SPL-1000 amp. released magic smoke...

[Nigel Goodwin]

ok, so what is the formula for correct wattage? and is there any way to add thermal compensation, or this too complex? i read somewhere about adding silicon diodes from transistors base to ground, but i dont want to fry anything from uncorrect tweaking and, you said not 4ohm loads, welll i used 4ohm speakers, not surprise then it burned....

The formula is pretty simple - W=V*V/R

R is simply the speaker impedance, the trick is to find V.

You said 46V on each capacitor?, so assuming this means a +/-46V supply, this gives a 92V supply rail (off load).

First assumption - the rails will drop under load (plus losses in the amplifier), so assume a 'rough' maximum peak to peak output voltage of 80V.

Convert p-p to RMS - 80/2.828 = 28V RMS

28 * 28 = 784 (V squared)

Now simply divide by the speaker impedance

784/4 = 196W (4 ohm load)

784/8 = 98W (8 ohm load)

Assuming the 46V is the total supply rail, those powers reduce by a factor of four, giving 49W to 4 ohms, and 24.5W to 8 ohms.

This last figures seem just about plausible for TIP41/42 (although 4 ohms is probably pushing things a bit), the first calculations (92V total supply) is massively beyond their capabilities, unless they are just drivers feeding some much more substantial output transistors.

 

[fezder]

Ah, this info comes handy, thanks again formula to add my notes. but, i need to gather those burnt 2w resistors, dont have those ATM yet, but luckily seems to be repairable

 

[unclejed613]

with speaker loads 4 ohms with an inductive component probably pushed those transistors over the edge. even if you were within the safe operating area with a resistive 4 ohm load, inductive loads like speakers often take output transistors outside the boundaries. in other words, with a resistive load, the relationship between current and voltage across the output transistors is entirely linear. with inductive loads (speakers), the voltage and currrent is out of phase, often resulting in taking the transistors into portions of their Ece/Ic curves where they should not be.

 

[fezder]

ok, i had some other things at hand, and had to make some component orders and stuff. But, i made some purchases that could be helpful, a 8ohm/100w resistor, should be enought for dummy load, and ESR meter, which has ability to test on-board (but not paraller) caps. i checked ESR's and none were too big, at least IMO, but i can be mistaken . but, at the moment, i have my orbinary testing equipment, that is 'scope, lamb-cord(you know, the one with light bulb in series with live cord) and for safety, i dont know what this is in english, but this ''fuse'' that trips power off when there is earth leakage current. oh and also IR-thermometer. So, any suggestions what to measure/inspect next? i'm more than eager, and willing to learn, without getting zapped...

 

[fezder]

sorry to be pain-in-the ass, but any help aprreciated. i asked from teacher, and he suggested possible oscillations....what do you guys think?

 

[fezder]

ok, i made some measurements which yielded some interesting results. I made these measurement without load.
TIP41 Base-emitter=0.45vdc
TIP42 Base-Emitter=-0.45vdc
they seem to be ok? the right, working channel gave same results.
But, you see the small connector which has R G + - L labeled next to it? well i measured those too and here's the results:
G L=0.5 Vdc
G R=8.7 mVdc
+ -=33Vdc
I did make other notices too, two big power resistor next to connector gets VERY hot, as well as faulty sides smaller transistors. There is about 16/-16v at zeners next to transistors.
There is huge difference between in G-L (which is the one that went S/C before) than in G-R
Does that difference matter, that connector has wire that goes to front pane, should i check there as well for faulty components?
In speaker terminals there is very little difference or DC/AC voltage whatsoever.

Does this information give you guys any help, with helping me (cryptic..) on looking where to look next? or should i just measure each component one-by-one?

 

[Tann-44]

Just to add on car amplifier's they use dc to dc converter within the psu..
12volts dc in then it gets split in to a daul rail voltage of plus and minus say from 20 to 40 volts on some car amps, I had a spl 800x years ago brought for scrap parts the main fault was dead! A failed psu ic plus s/c out put devices.. After fitting better devices it gave good acount of itself for a cheap amp.

 

[fezder]

Thanks for info, too bad thought, this 'aint car amp :s.....''luckily'' i'll try to make measurements with 'scope today, if anythink big comes up (hopefully not more smoke....)

 

[Tann-44]

Transistor power amp are straight forwards to solve for the more exp tech... Sometimes giving advice can be tricky due to not having the unit in front of you.

Using the scope is handy for full working tests running the repaired unit on soak test.. As out line check round the long tail pair(s) followed by the vas stage followed by the o/p ..

Common faults with amps are blown output devices this can lead to bias circuit faults such as the bias transistor going s/c or open..
A tip would be if one channel has gone use the working side to correct things.. Also check the voltage across emitter .33 5 watt resistors and re adjust setting.

I've seen some folks just pop in replacements and power up followed by over heating heatsink due to incorrectly set bias voltage...
The save damaging things use the bulb limiter tip limit current.. If all is well it should light them go out or glow very low...

 

[fezder]

yeah, it figures, pretty hard helping via forum only :S. i did compare the working side, as i posted before results. there was HUGE difference, but not now. havent tested it yet thought, now i should have time as easterns holiday (wohoo). err, i noticed couple terms i cant recognice due my language barrier, sorry. sould you explain:
long tail pair?
vas voltage?
soak test? just name for measuring?....
and, how to adjust biasing? if i understand, its correct resistor values to make transistor work on linear area, not as switch or something like that?

 

[Tann-44]

Long tail pair is the first set of transistors we see when looking at basic amp circuits and more complex in stages have current mirrors.. Vas stands for voltage amplification stage ie class A stage,,current mirror's are a simple design and work very well.. Bias circuit aids set up the out put stages..

There are many of us on here that carry out day to day service work within the audio field and will offer more Info on this subject plus there are many sites out there on amplifier designs to read up on..

Soak testing is done with scope/ sine wave generators dummy load on a repaired amplifier..

 

[fezder]

ok, thanks for info

 

[KeepItSimpleStupid]

Some quick indicators:
a) The voltage across the sandstone resistors in series with the output devices. Voltage vs time especially. The voltage should not continue to increase. In the 10-100 mV range. This is where the amplifier bias current is normally measured.
b) The voltage between the bases of the NPN and PNP outputs. about 2 Vbe Drops or about 1.2 V. Should change with heatsink temp.
c) The DC voltage across the speaker terminals. <0.5 V and preferably in the mV range

 

[fezder]

Some quick indicators:
a) The voltage across the sandstone resistors in series with the output devices. Voltage vs time especially. The voltage should not continue to increase. In the 10-100 mV range. This is where the amplifier bias current is normally measured.
b) The voltage between the bases of the NPN and PNP outputs. about 2 Vbe Drops or about 1.2 V. Should change with heatsink temp.
c) The DC voltage across the speaker terminals. <0.5 V and preferably in the mV range

hmm, thanks for info. i did measure dc voltage at speaker terminals, they were something like that. and i think i measured npn/pnp bases too, they were something like that. but, i havent checked those sandstone resistors yet, are those normally the big one resistors?

 

[KeepItSimpleStupid]

Base of the NPN to the base of the PNP.

Sandstone: Usually big and rectangular < 1 ohm. Measure w/ no signal. Monitor over time. Should not constantly increase.

 

[unclejed613]

that's the large grey resistors. 0.47 ohm/2 watt next to the output transistors standing up off the board. actually since there's no bias transistor (indicating fixed bias using diodes and no thermal compensation), you probably won't even have much more than a millivolt across these resistors, and maybe a 10 millivolts after the output transistors get warm. this is a very poor and unreliable way to make an amplifier. usually, even with fixed bias (such as in a Peavey CS-800), the diodes are usually thermally coupled to the heatsink. this amp has no thermal compensation, has underpowered output devices (i don't know what kind of math they used for determining the output power but it definitely is NOT 500W, and definitely NOT going to drive 4 ohm loads). at 33V rail voltage, if you clip the amp (drive the outputs to saturation) you will get currents that the transistor can handle without much difficulty (about 4A during clipping) but beyond the transistor's max current (8A during clipping, and the transistor is rated at 6A) into a 4 ohm load.

if we take the +/-33V, subtract a volt or two, the max RMS voltage out is about 21.9Vrms, which is about 60W into an 8 ohm load. it could be close to 120W into a 4 ohm load if it had proper output devices and power supply.... without thermal compensation, all bets are off for 4 ohms. looks like they did something fancy in the input/voltage amplifier stage like a current mirror or more likely a dual diff amp, but after that the designer cheaped out... this amp MIGHT actually be a good sounding 50W amplifier, but lacks a lot of things that could vastly improve it's reliability.

 

[fezder]

damm, you guys have huge acknowledge, i gotta say that . in good way that is, good that you are willing to help, even this seems to be unwise to repair....but for educational purposes, is there any stupid repairs anyhow?

 

[unclejed613]

yeah, there are...... like spending more than 1/2 hour fixing a $40.00 blu ray player..... which may have some educational value, but really isn't worth it financially.... you can pick one up at a thrift store for $10.00 or less... besides the almost-UV laser is a lot of fun to experiment with.....

 

[WTP Pepper]

In the UK (especially with PC based speaker systems) the problem is they quote PMOP - peak music output power over two channels (at least) and not per channel which is the peak-peak output voltage squared divided by the load. It's a meaningless measurement. Typically therefore on a 5 speaker system of 100W PMOP you probably get about 1W / speaker. Don't quote me on the exact figure as my calculator isn't at hand.
True power is the continuous RMS power.
Have you ever sat in a room listening to a stereo source with 10W RMS into 4ohms? It's very uncomfortably loud.

 

[fezder]

yeah, nowadays i dont care anymore about watts, i looks more into speakers density db/w and their output frequencies...i admit, i was fooled by this one .