Odd Sumo The Nine Plus

[thefragger]

Good lord AK has changed a bit since I've been gone...

I've got a bit of an oddball, here.

A disjointed album of photos can be found here: http://imgur.com/a/O95G4

I can't figure out what's going on, here.

I've had this amp for a few years and I'm now finding time to get into restoring it... but I've found a few oddities.

The transistors in the right channel are all Sanken 2SC1116A but the left channel is all Fujitsu (?) 2SC2522. The right channel has C2238 drivers while the left has TIP31C drivers. There are a couple extra diodes on the solder side of the left channel's PCB, and all the quick disconnects are soldered in place. The torroid is unshielded, from what I understand, indicates that this is a later model amp.

Of the three schematics I have, I can't find a single channel that matches anything of what I've got here.

My best guess is that this is a late-model amp, with the left channel being stock, and a repair was done with an old right channel swapped in as a module.

After verifying that replacement output devices are available (unlike The Nine, from what I understand) I tried powering the amp up while measuring the DC bias, no load. I saw -15 VDC in each channel, which slowly increased to -13.7 VDC, and that's when I smelt that burning smell, and turned it off. All the fuses are their suggested value, and all the fuses are intact.

Ultimately I'd like to thoroughly restore this amp, giving it the best I can buy, I just need a few pointers on where to start. I'm fine with stripping this back to bare PCBs and changing out all the output devices should that be what needs doing.

I appreciate any insight and input.


Thanks!
Philip.

 

[FootFungus]

I think your assessment on the amp disparities is correct. Bongiorno liked the Sanken, and Sumo switched to the Fujitsu (cheaper probably) later. The original 9 sans plus model uses Sanken 2SC1830 superbeta transistors for outputs. The gain is 900+, but they are not darlingtons.

Are the opamps AD823 on both sides? Have you tested all the outputs, drivers, MPS-UX5 devices for shorts?

The op-amp supply is 20v, so those are likely selected units. I know the OPA2604 is recommended for the non-plus.

 

[thefragger]

Making progress on this; looking at questionable soldering by some P/O and questionable part replacements. Additionally, quick disconnects for the thermal switches were removed and the wires soldered directly to the terminals.

​

I think your assessment on the amp disparities is correct. Bongiorno liked the Sanken, and Sumo switched to the Fujitsu (cheaper probably) later. The original 9 sans plus model uses Sanken 2SC1830 superbeta transistors for outputs. The gain is 900+, but they are not darlingtons.

Are the opamps AD823 on both sides? Have you tested all the outputs, drivers, MPS-UX5 devices for shorts?

The op-amp supply is 20v, so those are likely selected units. I know the OPA2604 is recommended for the non-plus.

Yup, AD823 on both sides. That said, on both channels they live in sockets, and the sockets are different styles between the channels. OPAMPs have date codes from 2003.

I have yet to test the MPSU55 devices yet. My plan is to replace all the output devices with new parts and I have a desire to replace the drivers just because.

 

[nj pheonix]

When you eventually power this back up I'd start with a DBT

 

[thefragger]

When you eventually power this back up I'd start with a DBT

Why?

 

[FootFungus]

If the AD823s are not original then you probably want to pop in some OPA2604's to be safe. The AD823s are only guaranteed to be good to +- 18V rails, and these are running at 20v.

The MPS-W55/W56 are the same parts as the MPS-U55 but in the TO-92L package rated for the same power dissipation. I'm not sure I'd use them as a sub on the regulator, but they will likely do the trick for the bias transistors.

If you don't use the DBT, then you probably should use smaller quick blow fuses when first powering up and biasing.

 

[thefragger]

I have no idea where they came from, but they will be replaced.

This is how JB selected OPAMPs, for future reference:

"Hook up a standard gain of 10 circuit with a variable tracking power supply (up to +/- 25V). Drive the opamp to say about 10 volts RMS at 1KHz with the supplies at +/- 18V to start. Null the output on a distortion analyzer. While watching the residue in X-Y mode, slowly turn the supplies up until the output shows cusps beginning to form on the residual. At that point, note the supply voltage. There's your limit."