Discrete is more flexible for design work.
Some say it sounds better too.
If that is not an issue then just build one of the many chip amps on this forum.
Some say it sounds better too.
If that is not an issue then just build one of the many chip amps on this forum.
I just look at the schematic of Nelson Pass AB100. This is supposed to be a famous high end amps. Look at the schematic, it is single LPT IPS, asymmetrical VAS with CCS on one end. Then using darlington that hfe is exponential to current. These are everything people sneezed on, books written against them.
My all time favorite amp is YBA. The schematic is not much better!!!! What's wrong with this picture? Here we are, heckling over 3EF, complementary LTP IPS with current mirror, symmetrical VAS, distortion cancellation............. Something is very very wrong with this picture!!!!
Are we trying to put everything under the microscope and missing the moon?
I came from the tube side to SS because there are so many "fancy" add on over there. Then I looked at some top end amps like Cary, they are as simple as anything. I thought that was ridiculous, that's the reason I come over here thinking SS is more standardized. Seems like same thing over here too!!! Are we really looking things too close and missing the moon?
Or is there any truth about there are more to sound quality than just distortion spec?
Last edited:
The graph is logarithmic in scale. @Ic=200mA, Hfe is ~2600, Ib=77uA. @Ic=1A, Hfe is 8500, Ib=118. @Ic=4A, Hfe=15,000 Ib=267uA. @Ic=7A Hfe=12,000 Ib=580uA. @Ic=10A Hfe=10,000 Ib=1mA. @Ic=12A, Hfe=8500, Ib=1.4mA.
Although the current gain is very high, it does not seem like a very linear d(t)Ic/Ib to me.Generally, bipolars for linear audio tend to have a much more linear current gain up until the point of electron crowding, indicated by the beta droop as Ic increases.
I gave this a lot of thoughts and I compute a table and look at the base current Ib vs collector current IC and look at the amount of change of Ib from collector current at quiescent to 5A.
I assume quiescent current is 200mA and speaker load is 4ohm for this calculation. Attached is the table of base current Ib vs collector current IC comparison between darlington 2SD2401 verses 2SC5200 driven by a pre-driver with constant hfe. The combined hfe is assumed 10500 and remain constant throughout the current range.
I also showed the delta Ib. As you can see, the change of base current in 2SC5200 combination is double than the darlington 2SD2401.
From my understanding, it's NOT the constant hfe you want, it's the constant Ib that you want!!! The less change of Ib, the less change of drive requirement needed from the VAS, the less distortion. In this case, the darlington kicks the butt of the two individual transistor in 2EF. The best case would be the hfe change linearly with IC so Ib remains constant!!!! Maybe Nelson Pass made use of this to design with darlington.
Please comment on this.
Attachments
Last edited:
Member
Joined 2009
Paid Member
There are many different types of output power devices, if the amplifier is designed appropriately there is no reason for it to underperform relative to other choices.
There is a lot of opinion and misinformation around too and in the end you may want to sift out the facts and understand which of them are relevant. And the final choice will be a compromise of sorts, you'll need to tradeoff things like cost, thermal properties, capacitance, speed, input impedance (hfe), parasitic stability, SOAR, circuit complexity, size, output impedance, etc.
I've built amplifiers with BJTS, FETS, Darlingtons, and Triodes - they all work fine 🙂
There is a lot of opinion and misinformation around too and in the end you may want to sift out the facts and understand which of them are relevant. And the final choice will be a compromise of sorts, you'll need to tradeoff things like cost, thermal properties, capacitance, speed, input impedance (hfe), parasitic stability, SOAR, circuit complexity, size, output impedance, etc.
I've built amplifiers with BJTS, FETS, Darlingtons, and Triodes - they all work fine 🙂
Last edited:
And you could enable/disable it with two very simple and low-current shorting headers. Trim the bias for option 1, listen+measure. Remove the shorts, trim the bias for option 2, listen+measure a second time. Choose a favorite.
No The OP question was generic in nature and now you want it to revolve around something that is incredibly specific to what is in yer mind.
Back to topic.
Im sure you can find a few devices (complementary?) that can work in an audio amps OP stage, but typically designers like to have options, like multiple sources for each and every component for product cost as well as long life. don't paint yeself in a corner in early product design unless you absolutely have to and even then it usually returns to bite you.
i have used the Motorola MJ11033/11032 in a leach super amp before many, many, many years ago, simply because i have them....been there done that....
i have repaired commercial subwoofer amps that used the tip122's....folks use them...
i have repaired commercial subwoofer amps that used the tip122's....folks use them...
Darlington in switch mode very Poor, saturation voltage twice as much. Don't be wasting resources in unneeded dissipation.
I think the difference in gain, input impedance, and the effect on the driving circuit would be relevant to why designers use one over the other. However I will bow out of the discussion if just opinion/speculation that's acceptable.
I don't get it. I thought one of the most important topics all those great solid state linear design minds put their focus on has been the non-linear Ib of BJT and taming it, as that is the very place where majority of non-linear distortion originates. Was I wrong?
Member
Joined 2009
Paid Member
All output devices are non-linear, Triodes less than most. I prefer to think about the base current of a BJT being simply a feature of finite input impedance. MOSFETs have non-linear gate impedance (mostly capacitance). All active devices are non-linear. What they have in common is that a control voltage (base-emitter or gate-source for example) determines the current flow through the device. Mostly it's an exponential relationship but degeneration helps linearize it. You can design the driver stage to suit BJT, MOSFETs, and combinations (Darlingtons, Sziklai etc.) of them so that you have a very good amplifier. For a BJT you need to be able to source that base current, for FETs you have to charge/discharge the gate capacitance, for triodes you have to drive the grid which can be a considerable challenge if you want to operate in Class A2. It's all doable with the right design.
Don't sweat it - the group think has already decided that Darlington's are no good - best just to agree, in case they turn on you 🙂
Don't sweat it - the group think has already decided that Darlington's are no good - best just to agree, in case they turn on you 🙂
Last edited:
Use a Darlington in a Class A setup like Nelson's A40 (discrete or integrated, who cares?). This obviates discussion concerning non-linear transfer characteristics, etc. I have a topology (sounds very nice, it kicked a** at Burning Amp a couple of years ago) that guarantees absolute Class A operation up to saturation no matter what the input. My first implementation used mosfets, but I see no reason why I couldn't substitute Darlingtons. Yet another project thrown on top of the teetering stack of those already in the queue....
Member
Joined 2009
Paid Member
Well if you're not having one side switch off, how are you getting crossover distortion? That's the largest non-linearity I'd worry about. At any rate, I'll probably try it out of perversity (a larger driving factor with me) sooner or later. I have both sexes of devices from the 2N6283 series and plenty of tunnel heat sinks drilled for TO-3. Thirty-forty watts of Class A per channel would be a no-brainer with the setup I contemplate.
Well, think in the point of view from the VAS, which is a current drive device. You think the VAS rather see a load current that vary less or more at different output voltage?
Crossover is one of the few sources of distortion. We are not talking about crossover distortion here.
One problem I do see is if you have 4 transistor in parallel, the 2Vbe are not match. Where as for discrete 2EF, you only need one driver to drive 4 big transistor in parallel. So the mismatch is only 1Vbe. That can be very important. That will kill the advantage of the darlington output transistor.
But how about using darlington driver transistor? Current variation of the driver stage is a lot less. Vbe matching is no different from 3EF.
So using simple BJT for power transistor, use darlington for driver transistor.
But how about using darlington driver transistor? Current variation of the driver stage is a lot less. Vbe matching is no different from 3EF.
So using simple BJT for power transistor, use darlington for driver transistor.
- Status
- Not open for further replies.