The BW is an issue with this type of bread-board. I just put an HP3314A functiion Gen and scope and measured the BW. Here with -7171 I limited the BW of the circuit to 1.2MHz and I also use low Z circuit values which help a lot to mimimize affect of strays'.
🙂
THx-RNMarsh
Last edited:
Speaking of oscillating breadboards.
If using one with the aluminum plate on the bottom .....
Make sure the plate is connected to the ground
of the circuit ! (and there is no apparent provision)
If using one with the aluminum plate on the bottom .....
Make sure the plate is connected to the ground
of the circuit ! (and there is no apparent provision)
Depending on the size of the SMD devices and the copper weight requirements of the power board you can have a mixed technology design as illustrated by Esperado (many boards are already done this way as connectors are still PTH) or where the copper weight for the power board is to heavy to allow the fine SMD detail to be etched then a daughter board is the solution.
As to power devices I wish I knew... There are new generation of power packages for SMD:
http://www.irf.com/technical-info/whitepaper/thermalpcim02.pdf
I don't think PTH TO-245 packages will totally disappear, but what is available in them will diminish, driven by market forces. For some designs where PTH is required it is fun trying to source components, the dreaded words "Not recommended for new designs" coming up more frequently.
Sorry for the late reply, these time zones are a bugger.
Thanks Marce,
I happen to still be up and was reading a book for once, 12:50pm. I'll have to go back and look at some of the VSSA threads and see where the latest development is on that. That may be a very good alternative to a class D amplifier in my case. I think I have more than enough heatsink area for that amplifier output.
I happen to still be up and was reading a book for once, 12:50pm. I'll have to go back and look at some of the VSSA threads and see where the latest development is on that. That may be a very good alternative to a class D amplifier in my case. I think I have more than enough heatsink area for that amplifier output.
I have thought it interesting that noise signals when added go by pythagorean, IE 1+1=1.414...
Am I right when I guess that this is because the phase components of the frequencies are randomly distributed between 0 and 180, and therefore, average to 90 degrees? So you can think of it as two sine waves 90 degrees apart being mixed?
Am I right when I guess that this is because the phase components of the frequencies are randomly distributed between 0 and 180, and therefore, average to 90 degrees? So you can think of it as two sine waves 90 degrees apart being mixed?
Because of EMC testing, signal integrity etc. these days a real PCB is done so that all these things can be tested well as the functionality of the circuit... And the nearer the prototype is to the finished product the better. Also apart from the SOIC like devices some of the newer packages are impossible to breadboard properly, especially on the digital side. Over the years I have actually done very few 2 layer boards, most have been 4 layers of more just to get better signal integrity, EMC immunity Power Supply integrity etc, you can get a much higher quality product.
Quite surprised to read this. I you add a current buffer in the feedback path of a CFA, you add a non linear element in this feedback path, and pole and zero. The low Z of the feedback resistance is just needed to reject the HF cut frequency created by the feedback resistance all along with the parasitic emitter capacitance out of influence ( with the associated phase turns).
So, we are back to the same negative effects of using an active device in the feedback path than in a VFA.
The only thing that remain the same is Feedback is applied to the emitter of the input transistor, with the associated dynamic reduction of the base-emitter voltage.
Is this the secret of the sound differences some believes between the two topologies ? I don't know.
Last edited:
That's because you are doing high performance mixed signal design rather than flooby!
Not always, I do and have done an awful lot of analogue over the years, the analogue was always multi layer even when some early digital was done on two layers. I am agreat believer in 4 layers giving a much better design, it was what I was taught as a fresh faced optimistic lad when I first moved into PCB design, especially by the analogue guys....
We are striving for audio perfection so should not compromise on the most important component in any assembly, the PCB.
🙂
I hate to compromise a layout for a couple of layers... The really funky stuff is 8 - 16 layers (the most I've done is 24 layer flexi rigid, deep pocket design... Hypertac connectors)
Last edited:
Designed as a CFA, the X1 buffer (Z converter) does not change that. You were the first one I saw to do it with discrete circuitry for a Power Amp. Should be great performance as per your usual.
THx-RNMarsh
At least the distortion added by the buffer that nothing will compensate. For what to gain ? The low Z feedback load is not a problem for a power amp, and make the feedback path less sensible to RFI RMI. The only negative point I can see is the need for a power resistance, here.
This said, If you had experimented this topology, you are in a better position than me to comment.
Perhaps if you take a look at the schematic of John's ML-2 circa '77, you will
see a similarity.
😎
You something lose and something gain. In mine CFA configuration distortion is going down with lower feedback network impedance but it's not good to go to low because the power loss and dissipation on the feedback resistors. I use 22R/560R as lowest value. When I added a buffer to the low impedance feedback input I've got even lower distortion then without, at least in simulation. I think there two reasons for that, very low buffer it self distortion (supper pair emitter follower) and even lower then 22R buffer output impedance. That is just simulation, but I hope as with other my simulations not far from reality. Don't be afraid to think out of the box and you could be surprised.
With that buffer in place you still can use quite low impedance feedback network.
And that opens other possibility I can't explain yet.
LT1363 VFA
From page 11 of the LT1363 datasheet:
"Comparison to Current Feedback Amplifiers
The LT1363 enjoys the high slew rates of Current Feed-
back Amplifiers (CFAs) while maintaining the characteris-
tics of a true voltage feedback amplifier. The primary
differences are that the LT1363 has two high impedance
inputs and its closed loop bandwidth decreases as the gain
increases. CFAs have a low impedance inverting input and
maintain relatively constant bandwidth with increasing
gain. The LT1363 can be used in all traditional op amp
configurations including integrators and applications such
as photodiode amplifiers and I-to-V converters where
there may be significant capacitance on the inverting
input. The frequency compensation is internal and not
dependent on the value of the feedback resistor. For CFAs,
the feedback resistance is fixed for a given bandwidth and
capacitance on the inverting input can cause peaking or
oscillations. The slew rate of the LT1363 in noninverting
gain configurations is also superior in most cases."
My point being that it is a VFA.
From page 11 of the LT1363 datasheet:
"Comparison to Current Feedback Amplifiers
The LT1363 enjoys the high slew rates of Current Feed-
back Amplifiers (CFAs) while maintaining the characteris-
tics of a true voltage feedback amplifier. The primary
differences are that the LT1363 has two high impedance
inputs and its closed loop bandwidth decreases as the gain
increases. CFAs have a low impedance inverting input and
maintain relatively constant bandwidth with increasing
gain. The LT1363 can be used in all traditional op amp
configurations including integrators and applications such
as photodiode amplifiers and I-to-V converters where
there may be significant capacitance on the inverting
input. The frequency compensation is internal and not
dependent on the value of the feedback resistor. For CFAs,
the feedback resistance is fixed for a given bandwidth and
capacitance on the inverting input can cause peaking or
oscillations. The slew rate of the LT1363 in noninverting
gain configurations is also superior in most cases."
My point being that it is a VFA.
Internally, it is a CFA circuit topology with a (diamond) buffer also on the invert input.... which gives it the best of both worlds.
From an application view, an engineer can still think in general terms of the VFA he/she already understands. And, still get the inherent high SR etal of the CFA.
THx-RNMarsh
Last edited:
- Status
- Not open for further replies.