PL resistors
Base stoppers are series resistors in the base of a transistor.
Use 1.5 - 3.3 ohm 1w base stoppers in the bases of all the output devices.
To stop all oscillations add base stoppers in the first drivers transistors of the output triplets....this gets rid of 95% of all PL oscillation issues.
Add a Miller capacitor (47pF/500v) on the VAS transistor. After these mods your PL will not oscillate PROVIDED you throw out those ugly PL909s.
The 10 ohm base-emitter resistor on each output transistor bank will dissipate about 0.036w assuming you bias the output stage hard which of course is not a good idea in these amplifiers.
So using a 2w resistor here is a waste of time, 0.5w is fine.
The emitter resistors used are WW and their inductance is so low that it makes no difference to the sound. The PL drive circuit is so bad anyway I do not see how these resistors in the output stage can make any difference to the sound.
Steve
Zed Audio
Base stoppers are series resistors in the base of a transistor.
Use 1.5 - 3.3 ohm 1w base stoppers in the bases of all the output devices.
To stop all oscillations add base stoppers in the first drivers transistors of the output triplets....this gets rid of 95% of all PL oscillation issues.
Add a Miller capacitor (47pF/500v) on the VAS transistor. After these mods your PL will not oscillate PROVIDED you throw out those ugly PL909s.
The 10 ohm base-emitter resistor on each output transistor bank will dissipate about 0.036w assuming you bias the output stage hard which of course is not a good idea in these amplifiers.
So using a 2w resistor here is a waste of time, 0.5w is fine.
The emitter resistors used are WW and their inductance is so low that it makes no difference to the sound. The PL drive circuit is so bad anyway I do not see how these resistors in the output stage can make any difference to the sound.
Steve
Zed Audio
The schematic shows the base emitter resistors on the pre-driver at 150 ohms.
The single 10 ohm bias resistor per bank is emitter to output.
I swapped all the 2 W emitter resistors with Mills non-inductive resistors and left them in for a couple days. It changed the sound dramatically. I did not like it... except that it was more detailed.
I swapped back!
The single 10 ohm bias resistor per bank is emitter to output.
I swapped all the 2 W emitter resistors with Mills non-inductive resistors and left them in for a couple days. It changed the sound dramatically. I did not like it... except that it was more detailed.
I swapped back!
I have swapped the 150 Ohm Rbe on the pre-driver. It appears stable. Sonically, a bit more sterile, but very enjoyable.
If it is safe to swap the 10 ohm bias resistor to a 0.75W, I will do so. I *think* the current flowing across the bias resistor is max when no signal is applied... but I'm still unclear on that. Anybody?
A quick question:
Why do the bias resistors on the negative outputs consistently read a higher voltage (about +0.1V more) than the bias resistors on the positive outputs? Should they not be the same? Can I make them the same?
Thanks.
If it is safe to swap the 10 ohm bias resistor to a 0.75W, I will do so. I *think* the current flowing across the bias resistor is max when no signal is applied... but I'm still unclear on that. Anybody?
A quick question:
Why do the bias resistors on the negative outputs consistently read a higher voltage (about +0.1V more) than the bias resistors on the positive outputs? Should they not be the same? Can I make them the same?
Thanks.
The current through the 10 ohm resistor reaches its max at full signal. It sees Vbe (about a volt) plus the voltage across the emittter resistor (another volt and a half at 4 ohms load), in a half sine waveform. And 0.34V (per spec bias) at idle. A 1/2 watt resistor will handle this. There is no reason for more in normal operation. The reason I use higher rated flameproof resistors is because of the $*** that happens when output transistors die, which isn't always predictable.
The resistor where bias is measured is across the emitter resistor of the pre-driver. That's the 10 ohm resistor.
As you've written, it will see less than 2 Volts -- the bias voltage (0.35) plus the full signal drive voltage -- 1.5 volts at 4 ohms. So, it is very safe.
I will try changing them and hopefully, no oscillation will occur.
Thank you!
Do you have any idea whay the bias resistors on the negative phase are always higher than the positive phase? It looks like about 0.1 Volts
I'm guessing that an incorrect value for the bias ( ie. driver emitter ) resistors could do it. Maybe with 1% parts, the difference will vanish....
Will post after replacing them.
As you've written, it will see less than 2 Volts -- the bias voltage (0.35) plus the full signal drive voltage -- 1.5 volts at 4 ohms. So, it is very safe.
I will try changing them and hopefully, no oscillation will occur.
Thank you!
Do you have any idea whay the bias resistors on the negative phase are always higher than the positive phase? It looks like about 0.1 Volts
I'm guessing that an incorrect value for the bias ( ie. driver emitter ) resistors could do it. Maybe with 1% parts, the difference will vanish....
Will post after replacing them.
So, I put a scope on the emitter resistor of the pre-driver, before the 10 ohm resistor.
It is easy to see 25+ volts above ground at this point in the circuit when playing music and pushing 1/2 power or a bit more.
How are 1/2 watt resistors ok here, when they generally see over 5 volts and can easily see over 20 volts or more, well before clipping?
It is easy to see 25+ volts above ground at this point in the circuit when playing music and pushing 1/2 power or a bit more.
How are 1/2 watt resistors ok here, when they generally see over 5 volts and can easily see over 20 volts or more, well before clipping?
Oh I get it.... the other side of the resistor is also high above ground, so the current across that resistor remains low. I need to get another scope probe, then chart X-Y.
The problem with using the scope probe's ground reference on the output side of the emitter resistor is that it make the amp oscillate on transients. However, I did not see the large voltage in that mode.... I'm too scared to try that again. The scope ground must mess that test up.
The problem with using the scope probe's ground reference on the output side of the emitter resistor is that it make the amp oscillate on transients. However, I did not see the large voltage in that mode.... I'm too scared to try that again. The scope ground must mess that test up.
I put a scope across the 10 Ohm emitter resistor on the pre-driver.
The emitter side was positive, the output side was inverted, the trace showed channel A+B -- in other words, the voltage drop acros the resistor.
At High output -- not yet clipping, there was 4 volts across this resistor, or
power = 4*4/10 = 1.6 watts, into a 4 ohm load.
No way is a 1/2 watt resistor or 0.75 watt appropriate in this location.
I would minimally go 2 watts, with 3 for some headroom.
The emitter side was positive, the output side was inverted, the trace showed channel A+B -- in other words, the voltage drop acros the resistor.
At High output -- not yet clipping, there was 4 volts across this resistor, or
power = 4*4/10 = 1.6 watts, into a 4 ohm load.
No way is a 1/2 watt resistor or 0.75 watt appropriate in this location.
I would minimally go 2 watts, with 3 for some headroom.
One resistor.
In my terms, the drivers have 0R22 emitter resistors, so Q11 is the pre-driver on the positive phase. Q12 for the the negative. Perhaps that is not the right way of describing Q11/Q12.
I'm measuring across the 10 ohm emitter resistor of Q11, on the positive side. That's R38. That is the resistor across which the bias voltage is measured 0.35 +/- 0.1 V.
That should minimally be 2 watts.
In my terms, the drivers have 0R22 emitter resistors, so Q11 is the pre-driver on the positive phase. Q12 for the the negative. Perhaps that is not the right way of describing Q11/Q12.
I'm measuring across the 10 ohm emitter resistor of Q11, on the positive side. That's R38. That is the resistor across which the bias voltage is measured 0.35 +/- 0.1 V.
That should minimally be 2 watts.
4 volts peak, and only for half cycles. Given even 4dB dynamic range of very compressed music, you won't even heat it up. The reason you may want a 3 watt resistor is if you accidentally get 200+ into it when something shorts out, not because of what playing it loud will do.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.