Look at Figure 6. https://www.ti.com/lit/an/snaa021b/snaa021b.pdf
Notice how both amplifiers share the same Rin. This solves a problem but it isn't foolproof. Input stages need to be reasonably close and parallel input capacitance needs to be taken into consideration. Issues multiply as more inputs are paralleled. I paralleled 3-1875s with no problem. Was a 47K resistor for Rin driven directly by an op amp buffer.
Keeping resistor values in the feedback loop as small as practical greatly mitigates parasitic issue that could unbalance the circuit. Values shown are fine. Notice how C1 is large with a low pole frequency with R1. This is necessary to keep imbalances from upsetting the circuit.
Keep in mind that electrolytics are very poor tolerance parts. Matching them is a moot point. It's a good idea to keep subsonic frequencies out of the amplifier altogether by blocking them at the input with a high order, high pass filter.
My triple parallel LM1875 amplifier used a single ended power supply. This offered two advantages. First, the issue of DC offset causing problems is completely eliminated. Second, since each leg has its own coupling capacitor to the speaker, smaller value capacitors can be used. And third, the tab can be bolted to the heatsink without an insulator or electrical isolater that way.
Notice how both amplifiers share the same Rin. This solves a problem but it isn't foolproof. Input stages need to be reasonably close and parallel input capacitance needs to be taken into consideration. Issues multiply as more inputs are paralleled. I paralleled 3-1875s with no problem. Was a 47K resistor for Rin driven directly by an op amp buffer.
Keeping resistor values in the feedback loop as small as practical greatly mitigates parasitic issue that could unbalance the circuit. Values shown are fine. Notice how C1 is large with a low pole frequency with R1. This is necessary to keep imbalances from upsetting the circuit.
Keep in mind that electrolytics are very poor tolerance parts. Matching them is a moot point. It's a good idea to keep subsonic frequencies out of the amplifier altogether by blocking them at the input with a high order, high pass filter.
My triple parallel LM1875 amplifier used a single ended power supply. This offered two advantages. First, the issue of DC offset causing problems is completely eliminated. Second, since each leg has its own coupling capacitor to the speaker, smaller value capacitors can be used. And third, the tab can be bolted to the heatsink without an insulator or electrical isolater that way.
Still waiting for the least intention of diy anything around your Parasound Z amps.Up until this post I'm the only one who posted a schematic in this thread and made a real diy suggestion...Discussing purely vague theoretical things are better suited for other forum sections!
[/QUOTE]
My triple parallel LM1875 amplifier used a single ended power supply. This offered two advantages. First, the issue of DC offset causing problems is completely eliminated. Second, since each leg has its own coupling capacitor to the speaker, smaller value capacitors can be used. And third, the tab can be bolted to the heatsink without an insulator or electrical isolater that way.
[/QUOTE]
I made similar suggestions here:
https://www.diyaudio.com/community/...instead-of-bridged.391167/page-2#post-7145610
I'd say Fig 13 of your document is more appropriate to the scope of this thread as Parasounds Z amps have dc offset correction .
Unfortunathely the OP corrupted almost everybody into a theoretical sterile discussion without telling anything around the very amps he's trying to work around...feeding a fake rhetorique around a Black Box like subject .
Truth is you don't need much real world hard data for today's PHD works... you can keep it vague and theoretical to the last bit .
Then just print it or let an AI find and compile the relevant text over the internet before autogenerating your next PHD work 🙂.
Was looking the other day for some schematic and review of a Pioneer deck and the only pages Google indexed were written by me a few years ago ...You get the ideea!
My triple parallel LM1875 amplifier used a single ended power supply. This offered two advantages. First, the issue of DC offset causing problems is completely eliminated. Second, since each leg has its own coupling capacitor to the speaker, smaller value capacitors can be used. And third, the tab can be bolted to the heatsink without an insulator or electrical isolater that way.
[/QUOTE]
I made similar suggestions here:
https://www.diyaudio.com/community/...instead-of-bridged.391167/page-2#post-7145610
I'd say Fig 13 of your document is more appropriate to the scope of this thread as Parasounds Z amps have dc offset correction .
Unfortunathely the OP corrupted almost everybody into a theoretical sterile discussion without telling anything around the very amps he's trying to work around...feeding a fake rhetorique around a Black Box like subject .
Truth is you don't need much real world hard data for today's PHD works... you can keep it vague and theoretical to the last bit .
Then just print it or let an AI find and compile the relevant text over the internet before autogenerating your next PHD work 🙂.
Was looking the other day for some schematic and review of a Pioneer deck and the only pages Google indexed were written by me a few years ago ...You get the ideea!
Last edited:
Servos work a lot better in a situation like this. The tutorial has a suggested servo circuit.
I think limiting the bandwidth is a path to limiting problems with the two amplifiers "fighting" each other. But I still think it's going to be kind of a mess. Making appropriate mods to the amp so it would work in parallel would be sketchy too I think.
I'd biamp them all day long. There's your most effective way to reduce distortion and "kick it up as notch." Bridging will certainly work better than paralleling in this situation.
I think limiting the bandwidth is a path to limiting problems with the two amplifiers "fighting" each other. But I still think it's going to be kind of a mess. Making appropriate mods to the amp so it would work in parallel would be sketchy too I think.
I'd biamp them all day long. There's your most effective way to reduce distortion and "kick it up as notch." Bridging will certainly work better than paralleling in this situation.
So pretty much you're saying that using a $300 MSRP Parasound Z-amp in parallel stereo configuration just.. isn't... a..... good.... idea.
I doubt they spent the money on 0.1% tolerances in there. And I would be halving the input impedance plus having to deal with mistmaches due to Rin/Cin for Left and Right channels so the input filter would be somewhat different as well.
True... but these are brand new speakers! Heck, I spent almost 500 bucks on them (on sale, yeah they are a steal... but still, 500 bucks ).
And if I were to biamp... and I've always wondered... the woofer will always take more power than the mid/tweeter, huh? I've always biamped with identical (*) amplifiers but I made it vertical... so that either amp drew the same both low and high duty.
Still then... transcending my own issues (**) how would one get around to building a custom parallel based amplifier and would it be of any benefit other than just a curiosity?
Note: yes, i see the servo discussion, and now that I looked at the note in more detail. interesting Figure 17.
(*) OK, I stand corrected here... there really is not much identical unless it was built by someone in these forums - he knows who he is... ;-)
(**) OK, I got many issues, hopefully I won't get another 50 posts on that one... 😛
Last edited:
Did you ask your money back for that course since it is obvious you did not learn even the basics?
I'm up to bat! I think it would only benefit if your were stuck with a 1 Ohm load somehow. A resistance value that could sink (make use of) the extra current available via paralleling - say - two, 2 Ohm capable output. A 6 Ohm load couldnt take advantage of the additional current capability.
I cant think of a reasonable scenario where speakers would be connected to form a 1 Ohm load. 4, 4 Ohm speakers in parallel? Why, when you could do a series / parallel connection and end up at 4 Ohms? Perhaps there's some advantage to it (1 Ohm) that I just dont understand, like damping factor...
Apogees? Wilsons? They drop to 1 ohm.... Maggies are 4 ohm and mostly resistive.
For a 1 ohm load, the resistors you'd need in the outputs to prevent any amp unbalance would negate any advantage to paralleling the amps.
As has been said, you could maybe design the amps from scratch to play nice with each other connected in parallel. But if you're going to that much trouble, there's a simpler way to do it requiring only one amp: double up on the current capability of the amp power supply, and double up on the number of, or current capability of, the output transistor / FET devices. I believe Krell subscribe to this design principle in their big monoblocks, hence their popularity with Wilson and Maggie owners.
As has been said, you could maybe design the amps from scratch to play nice with each other connected in parallel. But if you're going to that much trouble, there's a simpler way to do it requiring only one amp: double up on the current capability of the amp power supply, and double up on the number of, or current capability of, the output transistor / FET devices. I believe Krell subscribe to this design principle in their big monoblocks, hence their popularity with Wilson and Maggie owners.
(1) The only way to do it externally is by bridging one amp or using a different amp that has a higher voltage output.
(2) Okay, I hope your curiosity has been satisfied.
To summarize (HA!): Parallel amplifier outputs provide little to no practical advantage for the same speaker load. If you're designing something new, then sure, maybe. 🙂
I've had at least one commercial product in the shop that does this - a powered, bi-amped studio monitor that used a single 3886 for the tweet, and 3 of them, each with a low-value series output resistance, paralleled into a presumably low-impedance woof. Sorry I can't remember the brand (Event maybe?), but they were/are relatively well-regarded.
A more current (sorry) example is the miniDSP Harmony DSP 8x12 multichannel car audio amp. Unlike most car amps that have a bridge mode feature (double the output voltage), this unit allows 2 adjacent channels to be paralleled, providing the same output voltage as before but now with twice the available current. Of course this scheme is only effective if the speaker load impedance is halved, as described countless times above. As you might expect, this feature has caused no small amount of confusion in at least one car audio forum, with discussions very similar to this one!
When I built the current (final?) version of my home audio rig, I did kind of a variation of this (OK not really), using pairs of amp channels, inputs paralleled, to drive dual-voice-coil automotive woofers - one amp per voice coil. Each coil is 4 ohms, optimal loads for my power amps (block diagram attached).
To reiterate, the above examples start out on paper, not with an existing speaker design. Parallel amps don't seem to make much sense as a drop-in upgrade.
I've had at least one commercial product in the shop that does this - a powered, bi-amped studio monitor that used a single 3886 for the tweet, and 3 of them, each with a low-value series output resistance, paralleled into a presumably low-impedance woof. Sorry I can't remember the brand (Event maybe?), but they were/are relatively well-regarded.
A more current (sorry) example is the miniDSP Harmony DSP 8x12 multichannel car audio amp. Unlike most car amps that have a bridge mode feature (double the output voltage), this unit allows 2 adjacent channels to be paralleled, providing the same output voltage as before but now with twice the available current. Of course this scheme is only effective if the speaker load impedance is halved, as described countless times above. As you might expect, this feature has caused no small amount of confusion in at least one car audio forum, with discussions very similar to this one!
When I built the current (final?) version of my home audio rig, I did kind of a variation of this (OK not really), using pairs of amp channels, inputs paralleled, to drive dual-voice-coil automotive woofers - one amp per voice coil. Each coil is 4 ohms, optimal loads for my power amps (block diagram attached).
To reiterate, the above examples start out on paper, not with an existing speaker design. Parallel amps don't seem to make much sense as a drop-in upgrade.
i know - I'm no fun at parties either. 🙄
- Jim, a.k.a. "Debbie Downer"
- Jim, a.k.a. "Debbie Downer"
This is a long thread... I think we reached that point at around post 180 or so...
Imagine if I had proposed to use three amplifiers instead of two... would that increase the 3rd order harmonic distortion?
Last edited:
Why is anyone even entertaining this recalcitrant?
- Status
- Not open for further replies.