Sabaj A5 and A8 announced

[mdsimon2]

Done with my initial testing, overall very nice. I just wish it didn't have that dumb volume knob. The only surprise in testing is that the gain is the same in stereo mode and bridged mode which is nice as it allows you to use up to 3.2 V input voltage in bridged mode without issue.

All of the following measurements are done at roughly 24 dB gain for best performance. Noise performance is especially impressive as it is just a bit better than my Hypex NC252MP although that amp has higher gain and more power. I realized I didn't state it in my original measurement post but the bandwidth for the THD+N measurements is 20 Hz to 22 kHz. Also for the 4 ohm stereo measurements I can only drive one channel due to dummy load limitations (I am using two 8 ohm resistors in parallel). Here is a brief summary, again the 5 W THD+N measurements will be better in reality than what I show here due to DAC / ADC limitations. Overall seems very consistent with the ICEpower datasheet, as such I would expect THD+N to fall apart a bit at high frequencies.

8 ohm stereo
max power = 70 W / ch
5 W THD+N = - 88 dB

4 ohm stereo
max power = 130 W / ch
5 W THD+N = -86 dB

8 ohm bridged
max power = 250 W
5 W THD+N = -88 dB

4 ohm bridged
max power = 400 W
5 W THD+N = -86 dB

8 ohm stereo frequency response

8 ohm stereo residual noise - 129 uV (includes DAC, ADC and amplifier, see here for reference)

4 ohm stereo frequency response

4 ohm stereo residual noise - 126 uV (includes DAC, ADC and amplifier, see here for reference)

4 ohm stereo 1 kHz power vs THD+N - 1% THD+N power = 130 W, 5 W THD+N = -86 dB

8 ohm bridged frequency response

8 ohm bridged residual noise - 155 uV residual noise (includes DAC, ADC and amplifier, see here for reference)

8 ohm bridged 1 kHz power vs THD+N - 1% THD+N power = 250 W, 5 W THD+N = - 88 dB

4 ohm bridged frequency response

4 ohm bridged residual noise - 156 uV (includes DAC, ADC and amplifier, see here for reference)

4 ohm bridged 1 kHz power vs THD+N - 1% THD+N = 400 W, 5 W THD + N = - 86 dB

Michael

 

[Talisman]

Michael I am speechless, an outstanding job, you are our little private Amirm!
Performance in line with what was declared, incredible for the price, tons of clean power. I understand that bridge mode allows for a higher voltage input signal, would you recommend using them as power amplifiers with the knob at 12 o'clock?
Thanks again for your fantastic work

 

[TheBatsEar]

Done with my initial testing, overall very nice. I just wish it didn't have that dumb volume knob. The only surprise in testing is that the gain is the same in stereo mode and bridged mode which is nice as it allows you to use up to 3.2 V input voltage in bridged mode without issue.

All of the following measurements are done at roughly 24 dB gain for best performance. Noise performance is especially impressive as it is just a bit better than my Hypex NC252MP although that amp has higher gain and more power. I realized I didn't state it in my original measurement post but the bandwidth for the THD+N measurements is 20 Hz to 22 kHz. Also for the 4 ohm stereo measurements I can only drive one channel due to dummy load limitations (I am using two 8 ohm resistors in parallel). Here is a brief summary, again the 5 W THD+N measurements will be better in reality than what I show here due to DAC / ADC limitations. Overall seems very consistent with the ICEpower datasheet, as such I would expect THD+N to fall apart a bit at high frequencies.

8 ohm stereo
max power = 70 W / ch
5 W THD+N = - 88 dB

4 ohm stereo
max power = 130 W / ch
5 W THD+N = -86 dB

8 ohm bridged
max power = 250 W
5 W THD+N = -88 dB

4 ohm bridged
max power = 400 W
5 W THD+N = -86 dB

8 ohm stereo frequency response
View attachment 205255

8 ohm stereo residual noise - 129 uV (includes DAC, ADC and amplifier, see here for reference)
View attachment 205256

4 ohm stereo frequency response
View attachment 205263

4 ohm stereo residual noise - 126 uV (includes DAC, ADC and amplifier, see here for reference)
View attachment 205257

4 ohm stereo 1 kHz power vs THD+N - 1% THD+N power = 130 W, 5 W THD+N = -86 dB
View attachment 205266

8 ohm bridged frequency response
View attachment 205264

8 ohm bridged residual noise - 155 uV residual noise (includes DAC, ADC and amplifier, see here for reference)
View attachment 205262

8 ohm bridged 1 kHz power vs THD+N - 1% THD+N power = 250 W, 5 W THD+N = - 88 dB
View attachment 205268

4 ohm bridged frequency response
View attachment 205265

4 ohm bridged residual noise - 156 uV (includes DAC, ADC and amplifier, see here for reference)
View attachment 205260

4 ohm bridged 1 kHz power vs THD+N - 1% THD+N = 400 W, 5 W THD + N = - 86 dB
View attachment 205269

Michael

Those ICE modules are really sweet.

As for the volume knob, i think it's useful for the people that buy this one as a desktop amp.

Can you do a 2 ohm measurement to show they don't fall apart? Quite some speakers dip below 4 ohms and you don't want it to protect or release magic smokes.

 

[mdsimon2]

Michael I am speechless, an outstanding job, you are our little private Amirm!
Performance in line with what was declared, incredible for the price, tons of clean power. I understand that bridge mode allows for a higher voltage input signal, would you recommend using them as power amplifiers with the knob at 12 o'clock?
Thanks again for your fantastic work

No, you do not want the volume knob at 12 o'clock because that hurts noise performance. I recommend setting the gain such that it is about 24 dB. With nothing connected play a 0.5 V test tone and use a DMM to measure the amplifier output voltage and set the knob so it is around 8 V. This will allow you to drive the amp to full power at the best noise performance.

EDIT: If you are using this on its own I would use the volume knob as that will truly give you the best noise performance. If you are using it like a power amp set the knob like I described above.

Michael

 

[mdsimon2]

Those ICE modules are really sweet.

As for the volume knob, i think it's useful for the people that buy this one as a desktop amp.

Can you do a 2 ohm measurement to show they don't fall apart? Quite some speakers dip below 4 ohms and you don't want it to protect or release magic smokes.

I don't have the resistors to do 2 ohm measurements. 2 ohm stereo should be similar to 4 ohm bridged although ICEpower does spec a minimum impedance of 3 ohm.

Michael

 

[TheBatsEar]

2 ohm stereo should be similar to 4 ohm bridged although ICEpower does spec a minimum impedance of 3 ohm.

Well, if they say 3 ohms it's fine and robust enough.

 

[mario_rouge]

Done with my initial testing, overall very nice. I just wish it didn't have that dumb volume knob. The only surprise in testing is that the gain is the same in stereo mode and bridged mode which is nice as it allows you to use up to 3.2 V input voltage in bridged mode without issue.

All of the following measurements are done at roughly 24 dB gain for best performance. Noise performance is especially impressive as it is just a bit better than my Hypex NC252MP although that amp has higher gain and more power. I realized I didn't state it in my original measurement post but the bandwidth for the THD+N measurements is 20 Hz to 22 kHz. Also for the 4 ohm stereo measurements I can only drive one channel due to dummy load limitations (I am using two 8 ohm resistors in parallel). Here is a brief summary, again the 5 W THD+N measurements will be better in reality than what I show here due to DAC / ADC limitations. Overall seems very consistent with the ICEpower datasheet, as such I would expect THD+N to fall apart a bit at high frequencies.

8 ohm stereo
max power = 70 W / ch
5 W THD+N = - 88 dB

4 ohm stereo
max power = 130 W / ch
5 W THD+N = -86 dB

8 ohm bridged
max power = 250 W
5 W THD+N = -88 dB

4 ohm bridged
max power = 400 W
5 W THD+N = -86 dB

8 ohm stereo frequency response
View attachment 205255

8 ohm stereo residual noise - 129 uV (includes DAC, ADC and amplifier, see here for reference)
View attachment 205256

4 ohm stereo frequency response
View attachment 205263

4 ohm stereo residual noise - 126 uV (includes DAC, ADC and amplifier, see here for reference)
View attachment 205257

4 ohm stereo 1 kHz power vs THD+N - 1% THD+N power = 130 W, 5 W THD+N = -86 dB
View attachment 205266

8 ohm bridged frequency response
View attachment 205264

8 ohm bridged residual noise - 155 uV residual noise (includes DAC, ADC and amplifier, see here for reference)
View attachment 205262

8 ohm bridged 1 kHz power vs THD+N - 1% THD+N power = 250 W, 5 W THD+N = - 88 dB
View attachment 205268

4 ohm bridged frequency response
View attachment 205265

4 ohm bridged residual noise - 156 uV (includes DAC, ADC and amplifier, see here for reference)
View attachment 205260

4 ohm bridged 1 kHz power vs THD+N - 1% THD+N = 400 W, 5 W THD + N = - 86 dB
View attachment 205269

Michael

in the end could you give us a final judgment on the sound quality of this amplifier? I think you could post this review on the forum because I find it very professional I think Amir would promote it among the approved reviews!

 

[mdsimon2]

in the end could you give us a final judgment on the sound quality of this amplifier? I think you could post this review on the forum because I find it very professional I think Amir would promote it among the approved reviews!

Eh, not sure I would call it a professional review, to me it was a quick check to make sure the volume control didn't screw up the base performance of the 125ASX2 module (which is in at least two other amps reviewed on ASR). I've done a bit of listening with some old test speakers and sounds fine to me compared to my usual Hypex NC252MP.

Michael

 

[TheBatsEar]

Eh, not sure I would call it a professional review, to me it was a quick check to make sure the volume control didn't screw up the base performance of the 125ASX2 module (which is in at least two other amps reviewed on ASR). I've done a bit of listening with some old test speakers and sounds fine to me compared to my usual Hypex NC252MP.

View attachment 205280

Michael

Certainly looks nice.
In some pictures it was white, which looked even better and a bit different to the usual black, silver or champagne.
Only the font sucks. I feel a sans font would be more serious.

 

[mdsimon2]

Certainly looks nice.
In some pictures it was white, which looked even better and a bit different to the usual black, silver or champagne.
Only the font sucks. I feel a sans font would be more serious.

Yeah, not a huge fan of the looks and I agree the font is terrible. It really is amazingly small though.

I also really hate speaker binding posts that are stacked vertically but again I think that is a function of its size and how they built their RCA / XLR input board.

Also can’t get over how inexpensive this is, if you bought a carton of 20 modules at $143/ea and added a Ghent case for $100 you are looking at pretty much the same price but you don’t have XLR input or a stereo / bridged switch or a volume knob. Although it sounds like these are also being discontinued because of how inexpensive they are compared to production costs.

Michael

 

[Enkay25]

On a reflective note, I am liking Sabaj more……and feels the products does not get credits much like Topping. Hope @peterwen sends more products to @amirm to tests.

 

[mdsimon2]

I wanted to try and use my Cosmos ADC for measuring the Sabaj A8 as it is much lower noise than the MOTU M4 but it's low input impedance causes too much signal attenuation, at least with my voltage divider setup for amp measurement. As a result I tried to use a Neurochrome universal buffer in between the amp output voltage divider and the Cosmos and that worked very well. I am still limited as I am using the MOTU M4 as a DAC and when you attenuate digitally you lose SNR which limits THD+N measurements at lower levels (although representative of real world system performance).

Here are a few more measurements for 5 W in to 4 ohm in bridged mode at various gain settings, as expected lower gain results in better THD+N and the measurements are better than using the MOTU M4 as an ADC. The one thing that worries me a bit is the high frequency junk just outside of 20 kHz. Although this is outside of the audible range I certainly worry about modulation back down to the audible range and I wonder if that is part of the reason these modules have rising distortion at higher frequencies.

13 dB gain (1 V input), -92 dB THD+N

24 dB gain (0.288 V input), -90 dB THD+N

29 dB gain (0.158 V input), -86 dB THD+N

38 dB (max) gain (0.055 V input), -79 dB THD+N

Michael

 

[Talisman]

I wanted to try and use my Cosmos ADC for measuring the Sabaj A8 as it is much lower noise than the MOTU M4 but it's low input impedance causes too much signal attenuation, at least with my voltage divider setup for amp measurement. As a result I tried to use a Neurochrome universal buffer in between the amp output voltage divider and the Cosmos and that worked very well. I am still limited as I am using the MOTU M4 as a DAC and when you attenuate digitally you lose SNR which limits THD+N measurements at lower levels (although representative of real world system performance).

Here are a few more measurements for 5 W in to 4 ohm in bridged mode at various gain settings, as expected lower gain results in better THD+N and the measurements are better than using the MOTU M4 as an ADC. The one thing that worries me a bit is the high frequency junk just outside of 20 kHz. Although this is outside of the audible range I certainly worry about modulation back down to the audible range and I wonder if that is part of the reason these modules have rising distortion at higher frequencies.

13 dB gain (1 V input), -92 dB THD+N
View attachment 205484

24 dB gain (0.288 V input), -90 dB THD+N
View attachment 205485

29 dB gain (0.158 V input), -86 dB THD+N
View attachment 205486

38 dB (max) gain (0.055 V input), -79 dB THD+N
View attachment 205487

Michael

In practice, the lower the gain is, the more THD + N you get.
However, even at the maximum gain level there is a decent 79db which is still the average of the amplifiers tested by amirm, and basically inaudible (if you don't do things like hook your ear to the tweeter at maximum volume without signal)
I had noticed the increase in high frequency distortion already in the review of the module made by Amirm, considering that I already hear almost nothing above 17kHz, should I worry about this for some reason?

 

[peterwen]

Due to the limited quantity purchased by ICEpower, there are not many A8s produced, and only 7 units are left for sale now.

Sabaj A8 Icepower 125w*125w power amplifier

Sabaj A8 Icepower 125w*125w power amplifier

hifi-express.com

It is expected to be removed the listing by the middle of the month

 

[pickyAudiophile]

Due to the limited quantity purchased by ICEpower, there are not many A8s produced, and only 7 units are left for sale now.

Sabaj A8 Icepower 125w*125w power amplifier

Sabaj A8 Icepower 125w*125w power amplifier

hifi-express.com

It is expected to be removed the listing by the middle of the month

Oooh, for the moment three pieces left only. Sell like real hot stuff. BTW Can you give us a rough estimate about the "A10a 2022" and "A20a 2022" release dates?

 

[mdsimon2]

In practice, the lower the gain is, the more THD + N you get.
However, even at the maximum gain level there is a decent 79db which is still the average of the amplifiers tested by amirm, and basically inaudible (if you don't do things like hook your ear to the tweeter at maximum volume without signal)
I had noticed the increase in high frequency distortion already in the review of the module made by Amirm, considering that I already hear almost nothing above 17kHz, should I worry about this for some reason?

The worry about high frequency distortion is intermodulation down in to audible frequencies. The common argument against this being a concern is that levels are usually rather low at high frequencies so in practice the distortion levels will be low.

However in a case like the 125ASX2 where you have high frequency junk always present regardless of level I think it is a bit more of a concern. One thing to note with the ICEpower datasheet measurements is that they are using an AES17 filter which filters all of that junk out. I do agree that Amir's 45 kHz bandwidth power sweeps by frequency do a good job of characterizing the behavior which is definitely not as nice as something like a Hypex NC252MP (which itself is not great compared to a good class AB amp).

I should say I am not a huge believer in amplifiers having a huge impact on sound quality and I am very interested in doing some blind level matched testing between the 125ASX2 and a Hypex NC252MP to see if I can identify any differences. If they exist my bet is that they are very subtle.

Michael

 

[peterwen]

Specifications
Inputs .................................................................................................RCA / XLR
THD+N ..................................................................................................0.005%
SNR ..........................................................................................................103dB
Channel separation .....................................................................................90dB
low pass cutoff frequency ...................................................................70~200Hz
Input Sensitivity / Input Impedance ............................................. 430mV / 22kΩ
Output Power
Stereo................................................................110W x 2(4Ω) / 66W x 2(8Ω)
BTL..........................................................................................225W x 1(4~8Ω)
Power Consumption ....................................................................................40W
Size ...........................................................................200X187X59mm (WxHxD)
Weight ........................................................................................1480g/ 3.26Lbs

Here is the info on the new 2022 A10a.
If you're interested in this, click the file.

 

[Enkay25]

Specifications
Inputs .................................................................................................RCA / XLR
THD+N ..................................................................................................0.005%
SNR ..........................................................................................................103dB
Channel separation .....................................................................................90dB
low pass cutoff frequency ...................................................................70~200Hz
Input Sensitivity / Input Impedance ............................................. 430mV / 22kΩ
Output Power
Stereo................................................................110W x 2(4Ω) / 66W x 2(8Ω)
BTL..........................................................................................225W x 1(4~8Ω)
Power Consumption ....................................................................................40W
Size ...........................................................................200X187X59mm (WxHxD)
Weight ........................................................................................1480g/ 3.26Lbs

Here is the info on the new 2022 A10a.
If you're interested in this, click the file.

Thanks you.

And the 2022 A20a specs?

 

[Zek]

@peterwen Are there one or two Infineon chips in the A10a?

 

[mario_rouge]

Thanks you.

And the 2022 A20a specs?