
...
The pulses coming from TP020 are the switching pulses to drive TP023 and TP025. The 600v pp waveforms going in to TP020 are the Horizontal Blanking signals and also the switching regulator sync.
..
I would suspect the buzzing sound to be the flyback switcher either running off frequency (TP020 is wrong) or a bad flyback.
...
Do you have a High Voltage Probe to check the aquadag voltage (the lead coming off the side of the CRT)??? How far off spec is it?
...
By the way, by "dead" do you mean no picture? If so is the neck of the CRT lit?
....
A lot of people get bogged down with the High Voltage issues, only to find out that the horizontal burst info is missing (which means you won't have high voltage).
...
Most flyback power supplies have a tickler/regulator winding and a DC blocking cap to allow the AC to ground (which essentially keeps the chroma burst from affecting the high voltage) CP009 (right above and to the left of TP020) - check this component as well (it's cheap, so just replace it).
..
Dave
Phoenix, AZ
 

sheldons

Joined Oct 26, 2011
613
iv repaired a few of these .....you need to change the opto as well....its a good idea to check all resistors and capacitors etc to do with the regulation/feedback side of this circuit as a problem here will lead to instant destruction of your new fet etc....
 

Thread Starter

wayneh

Joined Sep 9, 2010
17,496
By the way, by "dead" do you mean no picture? If so is the neck of the CRT lit?
Dead means nothing happens when you hit the power button. The buzz appears shortly after plugging it in, but that's the only sign of life.
Question: Was the original MOSFET actually bad?
Unknown. The set was dead but buzzing when I started all this. One of the first troubleshooting tests is to check the voltage on the gate (currently less than 0.5V). I screwed up and touched something else also and shorted it out. Then it was certainly dead, and not buzzing anymore either. :eek:
Have you eliminated the flyback?
Sounds like you could have a bad flyback (the buzzing)..
No, I haven't looked at this. Is this relevant to a power-off set?
Do you have a High Voltage Probe to check the aquadag voltage (the lead coming off the side of the CRT)???
Nope.
CP009 (right above and to the left of TP020) - check this component as well (it's cheap, so just replace it).
Will do.
Thanks for your insights. I'll keep posting my findings. I really don't want this thing to beat me.
 
Top