Rail Voltage for one pair IRFP150/9150

Seen a review and under test at around 0.8 %

The amp did 84 Watts @ 8 ohms and 128 @ 4 ohms

Basic mosfet like this under load wont swing anywhere close to the rail.

So the safety margin for voltage, is very close but " possible "
But as mentioned when heat is involved that all changes.

Shocking as it is.
Almost amazing and scary what you can get away with.

This design really really hopes it sees a very low duty cycle around 15 to 30 watts.

Which is likely , and only how they got away with it.
Average user could drive most small to med " home" speakers
to full distortion no problem.

far as going over 30 watt duty cycle
then likely yes the 140 or 150 series mosfet.
just needs to jump up to the 240 series.

and the heatsink would need to be larger.
 
My point of view, one of my integrated amps is the Atoll IN80SE with "only" one 340VA trafo, I can tell you that even before my modifications (which include new larger one piece heatsink) this amplifier had no problem powering my Dynaudio Focus 110 (84db sensitivity) and Dynaudio Focus 140 for a long period of time with more than average SPL.
And if you are interested....I modified it just for fun.
To be honest to my ears it sounds better (clearer) than my Cyrus 2, previous Exposure 3010S(more detailed and less bass heavy) and alot better than the highly praised Naim Nait XS2 (better soundstage...actually better in almost any aspect) that I had before, I can totally understand why pepole talk trash about the Atoll build quality or eletctronics design but I like it's sound signature and I never felt any shortage in power.
 
My point of view, one of my integrated amps is the Atoll IN80SE with "only" one 340VA trafo, I can tell you that even before my modifications (which include new larger one piece heatsink) this amplifier had no problem powering my Dynaudio Focus 110 (84db sensitivity) and Dynaudio Focus 140 for a long period of time with more than average SPL.
And if you are interested....I modified it just for fun.
To be honest to my ears it sounds better (clearer) than my Cyrus 2, previous Exposure 3010S(more detailed and less bass heavy) and alot better than the highly praised Naim Nait XS2 (better soundstage...actually better in almost any aspect) that I had before, I can totally understand why pepole talk trash about the Atoll build quality or eletctronics design but I like it's sound signature and I never felt any shortage in power.
Sure industrial mosfets are very musical .... i trust you on that .... obviously when you compare this to a cyrus doesn't makes much sense sinc cyrus with 3 mhz transistors should sound more ore less the same NAIM also is more or less the same ...
i also trust that your main listening is classical rock blues and things that only find interest in the middle area ....
only music has afew other things except middle area
 
Sure industrial mosfets are very musical .... i trust you on that .... obviously when you compare this to a cyrus doesn't makes much sense sinc cyrus with 3 mhz transistors should sound more ore less the same NAIM also is more or less the same ...
i also trust that your main listening is classical rock blues and things that only find interest in the middle area ....
only music has afew other things except middle area
I actually listen to a wide variety of music genre form metal and rock to blues, "unplugged", electronic and epic.
Now I'm not saying it's the best amplifier I ever had but it's not as bad as it looks "under the hood".
 
I actually listen to a wide variety of music genre form metal and rock to blues, "unplugged", electronic and epic.
Now I'm not saying it's the best amplifier I ever had but it's not as bad as it looks "under the hood".
Read what is written in the other Atoll discusion The modification you made produces overcompensation and bias drops enough to produce distortion
i cannot take your opinion as valid You have altered the performance of the amplifier and your amp at some point works at clean class B
 
Read what is written in the other Atoll discusion The modification you made produces overcompensation and bias drops enough to produce distortion
i cannot take your opinion as valid You have altered the performance of the amplifier and your amp at some point works at clean class B
As mentioned in my first post my impressions about this amp were before I modified it so every thing you wrote is irrelevante.
I'm not here to "defend" Atoll and maybe their bad engineering, just to share my impressions.
 
As mentioned in my first post my impressions about this amp were before I modified it so every thing you wrote is irrelevante.
I'm not here to "defend" Atoll and maybe their bad engineering, just to share my impressions.
A) we dont take impressions as an instrument to our work ...thanks for sharing but it doesnt help match
we use scopes /genarators and spectrum analysers that cost thousands of Euro after we learned how to operate them
we use sophisticated software and tools to monitor voltage bias and temperature
b) as about the SOA it is enough to read datasheet of the manufacturer ( for transistors and heatsinks accordingly ) and one calculator to unserstand if something is wrong, there second grade maths is enough for that
c) we might descuss here what we think is wrong with this amplifier but if you want the opinion of an Electronics Engineer that has 40 years on the bench doing only audio, your approach of modification solves one problem and creates others
 
With a pair of IPFP240/9240 running at +/- 56Vdc, I'm getting 100W/8 ohms, 150W/4 ohms.
So running the IRFP510/9510 at +/-60Vdc will be slightly higher.
Whether it's advisable to run these mosfets at +/- 60Vdc is another matter.

What I find perplexing are the heatsinks. They are too small for a Class AB amp.
There is a real danger of the mosfets overheating when subjected to hard use.
 
  • Like
Reactions: JMFahey
Hello,
if there no additional rails for the VAS stage then the output swing is severals volts away from the rails. Plus additional sag under load and the devices may survive the voltage.
But the heatsinks are really small - plus the fact that the Hexfets like quite some bias which will be hard with this tiny heatsinks.
For me its also not a good design even if there is a overtemp-shut-off and a sophisticated SOA-protection.
For 100W/8R I actually use 8 pairs of 140/9140 in balanced configuration at +/- 25V...

Edit: Reviews in the web show high distortion / intermodulation and 12W consumption idling. Even if all is consumed in the output stage this will be a bias of 100mA. In praxis more going into direction of 0mA if you take into account all the input circuitry...

Regards, J-C
 
Last edited:
With a pair of IPFP240/9240 running at +/- 56Vdc, I'm getting 100W/8 ohms, 150W/4 ohms.
So running the IRFP510/9510 at +/-60Vdc will be slightly higher.
Whether it's advisable to run these mosfets at +/- 60Vdc is another matter.

What I find perplexing are the heatsinks. They are too small for a Class AB amp.
There is a real danger of the mosfets overheating when subjected to hard use.
Hello Michael Good to hear from you old friend !
 
Hello,
if there no additional rails for the VAS stage then the output swing is severals volts away from the rails. Plus additional sag under load and the devices may survive the voltage.
But the heatsinks are really small - plus the fact that the Hexfets like quite some bias which will be hard with this tiny heatsinks.
For me its also not a good design even if there is a overtemp-shut-off and a sophisticated SOA-protection.
For 100W/8R I actually use 8 pairs of 140/9140 in balanced configuration at +/- 25V...

Regards, J-C
No additional rails
No sag or swing with 640 VA power transformers behind just a couple of mosfet
No Vi limiter Just a string of zeners
no thermal shut off for this model ( they claim thta their new model has a function to turn the volume down if ubnormal temp is detected )
no overload detection
 
I also guess they got lucky because their market is made out of Home users, who don´t crank them, at all, maybe middle aged users up who listen to "mild" Music, never ever Heavy Metal, Punk, Reggae, loud Tecno, etc.
I bet that if they take the "speaker voltage test" suggested elsewhere the result will be quite low.
 
  • Like
Reactions: east electronics
If you have dynamic musik then the average power is low compared to the peak power.
And I cant see it perfectly on the pic, but I suppose the heatsinks also use the botton plate of the amp.

As a DIY developer I always try to learn learn from commercial products. Seen from this side its an interesting product. They are on the market, so it cant be too wrong. To chose big toroids but tiny heatsinks seems odd to me, what are the goals behind this decision?
At 12Watt idling and 10Watt standby the bias seems to be close to 0. Then there is no complicated bias schematic neccessary. 0 stays at 0, only the distortions change over temperature. And why are Mosfets chosen with the given cooling / biasing and no BJT CFP stage?
To me it looks like the amp has its customers because it has an specific sound. Perhaps you can learn from this that you dont have to be perfect, its sufficient if you have an "outstanding" product which may get its buyers because of special properties.
Perhaps a lot of customers prefer its sound over a well designed ppm distortion amp.

Regards, Jean-Claude
 
If you have dynamic musik then the average power is low compared to the peak power.
And I cant see it perfectly on the pic, but I suppose the heatsinks also use the botton plate of the amp.

As a DIY developer I always try to learn learn from commercial products. Seen from this side its an interesting product. They are on the market, so it cant be too wrong. To chose big toroids but tiny heatsinks seems odd to me, what are the goals behind this decision?
At 12Watt idling and 10Watt standby the bias seems to be close to 0. Then there is no complicated bias schematic neccessary. 0 stays at 0, only the distortions change over temperature. And why are Mosfets chosen with the given cooling / biasing and no BJT CFP stage?
To me it looks like the amp has its customers because it has an specific sound. Perhaps you can learn from this that you dont have to be perfect, its sufficient if you have an "outstanding" product which may get its buyers because of special properties.
Perhaps a lot of customers prefer its sound over a well designed ppm distortion amp.

Regards, Jean-Claude
Jesus it seems that you dont get it .... I am sorry but you are wrong top to bottom ....
1--One company that chooses to overdisgn one amplifier with 680VA of trafos must first learn to use them properly IE 😀O NOT conect them in parallel it only causes trouble,Just use the exact same cost and material one transformer per rail and 2 rectifiers you spend no money more and you have far cleaner PSU and no problems at all .
2--No heat sinks are not related with the enclosure for better heat relief , and they cant be conected since they have rails on them !!!!
Factory used this style because they noticed that with insulators transistor is for example at 65 degrees and heatsink is at 62 degrees . Obviously this has been a considaration for many manufactures its a known problem and the only thing that you have to do is to choose a better quality material ...Understand that none is perfect so you choose the best for your application .
2A-- then if you solve the above problem then of course you can use a better heatsink ( notice also that even with this style there is still some room if you redesign the board for larger heat sinks and you may have to produce them accordin to the efficiency you need, but they didn't !!! what they have is commonly available on the market the went with the easiest solution .
3--Choosing a thermal compensation on this style ( is written above again and again ) indeed is a bit more audiophile but creates trouble for the end user since this machine at idle might work almost in class B will work ok if and when its warmed and even better when its hot but if it gets too hot there is almost nothing there to stop it . For the end user this will be confusing since this machine will have a huge variation to its sonic signature depending on temparature ....Huge variation next to any amplifier that controls this issue better
4--The sound of industrial mosfets when cold is simply horrible and even worst if driven by clearly outdated transistors of 3MHZ ( vas on these amps is Tip31-32) with the astronomical ft of 3Mhz In reality of you look in the harmonics and distortion of this amplifier in a spectrum analyser as a picture is familiar with distortions often recorded in rock music alboums ....Meaning audience that is in rock music will love the sound of them in terms of presence and brillliance of middle area while by all means this sound has absolutely nothing to do with a Japanese amplifier which sound edgy next to Atoll .

Conlcusion
A) you cannot overlook SOA
B) you may not release in the market a machine that has habbits and behaviors, comercial consumer machines should be stable for the user regarding performance and sonic signature
C) Atoll CANNOT re write the handbook of class AB amplification to its own taste
D) you cannot predict dive of impedance on any given speaker of the market , you cannot predict crest factor of music on any given musical product of the market That is also why your machine has to be idiot proof up to a point .
the result is that these people had grate luck on the audience so far and listeners probably do a mild use of the product
so the got away with it beliving that they invented the wheel from scratch

That is my opinion