WTF-Roadster 66

[ttocs]

fets have solder in them?

[Eric D]

I am going to take a guess that no they don't have solder in them. The material which boils out is silicon. But, honestly I don't know for sure, so Jason is probably right and it probably is actually solder.

I have seen plenty of FETs with tiny silver balls formed near the mounting tab, and this to my knowledge is silicon which has melted and forced its way out. In this case these parts are trashed.

[ttocs]

I use to work in a semicondoctor manf facility and just figured that they "grew" the fets the same way they do the microprocessors. In that case there would be some solder in the packaging now that I think about it. Just never thought about it before ya know?

[Jacampb2]

Silicone melts at just over 2500 degrees F and boils somewhere north of 4000 degrees F. Not saying you couldn't briefly hit those temps during a fault condition, but you wouldn't be able to do it long enough and slow enough to get the balls. There is solder in the package, both the leads and tab are tinned.

AFAIK, the silicone wafer itself is "grown" as a long log, then it is sliced into wafers, and the normal process for actually making the die is more chemical "etching" via a lithographic mask than growing.

Later,
Jason

[ttocs]

they are actually made in wafers, then sliced up into chips. While I was working for intel worked both in the etch area, litho(which sucks) as well as thin films cmp. It was a strange job like working in a space ship but only 14 days a month was super nice............ I would be going for my 2nd sabatical now if I had not developed these digestive problems.

[Eric D]

The reason I figured the balls were silicon is my soldering iron would not melt them. In some other cases though I have had ones which the iron would melt.

[Jacampb2]

Eric, sometimes they use deposited aluminum for the "wiring" in the die, I was under the impression that that went out of vogue some time ago, but I am sure it still get's used, especially in low(er) tech components. Maybe that explains it? The lead frame, the place where the leads and "wire" join is a soldered connection in a traditional package. I guess wouldn't say with 100% certainty that solder always makes the balls when you find them, especially in light of you telling me you can't always melt them (I've never thought to try, lol).

Scott, I was talking about the wafers that you guys would have started with as a raw material. We make the liquid silicone raw material, grow the logs and slice them into wafers that the fabs use to make their dies (the chips)-- so we are on the same page I think When you mentioned growing them, I thought you were referring to the very early step that I mentioned above. The only other "growing" step I can think of would be one of the deposition techniques used on some dies, which in light of where you were working makes a lot more sense

Anyhow, sorry for the hijack on your thread Thunderdome!

Later,
Jason

[ttocs]

oh gotcha yea I went to our supplier once and saw how they "grew" the silicon tubes as you said, cool as shit. I was talking about how they grown the circuits on the board, but same yea same difference.

[Eric D]

Well now that I think about it, maybe some of the solder balls were just oxidized to a point my iron would not instantly melt them. I did not really spend much time trying to melt some, I just touched the iron to them and they were knocked off the FET without melting. In other cases they melted instantly.

[Thunderdome]

well sorry for slow update been pretty busy not much time for me lately but I did manage to get the fets off the board and was a piece of cake after seperating the legs from the fet body, I then checked the output fets and one was completely shorted and another was half way shorted if that makes sense (not fully open) but passing, havn,t got to the gate resistors yet but its next on my list and the fets should be here this afternoon,

now while checking my car over (Alternator which seems ok) for power issues, I noticed one of my subs is making a lot of mechanical noise at higher volumes, the question :could this sub be intermitently shorting out causing power surges which could cause a failure? this is the 3rd amp now that has had similar problems in this current install and yet in the past 15 years previously I have never blown one up ever so I,m starting to think maybe time for new subs

btw I checked them with a meter and they are correct and I pushed them in and out while testing but at higher voltages could this be happening?

thanks again for the help guys its appreciated

[Jacampb2]

I have not had the opportunity to work on any of the RSD line up, but, does the roadster have both a Class D amp and a Class AB amp in the one chassis? If it does, is it the Class D part of the amp where you are finding bad outputs? I only ask because you keep mentioning output fets, and I was under the impression that the AB RSD's were still TD BJT output section. If it is both, and the problem is in the class D side, you will probably need to pull the fets off board to be 100% certain of if they are in good shape or not. There is a possibility that the problem lies in the class D driver circuitry and it is actually switching the output fets incorrectly. And, if there was no design implemented to pull the fets back down at rest, they can maintain their state indefinitely. So if the gate charged when it was on last, unless you test it correctly, the output fet could still be partially enhanced and not actually be bad. This is all assuming that they are actually Mosfets and not BJT's, but either way, you really should not be finding a semiconductor showing any significant leakage when tested properly and out of circuit.

Good Luck,
Jason

[ttocs]

this is the 3rd amp that has blown on those same subs? Sounds fishy to me.

[Jacampb2]

Oh, and BTW, about your subs. you might have some kind of VC problem that shows up once it is hot, but that is the only thing I can think of. Beat the heck out of them and then check the DC resistance?

Later,
Jason

[Thunderdome]

I have not had the opportunity to work on any of the RSD line up, but, does the roadster have both a Class D amp and a Class AB amp in the one chassis? If it does, is it the Class D part of the amp where you are finding bad outputs? I only ask because you keep mentioning output fets, and I was under the impression that the AB RSD's were still TD BJT output section. If it is both, and the problem is in the class D side, you will probably need to pull the fets off board to be 100% certain of if they are in good shape or not. There is a possibility that the problem lies in the class D driver circuitry and it is actually switching the output fets incorrectly. And, if there was no design implemented to pull the fets back down at rest, they can maintain their state indefinitely. So if the gate charged when it was on last, unless you test it correctly, the output fet could still be partially enhanced and not actually be bad. This is all assuming that they are actually Mosfets and not BJT's, but either way, you really should not be finding a semiconductor showing any significant leakage when tested properly and out of circuit.

Good Luck,
Jason

hey thanks Jason and yes it does have class D and AB boards joined with wires see pic below, and yes they are irfz44's on the power supply and irfb42n20d on the outputs, I did pull the outputs as well as I was getting odd readings in the board

I checked the outputs per the bcae program with a jumper and 2 were the same and the other 2 were different one shorted and the other high readings so I figured that wasn't good just going through the gate resistors now, and wanted to make sure if I test them in the board but with the fets out I should be able to get accurate readings?

thanks again for the help

[ttocs]

hopefully you discharged those caps before doing that, I would be afraid they would short out on the sink.

[Thunderdome]

yep there drained, I generally keep towels on and around everything to avoid scratches but removed them for the pic as you wouldn't be able to see the amp layout with everything in the way is all but thanks for thinking about it

[Jacampb2]

You should be able to check the gate resistors in circuit as long as the damaged fets are out of circuit. Personally, I have stock of pretty much every commonly used gate resistor in the amplifiers that I tinker with, and I just replace them, it is not worth having one fail after the repair is set and taking out a bunch of fets when it happens.

Eric touched on it earlier, but if a gate is open, then when power is applied to the mosfet, it can and very likely will act like it has been enhanced. If the resistor is ok now, at room temperature, but has a odd temperature coefficient as a side effect of having been stressed so badly, or if it's value just shifts all together at a later date, then you are going to have some of the fet's sharing more of the load than the others and it is only a matter of time before they go again. It's not worth the few dollars in resistors for me to want to reuse them...

Good Luck,
Jason

[The Golden One]

^^^^^ he hit the nail right on the head with that one, i just have a few blown irfz44's in my 0.5 im just going to replace all of the resistors in that section. and maybe the others that are in the output fets section, it's just not worth giving it another chance to mess up again.

[Thunderdome]

ya I see what you mean, I have to order a bunch of caps for my ms amps so I can just order the resistors at the same time as I'd rather just do this repair once and resistors are cheap insurance thanks again

[Jacampb2]

^^^^^ he hit the nail right on the head with that one, i just have a few blown irfz44's in my 0.5 im just going to replace all of the resistors in that section. and maybe the others that are in the output fets section, it's just not worth giving it another chance to mess up again.

Your ZPA does not have fets for outputs, it has BJT's, and, more than likely the series base resistors to those outputs are fusible and flame proof. Don't mess with them unless they are bad, and if they are, you should replace them with the same exact thing. Finding the proper fusible resistors is a bit of a pain, so be careful about clipping leads and such.

Good Luck,
Jason

[The Golden One]

^^^^ it's like gi joe when they would say {and knowing if half the battle} the damage to my amp happend when it wasnt even on so it may just mostly be in the power supply anyway. i just need to look at things a little different,but that bit of info should come in handy.