Anyone willing to hold my hand to get me back on my feet with a car amp repair?

[OldRad]

Hello all!
First post here so a bit about me...
It has been a LONG time since doing anything electronics related.  I graduated in 93 with an associates in electronics engineering.  I did a bit of EE work for a few years after graduating, but then moved into IT.  Since then my electronics knowledge faded away.  I miss it though.  The challenges, the feeling of accomplishment when fixing or designing something, etc.   So here i am 20+ years later with blown car amplifier.  I could easily go buy a new one for 150ish but what do i do?  I go buy a bench top power supply, an oscilloscope (Man have they come down in price since 93!!) a couple of meters (amp and dmm), a soldering station, a cheap esr meter, and a digital smd soldering microscope.   3 times what a new amp would have cost me.  So needless to say, i am now committed and determined to fix my amp.    
I'm still waiting for some of these devices to come in.  PSU (tomorrow) and o-scope (next week sometime) are the big ones.   
This amp repair is really to get me back on my feet again, so i can start walking down the path of designing some things that have been in my head for a bit.

Now, a bit about the amp.  It's an infinity reference 7520a. It flashes the the protect LED and seems to power cycle itself.  It's a constant loop of this.  This happened shortly after jumping someone else's car.  My thoughts are that it will be on the power side of the amp and not the audio side.  Initial inspection of the amp does not show any burnt components or traces.  But does have that overheated silicon smell. (it's amazing how one doesn't really forget that smell)  I've already downloaded the schematic for this amp and have started to get familiar with it while i wait for the gear to come in.

Thanks for taking the time to read this and if anyone is willing to help this older geezer, i would appreciate it.
-OR

[Whales]

You can test many of the parts on the power input using a DMM.

Fuses, including any items that are in series with the power input (some types of fuse look like big ceramic caps or resistors): ensure they're open.

Transistors: ensure they're not shorted out.

Question: are there any tantalum capacitors in this unit?  Some pictures are worth a thousand words.

[OldRad]

Thanks Whales.
"Question: are there any tantalum capacitors in this unit?"
It does not appear so.  Electrolytic, Mylar and ceramics according to the schematic.

"ensure they're open."  I'm assuming you meant closed?

[oPossum]

The best resources I know of...

Forum:  https://www.diyaudio.com/forums/car-audio/

"Book": http://www.bcae1.com/repairtutorialcdinfo.htm

Video: https://www.youtube.com/playlist?list=PLS8D4j9zEOwp6m2rDR1p3xjT1E5Km3aML

[Mechatrommer]

do your audio amp looks like below? note the main ingredients:

1) big toroidal transformer driven by smps chip (center left) and...
2) power mosfets for smps switching (6 in a pack top left)
3) power diode rectifier (2 in a pack bottom left)
4) power bjt elements (4 in a pack both top and bottom right)

the count number and location of power elements may differ, but thats the general ingredients...

[oPossum]

Service manual: http://www.compendiumarcana.com/forumpics/infinity_reference-series_7520a.pdf

A good first step for car audio amps is to check for shorted power supply MOSFETs and output transisitors. (With power disconnected of course!)

[oPossum]

The protect LED will turn on for:
- Over temperature
- Output over current
- Output DC offset

Most likely suspect is shorted output transistor(s)

[OldRad]

Thanks for the links oPossum.  I will go give those a read.
Mechatrommer, here is a pic of the amp.  0 SMD's.  Which is perfect for my failing vision.

[Mechatrommer]

there are few trick i learnt. the easiest to isolate fault is disconnecting diode rectifier (i guess the 2 black smaller TO-220 below left)...

1) if short still exist, fault is on primary (left) side highly suspect is switching mosfets (4 in a pack top left), desoldering them out and check with transistor/diode/resistance tester (i found cheap china germany originated mk8 transistor tester to be magic in this task), replace bad units and turn on again, i advice only installing one for each mosfet type for the start to avoid risk burning unecessary mosfet in parallel, they are 2 pack there, each pack 2 mosfets iirc switching alternatingly.

2) if no fault while disconnecting diodes rectifier, secondary (right) side is faulty, remove all the 4 power pnp and npn bjts on the right top bottom and check for faulty unit, replace bad one (or two or more).

dont forget to check the diode rectifier themselve for faulty. if everything replaced with good unit, turn on and check functionality, if short still exist, the driver components is faulty, those little little but many components there, it will need more works, diagnostics and probing if those are faulty. transformer also can be faulty (burnt short inside) thats not impossible, if primary side is checked to be faulty but mosfets are ok, drivers are ok, transformer can be toasted. that in a few shorter words description, repairing can be more complex than this, if you lucky 10 minutes fix it is, if you unlucky, it will prove if you are the man. hope that helps ymmv.

edit: if you have thermal camera, that can be a golden too. i bought one after my audio amp fix adventure. and i forgot one important thing, i diyed a simple current limited 12V psu (iirc a simple comparator based) as my main workhorse during diagnostics (now i have a proper one longwei brand ). without it we can keep burning mosfet/bjt until the shop is empty. iirc above 1-2A current consumption during idle is considered fault, more like 0.5A idle is normal functioning operation.

[OldRad]

Ughh, It took me about an hour to decipher what you said.  But i think it makes sense. Please correct me if i'm wrong.

1.  The bottom left TO-220 packages are YG225D2 and obviously(now after my scratching my head for an our) on the secondary side of the transformer.
According to the schematic it looks like these rectify the L&R +33/-33V rails??  There is a 5T:12T transformer just before so i assume 14V from the car battery through the 5T:12T to get 33V??? (though i havent figured out where and how that is getting converted to AC before being sent to the transformer yet)
I have a MK-328 on the way. Not sure when it will get here though.
 
2. The 4pack of fets in the top left are FQP50n06.  all the gates are being fed by pins 9 and 10 of the TL494CN (this chip i assume sets up the PWM carrier of a class D amp?)

3.  These guys are TIP 35(6)C's and they carry an amplified sine wave from the +/-33v rails to the speakers??

I wasn't lying when i said it's been a LONG time since doing any electronics work at all.

[oPossum]

Everything on page 19 is the power supply. TL494 is the power supply controller. It is very common in car audio amps. The MOSFETs driven by the TL494 alternately switch ground to the primary of the transformer. YG225D2 are the rectifiers.

Everything on page 20 is the crossover and amplifier. TIP35/36 are the output transistors. Q151/251 detect over current. Q181/281 detect DC offset. All other transistors are part of the amplifier. It is a class AB amplifier.

[Mechatrommer]

Ughh, It took me about an hour to decipher what you said.

lol i try chinglish here, fewer words to get the most infos... i tried standard english (longer words) people dont understand me anyway. dig deeper into the circuit may reveal some of the senses...

1.  The bottom left TO-220 packages are YG225D2 and obviously(now after my scratching my head for an our) on the secondary side of the transformer.

yes refer to 1a.png below... a bit strange you have 2 rails for each Vcc and Vee, 4 rails all. the few i repaired only 2 rails, so i had the green rectifiers removed only (i guess thats the 2 black chips on the bottom on pcb. now you have another rectifiers purple in picture, but i guess thats for low powered rails to feed drivers and 15V ic's.. i draw each side secondary and primary for you so you dont have to watch hours of youtube video that usually have many unrelated rants to figure out what... red are highly suspects, those 4 mosfet top left on your pcb, have them removed all and checked/replaced for damage status. yellow are mosfet driver or protection something, less likely to be damaged, blue is the heart of smps TL494 (Vac that you have a headache of how is generated) very less likely to get damaged. you can probe those pins output to have square signals, if not, that probably dead and no power no over current warning, not your problem. your problem is dead short, have that red color mosfets checked asap.

According to the schematic it looks like these rectify the L&R +33/-33V rails??  There is a 5T:12T transformer just before so i assume 14V from the car battery through the 5T:12T to get 33V???

yes. thats the big toroidal transformer you are talking about, very hard to miss on the board. thats the heart of the power.

(though i havent figured out where and how that is getting converted to AC before being sent to the transformer yet)

you can study later it can be wonderful thing how 12Vdc can get electrocuting high voltage Vac on the output. but first you have to make sure components are ok first, check the red color and then we can continue talk.

2. The 4pack of fets in the top left are FQP50n06.  all the gates are being fed by pins 9 and 10 of the TL494CN (this chip i assume sets up the PWM carrier of a class D amp?)

yes thats what we keep talking about, have that checked. no, its not for class D amp, your amp is AB only (no digital), class D is different beast, but only difference on secondary side, primary side still the same as your AB, the PWM is what makes Vac, for higher rails to feed your class AB (and D, not yours) amp on secondary side.

3.  These guys are TIP 35(6)C's and they carry an amplified sine wave from the +/-33v rails to the speakers??

now you bring us to secondary side, refer 2a.png... red are those tip 35/36 pair giants (red colored) have those checked as well, another highest suspects on secondary region. forget about me talking to isolate secondary side. have them all checked is the easiest without moving a brain cell (all the red colored, and then later green and purple rectifier all those are hard to miss on the pcb, the all black box transistors 3 legged alike).

I wasn't lying when i said it's been a LONG time since doing any electronics work at all.

you ask dave to repair tv or something else he can get nuts. different device needs different experience to repair. we have people here expert in digital or RF, but know rat arse about power electronics. vice versa. this is very broad field, some are highly specialized into one area, some are jack of all trades, but not so expert, only armchair keyboard warrior like me

[OldRad]

lol i try chinglish here, fewer words to get the most infos... i tried standard english (longer words) people dont understand me anyway. dig deeper into the circuit may reveal some of the senses...

It wasn't your English that i was struggling with.  That was perfectly fine.  It was translating that, to what i was seeing on the schematic and trying to work out functions.
Picture a child sounding each letter in a word to form the word, and then trying to make a sentence.  That is what took me so long.
Thank you for breaking the pictures down in color codes.  That helped quite a bit.

I shall see what i can get done today, sadly after the wifes to-do list. 

-OR

[OldRad]

Rectifiers are out. Solid blue light on the amp.
Now i have to wait for the MK328 and oscope to show up before i can continue on.  None of the power fets pins are shorted (in-circuit). I do hear a  short continuity blip, then it goes away and values increase.  I assume this is the charging of caps.  I will probably replace all the fets anyway as the amp is almost 13ish years old.  I really am afraid it's going to be in the driver section.  We will see though. 


I really do appreciate the help ALL of you have given so far.  Thanks.

-OR

[oPossum]

An in-circuit test of bipolar (NPN/PNP) transistors can help determine what should be removed for further testing. Use the diode test feature of your multimeter to test base to emitter and base to collector. Should be 500 to 700 mV. Failed shorted and failed open can usually be detected with this simple test. Suspect transistors would be the TIP35/36 outputs and the C1027Y/A1023Y predrive.

[OldRad]

Ok, So my O-scope and MK-328 came in today.  I haven't had a chance to use them yet, but i plan on getting to some testing tonight.
The YG225D2's I tried to test these on the DMM while they are out.  Shouldn't these guys behave like normal diodes?  I do not get continuity with either polarity, in resistance mode it's about 1meg on each junction. But if i hook them up to a power supply with about 2 volts, 2ish volts come out.  What kind of sorcery is this?!?! 

[oPossum]

You have to use diode test mode, not resistance.

[SilverSolder]

They should act like regular diodes,  if you hook them up to your power supply make sure the current limit is set low enough that they don't blow up!

Say you set your power supply at 3V and a 100mA current limit.

You should see 0.7V one way,  and 3V the other way when you connect the diode in both directions.

[Mechatrommer]

But if i hook them up to a power supply with about 2 volts, 2ish volts come out.  What kind of sorcery is this?!?! 

YG225D2 diode will form 0.6V across it so if you put 2 volt into anode you should read 1.4V at cathode (practically this reading could be higher 1.7V something ie lower 0.3V Vf diode). with diode tester you can see this 0.6V reading when red probe to anode black probe to cathode (no external power source needed), on my DMM, diode tester is shared knob position with continuity and resistance tester, so i have to press a button to switch between mode. if diode damage shorted, 0V across diode and continuity tester should beep but in your case, no continuity Mohm resistance, but voltage reading is registered, i think the rectifier still ok, you just need to get that 0.3 - 0.7V reading in diode tester mode across it to confirm. if you put wrong probes across pins, no voltage reading ie 0V, ymmv.

[OldRad]

That's what im saying.  In Diode mode, regardless of polarity i do not get a reading.  But it's passing voltage. 
I truly hope these 2 parts are ok. They are obsolete and i cannot find a replacement.  I see some for common A or common K but not not a series one in the direction i need.   The series one that is in the wrong direction.... Could i turn it around?  Logically i would think so, but it wont mate with the heat sink properly.

I was hoping to get some readings done last night, but most of the night was spent calibrating the scope.  Hopefully tonight.

[Mechatrommer]

you can try asking your electronics local store for compatible parts (equal or more than 10A) but in my place, if i ask can this be replaced by this? they will shrug. they can only sell but can give zero advice, so i have to work it on my own. they only want the part number and will happily to take the part out of their drawer, just like highly advanced autonomous robot that can only recognize numbers and letters. what they sell is what usually found in car/home/hifi audio amp/player and regular consumer electronics parts.

https://www.ebay.com/itm/YG225D2-FAST-RECOVERY-DIODE-TO-220F-YG225D2/323738430567?hash=item4b60506c67:g:xysAAOSwd4tT-gyN
but quite expensive...

its surprising difficult to find series configuration in ebay or digikey, you can always improvise, this is just diodes. you can connect together the 2 legged TO-220-2 version such as...
https://www.digikey.my/product-detail/en/on-semiconductor/MBR10100G/MBR10100GOS-ND/918577
and then glue or screw onto the casing/heatsink, use flexible wires for conection to pcb, but this requires artistic work if we want to resurrect obsolete item.

another compatible replacement i can find in my database is F12C20D, F12C40D, F12C60D but its difficult to find either... but these parts, including power mosfets and bjts are easily replacable with equal or higher spec if we know which spec to look at, max current, Vce max etc...

The series one that is in the wrong direction.... Could i turn it around?  Logically i would think so, but it wont mate with the heat sink properly.

yes certainly, but the problem is as you mentioned. you can make another artistics work by drilling 2 holes next to the inverted package and put bracing aluminium bar pressed by 2 screws into the newly drilled holes, ymmv.

[OldRad]

Ok here is the latest.
I ordered the Fets for the power side, and the transistors for the output side.  Before replacing them i decided to power the amp up and get my baseline. (rectifier diodes still out at this point)  Now the amp is going into protect again.   Ok fine. i have the parts i'll just swap them ayway.  All Fets and transistors replaced.  Still going into protect (sill with the rectifier diodes out).  So i begin to look for less obvious things. I went to the tl494.  Pin 5 has a sawtooth wave, but man it's NOISY.   When i looked at this a few days ago it was a perfect sawtooth.  I dont remember freq exactly but i think it was about 25K.  Anyway,  So i look at pins 9 and 10.  It starts to build a (Noisy) square wave, and that is when the amp goes into protect.  I see the same thing at the fets. 
While it's cycling through this protect loop, i've been feeling for anything that is getting hot.  Nothing besides a couple of output 1 watt 330ohm that get hot but not scalding to the touch.
I think  at this point i want to narrow the power supply side first, then move on.
Ideas?

[Mechatrommer]

now we get deeper... lets go to "non-memory" item (attached picture, page 6 troubleshooting section from amp manual)... tracing the schematics, maybe your amp is protected (led on) when there is no ±33V rail on 2ndary side. i guess PROT node on the bottom of primary side is relevant, tracing it probably ceases TL494 duty cycle through deadtime pin (HI) and turning PRT-LED on through R90, Q90 on TL494 pin 4. the schematics gave useful infos throughout the nodes on voltages that you can check (i guess during idle, ie no sound present). for example in normal operation, pin 4 should be 0V, and 5V should present at TL494 pin 13-15 etc. not sure how 5V rail is generated. but now, the diode rectifier is highly suspect, replacing it with new part should restore ±33V rail and disable protect led.

make sure your newly installed mosfets are not hot or damaged in case problem is from different source. you can confirm the mosfet's drain not shorted (top of red mosfet Q44) is solid 12V during no PWM or during protect led on, but going low when there is pwm. and also check the diode rectifiers (purple colored in my earlier post), to ensure proper operation on 2ndary side. since 2ndary and primary side is highly coupled, they need to work altogether for protect led turned off. trying to decouple them is a bit of work probably cutting traces, so its not advisable.

if you dont have the YG225D2 (green) part at hand, maybe you can decorate 2 series diodes together (with current capability up to the amp spec) and connect to respective traces on pcb temporarily to get back ±33V rail to see if protect still activated. be careful

about your noisy pwm output from TL494. make sure you have short and solid gnd wiring connection to your oscilloscope probe, long gnd probe may show excessive noisy or ringing signal, ymmv.

[OldRad]

I was actually doing a bit of this before i read your reply.  So it seems my thought process is Ok.
So starting from the top.
My supply is set to 9v for testing So numbers are a bit off but Close'ish enough.
Test points around q1 and q2 OK. R05 and R60 OK, Q11/12 Ok, Pins 14,15,8 TLN ok.  Pin4....... No.   +1.5 volts.  If i short this to ground the power light goes solid and protect loop stops.  Though i have not tried to scope it with the pin grounded.  (not enough hands). With that my thought process was to see what rides on pin 4 and look at those circuits,  No short on c12. My next thought was the thermistor, It's a 50K NTC it's reading 5.1K, it drops when i put my thumb on it. 
but i cannot remember if you can accurately read resistors in circuit or not. 
If my theory is correct even with all the other stuff in the circuit, i get 50K. (TH1 50K + r25 68K + r21 24K +r90 47K /4) + 750 of R16.
Mosfets seem OK.  I will confirm again shortly.
I will upload some pics of the waveform on pin 14, as well as a shot of what the MK-328 says about the rectifiers.

Thanks mech

[oPossum]

My supply is set to 9v for testing

That may be a bit too low. The amp will go into protect if it detects that the supply voltage is too low or too high.