Can anyone recommend a DIY build with PCBs etc for something similar to ATC Amp Pack amps. The actives power amp specs are
Treble 50 Watts RMS
MID 100 watts RMS 16 ohms
Bass 200 Watts RMS 8 ohms (so mid and bass would be same power)
Apparently the originals are Mosfet Class A up to 2/3 of output power. The 3 amps share a power transformer. Is there anything out there for DIY? Or a sleeper amp that I can use or modify that has a similar spec? I would like to use at least as much power as the originals if not more.
Treble 50 Watts RMS
MID 100 watts RMS 16 ohms
Bass 200 Watts RMS 8 ohms (so mid and bass would be same power)
Apparently the originals are Mosfet Class A up to 2/3 of output power. The 3 amps share a power transformer. Is there anything out there for DIY? Or a sleeper amp that I can use or modify that has a similar spec? I would like to use at least as much power as the originals if not more.
Can anyone point me in the right direction
I have found the Quasi amps. 200w, 350,500. I have seen plenty of pages on these amps but no one using them with large studio type midfield monitors.
The ampacks don't seem to have a great deal of capacitance filtering and all thee amps share the transformer. Would that be separate windings? separate rectifiers and filters? There are 3 sets of 2 capacitors on the am packs suggesting that they each have their own filtering and possibly rectifiers.
Do all Class AB amps operate Class A up to a point? Did I read up to the point where their output current is twice the bias current?
I should be able to get some closeup photos of the ampacks. Would anyone be interested?
I have found the Quasi amps. 200w, 350,500. I have seen plenty of pages on these amps but no one using them with large studio type midfield monitors.
The ampacks don't seem to have a great deal of capacitance filtering and all thee amps share the transformer. Would that be separate windings? separate rectifiers and filters? There are 3 sets of 2 capacitors on the am packs suggesting that they each have their own filtering and possibly rectifiers.
Do all Class AB amps operate Class A up to a point? Did I read up to the point where their output current is twice the bias current?
I should be able to get some closeup photos of the ampacks. Would anyone be interested?
Is it the circuit board layout you want?
I have a component layout, transformer wiring, and schematic, but not the track design for the circuit board for an ATC AmpPack.
I have a component layout, transformer wiring, and schematic, but not the track design for the circuit board for an ATC AmpPack.
Last edited:
I doubt the amps were class A. 100w and 200w class A amps are huge home heaters.
There are plenty of good 50w, 100w, and 200w class AB amps in this forum. Check for amps by Apex. I bet he had one for every case there.
I wonder if anything special about ATC amp. Their active do sound good, but I think it is mostly due to well designed acive crossover, not the amp...
Hi XRK...
Personally I love to see strongly built metal boxes piled high, knowing that there is a ton of power inside them. I have used Naim amps, tube amps, and quite a few others like this and thoroughly enjoyed the result. When I tried an active system for the first time, it was a huge improvement on anything I'd had before. That was using a home made Elektor magazine design and 3 Naim power amps, which I then replaced with some Leak Stereo 20 valve (tube) amps.
Since then I've been using a more compact system...not active....but with very expensive Krell and Audio Research components.
But when I tried these active ATC amps for the first time, it was a bit of a revelation.
I can confirm that the ATC amps are VERY heavy plate amps. There are 3 amps, a 200w bass amp, a 100w midrange amp, and a 50w for the treble. The crossover is integral. The amps work at class A up to ⅔ of their output power, then switch to (I think...Class B).
They are HUGELY powerful, and have a MASSIVE heatsink on the rear. Don't underestimate these amps, they really are massive but sensitive beasts. At my normal listening levels, because of the size of the heatsink, they only get WARM. I have had many amps including a 211 tube amps, a Krell FPB200, an ATC integrated amp and an Audio Research amp which all got hotter.
I make no claims for ATC. I don't need to. But I AM enjoying their residence in my house.
Google of a few reviews of the ATC active speakers and you'll see that they are a pretty formidable and acclaimed species.
Why does the Library of Congress use ATC active speakers and turntables by Simon Yorke Designs when there are SO MANY alternatives available?
Have a look at some ATC users over the years: Pink Floyd, Roger Waters, Eric Johnson, Mark Knopfler, Calvin Harris, TomPetty, Rolling Stones, Jack White, Lou Reed, James Newton Howard, Diana Krall,Joe Satriani, Sting, Ronnie Scotts Club London, Royal Opera House London, Sydney Opera House, Walt Disney Concert Hall, BBC, CRC Canada, London Studios, Doug Sax, George Massenburg, Michael Bishop, Robert Miles, Mark Ronson, Bob Ludwig, James Guthrie, T-Bone Burnett, Ziggy Marley, The Front Stage at Nashville, recording studios across the world including loads in the USA such as Telarc, Sony SACD Studios, Backstage Nashville, Electric Lady Studios in NYC were using some recently, The London College of Music, Cambridge University, Tate Gallery, and the good old US Airforce as well. This must mean that if you use ATC gear, you're probably using something which is significantly better than many alternatives...and maybe something certainly one of the best options.
Last edited:
It doesn't matter how big the heatsink is, a class A amplifier will still be dissipating something like 2-4 times its nominal power rating as heat constantly. In stereo, that's around 2kW of heat and would soon become unbearable without generous air conditioning, forced draught etc. Your transformer might seem large but would be only be fraction of the size required for 350W of true class A amplification.
Krells generally dissipate a lot less heat than a basic class A amplifier because they use a variety of modulated bias control systems that track bias needs as they arise. If you find a Krell of similar total power to the ATC amplifiers is running so much hotter, your ATC amps, or at least the bass amp, is not actually class A.
I think that a total 350W of class A amps would have to be dynamically biased class AB in some form, to run as cool as you suggest. Blurb can be misleading - check the facts.
Krells generally dissipate a lot less heat than a basic class A amplifier because they use a variety of modulated bias control systems that track bias needs as they arise. If you find a Krell of similar total power to the ATC amplifiers is running so much hotter, your ATC amps, or at least the bass amp, is not actually class A.
I think that a total 350W of class A amps would have to be dynamically biased class AB in some form, to run as cool as you suggest. Blurb can be misleading - check the facts.
Hi Ian.....Thanks for your reply.
I bow to your greater experience in this matter....As you say, blurb can be misleading, I'm just quoting direct from ATC....
I bow to your greater experience in this matter....As you say, blurb can be misleading, I'm just quoting direct from ATC....
that's possible for the 50W treble amplifier.
The 100W mid amplifier will be a ClassAB
The 200W bass amplifier could be ClassAB or even a ClassB.
Both the mid and bass amplifiers will operate from the same voltages and may even share the same PSU.
The treble amplifier will operate at roughly half the voltage of the bigger amplifiers. This will almost certainly be from different secondary windings.
They are class AB amps and they run with a fairly high bias. I have the circuit diagrams somewhere and will work it out at some point. The heat sinks do get very warm!
The amps themselves are also a little different to the norm and utilise independent floating power supplies - hence the need for separate windings. The MOSFETs are not "grounded sources" as indicated in the ATC literature but are offset from ground via small source resistors. Nevertheless the amp output devices run in a "transconduction stage" configuration.
The ic input stage provides much of the open loop gain and drives the output stage directly. One unique feature of this amp design is that a local feedback path around the output devices (aligned via a trimmer) is used to null the differences of the N and P MOSFET devices. This is a very simple and very effective circuit that deserves a lot more praise!
The packaging of the output devices and power supply capacitors is also unique in my experience. A very neat, reliable and well-engineered package all-round...
Likewise, AndrewT, I bow to your greater knowledge. I am sure you must be right. Thank you for the information. Cheers, David
Quick correction to my previous post after checking... The ATC amp pack does not use source resistors and is therefore a genuinely grounded source amp. It also appears the MOSFETs are sandwiched between a ground plane and the grounded chassis plate - and presumably the device cases connected to source!
All three amp sections are biased with a maximum of about 1.1V at the gate. If we assume a MOSFET transconductance of 1A/V (Id/Vgs) - which is not a bad approximation for higher drain currents at least then:
For the HF, we have 1A which approximates to about 10Wrms in class A into the 8R tweeter;
For the MF, we have 1A which approximates to about 20Wrms in class A into the 16R midrange driver;
And for the LF we have two parallel MOSFETs which approximates to about 40Wrms in class A into the 8R woofers.
These figures are therefore some way short of those quoted. The peak powers appear to be justified, however.
BUT given that the maximum powers appear to be distributed according to some pink noise criterion, the lower powered sections would not be a fair comparison to adopt in comparing to a normal class AB amplifier driving a passive crossover. Accounting for the passive crossover (VA) losses, the amp pack would then appear equivalent to a class AB amplifier with some 80Wrms class A capability - or 160Wrms on music programme.
Using a "real world" analysis such as this appears to get us to the figures ATC quotes...?
Also to report is that the circuit diagram I have indicates second-order crossovers. If you are using these drivers my experience would be to always adopt fourth order crossovers. Presumably, however, the amp pack suffices perfectly well in many applications?
All three amp sections are biased with a maximum of about 1.1V at the gate. If we assume a MOSFET transconductance of 1A/V (Id/Vgs) - which is not a bad approximation for higher drain currents at least then:
For the HF, we have 1A which approximates to about 10Wrms in class A into the 8R tweeter;
For the MF, we have 1A which approximates to about 20Wrms in class A into the 16R midrange driver;
And for the LF we have two parallel MOSFETs which approximates to about 40Wrms in class A into the 8R woofers.
These figures are therefore some way short of those quoted. The peak powers appear to be justified, however.
BUT given that the maximum powers appear to be distributed according to some pink noise criterion, the lower powered sections would not be a fair comparison to adopt in comparing to a normal class AB amplifier driving a passive crossover. Accounting for the passive crossover (VA) losses, the amp pack would then appear equivalent to a class AB amplifier with some 80Wrms class A capability - or 160Wrms on music programme.
Using a "real world" analysis such as this appears to get us to the figures ATC quotes...?
Also to report is that the circuit diagram I have indicates second-order crossovers. If you are using these drivers my experience would be to always adopt fourth order crossovers. Presumably, however, the amp pack suffices perfectly well in many applications?
Last edited:
ATC's documentation says:
The amplifier pack consists of three grounded source MOSFET power amplifiers mounted on a common heatsink. (There are 12 Mosfets in each tri-amped unit)
Each amplifier has its own supply which floats at output signal level. The positive terminal of each drive unit is connected to the centre tap of its power supply. The driver circuitry and the crossover use a common 12-0-12 volt zener stabilised power supply. The Mosfet output devices are connected directly to the heatsink. All three amplifiers are class A/B with the middle and high frequency ones designed to stay in class A up to very high listening levels.
However, In the specification it says:
The crossovers in their 3 way amp packs cross over at 380Hz and 3.5kHz, and are 4th order, critically damped with phase compensation.
The circuit diagram I have for the Mk4 circuit (the latest version) definitely says on it '4th Order'.
Which version I have is not indicated but the quote "All three amplifiers are class A/B with the middle and high frequency ones designed to stay in class A up to very high listening levels" appears to be a valid description given the above calculations.
I would be grateful if you could post your copy - or just send me one...
The crossovers are definitely shown in my old (?) version as second order however, at 380Hz and 3.8kHz (not 3.5kHz). The "phase compensation" refers to time alignment for the different driver to listener delays implemented as a first-order all-pass as first shown by Linkwitz.
Hi Soundbloke. That is VERY interesting. ATC is a fabulous company. Absolutely superb products.
Re: the diagram, No problem. PM me your email address....I can't seem to PM you...you must have set it so that annoying people like me can't pester you!
Re: the diagram, No problem. PM me your email address....I can't seem to PM you...you must have set it so that annoying people like me can't pester you!
Thank you vintageaxeman for the Mk4 diagrams. I can confirm this later version does now adopt fourth order filters. The revised biasing also indicates approximately 20W class A working for the mid and high frequencies but as little as 2W for the low frequencies.
However, remember that the class AB performance is better than expected because of the unique (?) local feedback applied around the output devices. But it still seems a rather odd arrangement compared to that in the earlier versions...
However, remember that the class AB performance is better than expected because of the unique (?) local feedback applied around the output devices. But it still seems a rather odd arrangement compared to that in the earlier versions...
- Status
- Not open for further replies.