Threshold 400 help

[Nelson Pass]

Temperature compensation for the bias.

[P901]

Good afternoon Mr. Pass !

Thank you for that feedback.

Not being an electronic engineer or designer it was
curious that the device was across the E/B pins on the one
transistor on the circuit board I have but in the three schematics it
was installed in another position across two transistors.

Was it found to be more effective or efficient in this position as
opposed to what is shown in the schematics?

Thanks again for the feedback, andrew

[P901]

So, got a challenge here with the repair.

The right channel still works fine, the left channel failed.
Had three failed output transistors, two A92`s and one
2% tolerance 1K resistor had "drifted" to 4K!

Going through the components one by one we have been able to
get to the point where the unit no longer blows the fuse but it slowly
"cooks" the two A92`s, the LVA and the one next to it.
Prior to energizing the unit it was carefully checked, one side against the other for all values of the components, "in situ", for resistance, capacitance and diode function. The numbers all matched, (trying to avoid removing all of the left channel components).

Anybody have this happen? Any suggestions would be helpful.
Possibly the new MPSA92 transistors or the OnSemi MJ 15025/4`s are not a close enough match? Could the mismatch cause a need for readjustment of the bias resistance? How to do that without burning transistors if that is the case.
My next step is to go over all the components again one by one before trying one more time. If it cooks these transistors again then I guess I will be forced to remove all the components individually for testing.
The unit had been recapped about 6 months ago and had been working fine.
The right side still works perfectly.

Thank you in advance, andrew

[m2003br]

Hi Andrew,

You have too much current running through LVA. I guess this is caused by defective constant current source who supply the LVA (the MPSA42 on the lower side of LVA, togheter with 2x 1n4148, 150R and 470R) When all is OK with front end, you have aprox. 4.6mA in the LVA (0,69V over the 150 ohms resistor).
When you change output transistors, always substitute all 3 pieces from the same parallel group, at least, and with components from the same batch. There is no need for tight match between them, thanks to a high value for the emmiter resistors (0,68R). But, Never mix togheter new MJ's with old 2N's in the same parallel group.

All the best,
Marcos

[P901]

Good evening Marco !

Thank you so much for that information !

I had suspected that something was out of balance but never having
worked on such a unit with my level of experience I was at a loss.

I will follow your guidance, this must be the problem since everything else
has been thouroughly checked and re-checked.

I have read that the current across the .68ohm resistors should be set at
90ma across the resistor. Would there be any benefit to lower this value so slightly as to cool the amp just a little bit to assist in the life of the components ?

Thank you again ! andrew

[JamesBrennan4]

Originally Posted by P901

Ok, making progress,

Can anyone please give us some idea why there is a thermistor across
the Base/Emitter pins of the "first" power output transistor.

All the schematics we have show the thermistor going from the base pin
of the "first" transistor to the emitter of the "second" transistor with the emitter of the first transistor going to the base of the second transistor
so its not as if the connection shown on the schematic was physically achieved by an alternate execution.


Thank you, andrew

The schematic made by m2003br in post #14 shows the thermistor connected like you are describing.

All of the schematics show two thermistors. Do you also have two, both connected across the base/emitter pins?

[JamesBrennan4]

May I suggest that the 5K bias adjust resistor has been turned or become shorted. Either could result in excess output idling current, such that the output transistor will over heat. When the first output transistor goes south, then the other two are additionally stressed, finally resulting in the other two also going south.

Verify that the bias adjust resistor is set the same for both channels to start with. The position should be near center.

[JamesBrennan4]

On second thought, the 5K adjustable resistor may have become open, not shorted. Check for "cold" solder joint.

You can read a complete description of the bias circuit by reading Nelson Pass US Patent 3995228 - Active bias circuit for operating push-pull amplifiers in class A mode.

[P901]

Good morning James,

Thank you for those great suggestions !
I will verify the action and connection of the 5K bias resistor.
During our ownership of this amplifier we have not moved the position
of the 5K bias but I have made very small pinpoint ink marks of the
position.
My thought was that if someone with experience could tell me of a a "conservative" starting point in the controls resistance
then I could slowly ramp up near the recommended current levels until the
amp came up to temperature and then follow the final setting procedure.
As you suggest, not knowing if this control had been previously adjusted
I thought that it would be good technical practice to reset/reverify
the setting.

Regarding the thermistors, there is only the one as shown on Marco`
schematic. There is another device wired into the circuit board and
"glued" to the output transistor base plate. It must be the thermal
overload which appears to open the line between the A92 LVA and
the A42. No other thermistors.
I have included a photo of the output section, you can see the two blue thermistors and the "overload" device glued to the plate.
(Also note that a couple of the resistors have been "lifted" while undergoing testing)

In the photo of the circuit board you will see the one transistor that I have marked with an orange dot. This was the other transistor that had partially failed along with the LVA A92. On the partially failed transistor once removed from the board I had from B to E .609 and from B to C .249 setting was on diode function with the Fluke 87V.
The failed devices have been replaced and you can also see the beige colored 68ohm resistor which is the one that had burned and split.
Awaiting a shipment of output transistors from Mouser and then we will begin again.

Thanks much to all for the help, andrew

[P901]

A bit of feedback,

Checked the 5K adjustable bias on both sides, Ok, good connection.
They were both set at 1.396K and had good function when
dialed up and down.
Would anyone know what an acceptable reading would be on the
"back to back" diode that it mounted in the center of the output plate
on one end? When removed from the plate each diode in the pack reads
.409 and a total of .811 through each other.
It says VARO R711XE on the case.
Good, bad ??

Thanks in advance, andrew