Roly, perhaps you could comment on the origin of the term "half rail".
To someone who has only ever seen split supply designs, it may not make sense without context.   

Quote from: phatt on March 19, 2015, 04:57:14 AM
Re the deebee watts thang.
I've read that plus 3dB is x2 louder.
Q, From that rule of thumb, Would I be right in assuming that a speaker with a 103 dB is twice as loud as a speaker with a 100 dB rating?

10db is generally considered as sounding "twice as loud" to our ears.
I think maybe you read that a 3db gain was equivalent to having an amp of 2X the power?
That is true, doubling the power of  amp (ex.50w to 100w head) will give 3db increase, the equivalent of a speaker that has 3db better sensitivity.

  Have a real good look at the solder connections on the pots and jacks.  Maybe just resolder them anyway.

 In that spider 3 manual, look at page 4 of the pdf.
It shows how the output IC is mounted.  If it has a grey silpad like that (no rips or damage) you don't need heatsink grease.  The silpad insulates the back of the IC from the chassis.
  Now the tab of the output IC needs to be insulated from the mounting screw.  This is what the plastic shoulder washer does.  The narrower part of the shoulder washer fits into the hole in the tab of the output IC.
After the shoulder washer, there should be a flat washer, split washer, and nut.
After it is all mounted, check that there is no low resistance between tab of output IC and chassis.

What are you using in the loop?  Try a plain patch cord.

  If it's not triple layer pcb, consider yourself lucky. 

Attached is the very extensive service manual for the III:

Yes, you should be able to measure ripple on the AC range, unless it is a real cheap meter that can't measure AC on DC.  You can check by measuring a AA or 9V battery with the AC range, it should read 0V AC.  If so, you should be able to measure ripple with it.

Quote from: Roly on March 11, 2015, 11:51:44 AM
Even with a serious blow up this class of amp is normally quite repairable.

If you mean the power amp, I agree  .  But there is still the original fault, which I don't think was power amp, and in this regard Line6 stuff can be non-repairable.
So there remains the possibility of going to all the trouble of repairing the power amp only to find you need a new "mainboard"  :grr.

  As far as the heatsinking goes, a fan will always help.  Even if there is no exposed heatsink, just pointing a fan at the chassis will help remove heat.  Keeping it cool is never a bad idea, but not really any thing to worry about unless you are noticing heat issues.

As to the age-old capacitor sonic debate, a couple comments.
Changing old caps will often alter the sound.  Is this because of the particular brand/type installed, or just that they are new?
Brymus, when you changed to new caps and your drummer noticed, might he have noticed any brand because they were new?  If you had then removed the new UKW's and replaced them with some other brand of new caps, would he have thought the sound was noticeably worse?
  The arguments are not so much about whether the sound was changed, but whether the change was due to materials (cap type), or capacitance values.
  No one ever seems to go to the trouble of making precise measurements and strict scientific method when comparing cap types.  If people did do this, and a sonic difference was provable, we would see frequency response charts on the manufacturer data.  It would gain them lot's of sales, so obviously they would make a big deal about it, with actual data.  If it can actually be heard, it can be measured.  Why does no one do this?  This controversy is very old, if there was measurable data, it would be everywhere.
  So it comes down to comparing apples to apples, as in the caps to be compared must measure exactly the same capacitance and ESR, not just within some tolerance.  Then precise scientific method must be used, double blind tests with recorded program material re-amped, with frequency response and spectrum analysis plotted as data.
  If there is found to be differences, then us skeptics will acknowledge the differences.
  If not, then it must be psycho-somatic.  And I personally don't have such a big problem with that, if it sounds better to you, you will play better, even if the difference is imaginary.
  But I think in a lot of cases, from the fact that the circuit sounds better, it is assumed it has to do with "brand/type x" caps, when really it has to do with properly working fresh caps, or proper/exact value caps.

  A plate to heater short would have the same effect on a 6L6.
Now the trick is to figure out if it was caused by a tube currently in the unit, or was the damage done long ago.
  The amp will continue to function fine without those resistors, it will only affect hum levels.

The heater resistors usually fail if there is a horrible fault in a power tube, or the locating pin in the power tube gets broken off and the tube gets installed with pins offset from correct location.
  The "screen stoppers" as you called them, serve 2 functions.  For their current limit functions, they can be located anywhere.  But for their "stopper" function (preventing oscillation) they are best located right on the tube socket, just like the grid stoppers.
  The vintage Fender way is good for this.

  And is the +15V holding up or not?  What does it measure?

Quote from: gbono on February 28, 2015, 03:39:55 PM
found Q20 (+15 rail series regulator) getting hot. It appears there is an issue in the mute/fault circuit since I can disconnect J2 and still have Q20 getting hot.

Are Q13 or U1 getting hot?  With J2 disconnected, what kind of resistance do you measure from +15 to ground?
Everything else in the Q20 circuit is ok?
The protect circuit sends the mute signal to R8 (near Q4).  And turns on the protect light.

 A few comments.  You mention "power supply rating", but that is the rating of what the amp uses from the wall outlet.  Just ignore that number, there is no way to say how efficient the amp is, only we know it will be less than 100%.  You can't really get any useful info out of the 120V rating as far as power output.
Now, as to the RMS values given for various outputs, it will change depending on the load impedance.  The RMS into the speaker is a function of the DC supplies to the output transistors.  This will change depending on how much current is being delivered to the load.  You won't get the same RMS into a 4 ohm load as you will into an 8 ohm load.  This is because the power supply will "sag" when asked to deliver more current.
  Look at the Bandit for example.  80W@8ohms.  This works out to 25.3VRMS as stated.  But they rate 100W@4ohms.  Working backwards from that, you see this is only 20VRMS.  This is because the power supply sags when asked to deliver more current.
  The RMS value is a function of the DC supplies.  Converting 20VRMS we get roughly 56V peak to peak.  That would be the voltage available to the output transistors from the DC supplies at the time of measurement.  So at full power into 4ohms there would be 56VDC between the + and - supplies, or +/-28VDC.
That is what you would measure at C72 and C87.
When the Bandit is running into the lighter 8ohm load, the supply does not sag as much.  The 25.3VRMS works out to approx. 70.8V peak to peak.  So there would be +/-35.4V at the supply rails at this time.
  I hope this shows how the RMS voltage does not stay constant as the load changes.
  So the Bandit will run higher supply voltages to the 2 transistors, that is why it can deliver more power than the other amp that has the same transistors.  To get the extra supply voltages you need a bigger power transformer.
  To run the transistors at higher current and power, you need better heatsinking to keep them from burning up.

Yes, you already have a send jack, you just need a return with a switching jack so you can run straight in to the power amp if required.
Also, the protection diodes like CR7&8 in the express/studio are a good idea.
In the attached diagram, I've cut the circuit and added the return jack and protection diodes from the express/studio.

** If anyone sees errors or omissions, please speak up!

  If the amp could be killed by a 10Khz signal, then it can be killed by a high note.
Putting a cap in to kill 10khz is going to hurt the treble response of the amp.