Hi Bob,
I have a different understanding of his explanation, really mosfets have high FT, but with increasing frequency the capacitance Gate tends to increase the use of stopper resistors mitigate the problem because increased the time of the capacitor(RC), I believe that the use of an inductor in Gate instead of a resistor can be better (but can cause a resonant circuit).
Increasing the capacitance Gate of MOSFETs in function of frequency is the main argument for creation of IGBT,s for high power switching.
Hi Chris,
Yes, the whole thermal thing is a tough nut to crack, and it was one of the toughest chapters to write in the book. Over the years, I have learned sometimes the hard way about the consequences of creeping over-bias. MOSFETs are much better this way, although not completely immune (at least for verticals).
Cheers,
Bob
Hi Andrew,
I'm not sure if this answers your question but here goes: First, I usually bypass a Vbe multiplier spreader from top to bottom with a 10 uF aluminum electrolytic; I like the bit of ESR to damp the circuit at HF. You are correct that the capacitor will not do anything for reducing the bias spreader impedance at low frequencies.
I am also usually mindful of the possibility that a Vbe multiplier transistor placed remotely on a heat sink may oscillate due to the wiring inductances going back the the VAS circuit. For this reason, I often connect a 0.1 uF bypass from collector to base right at the Vbe multiplier transistor. This is often physically inconvenient, but I tend to be paranoid about parasitic oscillations. The capacitor makes the transistor think it is doide-connected at high frequencies. On top of that, always twist the three leads from the remote transistor together.
Cheers,
Bob
Thanks. I guess this leaves us poor slobs without a distortion analyser the use of a calculator and our ears. (and a thermometer)
Here's a thought off the top of my head if you don't have a distortion analyzer.
Buy a good sound card like the Juli@ and get some audio analysis SW for it that will show spectrum or THD. Virtins is but one example. Close the global feedback loop on the amplifier from the driver instead of from the output, so you can isolate output stage distortion. Test the amplifier at 1 kHz; this frequency, under output stage open-loop conditions, is fine for looking at static output stage crossover distortion.
Cheers,
Bob
Hi Bob,
>The capacitor makes the transistor think.....
How high is the IQ of this tranny?
Cheers,
E.
Mind you , it can be considered superior to 0..
Hi Bob, pooge,
Sound cards often have a restricted voltage input range compared to most amplifiers. The common input impedance seems to run around 10 K ohms with a peak level of +/- 5 VDC. This can vary from card to card, and a USB card may have a wider range - or not. Time to very carefully look at the specifications first. That's before you even try a signal to see if it works folks! The outputs from an amplifier may run 25 VAC to 86 VAC (for a Carver - say), RMS readings! The Carver may deliver peaks around 125 V. Sound cards may not survive any of this, or even lower. That means that you must set up something called "input conditioning", or "signal conditioning". That is what makes test equipment expensive, but since you're only worried about sound card survival, the limits are relaxed. You have to do something in order to protect the sound card.
That's all I have to say about that.
-Chris
Errr, please be careful there. I know you didn't mention the following things that seem obvious to you, but a major pitfall here exists. Let's avoid some tears ...
Sound cards often have a restricted voltage input range compared to most amplifiers. The common input impedance seems to run around 10 K ohms with a peak level of +/- 5 VDC. This can vary from card to card, and a USB card may have a wider range - or not. Time to very carefully look at the specifications first. That's before you even try a signal to see if it works folks! The outputs from an amplifier may run 25 VAC to 86 VAC (for a Carver - say), RMS readings! The Carver may deliver peaks around 125 V. Sound cards may not survive any of this, or even lower. That means that you must set up something called "input conditioning", or "signal conditioning". That is what makes test equipment expensive, but since you're only worried about sound card survival, the limits are relaxed. You have to do something in order to protect the sound card.
That's all I have to say about that.
-Chris
Drivers and pre drivers.
HI Bob,
There has been some discussion on the optimal bias of bjt opt. But what about our drivers and pre-drivers if used. Is there an optimal bias for the drivers in a darlington or drivers and pre drivers in triple darlington configurations. If so in what proportion would that be to the opt bias current.
David.
HI Bob,
There has been some discussion on the optimal bias of bjt opt. But what about our drivers and pre-drivers if used. Is there an optimal bias for the drivers in a darlington or drivers and pre drivers in triple darlington configurations. If so in what proportion would that be to the opt bias current.
David.
Hi SY,
Yes, it sure can. My point is that no one should go about merrily connecting the input of a sound card to the output of any amplifier. Not even with the best of intentions.
You need something between the speaker output and the sound card. No, somehow I don;t think a Bybee device is going to help here at all! How anyone goes about it is entirely up to them.
-Chris
Yes, it sure can. My point is that no one should go about merrily connecting the input of a sound card to the output of any amplifier. Not even with the best of intentions.
You need something between the speaker output and the sound card. No, somehow I don;t think a Bybee device is going to help here at all!
How anyone goes about it is entirely up to them.-Chris
Hi SY,
Yes, it sure can. My point is that no one should go about merrily connecting the input of a sound card to the output of any amplifier. Not even with the best of intentions.
You need something between the speaker output and the sound card. No, somehow I don;t think a Bybee device is going to help here at all!
-Chris
Even something as minimal as a low noise low distortion opamp set up as a buffer with an attenuator in front would provide some protection. Something like one of National's ultra low noise low distortion dual opamps would do. These opamps will run off batteries so the supply voltage can be keep low. This way one is risking a $5.00 opamp instead of the sound card.
David.
I hate to think what this thread, and the many others it will spawn, will be like when the book actually ships and gets read.
It'll be a very deep read for me; I've slogged my way through the other 'cookbooks', so what's a mere 600 pages? Plus learning SPICE. Oi...
I've been trying to wrap my small intellect around the Hawksford error correction concept. That seemed like the next logical step beyond common negative feedback amplifiers designed mostly in the 70s and 80s.
It'll be a very deep read for me; I've slogged my way through the other 'cookbooks', so what's a mere 600 pages? Plus learning SPICE. Oi...
I've been trying to wrap my small intellect around the Hawksford error correction concept. That seemed like the next logical step beyond common negative feedback amplifiers designed mostly in the 70s and 80s.
Hi Damon,
I can't speak for something I haven't received yet, but many audio books out lately have been an easy read. Just don't try to get through it in one sitting! The light overview that's coming out in this thread simply re-enforces the desire to have the book in my hands - like NOW!
Thankfully, it's not a "cookbook" either. That's all I know beyond reading other articles Bob has written. They should give you a reasonable idea of the writing style in this book.
Hi David,
Starting to get stickier now.
-Chris
I can't speak for something I haven't received yet, but many audio books out lately have been an easy read. Just don't try to get through it in one sitting! The light overview that's coming out in this thread simply re-enforces the desire to have the book in my hands - like NOW!
Thankfully, it's not a "cookbook" either. That's all I know beyond reading other articles Bob has written. They should give you a reasonable idea of the writing style in this book.
Hi David,
Yes, it certainly would ... as long as the rails are clamped from rising (being charged up to a higher level) due to the low impedance of an amplifier's output. You just want protection and input scaling without affecting the signal too much.
Starting to get stickier now.
-Chris
Hi Chris,
Thanks for this very important reminder. I've got a chapter in my book on audio instrumentation where I discuss a sound card interface box. These are both important and handy, and anybody can build one.
Cheers,
Bob
Hi David,
I like driver arrangements that run in class A, like the Locanthi T circuit. As you probably know, that is a triple EF where the emitters of the predrivers are connected together with a single resistor to establish the bias current, largely independent of signal swing. Same thing for the drivers.
I tend to like to run 10 mA in the pre-drivers and 50 mA in the drivers, but the 50 mA number can be over a range depending on how many output devices are being driven and how much performance at high output current slew rate is desired (i.e., how much base suck-out current you want to make available to the output transistors for turn-off). The required turn-off current also depends on the ft of the output transistors. I like to err on the hot side, but that is inter-twined with how you want to heat sink the drivers (e.g., whether you put them on the big heat sink).
The needed current to discharge the output transistor collector-base capacitance also needs to factored in, especially with designs employing many output pairs in parallel. The current needed for that is proportional to output voltage slew rate.
Cheers,
Bob
Hi David,
You're on the right track. The necessary sound card interface box can be quite simple, and something anyone can build. The box combines protection and convenience; the amount of convenience is up to you.
The key is input protection for the sound card that does not add distortion to the signal. One approach is to run the op amp at normal voltages so that you get the best performance from it, then incorporate passive limiting devices at the output to protect the subsequent sound card input circuitry. Diodes connected to threshold voltages that reverse-bias them for normal signal levels will work. The resistor in series with the diodes can be fairly low in value to minimize any distortion introduced by the diodes due to nonlinear diode capacitance or very weak turn-on.
Some amount of passive attenuation at the input to the op amp circuit, perhaps also with some protection diodes, will serve to protect the op amp. An input attenuator with a 1,2,5 or 1,3,10 sequence will add convenience.
Don't forget the other side, the output of the soundcard feeding the amplifier. It probably needs buffering and maybe a combination of attenuation and maybe even gain for some tests.
Cheers,
Bob
Hi Bob,
Yes, the normal output of a sound card should drive an amplifier without too many problems. Messing around with the "volume control" is a pain in the behind, and most sound cards can put out an almighty "THUMP" when powering up or down. Changing sources can also create a heart stopping sound. An external volume control is therefore much easier to live with and places a control nearer to your work for adjustments. Since you are adjusting the input level to the sound card as well, not a bad thing.
If you end up with a buffer to drive the input of the sound card, you should have a lower noise floor as well. That would also keep the attenuation factors closer to what you calculate without trying to correct for an unknown input impedance. We know it's probably 10K, but what is it exactly?
-Chris
Yes, the normal output of a sound card should drive an amplifier without too many problems. Messing around with the "volume control" is a pain in the behind, and most sound cards can put out an almighty "THUMP" when powering up or down. Changing sources can also create a heart stopping sound. An external volume control is therefore much easier to live with and places a control nearer to your work for adjustments. Since you are adjusting the input level to the sound card as well, not a bad thing.
If you end up with a buffer to drive the input of the sound card, you should have a lower noise floor as well. That would also keep the attenuation factors closer to what you calculate without trying to correct for an unknown input impedance. We know it's probably 10K, but what is it exactly?
-Chris