What do you mean when you say TR4 and TR5 continue to fail ? Do they measure correctly on a DMM Ohm test ?

R46 and R47 will run hot. It is best to mount them up off the PCB for better heat dissipation.

Reconnect R29 and R30. With TR4 and TR5 removed, check that there is roughly 1.2V to 1.4V across R74. This verifies that D8 and D9 are doing their job to bias TR4 and TR5.

Check R32 and R33.

Remove that resistor across the tank's input. With a good TR4 and TR5, Pin 1 of IC4 should measure close to 0V.

Check the ripple on the filter caps. Is it 100Hz as it should be, or 50Hz if the amp is operating as two half wave power supplies.

What I find curious about the scope photo in post #9 is that the little peaks are 10mS apart but of alternating polarity. The only thing I can think of is that one or two of the diodes have failed open in the bridge rectifier effectively making it two half wave rectifiers. The plus peak of the 50Hz makes a little positive bump in the ground, then on the next half cycle the negative peak makes a little bump of opposite polarity.

Quick edit: Could be that one side of the transformer winding has failed open.

The tab on a TO-220 or the case on a TO-3 is connected to the Collector on a PNP or NPN power transistor. That's because the backside of the Silicon Die is mounted on the metal for best heat conduction. (The same applies to the Drain of a power MOSFET.) On a Positive Voltage Regulator chip, the backside of the Die is the Voltage Adjust Pin so that may be grounded depending on which Regulator you are using. On a Negative Voltage Regulator, the backside of the die is the Incoming Negative Voltage which is almost never grounded.

There are exceptions to these rules, you must consult the part's data sheet for the best information. On an Adjustable Positive Voltage Regulator, the Voltage Adjust pin goes to a couple of resistors connected in series between the Output pin and ground so the case must be insulated from the heatsink.

Connecting a big heatsink to anything except ground is considered bad design because people want to touch them to see if they are hot. They don't want to be shocked or killed. Little PCB mount heatsinks may not be grounded, don't touch unless you verify that there is no hazardous Voltage present.

Is this one of those transistors that looks like a TO-5 welded to a small heatsink ?

https://www.ebay.com/itm/364778815801?_skw=40409+transistor&itmmeta=01JT2DBRD9SAA9TZ9G682QKNNS&hash=item54ee830139:g:qXsAAOSwxj5l7SZo&itmprp=enc%3AAQAKAAAA8FkggFvd1GGDu0w3yXCmi1ebO2I4P7B0y9cqe%2FLpuc%2Bcnd4BcCLtZN5aDO3gCTpR4j2Ux3uXDH%2FqkdI41qdvCdP7JuQwV7hz%2F1EnMO%2FJzzCaRnfkXHRkL1eRogNyyy3RDCu8EnNiZ3eK502CevyOHC6OLfQMxihEPEFIFiH9gAuzevo4y5AsbW%2FlRe9ZSgTfo0QHfFea29kNXL4m7n2NTNK9avTWxRByjIPHnmdu4%2FIjO5yJ6u4FCDPBPxezVT8ggb35hV1igphYxuPOTY0iE--sLBXmHuTQa9R%2Fjnm88p%2FOaxLeXtlppPy2WdJ2%2FSsg7g%3D%3D%7Ctkp%3ABk9SR96Gr83QZQ

What are the little circuit boards attached to the lid ?

Operating the CMOS without feedback may emphasize higher frequencies where the impedance of the ear buds is higher. Try setting the gain of the output stage to something between 1 and 10.

The output stage doesn't use any feedback to lower the gain. Check the old National Semi application note in post 4 of this thread:

https://music-electronics-forum.com/forum/amplification/solid-state-hybrid/922490-what-does-this-ic-do-in-the-circuit

I was wondering what it would be like volume wise to have four 16 Ohm speakers and be able to select one, two or four in parallel with a footswitch ?

Most of these things use a regulated SMPS so they can come pretty close to doubling power with half the load impedance.

The 2N5484 should work fine in this application, and the pin configuration is the same.

Did you check the resistance of all the resistors before you installed them ? If that 1M5 resistor is 1K5, the generator will drive it fine but a guitar will barely be heard. You can probably measure it without removing it from the build, just reverse the meter leads and take the highest reading.

Quote from: gerry113377 on January 21, 2025, 02:59:20 PMThen I plugged in a 450w 8ohm AMPEG cabinet, with a bass guitar, struck my first note and it faulted immediately.

When you plug or unplug a speaker cable equipped with phone plugs on the speaker cabinet end, you are briefly shorting the output of the amplifier. If the amp was powered on, the brief short could have killed the amp.

Sorry for the delay but here's what I got from my guy at TI. He's not an actual employee as far as I know, more of a paid consultant. The TI engineers are talking about those designs where you use two chips in parallel or two chips in a bridge drive configuration. I think Marshall has a design that uses four chips for amp. Hard to keep working blowing up chips all the time.

Yea LM3886 is back in stock. I've printed a few PCB batches
for a guitar and Bass power amps to build with them but after
talking with TI engineers they steered me away from them to
use Class D. I've yet to build any amps with the LM3886.  They
pretty much implied it was designed for single module application
or dual in stereo and not recommended for multi modules designs
to double the power. They said rail voltages needed to achieve the
result I expected would exceed the maximum continuous voltage
and cause them to fail. So I get they exaggerated the true power
output and adding additional chips don't increase total power
output in expected quantities. They also exaggerate about the
Class D chips. In reality the advertised output is at 2 ohms and
maximum rail voltages before it burns!! So if you want 200 watts
in the real world you must get 700watts to come close.

Quote from: J M Fahey on December 30, 2024, 07:31:41 PMI love LM3886 but they are out of production for years now.

Only non original ones freely available.

I don't think this is true. There was some availability issues around the time of Covid, and there was a switchover from National to TI when TI bought National, but I think TI is still cranking them out. I will check with a guy I know who works at TI, may take a day or two.

That's not to say there aren't fakes out there. Like always, buy from a known good source like Mouser or Digi-Key.

I was thinking that a Cliff type jack might briefly short the cable as the Ground contact of the jack slides over the insulation between the tip and sleeve of the plug.