ICEPower 125ASX2 Class D Amplifier Review

[hyperknot]

I think this review is very interesting, as it might finally answer the big question asked online about ICEPower modules: that is, are the cheap Aliexpress / eBay modules the same as the more expensive genuine ones from known US / European resellers. The price difference can be quite a lot:

Chinese seller: 116 USD
Official European reseller: 245 USD (187 GBP)

The previously tested Yamaha WXA-50 contains the exact same module, and did much better on the review.

@amirm is it possible to overlap the two measurements on each other?

 

[Biblob]

let's go for 110dB Phon (90dB/W/m speaker)

View attachment 42894

I have trouble interpreting this graph. Could you explain it to me a bit?

 

[hyperknot]

It's like a totally different amp, just because of Chinese vs. genuine modules:

 

[Matias]

It's like a totally different amp, just because of Chinese vs. genuine modules:

Probably was measured at different temperatures, as they influence a lot in distortion.
Only case and cables are Ghent DIY, the module is the same.

 

[Deleted member 65]

I am serious and no, it is not. Most of us can't even hear 20 kHz let alone its harmonic distortion that is included in those graphs.

So, judgement from a hearing perspective as opposed to engineering/technical.

 

[Matias]

So, judgement from a hearing perspective as opposed to engineering/technical.

Good engineering practice considers hearing perspective for balancing compromises.

 

[Deleted member 65]

Good engineering practice considers hearing perspective for balancing compromises.

True & agree, splitting the two it would look like something like this:

Technical performance = Fail
Hearing = Pass
Verdict = Ok, let’s give it a pass ... ;-)

 

[CerealKiller]

It's like a totally different amp, just because of Chinese vs. genuine modules:

The suposedly chinese modules replicates the data sheets very well. Yamaha amp measurements does not. It actually does better! What might be the cause?

Edit: does better at full vol butt starts far worse at lower levels, i'm I correcto?

 

[jtwrace]

The case is from Ghent Audio who is fantastic to deal with.

 

[CerealKiller]

Reading the thread on the Yamaha WXA-50 Review, that amp comes with some pre installed heatsinks with Thermal pads to Transfer heat from the IC to the top of the chasis. Will that be suficient for making diference and improvement?

Guess Will never know...

 

[solderdude]

I have trouble interpreting this graph. Could you explain it to me a bit?

Given the S/N ratio I would say the 'grass' would only be there when a signal is there.
When looking at a 90dB/W/m speaker one would say the grass falls in the audible range...
BUT and here comes the more difficult part.. whether or not that may become audible depends on the music content and how loud other frequencies are there and are masking the 'grass'.
In other words with some particular music one may or may not hear something that should not be there.
And now for the final... this is measured with a resistive load. It may look worse on a more complex load.
We don't know... so whether one calls this a pass or fail cannot be concluded from this particular graph.
Also not because speakers also add harmonics and stuff.

I give up ... draw your own conclusions dammit.

 

[ta240]

....And now for the final... this is measured with a resistive load. It may look worse on a more complex load.
We don't know... so whether one calls this a pass or fail cannot be concluded from this particular graph.
Also not because speakers also add harmonics and stuff.

I sometimes wonder about the ethics of branding items as good or bad with non-real world tests. As entertainment the graphs are fun but these posts can gain or lose a lot of sales for a company.
It would seem possible that an amp that got a recommendation based on a resistive load test amp could perform worse than a not recommended amp when both are presented with a complex load.

Good engineering practice considers hearing perspective for balancing compromises.

Was it a double blind listening test? Because I've read somewhere that if it isn't double blind then it is of no use.

Otherwise preference by some to class A or AB amps due to compromises in design based on sighted listening is being dismissed and preference by others to class D amps in spite of compromises in design is being accepted based on sighted listening .

 

[solderdude]

double blind is not specifically needed.
Blind is enough when all the needed controls are in place.

I don't think any brand will be loosing sales over what is said here. In some cases I would hope it is the case but the reality is and always will be:
The vast majority does not care about measured results. The vast, vast majority of people cannot draw any valid conclusions based on a certain set of measurements.
Nope.. the vast majority of buyers simply buy based on recommendations on subjective sites/magazines because they WANT to believe and their ears tell them they are right.

In some rare cases a few manufacturers will scratch their heads and possibly have another look at their devices. The world keeps on turning.

Does that mean measurements are worthless ? Hell no.
Independent verification is important and those understanding what and how certain things are measured and what audible consequences could possibly be will appreciate the measurements posted here and discuss them.

 

[SEKLEM]

I think this review is very interesting, as it might finally answer the big question asked online about ICEPower modules: that is, are the cheap Aliexpress / eBay modules the same as the more expensive genuine ones from known US / European resellers. The price difference can be quite a lot:

Chinese seller: 116 USD
Official European reseller: 245 USD (187 GBP)

The previously tested Yamaha WXA-50 contains the exact same module, and did much better on the review.

@amirm is it possible to overlap the two measurements on each other?

The stuff coming direct from China at cut rates is almost undoubtedly an inferior product of questionable origin (factory seconds that didn't pass inspection or worse, flat out knockoff). I may not know where to look, but I cannot find a source for the 125asx2 outside of Chinese run outfits or eBay.

 

[Matias]

ICEpower modules are manufactured in the same factory regardless where they are sold. Performance should be similar.

 

[SEKLEM]

ICEpower modules are manufactured in the same factory regardless where they are sold. Performance should be similar.

It's not possible that somewhere in China someone is copying them? It's not possible that the factory that makes the genuine ones has units that don't meet spec that they then sell at discount through 3rd party sellers?

 

[CerealKiller]

The stuff coming direct from China at cut rates is almost undoubtedly an inferior product of questionable origin (factory seconds that didn't pass inspection or worse, flat out knockoff). I may not know where to look, but I cannot find a source for the 125asx2 outside of Chinese run outfits or eBay.

What about the ones for sell on partsexpress?

 

[Deleted member 65]

Swedish AES has finalized build of their F/E device. To be used as a device supporting Swedish AES method of evaluating/testing amplifiers, DAC’s etc. Supports both A/B listening as well as the F/E (Before/After) method. Levelmatching within 0.05 dB.

Edit: F/E device is controlled by computer thru an Arduino box.

 

[gattaca]

ICEpower modules are manufactured in the same factory regardless where they are sold. Performance should be similar.

They might be manufactured in the same factory. The leap is assuming they use the same components with the same specs from reliable suppliers for batch A, B, C. That is likely not the case unless they building to some customer's spec that calls out every component, spec in the BOM. They might buy cheaper capacitors from supplier X for batch A but then use caps from supplier Y for batch B b/c they demand using better graded, tighter tolerances, higher voltages as Amir often points out. Some of these manufacturers will do anything to save a 1/2 US cent per component. They will also sell final goods which may not pass the customer's spec as "factory seconds" or even label them as good parts, which they may be, but they they just didn't meet customer X's specs. Remember, any $$ for a made piece is better than no $$ in the recycling pile. Peace.

 

[Trdat]

This is a review and detailed measurements of a DIY amplifier based on ICEPower 125ASX2 module in Ghentaudio case. It is on kind loan from a member. The owner says the total cost is only US $220. I can't find the amp at the price he bought it so maybe it costs a few dollars more but still, it is a good price

Isn't this sold for $650 at GhentAudio?