Onkyo TX-4500 Rebuild

[BadChile]

Following my rebuild of an Onkyo TX-2500 (mentioned in the thread here and featuring the lamp housings) and its displacement of the Harmon Kardon 430 for driving my Heresy IIIs, I began my search for the bigger brother - the TX-4500.  After eight months of casual searching I found one locally that met my requirements - not working, throwaway low price, and the front glass was intact.  Picked it up in early October with caution from the seller that the unit had spent decades in a basement and then a storage shed after it stopped working early in its life.

 

The life this beast had led to rust on parts of the interior, but more concerning was the amp failing to come out of protection mode.  Found a dead transistor and a couple resistors that went high; enter Mouser order #1.  Figuring I was now in the clear....nope.  Still not out of protection.  But more interestingly enough, no matter the adjustment for the idle current, the idle current would always read 0.000mV +/- 0.005mV despite adjusting the full range of the variable resistor R518/R618 on both channels.  Figured at this point something was really off, so started methodically removing components from the circuit to test.  Remove-test-reinsert-check...most of the metal oxide film resistors would crumble when removed.  Mouser order #2 time.

 

Blanket replacement of all the metal oxide film resistors should have done the trick I figured to solve the whole protection issue and the inability to set the Idle Current, but no matter what would happen, the Idle Current would not budge and the amp would come out of protection sometimes - and sometimes not.   More concerning was occasionally a legacy resistor would go up in flames, a transistor would croak, or the heatsink on the drivers would get quite hot.  Wonder would could cause that other than the idle current being too high?  But it is reading 0.000mV?

 

Sunday night I decide to sit down and trace the circuit.  And what do I find?  Resistors R524 and R624 were never installed from the factory.  Without those resistors, the Idle Current (ID) to Center (CT) test point results in an open circuit, leading to the low reading on the multimeter while the variable resistor was adjusted - sometimes low, sometimes high, but never to any effect.  Take out the R524 circled below and....open circuit!

 

 

Looking at the underside of the trace, the ID is completely isolated without R524/R624 in place (the TX-4500 does not use resistor RX25 in the circuit).

 

 

Grabbed a couple 56 Ohm, one Watt resistors from the parts bin for each channel (wonder where the 100 Ohm resistors went), wire them up in series, do some math to determine with the 111 Ohm resistance (wow, all four read low) what the ID should read (22mV) and.....it works!  It works awesome.  Still need a couple more lamps to replace the missing ones and do a better glamour shot.

 

So, for future generations working on an Onkyo TX-4500....check to see if R524 and R624 are installed.  If not....install them before starting your troubleshooting journey to save yourself the headache!!!  In the meantime, it is crudely wired up to test in the main system with the TX-2500 looking on, playing a few songs by the Cowboy Junkies as is customary when a new (old) piece of gear shows up.

 

Edited November 5 by BadChile
Typo Corrections

[babadono]

Cool...where do CT and ID go? Were these just test points for the factory?

[BadChile]

Yes 

18 minutes ago, babadono said:

Cool...where do CT and ID go? Were these just test points for the factory?

Yes - each channel has three test points (E = earth, CT = Center voltage, ID=Idle Current [aka, Bias]).  The Center Voltage is set measured between E and CT, the Idle Current measuring between CT and ID.  In my board photo above you can see the E CT and ID point right at the top of the red line.

 

Interestingly enough the TX-2500 has pins for the test points, whereas on my TX-4500 there are no pins, just through-hole solder pads with solder on them.  First thing I did was add pins to each of the three test points per channel to make measuring easier.

 

Fun footnote: after everything was working on Sunday I was setting the Center Voltage and Idle Current using hook probes and when removing my hand slipped and somehow shorted the protection circuit, blowing the small signal transistor in that circuit.  

 

 

[babadono]

Hmmm...If that 100 ohm r524 doesn't go anywhere I do not see how installing it fixed the amp. Perhaps you cleared some corrosion off of something that was not making connection? CT looks like the amp output before the zoebel network where the feedback signal comes from. And yes ID looks like a place to measure the bias.

[BadChile]

12 hours ago, babadono said:

Hmmm...If that 100 ohm r524 doesn't go anywhere I do not see how installing it fixed the amp. Perhaps you cleared some corrosion off of something that was not making connection? CT looks like the amp output before the zoebel network where the feedback signal comes from. And yes ID looks like a place to measure the bias.

It made it so I could measure the bias. Without the resistor the DMM would always read 0.000mV no matter the setting. But I would spin the variable resistor to set the bias without being aware it was changing. Lol.

[babadono]

Yes of course, sorry. But you could have measured it at E1 or the top of R526.

[Deang]

Impressive. But I'm easy to impress.