comments on my ongoing project appreciated

[Jef Patat]

Dear reader,

I have some experience in building kits and repairing circuits. Now I would like to design and build my own high quality amplifier.
For this purpose I designed a website, but as I am quite new, I'm not sure if I interprete everything correct and complete. I would like my website to be a resource for other newbies as well.

All comments are appreciated, many thanks, Jef

 

[mechie]

Three Stage Audio Amp

Jeff
Congrats on your website - looks good 8)

A question about your **broken link removed** -

The output stage looks like a 'Push-Pull' driver, requiring two signals to it, in-phase but with a constant bias between them.

The voltage amp seems to cater for this bias with V1 at 1.4v but the connections between the two stages seem wrong, the lower push-pull half is connected to the -50v supply rail :?:
Won't this cause Q8 and Q9 to be turned permanently 'hard on' :?:

 

[Jef Patat]

Indeed

Hi,

and thanks for the reply mechie.
You are indeed right, as the image was it would constantly turn hard on.
It was a mistake in my drawing which I now corrected.

Thanks a lot for your critical view.

Kind regards, de jef.

 

[audioguru]

Hi Jef,
Your choice of little TIP31C and TIP32C transistors as output devices will probably blow them up quickly. The peak current is nearly 100V across 8.2 ohms (many speakers have an even lower impedance at certain frequencies) which is 12.2A!
Also, these little transistors have a very low Safe Operating Area.

Now you are using full darlington arrangements for the outputs along with their double-junction voltage loss. A better arrangement would probably be is to use NPN-PNP combinations, with an NPN emitter-follower driving a PNP common emitter and vive-versa. You know what I mean?

 

[Roff]

Hi Jef,
Your choice of little TIP31C and TIP32C transistors as output devices will probably blow them up quickly. The peak current is nearly 100V across 8.2 ohms (many speakers have an even lower impedance at certain frequencies) which is 12.2A!
Also, these little transistors have a very low Safe Operating Area.

Now you are using full darlington arrangements for the outputs along with their double-junction voltage loss. A better arrangement would probably be is to use NPN-PNP combinations, with an NPN emitter-follower driving a PNP common emitter and vive-versa. You know what I mean?

The peak current is really ≈50V/8ohms (≈6 amps), but it is still too much current and power for these transistors.

 

[Nigel Goodwin]

The peak current is really ≈50V/8ohms (≈6 amps), but it is still too much current and power for these transistors.

I don't think he's ever suggested using TIP31/31 as the output devices?, he hasn't got the output stage designed yet (or at least not the webpage for it!) - I presume these devices are simply being shown in the crude overall theoretical diagram, not with any thought of their use!. Probably to allow a simulator to run?.

 

[audioguru]

[quote="Ron H"
The peak current is really ≈50V/8ohms (≈6 amps),
[/quote]
Hi Ron, Happy New Year.
Am I wrong to believe that if the amp drives a speaker cone and pushes a certain amount of air from 0V idle with a peak of 50V, then the power must be double when the amp must push the speaker cone and certain amount of air twice the distance using twice the voltage?

It might use more than 12A since the inductance of the speaker's voice coil will make a back-EMF that will try to resist motion in the opposite direction.
Does anyone have a sim that is smart enough to figure this out?

 

[Roff]

Hi Ron, Happy New Year.
Am I wrong to believe that if the amp drives a speaker cone and pushes a certain amount of air from 0V idle with a peak of 50V, then the power must be double when the amp must push the speaker cone and certain amount of air twice the distance using twice the voltage?

It might use more than 12A since the inductance of the speaker's voice coil will make a back-EMF that will try to resist motion in the opposite direction.
Does anyone have a sim that is smart enough to figure this out?

Happy new year right back atcha!
I found a speaker model **broken link removed**, which I sim'ed on Jef's amp. I ran an AC sweep to find the speaker's minimum impedance frequency, which was at DC and around 160Hz, and the maximum impedance frequency, which was about 27Hz. The maximum phase shift points were at 20Hz and 35Hz. I ran the amplifier at a gain of 11, with 80V p-p into the speaker (probably not possible with real transistors of the types in the circuit). Peak currents varied from 800ma to 6.2 amps. Even at the extreme phase shift points, currents in the output transistors were well-behaved class B waveforms. I can post them if anyone is interested.

This model obviously doesn't represent all speakers, but the waveforms were so clean that I believe the results would be similar for other speakers.

 

[audioguru]

Hi Ron,
Ah-ha! Your speaker doesn't have a cabinet so it is just flopping around, the air pressure in front of it sneaks around its edge to fill the vacuum behind it. It isn't doing any work! Please tell your sim.

With a cabinet, it will resonate with a 2-pole response if sealed, or a 4-pole response if ported. Those extra poles will make it appear highly inductive and capacitive near resonance. At clipping, the transistors will have the 100V supply across them at the same time as a peak current of more than 12A. Those little TIP transistors will release their majic smoke, and probably short, blowing the speaker!

That is why 140W RMS into 8 ohms amps like this one is supposed to be, have many high-voltage and high-current transistors.

 

[TheOne]

The model seems to be OK. Here is a model for a typical enclosed 2 way speaker system.

One other point on amplifier HF instability: In a study it was stated that it is very difficult to model theoretically apart from global feedback (Nyquist) stability calculation and simulation. It showed that the higher pole frequencies that caused phase shifts to accumulate are most of the time very mysterious and difficult to account for.

The other oscillations, however, such as parasitics and other more obscure oscillatory misbehavior, seem to depend on various unknown or partly-known second-order effects that are difficult or impossible to deal with quantitatively and are quite reasonably left out of simulator device models. This means we are reduced to something not much better than trial-and-error when faced with a tricky problem.

 

[Roff]

Hi Ron,
Ah-ha! Your speaker doesn't have a cabinet so it is just flopping around, the air pressure in front of it sneaks around its edge to fill the vacuum behind it. It isn't doing any work! Please tell your sim.

With a cabinet, it will resonate with a 2-pole response if sealed, or a 4-pole response if ported. Those extra poles will make it appear highly inductive and capacitive near resonance. At clipping, the transistors will have the 100V supply across them at the same time as a peak current of more than 12A. Those little TIP transistors will release their majic smoke, and probably short, blowing the speaker!

That is why 140W RMS into 8 ohms amps like this one is supposed to be, have many high-voltage and high-current transistors.

Audioguru, I am no audio guru. I'm just an old analog engineer with a lot of experience, but very little in audio. I was aware that I had no enclosure model in my sim, but had concluded from some web sites I visited that they were pretty much second order effects. If you know that they are important to amplifier power ratings, could you point us to some references? I have trouble imagining a speaker/enclosure network that would have 180 degrees phase shift between current and voltage, and still have 8 ohms impedance (I think that's what you are implying). Perhaps the back EMF effect you mentioned could explain it? When one measures the inductance if a speaker, does the effect of the magnet get included? I would think so, but I never thought about that before.

I agree that TIP31C and TIP32C are too wimpy for this application, and I also agree with Nigel that they were probably not intended to be the final selections.

 

[Jef Patat]

Sorry for all the confusion

I agree that TIP31C and TIP32C are too wimpy for this application, and I also agree with Nigel that they were probably not intended to be the final selections.

Let me take away all doubt: I had no intention of using this transistors, i never looked up a datasheet, i just chose a random transistor for drawing purposes. I never thought this could lead to this kind of a discussion.

I MUST say that I feel it woul be a very bad idea to build this schematic an actually use it, its only pictured there to give a global theoretical overview of a three stage amp.

I feel that music will always contain short term peak currents that may exceed what one can calculate out of Ohm's law, independent of speaker configuration. For the moment I think i always used transistors that were able to deliver a minimum of 2 to three times the current calculated with Ohm's law.

I will change the image on the particullar page, as i see it can cause much confusion.

Kind regards, de Jef