dfdye said:
The MJ11032 is rated for only 120V. Keep in mind that the OFF transistor can see the full dual rail voltage under worst case. So if you want supply rails higher than +/- 60V you need higher voltage transistors.
You can also try MJ11016 and MJ11022. The hairy part is that you need at least 1000 hfe for the device to output full power with a 5mA drive. And most Darlingtons lose hfe rapidly as collector current goes up. You're working right on the edge. There is no extra current gain to spare.
I genuinely can't understand why someone would make a power amp driver, and give it output current capability of only 5mA !?
Andy
It is also possible to make your own darlington transistor.
One driver + one power transistor in darlington config.
This way you can make a custom output device with very good qualities and high gain.
It is even possible to use Sziklai configuration.
Figure 5.20:
- a) The Darlington connection of two transistors to obtain higher current gain and
- b) the Sziklai connection of an NPN and PNP transistor to obtain the equivalent of a high-current-gain NPN.
Darlington and Sziklai Connections
One driver + one power transistor in darlington config.
This way you can make a custom output device with very good qualities and high gain.
It is even possible to use Sziklai configuration.
An externally hosted image should be here but it was not working when we last tested it.
Figure 5.20:
- a) The Darlington connection of two transistors to obtain higher current gain and
- b) the Sziklai connection of an NPN and PNP transistor to obtain the equivalent of a high-current-gain NPN.
Darlington and Sziklai Connections
Thanks for the info, lineup.
Making your own Darlington frees you from having to purchase dedicated parts, but it doesn't solve the hfe problem.
Even with discrete, it's not easy to get an hfe above 1000 when you use high currents, and that's what this circuit needs.
Keep in mind that to get the powers that this driver suggests (with it's +/-75V rails -->250-300W ), you'll likely need to parallel some output devices. That places even greater demands on hfe. For any power under 200W, it's not worth messing around with the 4702, when the 3886 (bridge/parallel combinations) is so reliable and flexible.
Now a triple Darlington, (a la Leach) will solve the hfe issue and is simple to implement, but I'm not sure if that won't introduce other issues, like linearity or instability. Any thoughts on the triple?
Andy
Making your own Darlington frees you from having to purchase dedicated parts, but it doesn't solve the hfe problem.
Even with discrete, it's not easy to get an hfe above 1000 when you use high currents, and that's what this circuit needs.
Keep in mind that to get the powers that this driver suggests (with it's +/-75V rails -->250-300W ), you'll likely need to parallel some output devices. That places even greater demands on hfe. For any power under 200W, it's not worth messing around with the 4702, when the 3886 (bridge/parallel combinations) is so reliable and flexible.
Now a triple Darlington, (a la Leach) will solve the hfe issue and is simple to implement, but I'm not sure if that won't introduce other issues, like linearity or instability. Any thoughts on the triple?
Andy
Well, I got the LM4702's in today, so I guess I will have to really start bearing down getting some supplies together to try. 😀
Thanks for pointing that out, though! I would never have thought it would have been an issue.
(You are talking about the Darlingtons when taling about the OFF transistors, right? Bear with me! If OFF stands for somethinng, I don't know the abbreviation--still new to DIY audio, but I am guessing it starts with "output". . . 🙂 )
Again, I may be missing something. . . .
Vectorplane, thanks for the comments, but I thought that in the output stage of the typical application diagram in the LM4702 data sheet each of the darlingtons would see at most 1/2 of the rail to rail voltage, making the 120V more than sufficient. I may just not understand how that section of the circuit is behaving. . . 😕
Thanks for pointing that out, though! I would never have thought it would have been an issue.
(You are talking about the Darlingtons when taling about the OFF transistors, right? Bear with me! If OFF stands for somethinng, I don't know the abbreviation--still new to DIY audio, but I am guessing it starts with "output". . . 🙂 )
Well, I think that is why they recommend the Darlingtons. . .
Again, I may be missing something. . . .I was asking myself why no one had already suggested that idea. I like it and it's a heck of a lot easier than settling for just any old transistor. Now that i've said that, does anyone have suggestions for the pair?
i will not hesitate to use a leach type output stage for this chip amp, you can even adapt this chip to a leach super amp configuration. afaik, leach VAS also operates in the 5mA level.
dfdye said:
Sorry...I take shortcuts to avoid typing. The OFF transistor is the output transistor that's not conducting in a particular half-cycle. (You have the ON transistor and the OFF transistor, meaning one has low resistance and the other has high resistance) the ON device may have as little as 4-5 Volts accross it when conducting at the peak of a cycle. At this point the full positive plus negative voltage rails (minus 5 volts) will show up accross the other transistor which is just barely biased into conduction.
So whenever you see a class B output stage, add up both rails to know what each transistor needs to withstand.
Which idea is that? the triple Darlington?alexcd said:
Tony said:
True enough, but Leach designed the VAS knowing he'll have triple Emitter followers. National did not.
It's just a personal opinion. I think they could have given the 4702 some 20-40 mA of output drive. Some opamps have 80-100 mA capability on the output. It just seems odd. You can't really call this a "driver", since the user has to supply the driver. It's just a front end and a VAS.
On the other hand, you could call the 4702 an "experimenter's playground", since it gives you the freedom to experiment extensively with power stages without worrying about the rest.
Andy
vectorplane said:
If up to +/- 30V is ok, OPA551 or OPA552 does it.
If higher PSU voltage is needed (up to +/-45V) and ~15ma max. output current is ok, OPA445.
well the chip is out now, and it is up to us to find ways to make use of thse chips.
i will use MJL1302/MJL3821 as ouputs, MJE150031/32,MJE34-/350 transistor compliments...on a triple darlington output stage. rails at +/-55volts with front-end regulated supplies of +/- 65volts.
i will use MJL1302/MJL3821 as ouputs, MJE150031/32,MJE34-/350 transistor compliments...on a triple darlington output stage. rails at +/-55volts with front-end regulated supplies of +/- 65volts.
No problem at all! I just didn't follow at the time. Thanks for taking the time to explain.vectorplane said:
Following up on the theme, won't the ouptut resistors and the "ground" of the connection to the speaker prevent the full rail to rail voltage from being seen by the OFF transistor 🙂) I like the description a lot!)?
I know this is a basic question, so thanks for taking the time clear up my misunderstanding!
Tony said:
Awesome..
You'll probably get an hfe of at least 30-40 from each of those devices, giving an overall beta of 30,000+. That's enough drive to parallel more output devices.😀
What values of base pull-down resistors would you use at each stage?
Andy
dfdye said:
Assume for a moment the top transistor is conducting (full ON) and bottom transistor is quiescent (OFF). Then the entire positive rail (minus a few volts) will be present at the output node (speaker hot terminal). If the output node is at V+, then the lower transistor sees V+ to V- accross it.
Surely we hope the 8ohm speaker load will not lower this voltage, or else we have a cr@p power supply. When we use a 60 V rail, it's because we want to get something very close to 60V (peak) right across the speaker. We need that to happen for our power output to be what we calculated.
A few volts will unavoidably be dropped across the upper transistor (3-8V) and the emitter degeneration resistor (1-2V), (since this is not a Rail-to-Rail amp), but not much.
Doing what Tony suggests and making the front end supply higher than the power stage, allows the amp to come closer to being R-to-R, reducing to some degree power waste in output stage.
You cannot possibly believe that the emitter resistors will drop enough voltage to hide the V+ Rail ftom the bottom transistor!
You may want to read this book:
"Audio power ampifier design handbook, Douglas Self" ,
available as e-Book here:
Link removed by moderator.
(scroll down and press the "free" button, then wait the number of seconds mentioned at bottom of page)
Andy
Pirated ebooks....
vectorplane wrote :...
MODERATORS PLEASE NOTE !!!
vectorplane wrote :...
Illegal !!! to post links that leads to pirated stuff !!
MODERATORS PLEASE NOTE !!!
Looks like someone was trying to help...and was bringing a lot of valuable insight to this discussion.
What are YOU bringing? Who are you helping?
You have not contributed anything to this converstion other than criticism.
The EDA Forum has some additional learning material for those who want to educate themselves in electronics.
(EDA stands for Electronic Design Automation)
http://www.edaboard.com/forums.html
The fourth forum from the top, is a place you should definitely not look at,
because it may upset someone....
-----------------------------------------
If you mind other people's business,
you should be ready to also inherit their problems.
---Confucius---
What are YOU bringing? Who are you helping?
You have not contributed anything to this converstion other than criticism.
The EDA Forum has some additional learning material for those who want to educate themselves in electronics.
(EDA stands for Electronic Design Automation)
http://www.edaboard.com/forums.html
The fourth forum from the top, is a place you should definitely not look at,
because it may upset someone....
-----------------------------------------
If you mind other people's business,
you should be ready to also inherit their problems.
---Confucius---
Guys...
Relax, Cool down....
I don't think it's a pirated book. There are a lot of eBooks out there for free on the web.
There is nothing on that link that indicates this is pirated.
We should get back to discussing the LM4702 instead of making negative statements.
I'd be interested in knowing the resistor values that Tony used also.
Don't worry. Be Happy...
Relax, Cool down....
I don't think it's a pirated book. There are a lot of eBooks out there for free on the web.
There is nothing on that link that indicates this is pirated.
We should get back to discussing the LM4702 instead of making negative statements.
I'd be interested in knowing the resistor values that Tony used also.
Don't worry. Be Happy...
vectorplane said:
for pull down resistors, i am using 470ohm between bases of MJE340/350, and 220 ohms between bases of MJE150031/150030 drivers, pretty typical.
😀
Tony said:
Tony,
Any reason why the 470 Ohm resistor can't be higher?
Typically when a Darlington chain has a 220R on the output base, you see a 5-10x higher R on the previous leg, something over 1K.
Just concerned about power dissipation.
-------------------------------------------------
Vivez, riez, construisez des amplis...
Right on. Thanks again for helping out. Everything now makese sense. I hate it when I "figure out" how something works and it is a little off. It's always hard to get your head out of the dark place when that happens. 😀vectorplane said:
I will have to check out the book and avoid a few more dumb questions.
David
PS Just glancing through the book, it seems like it will be a great resource. Thanks for pointing it out!
- Status
- Not open for further replies.