Aguilar DB 751 Bass Amp output RC network fried

[nevetslab]

Usually after Cochella Music Festival & Stagecoach Festival in Calabasas, CA the broken gear begins to flow back into the shop. Among the casualties is this Aguilar DB 751 Bass Amp. I pulled the cover off, as this one had a prize or two rattling around inside. The prize was two halves of the ceramic housing off a 10W 4.7 ohm power resistor that overheated and blew the housing in half. This is the RC shunt network (4.7 ohm in series with 220nF to ground) following the LR output network on the output of their MosFET power amp.

At least Aguilar leaves copper on their double-sided PCB's, and has healthy traces everywhere. Though the fat trace between this 4.7 ohm resistor and it's 220nF companion overheated enough to burn the solder mask off, cooked the fiberglass below the copper, which is now soft and spongy, and, of course, broke down the adhesive bond. Circuit is still intact, while the cap is no longer a capacitor. The MosFETs (Exicon ECF20N20 & EDF20P20's, all TO-3 metal can) test ok, (being able to turn them on & off with an ohmmeter), and the only damage found is this network and the PCB area where they were mounted.

I sent an email to Aguilar, with photos of the damage, requesting the service doc's on the DB751 (I only have the doc's for the predecessor DB750). Usually, the common fault I find on the DB 751 is their 35A Bridge Rectifier failing. I'm guessing the amp somehow began oscillating, and the cap ended up shorting briefly? That would put the output current thru that 4.7 ohm 10W resistor and the cap. I didn't get any information on how the amp ended up on stage. I haven't yet powered it up, now waiting for parts, though concerned about the PCB foil damage & the fiberglass. I've thought about applying an epoxy underneath the foil to secure it, then insulating the exposed copper with some 3M Yellow Polyester tape, or maybe some of their Type 63 PTFE tape.



I'm not getting any resistance readings on the charred fiberglass in the region of the burn. I'm waiting to hear back from Aguilar as this is one of their amps, not owned by CenterStaging....one of the many companies that park their gear with us for industry visibility, as we mostly have major groups and artists flowing thru using our facilities as well as our supporting their tour needs.

Any thoughts?

[J M Fahey]

Yes, clearly result of HF oscillation.
Capacitor does not actually need to physically short, at high frequencies its impedance is real low.
Just calculate at what frequency it has, say, 20 ohm impedance and that is low enough to fry a 10 ohm series resistor.

Mosfets can happily oscillate at *many* MHz.

[nevetslab]

I heard back from Aguilar on the damage. They're having me cut out the overheated foil, scrape away all of the conductive burned fiberglass and then coat it using Corona Dope (which I didn't have, now on order). And, assuming the solder pads are still usable and can adequately support the replacement resistor & cap, wire those up using 12AWG PTFE insulated wire from the bottom of the PCB.

All I have here is 22AWG Teflon insulated wire (stranded), as well as 3M Type 57 Polyester Insulating tape, Fish Paper and 3M 1/2" wide 1 oz copper foil, adhesive backed, which I've used often in repairing PCB foil damage. The charred fiberglass was quite soft to the touch. I haven't yet gone back to remove the PCB from the heat sink, and see what I get in measuring the conductance of that material. I didn't get any reading on the 300M ohms scale (Fluke 8060A), while the Conductance scale reads up to 10,000Meg, when converting back to ohms).

The parts arrived, only I didn't know I'd be needing the Corona dope. Should have had some on hand all along, though.

I had to sign a NDA to receive the service doc's on the DB751, so, can't share those, nor the DB 750 docs.

[Leo_Gnardo]

Rather than try to re create the PC board traces, what about "floating" the components over the affected area?

[nevetslab]

Rather than try to re create the PC board traces, what about "floating" the components over the affected area?

Aguilar was quite specific about what they wanted. These still belong to them, so I'm following their requests. Yesterday, having received the Corona Dope, I moved the module back to the bench, sliced off the burnt wide PCB trace from the resistor's solder pad to the cap's solder pad. Only the remainder of the resistor's pad pulled right off as it had already broken down the adhesive. The caps' solder pad fared much better, and did allow cutting. I also had to cut away all of the ground plane that was underneath the resistor, and sliced it's path along the jouney towards the end of the PCB. Then, spent a good deal of time scraping away and down one full layer of glass cloth to the second layer, finally getting all of the charred remains out. The solder pad of the resistor at the edge of the board remained scorched, and I eventually stopped going deeper. The solder pad on the back side of the board DID remain intact, along with what's left of the plate thru. I then sanded down the new 'desert' area that was charred, and it looks like I can get this to recover.

The RC network is on the MosFET side of the output RL network. They wanted PTFE insulated 12AWG wire to connect between the resistor and the cap, and it being glued to the bottom of the PCB with silicone adhesive. I didn't have any PTFE wire that large here (20AWG is as large as I have in that, and they weren't buying it). I didn't want to get into a debate about just WHAT IS the effective AWG of that original 0.2" wide 2 oz copper, which works out to 400 x 10-6 sq in, or 24AWG). I used some shielding braid, with some 600 V rated FIT 221 shrink tubing over it, though I haven't yet installed it. I've just applied a second coat of Corona Dope over the 'desert'.



The third image is looking from the bottom at an angle, with the board backlit by my work lamp. I still have to pull the PCB assy off the heat sink in order to install that insulated braid. I had suggested using some copper foil tape, applied on the top side, over a couple layers of 3M Type 56 insulating polyester tape, but that too was struck down.

Once the repairs are in place, I still have to see if this amp is a power oscillator or not. No idea what the conditions were when it became one and pitched a fit.

[Jeema]

Once the repairs are in place, I still have to see if this amp is a power oscillator or not. No idea what the conditions were when it became one and pitched a fit.

I'm not real familiar with the conditions that could cause this, but would it maybe be a good idea to install a 1A quick blow fuse inline there at least for testing purposes? ...that way if it happens again during testing you aren't back to square one.

[nevetslab]

I'm not real familiar with the conditions that could cause this, but would it maybe be a good idea to install a 1A quick blow fuse inline there at least for testing purposes? ...that way if it happens again during testing you aren't back to square one.

Good idea, but as the nominal AC mains current at idle on these is 1.5A, I'll go a bit higher. Normal practice in the uncharted territory region.