| kelticwizard |
Hello, I'd just like to ask a couple of elementary questions if I may.
An 8 ohm speaker rated at 500 watts is hooked up to two amps in parallel. In both cases, the amps are delivering the same program material.
A) In one case, amp A is 200 watts, the second amp is 20 watts. What happens when they both deliver full power? I mean to both amps and speaker.
B) Both amps are rated 100 watts. What happens to them and the speaker?
I'm interested if both amps are tube amps instead of solid state, would that change anything? |
|
|
| theAnonymous1 |
A) :hot: :hot: :hot: :hot: :hot: :hot: :hot: :hot:
B) Assuming they are identical amplifiers that can even be paralleled, you won't gain anything. You will still have the same voltage swing on the outputs, but twice the current capability.
 |
|
|
| ACD |
I can only agree with theAnonymous1 .....
A) The 20W amp will most likely get fried :hot: and proberly cause the 100W amp to get fried too :hot:
;) |
|
|
| XEAGLEKEEPER |
| Have fire extinguisher ready at hand:hot: |
|
|
| 4fun |
Hi,
Also..
If we are talking about paralleling two amps they must be identical, very much so infact.
Imagine the amps has slightly different phase response, usually common in the ends of audioband. Large currents will then travel betwen amps and......:hot: :hot: |
|
|
| anatech |
Hi kelticwizard,| quote: | | A) In one case, amp A is 200 watts, the second amp is 20 watts. What happens when they both deliver full power? I mean to both amps and speaker. |
The 20 W amplifier bocomes another load, a low resistance one. The excess voltage will dump into the lower supplies through the outputs or drivers (breaking down in reverse) or the flyback diodes.
| quote: | | B) Both amps are rated 100 watts. What happens to them and the speaker? |
If the gains are identical, there will be some form of current sharing. If even so much as the phase shifts between them, one becomes an unhappy load. Excessive currents will flow.
| quote: | | I'm interested if both amps are tube amps instead of solid state, would that change anything? |
Yes, if the amp is made to do this and the gains are the same, you change taps and the relatively high output resistance will current share. A tube amp with very high damping factor may have issues with this, or if the channels do not have the same gain (ie a fault with one side).
-Chris |
|
|
| burnedfingers |
I sincerely doubt that this would work with any type of amplifier.
What would be the advantage of even trying it? |
|
|
| anatech |
Hi Joe,
Many tube amps were set up to do that. Connect both together on the 16 R tap and connect an 8R speaker there also.
| quote: | | What would be the advantage of even trying it? |
More power! :devilr:
-Chris |
|
|
| burnedfingers |
Chris,
Wouldn't it be simpler just to use a higher powered amplifier? When one looks at the possible problems involved it just isn't logical to do it. What happens when/if one amplifiers lets out some magic smoke? Damage will result and this damage would be more extensive than it would be with just one amplifier in the mess. Well, there is driver damage also that more than likely will result. Call me cautious but I will opt for a larger amplifier every time but then again thats just me. |
|
|
| anatech |
Hi Joe,| quote: | | Wouldn't it be simpler just to use a higher powered amplifier? |
Absolutely!
There was a time when high powered amps were terribly expensive and rare. The solution was to use a couple stereo units as mono units. I'm talking about the 50's and possibly early 60's. Not now.
-Chris |
|
|
| Sonusthree |
A long time ago and before I new better...........
I once connected both channels of a Rotel integrated to a 12" driver. Both channels were fed identical signals. Worked OK at very very low volume.
I turned it up and the fuses blew everytime. Come to think of it, that Rotel was virtually indestructable. ;)
Come on..... admit it...... I bet we've all done stupid things like this before we learnt the hard way? :angel: |
|
|
| kelticwizard |
I would like to thank everyone for their clear answers.
Believe it or not, I was not after more power so much as something else. After all, there is little hearable difference between 100 watts and 120 watts.
Now, I would hope you would answer another question. I have a hunch the answer to this one is just as firm a No as the first, but I just want to make sure.
Both amps hooked up are 100 watts. However, one amp can deliver 28 V and 3.5 A, while the second delivers 57 V and 1.75 V.
I hook them both up as before, as I illustrate below.
Could this work? |
|
|
| anatech |
Hi kelticwizard,
As soon as the voltage exceeds the supply of the smaller amp, currents will flow in a way the designer never envisioned. The 28 V unit will have it's outputs and drivers reversed biased at some point.
-Chris |
|
|
| 4fun |
Hi kelticwizard,
| quote: | | Believe it or not, I was not after more power so much as something else |
What were you after?
No, as explained earlier in this thread. |
|
|
| burnedfingers |
kelticwizard
What is it that you are really after anyway? Something different or some serious power? What are you trying to drive anyway?
Please keep in mind that when you double the power you only gain 3db. |
|
|
| kelticwizard |
I was going to keep this secret, but if you are going to ask people questions and expect them to answer in good faith, you have to come clean yourself, so here goes.
I really wasn't after hooking two amps up, what I am trying to figure out is some way to hook an amp to an output transformer and have the amp drive the speaker for most of it's range, and have the output transformer drive the speaker in the bass range.
The output transformer would have a higher voltage and lower amperage than the speaker, so in the bass range where the speaker's impedance is high, more power can be delivered to it.
As a second choice, illustrated below, would be to have an output transformer with a network integral to it which gives higher voltage/lower impedance in the bass range. Haven't wuite worked out the details on that, (amateur, no tech training), but I do think a resonance network can be worked out.
The idea is to give a bass boost in the high impedance bass region of the speaker while leaving the rest of the frequency range pretty much alone.
All this is for a single voice coil speaker. I suspect a double voice coil speaker would present better possiblilites, but I want to work it out for a single voice coil speaker at this time. |
|
|
| anatech |
Hi kelticwizard,
The real problem in extending bass response is that the resonant frequency of the box / woofer is fixed. It forms an acoustic high pass filter. The response will fall quickly after that.
The only thing you will accomplish in driving the woofer harder at low frequencies will be to exceed it's maximum excursion. If you are lucky, it's just a "clack" as the voice coil strikes the back plate (flattening the rear of the former). The distortion will rise long before this.
The way I see it, your quest for more bottom end will best be answered physically. It's the box and speaker combination. Some boxes are better at extending the low frequency range than others. The speaker must have characteristics suitable for the box.
I think Dave and Cal could help you here better than I can for sure.
-Chris ;) |
|
|
| kelticwizard |
| quote: | Originally posted by anatech
The only thing you will accomplish in driving the woofer harder at low frequencies will be to exceed it's maximum excursion. If you are lucky, it's just a "clack" as the voice coil strikes the back plate (flattening the rear of the former). The distortion will rise long before this.
-Chris ;) |
I disagree. If I had made the box bigger, the resonant frequency (closed box with Vb well under Vas) would be moved down and more power would then be delivered at the same frequencies which are now cut off with the smaller box. Of course, I would want the network to give a shape and amplitude commensurate with the rolloff of the box/speaker combo-not send it up to some 20 dB above midpoint or whatever.
As for distortion, I would like to try it and see. |
|
|
| BWRX |
Regardless of the driver and box behavior, it's wise to heed the advice of everyone regarding the interaction between the output stages of two different amps. If they're not intentionally identical or within the same feedback loop then don't expect anything less than sub-optimal results.
What you're trying to achieve is sort of an active crossover with two amps and a single driver. It would be better to try with two amps and two drivers ;) |
|
|
| kelticwizard |
Thank you, BRWX.
The two amps bit was just to try to get a feel for what happens when two sources feed a single voice coil. Like I said, if I had my druthers, I druther have a transformer hooked up for the lowest octave and leads from the amp feeding it for the rest. Only as a second choice would I have the amp driving the single voice coil speaker throughout it's range, with a network boosting voltage and suppressing amperage the bottom octave. So I never really intended to hook up two amps, but am wondering about hooking up two sources-amp and transformer driven by the same amp.
But your cautions are well taken, and certainly should be stated for other diyers who might actually intend to hook up two amps to the same single voice coil driver. |
|
|
| BWRX |
Don't thank just me. I only reiterated/reinforced the cautions of others.
The situation of putting amplifiers in parallel is similar to putting power supplies in parallel. You want the output voltage of both to be exactly the same or they will not share the current equally. An amplifier is similar except the voltage is constantly varying!
The situation is quite a bit different when one amplifier is driving two loads (i.e. a midrange and a tweeter separated by crossover). A properly designed crossover will present the amplifier with a fairly constant impendace over the required frequency range. Two amplifiers in parallel with any small voltage difference in output will present each other with a low impedance path for current to flow. I believe anatech mentioned this earlier. |
|
|
| XEAGLEKEEPER |
| Have you given any thought to a electronic crossover? |
|
|
| planet10 |
| quote: | Originally posted by burnedfingers
Wouldn't it be simpler just to use a higher powered amplifier? |
Actually 2 amplifiers are often used to make a more powerful amplifier. But the amps are identical (or as close as practical.
Most common today is where 2 amplifiers are connected in series, one fed an in-phase signal, the other out of phase. Most commonly called a bridged amplifier, Many stereo amps can be bridged into mono (outputting 4x the maximum power but cuurrent limited to twice the impedance of a single channel alone). And sometimes amps are just built this way. I'd guess, in some sense, a push-pull amplifier is athe simpliest version of this scheme.
The other way is to parallel 2 amplifiers. This gives no more power at the original nominal load but with double the current is capable of driving a load with full power at half the impedance of a single amp. Probably the most common current example are LM4870 amps where both amps in the chip are run in parallel. Being 2 3886 chips, which if run at close to max voltage don't really like anything below 8 ohms, the parallel amp allows full power at 4 ohms.
This last was typically how tube amps when run together with the added complexity/versatility of a multi-tapped transformer.
dave |
|
|
| planet10 |
Now an answer to Peter's real question: you drive the speaker with a current amp (aka transconductance amp). Examples are 1st Watt F1 & F2, and any pentode amp with little or no feedback. SE triode amps tend also to act as mild current amps with their low damping factor.
These are most usful with FR speakers where the impedance rises as their output tends to fall off (at least when driven with the ubiquitous voltage amp).
dave |
|
|
| kelticwizard |
| quote: | Originally posted by XEAGLEKEEPER
Have you given any thought to a electronic crossover? |
Not on this setup.
But even if I used the electronic crossover, wouldn't that still be two amps powering one single voice coil speaker? One might take over for the mids and one for the bass, but near the crossover frequency both amps will be powering the speaker nearly equally. Wouldn't that present the same problems? |
|
|
| kelticwizard |
| quote: | Originally posted by planet10
Now an answer to Peter's real question: you drive the speaker with a current amp (aka transconductance amp). Examples are 1st Watt F1 & F2, and any pentode amp with little or no feedback. SE triode amps tend also to act as mild current amps with their low damping factor.
dave |
I was very high on the idea of a transconductance amp, but the problem is that if I have a high output impedance, I wouldn't be able to deliver full power over the frequency where the impedance is 8 ohms.
Suppose I am driving a single speaker in a closed box. My amp is rated 32 watts. From 2,000 Hz down to 80 Hz, the speaker averages close to 8 ohms. So if I give the speaker 2 amp current, over this range my voltage will be 16 volts, and I am delivering full power for a 32 watt speaker.
Now I hit the bass range, and the impedance rises to 32 ohms at resonance. The transconductance amp sends 2 amps to the speaker, and now the speaker requires 64 volts to cover it. In effect, I would need a 128 watt speaker-not a 32 watt.
The voltage amp delivers full power to the mids and treble, but not the bass. The transconductance amp delivers full power to the bass, but not the mids and treble. :xeye: That is why I am looking at these alternatives. |
|
|
| burnedfingers |
| quote: | | The other way is to parallel 2 amplifiers. This gives no more power at the original nominal load but with double the current is capable of driving a load with full power at half the impedance of a single amp. Probably the most common current example are LM4870 amps where both amps in the chip are run in parallel. Being 2 3886 chips, which if run at close to max voltage don't really like anything below 8 ohms, the parallel amp allows full power at 4 ohms. |
For the sake of arguement here lets consider SS home or commercial grade amplifiers and let the clones and car amps alone.
I understand fully how bridging works and also the concept of parallel amplifiers. I just don't understand the logic behind trying to make two nearly identical amplifiers work together. I have seen
what was left when someone tried to do this and it ain't pretty. I believe logic would dictate that common sense be employed. Granted it does work to some degree with tube amplifiers and had been used in the past due to the lack of large tube amplifiers.
Also one must consider the increase in THD and IMD in doing it. Granted there are amplifiers out there that do employ the capability to parallel amplifier channels and still stay together. Crown makes a few models that allow the user the ability to flip a switch and parallel up amplifiers. Please note that they were specifically designed to do this task.
With respect to the bass boost idea...I don't think it will work because I have seen a similar idea that was tried about 15 yrs ago by a engineer friend. I would however suggest the idea of using a electronic filter to tune. I had a crossover once that I had a 27hz variable bass boost circuit. I used this and still do on my 18" sub in my theatre. Everything below 27 is kicked to the curb, the boost is at 27hz and the maximum frequency is 74HZ at a 24 db crossover. My box is 4th order tuned to 27HZ partly because I don't need wasted cone motion and partly because of my standing wave in the room. This has worked very well for over 10yrs and continues to work today. |
|
|
| AndrewT |
Hi Keltic,
the increase in impedance comes in the range where the driver is at or near resonance.
A fortunate consequence of this resonance is very high efficiency.
Feed the speaker with a constant voltage signal at various frequencies and the increase in impedance is compensated by the increase in efficiency. The manufacturer strives to get the result to be a flattish response at all frequencies (some are better than others). It seems the reduced power delivered gives the same acoustic output (by design).
The same cannot usually be said of treble drivers. At resonance the increase in efficiency is more than the reduction in delivered power (due to the increased impedance) and the result is a bump in the response that often sounds terrible and various compensations are implemented to reduce acoustic output around this resonant peak. |
|
|
| kelticwizard |
| quote: | Originally posted by burnedfingers
With respect to the bass boost idea...I don't think it will work because I have seen a similar idea that was tried about 15 yrs ago by a engineer friend. |
Details would be greatly appreciated.
| quote: | Originally posted by burnedfingers
I would however suggest the idea of using a electronic filter to tune. I had a crossover once that I had a 27hz variable bass boost circuit. I used this and still do on my 18" sub in my theatre. Everything below 27 is kicked to the curb, the boost is at 27hz and the maximum frequency is 74HZ at a 24 db crossover. My box is 4th order tuned to 27HZ partly because I don't need wasted cone motion and partly because of my standing wave in the room. This has worked very well for over 10yrs and continues to work today. |
A nice use of a vented box and electronic crossovers, which can give steep slopes easily, thereby protecting the woofer from overexcursion below box tuning frequency.
On this idea, I do plan to use a vented box eventually, but first I want to establish how things work in the simpler closed box before adding the variable of box tuning.
And about your engineer friend's idea-the first application I have in mind is PA speakers, so if there is an increase in distortion, within reason, it would still be acceptable. PA systems put a greater emphasis on output than extremely low distortion compared to home systems. |
|
|
| anatech |
Hi kelticwizard,
As has been said earlier, the efficiency goes up at resonance and the distortion drops (in ported enclosures) due to the extra damping on the woofer.
The fact remains that you can not connect two conventional amps in parallel. If you connect a transconductance amp in parallel with a normal amp, the high voltage at resonance will damage the conventional one. This is not the way to accomplish what you want to do.
The way I see it, you are still stuck with a physical speaker / box solution. Especially if you are limited to a 30 W speaker in what appears to be a full range design - or a 2 way with a high crossover point. Why not consider a sub in this case?
-Chris |
|
|
| kelticwizard |
Chris:
I used the 32 watt amp just as an example to illustrate the point. The math was easier with it. The point is, transconductance amps give full power to the bass but not the mids or above, voltage amps give full power to mid/treble but not bass range, and those with fixed output impedances in between are exercises in splitting the difference somehow.
Now an amp or amp/transformer combo which changes it's output impedance in the bass range from where it is in the mid/treble range, when fitted to the speaker/box combination with characteristics designed for it, would give more bass in less space. Without increasing the max power requirements of the amp. Which is kind of the point. |
|
|
| burnedfingers |
There is no solution for what you want to do here other than frequency limiting the "main" speaker as to take the low frequency load off it and shift it to a subwoofer. You build a circuit to kick everything from say 100hz to the curb for the "main speaker" and use the sub from the usable lower limit with consideration given to X max and tuning to an upper range of say 100hz. This when used correctly will only add to the system. The system will be able to be run harder with more overall clarity.
I am employed in pro/commercial sound. |
|
|
| kelticwizard |
Yes, using present equipment, that is the only way.
What I am trying to do is to modify a few things to possibly find a better way.
You mentioned an engineer friend who tried something awhile back along these lines. Could you be so kind as to tell me what he tried, and what came of it? :) |
|
|
| AndrewT |
| Probably a pair of destroyed amplifiers! |
|
|
| Geoff H |
kelticwizard, First of all, let me commend you on your thinking outside the square.
What you are trying to achieve is an electronic solution when perhaps the problem should be addressed at the enclosure, acoustically.
Rather than trying to mix the output of two amps, it may be better to redesign the amp so that below a certain frequency it behaves as a constant current source, and above that frequency a constant voltage source. That could be a nightmare resolving stability problems.
The same result can be obtained with equalisation. The freq response is tailored so the driver gets an increase in drive near resonance.
"The fact remains that you can not connect two conventional amps in parallel"
That statement is incorrect. It has been done many times. The trick is to use low R load balancing resistors. The amps must be as identical as you can get, and the inputs fed exactly the same.
It's probably not the best way in Hi-Fi, either.
In fact, I was doing that just yesterday, driving a common channel bass driver. I was running it full time for several weeks. No blown amps. Just takes a bit of care.
As Andrew pointed out, the efficiency increases at resonance, so it's not a requirement to have a flat z curve. At one stage I had the z peak on my full range system at 11 ohms, from a high Q driver. It sounds more lively with a Zo at 16 ohms. The proof of that is 35 Hz at -5dB, in 29L cabs.
But keep searching ideas, you may come up with something.
Geoff. |
|
|
| burnedfingers |
| quote: | | What I am trying to do is to modify a few things to possibly find a better way. |
Don't you think that if there was a better way it would have been done long before this?
| quote: | | You mentioned an engineer friend who tried something awhile back along these lines. Could you be so kind as to tell me what he tried, and what came of it? |
I will be kind here and sum it up. He blew the hell out of two amplifiers and some drivers. He learned rather quickly that the uneducated like myself sometimes possess vital information and knowledge as a result of experience and this should be weighed along with textbook theory.
If I may suggest... Try to pick up a copy of "HI-FI Loudspeakers and Enclosures" by Abraham B. Cohen
This book sums up the experience laid down by the fore fathers of HI-FI. Chapter 21 is all about the "Electrical Third Channel" and this may give you some ideas. |
|
|
| Geoff H |
On page 380 of that good book, you will see two amps connected together via chokes, feeding the one driver. That is similar to how I do it.
Bear in mind this goes back to the days of forgiving tubes. On a solid state amp, it would be wise to check the DC offset first. My concern with this scheme was to do with bass energy not being balanced between the two channels.
"Don't you think that if there was a better way it would have been done long before this?"
If we all thought that way, we would be feeding the horses, not filling up at the gas station. It's imagination, creativity, and good discussion that maintains progress. Let's not kill it.
Geoff. |
|
|
| kelticwizard |
| quote: | Originally posted by burnedfingers
If I may suggest... Try to pick up a copy of "HI-FI Loudspeakers and Enclosures" by Abraham B. Cohen
This book sums up the experience laid down by the fore fathers of HI-FI. Chapter 21 is all about the "Electrical Third Channel" and this may give you some ideas. |
I have thqt book. I love it. Though pre-Thiele-Small, it is one of the most underrated books in audio history. I will indeed re-review the Electrical Third Channel chapter, it has been a few years.
Actually, though, from the outset I was thinking less in terms of two actual amps and more in terms of using the direct amp for mids/highs, and a transformer output hooked up to that selfsame amp for the bass. Or some other alternative. |
|
|
| Geoff H |
Do you mean something like this? It works.
The upper transformers are hi quality units, ex Martin Audio (UK) Concert Series. The lower unit is a bifilar wound unit, providing a 2:1
ratio. Below the xo freq, the 8 ohm load transforms to a 2 ohm load, which loads each chan at 4 ohms.
A very easy way to try a bass unit with fullrange drivers with a big difference in sensitivity.
Regards,
Geoff |
|
|
| kelticwizard |
Geoff:
Thanks for posting that schematic. Is that from Cohen's book? I'll have to dig it out to review.
At first glance, I'm having some trouble understanding that schematic. Let me look it over a couple of days to get an idea what is going on with this. |
|
|
| Geoff H |
No, not quite. It's based on the series crossover used by Scanspeak and Peerless at times, with the low frequency sections merged.
It provides a gentle slope in the crossover region, but then goes to 12dB/oct. At the crossover freq, the drivers are in series, either side of the Fo the Xl or Xc shunt the driver, providing full drive to the other driver.
The Auto transformers are used as chokes, in the 10mH region, with low R.
Geoff.
edit: The lower transformer changes the Z seen by the amp/s to 2 ohms, which is shared between the two channels. I.E. 4 ohms load per channel.
No smoke, blown fuses or 3 way short circuits.
|
|
|
| kelticwizard |
| Thanks again, Geoff. I will examine the circuit in more detail, and see if I understand what is going on. If I don't, I'll prepare a list of cogent questions to see if I can figure it out. |
|
|
|
