I'm working on a Pioneer VSX-454 that has a shorted 2SA1302/2SC3281 output pair. I've been reading that there is a big problem with counterfeits of these transistors. Does anyone who does a lot of audio repairs have either 1: a reliable source for these transistors, or 2: good substitute part numbers?
Pioneer has them, but they want $18 for the 2SA and $16 for the 2SC. This would make the repair too expensive, since these (and the bias transistor) blew while bench testing after fixing the original fault.
The new Toshiba replacements are 2SC5200 and 2SA1943.
The most reliable source I know of is B & D Enterprises. You can Google Bdent.com for their phone number.
The newer Toshiba numbers are also being counterfeited. MCM sells replacements for the 1302 and 3281, same as Pioneers', made by Onsemi using Motorola jigs, which Motorola had previously got from Toshiba, so the story goes.
There's also a local place here in Denver called Electronica that sells ICs and transistors and he only will sell original parts. The best thing he is also a tech, so he knows the difference between an original manufacturer and a counterfeit and he will refuse any transistors that are not original. Other darn places don't know the difference and don't give a damn, just want to buy the cheapest crap they can buy. According to my price list your transistors are around $3.50 each. This is the only place I buy from other than direct from the manufacturer. Electronica
2828 E. Colfax ave. Denver, CO 80206
1-800-722-5578
303-322-3037
This would worry me. Even if you could get the parts for a dime a dozen, don't waste silicon. I bet you did not fix the original fault.
Now you're holding up my dinner but that's OK because I got beer. This might get a bit longish. Years ago I worked at a rental place (I won't again) and they rented alot of audio stuff. I developed a method of testing these things so you wouldn't keep blowing outputs.
The amps we had were Fisher CA270 discete, and the PCB design easily facilitated my process. You'll have to figure out how to do it with the architecture of the PCB you're working on now. This applies to most discrete component audio output circuits, except for class D.
Disconnect the outputs. After you check the drivers and things of course, wire the circuit point which would be connected to the base, to where the emitter should go. A pair of 220 ohms should work fine, but you can lower it, not too far. Actually higher is better. During this operation you do not connect any load.
If your driver, bias, and voltage amplification stages are working it should achieve DC palance and let the relay kick in. This is why you don't want a load on it.
Now to find out if it's going to poof when you install the outputs, of course you got very low or no DC the the actual output, the junction of the emitter resistors. Now measure what is at those base terminals. In fact just measure point to point !. The bias circuit should be floating a volt or so across those two base connections. Now DRIVE it, not to clipping, but let it rip a bit. With no load the voltage shoould not increase. If it does something is leaky.
With no input nothing should be getting hot. At this point it's time to apply an input signal and look at it with a scope. It should be going both ways if you know what I mean. Put the trace in the middle and keep cranking the input level, but after a first quick test at a low level. Higher power amplifiers with a low impedance main feedback network might load the drivers too hard and cause you to waste silicon.
Keep checking for the drivers overheating, but you should be able to get it to drive all the way to the PS levels. If you got + and - 60V you should be able to get a 120 V P-P waveform on the scope. You should even be able to plug some headphones and hear the audio since they don't load that much. Don't leave it running that way though, even that light load could be more that the drivers can handle.
There are several variations. If you really need to do a heat run you can short the bias network out and short base to base of the outputs. Shorted base to base, the drive circuit has two chances of frying your new silicon, slim and none, but slim didn't leave town. If you have a problem that causes it to generate RF or any ultrasonic frequency, if you got the bases shorted together this is almost the only way it can happen. (opinions from top techs here please, overloaded by speakers that is not the situation on your bench, question is what else can do it after you don't have any DC offset ?)
I once had to fix a CA270 discreet that would keep coming back, and blow one of their woofers on occasion. It was all on us, rent to buy. You got it turned up and it would blow kinda quick, it had to be doing about 50 WPC or so (thing was rated 110 but did actually almost 200), thing was we bought the right woofers so they COULD crank it up. This one we were pretty sure they weren't throwing six more speakers on it or anything and they were getting pissed. Managment actually thought they were scamming us somehow.
By disabling the outputs and running it unloaded, I could see on the scope that at higher power levels it was getting a significant DC component on the output. It got worse as the level increased.
As I remember it did some very light clipping at the top (positive) and the whole trace moved downward (negative). I think I'm onto something here. The DC wouldn't stay on sustained high output, it came and went with the changes in the music. But the cause ?
HAHA, we were authorized and we got prints. What do I see ? A muting transistor at the main amp input ! Stupid idiots didn't use an FET like any respectable company.
I disconnected the collector and the abberation disappeared from the scope. No more DC offset.
Of course the rest of the job is obviously a piece of cake.
Incedentally, don't ever allow a unit to be shipped with the bias off (shorted). I found out the hard way, doing so allows the power supply voltages to climb too high in some units. If a period of quiescence is followed by a full power blast it increases the likelyhood that the current capabilities of the outputs being exceeded. This is likely to happen more in urban areas where the power line is likely to be dirtier. (spike, harmonics)
Anyway, I've addressed certain other things in this forum, such as how to REALLY set the bias for best performace and efficiency, and a few other things, but I've not before gone into exactly how to find (or lose ? lol) some of those pesky recalls.
Being DC coupled modern amps pose their own set of problems. Too many components make it nessecary to disconnect certain things for troubleshooting, but to know what to connect then is important. In such circuits sometimes even a scope can't tell you the problem without setting up a test configuration.
Last but not least, if those outputs blow with NO load and NO input, you got an RF problem. In that case you are looking for an open capacitor, most likely a teensy lil thing.
When shorted these teensy lil things cause a quick failure and you find it with an ohmmeter, when open it sticks up the works a bit. Sometimes they decide to be a resistor instead. In that case they might not even cause a big failure by being a resistor, it might happen because they are NOT a capacitor. Get it ? If a .0147 opens up that's tuning the output of an HOT, HVOT or SMPS output what happens ?
If shorted all other components should be good except maybe a fusible, or in the case of our wonderful Sony, the SMPS choppers. Fuses are that expensive Sony ? Come on !
In older Sony's shit would short out and there were literally no problems with the power supply. Early XBRs were a great example. Things are different now.
Back to the subject, this test mode can be used at the output of the voltage amp too, but expect a clipped waveform because it is alot less likely to be able to even drive the feedback. Learn to desolder and setup test setups, it is the only way to truly effectively service them. If you ever need to go there, after it works you get the driver right, then finally install the outputs.
As Murphy himself told me, you don't get luck, you make your own. I think he was right.
I'm going to take JURB's advice and do some more testing before I order those transistors. The reason I was concerned about the counterfeits was that I only wanted to make one more parts order for this unit. Having to meet a minimum order and pay $8 shipping each time starts to add up, so some scope testing with the outputs removed makes sense.
This Pioneer VSX-454 is a Pro-Logic receiver with a 4 channel amp. My father-in-law gave it to me to repair, but it looks like he got it from a flea market or yard sale. ("$7.00 as-is" written in black marker on the top cover.) So I had no history of what was wrong with it. The only symptom was that it smoked when it was turned on.
The power amp in this unit is a "module" of sorts: It's like a box, made of two circuit boards (L/R and C/S amps) and two heatsinks (all NPN outputs from both boards on one sink, all PNPs on the other sink). The heatsink fins point inwards and there is a fan at one end that blows thru the box. The heatsinks are "live" - they are at the
+/- rail voltages, and no insulators are used on the transistors. This makes testing the amp outside the receiver very risky, as it can't touch anything, not even the chassis, and the multiple ribbon cables connecting it to the main board aren't long enough to get the amp completely out of the receiver.
3 of the 4 channels had already been repaired before: The Right channel has an off-brand 2SA1302, the Center channel outputs were subbed with Matsushita transistors, and the Surround outputs were subbed with physically smaller Toshiba transistors. A few of the flameproof resistors on the L/R amp board had been replaced by cutting the old ones off and soldering new ones to the "stubs". (In order to get to the bottoms of the boards to do any soldering, all the outputs & drivers & bias transistors have to be removed from the heatsinks to free the board up.)
Anyway, what I found was a few 1/8 W resistors on the L/R amp board burnt to a crisp, and a browned area on the board around them, which covered a few unburned resistors. This is what smoked when powered up. The amp circuits looked pretty simple, and I got the burnt resistor's values from the other amp channels, so I decided not to order a schematic, and just test parts. Since I didn't know the extent of the damage, or if more than one channel was faulty, I checked everything on both boards. I found the damage was confined to the Left channel, and the Outputs & Drivers were about the only parts NOT blown.
3 of the 4 "tall TO92" transistors as well as the little bias transistor on the heat sink were shorted, 4 47 & 4.7 ohm flameproof resistors connected to them were open, and 4 zener diodes were shorted.
I replaced all the bad parts, all the resistors in the "burned area" even the ones that looked & measured OK, and all 4 of the "tall TO-92" transistors. Powered it up with load resistors on L&R outputs and speakers on center & surround outs, at 90V on a Variac. It seemed to work OK, no DC on the outputs, and I was able to get sound form the Center & surround from its FM tuner, so the rest of the receiver was working, so I brought the line voltage to 120. I checked the voltages across the .33 ohm emitter resistors, and it was only 1 mV on all 4 channels. The bias is not adjustable on this model, and all channels read the same, so I assumed it was OK.
When I turned up the Volume to about 10:00, a flameproof resistor in the left channel immediately lit up like a sparkler, then the line fuse blew a few seconds later. I never got around to putting the scope
on the outputs, but the air blowing thru the heatsinks was coming out cool, so there probably wasn't any high-level oscillation taking place (at least until I turned up the volume.)
So I tore the amp module down again, retested everything in the Left channel, and the only bad parts were the resistor that lit up, and both outputs and the little Bias transistor on the heatsink were shorted. Since I replaced nearly everything else, and those parts were still good, I figured replacing the outputs (and drivers for insurance)
should take care of it. But there is a possibility that it took a certain signal level to kick off an oscillation...
The burnt resistors should point you to the root cause of the failure. With no load and no outputs there is no reason for anything to smoke if the circuit works properly. Had the outputs been present they would probably short first and blow the fuse before anything else could happen.
Sorry about the eight resistors, but they should point you to a component which might test good on an ohmmeter but gets leaky or shorts at actual operating voltage.
Remember DO NOT connect speakers or a dummy load when you do this, it WILL blow the drivers.
This one is a subwoofer amp, and probably had a bit of a lightning surge. Some op-amps were bad, but the amp part seemed OK. Replaced the op-amps, but the amp was railing out. Looked closer at the amp circuit itself - an MPSA06 transistor was bad. Replaced it, the offset was cut in half, but the bias was hot. Since disassembly was major just for access, I thought I'd try your method. Didn't work in this case. Going to order a service manual and get a better look at the big picture.
I suspect the OP is having access difficulties. If I'm not mistaken that unit has the amp built around in like a "box" and is fan cooled through the middle. Ugh.
A shop I worked at previously actually took to replacing the whole assembly, though I have no idea how they worked the cost of it into the estimate. I think MCM sells those assemblies. I don't remember how much they cost.
I've also noticed class D amps in subwoofers, and if that's what you got this method won't work. You can tell by a pi filter at the output. Class D amps lend themselves well to subwoofers because they can get away with a lower chopping frequency. With their efficiency they are singularly suited to the massive current needed to drive some woofers.
You mentioned bias, so it's not class D, consider yourself lucky, class D amps are a whole different animal.
This is a pretty conventional design, a KEF PSW-2150. 2 ea. 2SC 5200 and
2SA1943. I was just trying to do some troubleshooting with the outputs removed. The board was upside down at the time or I would have seen the smoke sooner. :-(
This one just gets worse and worse. I suppose I'll just order a replacement amp plate from KEF, cost permitting.
35 volt DC offset. Outputs NOT biased on, measure NO millivolts across the emitter resistors, yet the outputs get hot. No HF oscillation, no AC signal, no load, no DC bias (B-E voltage no more than 50 mV), yet they get hot. Replaced the outputs just in case there was some leakage I had missed - no change. I'm pretty much out of ideas, and I'm way upside down time-wise on this one. Time to cut my losses.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.