My P68 sub amp just went poof - help me fix/troubleshoot it!

[phase_accurate]

Hi Jonathan

Just two things:

How does your powersupply look like (i.e. it's schematic) ? Do you use a centere-tapped transformer ?
I ask because there are also possibilities to make a symmetrical PSU with a single output winding only, but this would be inferior to the centere-tap version. I.e. it could lead to situations where the output voltages can get heavily asymmetric and components can be killed of overvoltage.

In Rod Elliotts article the following statement can be found:

The S300 is intended for intermittent duty on 4 Ohm loads, as will normally be found in a subwoofer.

Now I don't know how long your 20 Hz test took ??

Regards

Charles

[Rarkov]

Hi,

I had exactly the same problem with my amp. It turned out that a cap blew (between rail and GND) because of it's rating (about 2V below input). It took about 10mins of playing to blow up.

On a seperate occasion, two resistors blew up (?!)...I replaced them without changing a thing, and it has been running flawlessly ever since...About a month later, I measured some '1k' resistors only to find that the shop had labelled them wrong...They were 100R! I should have checked them - but I have to agree that 5 band resistors are hard to read!

Anyway - I think that accounts for both blow ups...Not sure if the first one is relevant at all. The caps never blew under test - only in use...Try using well overrated (> 2xrail voltage) if you can get them and run them like that for a while...
Gaz

[Jonathan M]

Hi all.

First off, the cap was certainly connected around the right way. Secondly, the DC protection is for the speaker only - this is an expensive driver (for me, anyway) so I figured it was worth the $10 or so it took to make sure it was never connected direct to up to 110V DC!!

The power supply is from a centre tapped transformer (40-0-40 secondary) to a bridge rectifier and then to 6000uF/100V per side caps. Rod suggests 4700uF per side in the S300 article.

The 20Hz "test" ran for a Maximum of 30 seconds. I started it at a very low level, then slowly turned it up, watching the woofer pump in and out. After a very short time (4 seconds max) the woofer stopped moving. There was then some relay clicking (Likely due to the faulty signal detect circuit turning the mains on and off - now remedied), and around 5-10 seconds later the cap went pop. I unplugged the power cord as quick as I could at this stage! I should, ofcourse, removed power immediately that the woofer stopped moving. Lesson learned.

I have tested all components - they are all the correct values (within tolerances). The amp also tested fine (correct bias current/little or no DC on the outputs etc.) when I first built it.

The transistor that blew is on the -ve supply rail side. The cap is on the +ve rail. It's the +ve rail fuse that blew. I have thought that perhaps the cap was just running too close to it's max operating voltage (56V versus 63V) but I would think there would be a little bit of head room there. The replacement I've got is another 63V one at present - I may see if I can get hold of some 100V parts locally.

I'll replace everything tonight (Assuming the postie arrives with my replacements) and recheck everything out. I'll probably test without the protection circuit first. The driver is a nice, solid unit, and I'm sure it could withstand plenty of DC for short periods if it has to. (It's rated at 650W)

[jarek]

Hi Jonathan,

Are you sure that the contact between transistors and heatsink is good? I had the same problem when I tested one of my amp with a signal. After a some time running transistors got overheat and blew. I find, that these transistors weren' t properly installed on the heatsink.
Maybe it helps.

Regards,
Jarek

[Jonathan M]

Hi all,

I've replaced Q6 and C13 (The dead components) and repowered with 22ohm safety resistors on the rails. They got hot so there is clearly still a fault. After reading a couple of other threads, I've removed all the output transistors and checked them using an ohm meter between B-E and B-C.

All measured lowish readings in one direction and high readings in the opposite which indicates they are fine - EXCEPT Q7 and Q12.

Q12 was shorted between B-C and Q7 had the same reading in both directions between B-C and B-E. This indicates they are dead. I will check the rest of the transistors later today.

Also, do you guys think there is a way to test the rest of the circuit with power applied with all the output transistors removed from circuit? (Q7 through Q12?) If so, then at least I can verify that section.

Thanks for your help so far - hopefully I can get replacements before Christmas!!

I've attached the schematic...

Cheers
Jonathan

[moamps]

Hi,

if C13 goes down, then voltage on C13 was to high, this happen when - pol of C13 is not connected to middle- ground point on power supply and is floating together with + pol of C12, check that all ground point (-C13, +C12) are connected to ground point of power supply, check R23 (for testing could by shortened)

one output section of transistor must be connected for testing,

try bypassing one diode D2 or D3, idle current must be smaller(check voltage on R16,17, or on safety resistors), output voltage must be around 0V

you must check and C-E resistance (both polarisation, high R) when you test BJT

regards

[Jonathan M]

Hi all,

Got my replacement parts, and everything is back running. The amp runs (with no load) with supply rails at +- 53V. Output DC is less than 30mV.

I'm a little worried about the bias current. Supposed to be around 55mA, but is running high at over 85mA. The driver heatsink (on which Q4, Q5, Q6 are mounted, and D2 and D3 are in contact)
gets quite warm - even after a few minutes. Bias current increases as the heatsink gets warmer (Logical) but everything is getting a little warm for my liking.

Moamp - will bypassing one of the diodes D2 or D3 lower the bias or increase it? I might try this if I'm sure it will reduce the bias current.

I'll recheck all the thermal contacts tonight to make sure everything is staying cool, and try and identify which components are warming up the most.

Cheers
Jonathan

[Jonathan M]

Hi All,

Well, as I said in my last message, the amp is testing OK as far as the drop to the rail voltage (via safety resistors) and DC output is going. Unfortunately, I'm still having problems.

R14 (220R collector resistor of driver Q6) is getting warm and starting to burn up. Has around 20V or so across it so is dissapating far more than it should be. R13, the resistor from the other rail to the other driver (Q5) has only 0.65V across it, so is all good.

I'm thinking that perhaps Q8 is to blame?? R14 goes between the -ve rail and the collector of Q6, or between the base and emitter of Q8. Should there only be a 0.65V drop across B-E of Q8 if it's working properly??

Any help/guidance would be fantastic.

Cheers,
Jonathan

[jarek]

R14 (220R collector resistor of driver Q6) is getting warm and starting to burn up. Has around 20V or so across it so is dissapating far more than it should be. R13, the resistor from the other rail to the other driver (Q5) has only 0.65V across it, so is all good.

Hi Jonathan,
I see you have a problem.
I think you can run this amp without Q9, Q11, Q10, and Q12. Also short two diodes D2 and D3. This will minimize bias current.
Before next running I would check Q6 and Q8 (can be blown). The same for positive rail transistors.
Check values R10, R11 and R14. Aren't the diodes D2 and D3 installed in backwards?
Good luck
Jarek

[moamps]

Hi,

Q8 is blown, V across R14 is must be 0,65V
If you short D2 and D3 , Iidle will be around zero.
Are collectors of Q4, Q5, Q6 isolated?

regards