Not the SOA curve I have for a 2SA1943 I have sitting here in front of me. Vce would be around 27V including the sag on the rails, which hits the DC line at 5A. Where's the space for 60 degrees of phase shift? That would almost triple the peak current! Under-rate and be safe.
I'd like to show loadlines not only with resistive, but reactive loads, but my gnuplot fu isn't that good as well as it being a bear for me to post-process my scope's raw sample data.
Maybe this helps?
look at fig2.3
ESP shows a peak current of >8Apk while the Vce~25V.
Now compare those data points to the temperature de-rated SOAR and see how long the transistor can pass that current at that temperature at that voltage when you ask it to do that intermittently (single shot transient)
Your protection must allow this current transient to pass or you will interfere with the audio signal going to the speakers. Not a fault condition, a real operating condition.
Not only must the current pass but the protection must not start it's triggering or that will also interfere with the wanted audio signal.
I have said in the past, that an amplifier protection system must allow valid audio signals to pass to valid speaker loads (without interference).
ESP shows a peak current of >8Apk while the Vce~25V.
Now compare those data points to the temperature de-rated SOAR and see how long the transistor can pass that current at that temperature at that voltage when you ask it to do that intermittently (single shot transient)
Your protection must allow this current transient to pass or you will interfere with the audio signal going to the speakers. Not a fault condition, a real operating condition.
Not only must the current pass but the protection must not start it's triggering or that will also interfere with the wanted audio signal.
I have said in the past, that an amplifier protection system must allow valid audio signals to pass to valid speaker loads (without interference).
It doesn't change the direction of current. The limiter won't kick in, of course. Set the value of the emitter degeneration resistors for the short circuit limiting threshold. For me, that's 0R33 with 53V rails that sag to around 48 to not hit the DC SOA line. For me, that's 2.5A and thresholds at 2.15A. Here, heat for the OPTs are dominant.
Next, assuming a 4 Ohm designed output stage, select the padding resistor (the one with the diode) to limit at a point inside the DC area. For me, that's 4.75A, but went with 2.75A (I think it was) because the power transformer's heat was rather high. Yeah, would have been safe with resistive loads to limit at 4.75, but it not only gave no safety margin for reactive loads, but over-heated the power transformer as I was above the VA rating.
If the phase angle is rather high, take a sub horn for example, I got one here that does -63 degrees @ 112Hz, the stress on the output stage is pretty extreme. See this for fun.
I strongly disagree with your conclusion. Without giving away too much personal info... <umm>.. We get quite a lot of returns for this one amp that is made here. All with blown output transistors. Since I changed the SOA protection so that it now actually operates for loads under the rating of the amp, we don't get any returns anymore.
If the amp must limit (clip early) to protect itself under extreme loads outside it's safe area, so be it... If it bugs ya, add more output devices.
That the protection must not interfere with a valid audio signal going to a valid speaker load. Surely you cannot be advising that the protection should be allowed to interfere with valid signals!!
I cannot be responsible for a badly designed and badly specified amplifier.
I definately agree with designing the amplifier to meet it's specification.
Why should it not limit extreme signals that are outwith specification?
I certainly did not say that the protection should allow all signals to pass to all loads.
9A peaks under resistive conditions is not okay when you take into account a phase shift safety margin. bang goes the magic smoke.
Enquiry for pcb
Hi friend,
I appreciate your job.
So, do you have this design of pcb's in a Protel file ?
Do you can share it with us or show it as an image in a pdf file?
best regards
Ted Jako - tajako(at)yahoo.co.uk
And here is the ProJet Power Module
An externally hosted image should be here but it was not working when we last tested it.
Hi friend,
I appreciate your job.
So, do you have this design of pcb's in a Protel file ?
Do you can share it with us or show it as an image in a pdf file?
best regards
Ted Jako - tajako(at)yahoo.co.uk
You are taking different statements together and out of context.
I did not say 9A continuous into a resistive load.
I did not omit reactive loading and what that does to increase the stress on the output stages.
Need advice
I just find this on net which suits the parts on my hand. (2sc5200/2sa1943 2sc5171/2sa1930 BD139/140)
http://www.elektroda.pl/rtvforum/topic1618798.html
I wonder why there is a 68R between B-E of power transistors? If i attend to use LME49811 instead of 810, should I avoid 100R front of drivers?
I'm planning to use 36VDC rails, then, can i use BD139/140's as driver in same values of resistors?
Thanks.
I just find this on net which suits the parts on my hand. (2sc5200/2sa1943 2sc5171/2sa1930 BD139/140)
http://www.elektroda.pl/rtvforum/topic1618798.html
I wonder why there is a 68R between B-E of power transistors? If i attend to use LME49811 instead of 810, should I avoid 100R front of drivers?
I'm planning to use 36VDC rails, then, can i use BD139/140's as driver in same values of resistors?
Thanks.
Last edited:
the driver base stoppers are a good safety/stability measure.
Worth adding base stoppers to the outputs as well.
You can delete these after testing, or just change the values, depending on results.
The 68r runs the drivers at near constant current at very low to zero output currents.
Expect ~9mA + output base current as the driver current.
Worth adding base stoppers to the outputs as well.
You can delete these after testing, or just change the values, depending on results.
The 68r runs the drivers at near constant current at very low to zero output currents.
Expect ~9mA + output base current as the driver current.
With same value of 68R or 0.22R as emitter resistor?
Are these base stopper resistors are applicable to LME49811 which has only 7ma output current? Or just decreasing the value in case of 49811?
I haven't any oscillator for detecting possible oscillation beyond of audio band. I can do only basic measurements, so, can you suggest me to respect the scheme as what it is or can we claim that avoiding these resistors don't touch final sonic result too much? I rather prefer avoiding resistors for simpler circuit.
I apologize for asking these silly questions, i'm just trying to understand what is it beyond of these kind of variations.
Oh, i got what Panson mean now..that was what does it matter between 68R and 68.22R ??
Oh, i got what Panson mean now..that was what does it matter between 68R and 68.22R ??
Sure, there is no different for 68R and 68.22R. What I meant is the configuration. The upper/lower driver is still cut-off in negative/positive cycle as that with the center tap connected to output.
if the center tap is not connected to the output, driver will not be off. I prefer this one.
Regarding oscillation, it does not a matter of the driver output current. You will see oscillation in small signal circuit. It is about gain and phase.
If we have base stopper at the output, we may be able to omit the stopper of driver. However, it is also layout/wiring dependent. If test confirms the base stopper can be removed, then take them out in the final version.
The base stopper ensure negative resistance never happens.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.