Bias current for 2sa1216? 100mA Ok?

[bastek]

I' m adjusting bias for a Kenwood KM-X1 6ch power amp. It uses a pair of 2sa1216-2sc2922 per channel.
The amp. was running very cold with only 10mv across test points (emitter resistors). Going up to 60mv it became too hot to touch after 20min. so now i have it at 35mv for a nice warm heat sink. This gives me .035V/.33Ω = 106mA.
Is that a safe value for these transistors?

[maxpou]

Hi Bastek,
if you can put your hand on heatsink 5 seconds without burn it, i think that's OK. Maxpou

[ostripper]

Every OP stage I have either repaired or built had a bias
of 50-100ma. In simulation almost all class B stages
exhibit the least distortion nearer to the 100ma range.
(I set most of mine to 70-75ma , crank it for a while
see if it stays within 10-15 % of that(80-90), turn it off ,see if it
is at least 60mA at a cold start..etc)


You seem to have it right.. .
OS

[fjr]

I would suggest not to increase the idle bias current!
30mA (10mV on 0.33r) is just fine for a BJT EF OP stage.
Increasing bias won't give better sound automatically.
Other reason is the thermal stability (and possibly runaway!) problem.

[bastek]

Would it make a difference in the amp handling low impedance speakers better? My speakers are like 3 ohms and trigger the protection at higher levels.

[ostripper]

Other reason is the thermal stability (and possibly runaway!) problem.

You might want to check to see if your Vbe (small tranny or
diode array close to outputs) is properly "greased" (t-compound)
Some manufacturers just "prop up" a to-92 on the main HS
and "blob" some thermal compound on it (white greasy blob
with 3 leads) If the blob dries out the circuit will not track
temp. and ultimately fail.

The best approach is to carefully "lift" all power devices with a
razor blade or x-acto knife, clean mica washers ,devices, and
heatsink with rubbing alchohol , apply new compound
to all surfaces... reassemble.
OS

[Alexamp]

The optimal bias current for BJT output stage from minimal crossover distortion point of view exists. It lies in range 10-30 mV for different BJTs. The sweet spot easy may be found!

[nigelwright7557]

Originally posted by bastek
I' m adjusting bias for a Kenwood KM-X1 6ch power amp. It uses a pair of 2sa1216-2sc2922 per channel.
The amp. was running very cold with only 10mv across test points (emitter resistors). Going up to 60mv it became too hot to touch after 20min. so now i have it at 35mv for a nice warm heat sink. This gives me .035V/.33© = 106mA.
Is that a safe value for these transistors?

I turn the bias up slowly until the crossover distortion goes.
This gives just enough without wasting power in heat.

[Alexamp]

It lies in range 10-30 mA for different BJTs. This is correct.

[AndrewT]

Hi,
the optimum bias for a ClassAB emitter follower (EF) output stage is 26mV across the total effective emitter resistance.
A compound follower pair (CFP) requires much less bias.

But, some of this emitter resistance is inside the transistor.

The Vre must be set to <=26mV to be within the optimum bias range.

Reduce your bias immediately. If your output stage is already triggering the protection circuits then running the amp hotter than cold reduces the safety margins and the protection is likely to come in too late to reliably save your amp from abuse.
Get it running cooler.

Get the speaker impedance in spec for the output stage.
If it triggers protection with 3ohm speakers then replace the speakers with something more appropriate or accept the consequences.