[Gallium Nitride amplifier] Impressions of the Mini Gan 5 by Premium Audio, pictures / video inside

[audio2design]

Copper would be more preferable.

Painted black (or any color) would be best, with black anodized okay. Anything shiny is a failure.

 

[audio2design]

Hi guys,

Did you hear about SiCFETs ? They claim to be better than GaNFETs

"Launched in 2020 the Singularity 3 Class D modules use the newest technology, and are among the best sounding amplifiers in the world. Even in competition with Class A amplifiers. This is leading edge technology. The switching devices in other Class D modules are regular MOSFET's but we use SiCFET's instead, which are about 10 times faster, giving better resolution in the top region, and less switching noise. Due to the latest enhancements in SiCFET technology, these are even significantly faster than the competing GAnFET technolgy in this power class. This may change as technology progresses though."

NewClassD Singularity3

Sounds like BS to me. SiC provides almost no or no benefit at the voltages in audio. Small advantage for the power supply perhaps. The claimed percentage advantages just are not there to be had.

 

[audio2design]

Is this product UL certified? If it is than those heat sinks are worthless. UL testing usually runs things way above any normal operating temps, and the product has to take it in stride. Unless UL has changed its requirements. In the future they should leave the heatsinks out and save the money. Use more white cementing compound on the inductors, it will anchor them with no problem. Either way, the heat sinks have nothing to do with the sound performance. Everyone knit picking it are just bored. Like me!

Here's the crux of the matter. SUBJECTIVE OBSERVATIONS HAVE NO VALUE WITHOUT CORRELATION TO MEASUREMENTS.




.

Or blind tests. You don't need measurements with blind comparisons.

 

[daniboun]

Sounds like BS to me. SiC provides almost no or no benefit at the voltages in audio. Small advantage for the power supply perhaps. The claimed percentage advantages just are not there to be had.

Thanks for the answer. Helpful

 

[Ordin Aryguy]

Sounds like BS to me. SiC provides almost no or no benefit at the voltages in audio. Small advantage for the power supply perhaps. The claimed percentage advantages just are not there to be had.

So, I've got to make my own quick blurb here.

The real advantage to SiC FET's are in their thermal performance. SiC has vastly better thermal transfer characteristics than the usual Si layers. Getting the heat out of the junctions is key.

As the junction temp rises, so does the RDSon resistance. And when the RDSon resistance rises, so does the junction temperature. It's a vicious cycle. This is true no matter where the actual FET is living circuit-wise, be it in a SMPS or the final switch of a Class D amp.

Can we hear RDSon, or thermal resistance? Not a chance, but we live with the effects of both in the form of big heatsinks, fans, and shorter MTBF's.

 

[audio2design]

As the junction temp rises, so does the RDSon resistance. And when the RDSon resistance rises, so does the junction temperature. It's a vicious cycle. This is true no matter where the actual FET is living circuit-wise, be it in a SMPS or the final switch of a Class D amp.

Haven't designed a lot of Class-D power stages for audio I will assume? I will also assume you don't know a whole lot about practical devices for this application either?

The lowest voltages for SiC MOSFETs is 600V. For a combination of RDSon, Voltage, and switching speed they are excellent (but require significant gate drive). I don't need 600V FETs for a Class-D. 100V gives me 1000 watts into 4 ohms. I can get a better 100V silicon FET than I can a 600V SiC FET, and way way easier to drive and 1/10th - 1/20th the price. Now if I am designing a power supply where I have up to 265VAC input, 375V peak, and I want it 1000 - 1500 watts, than a great 600V SiC FET can make a lot of sense. I may be able to extract 1 or 2% more efficiency out which drops the thermal load nicely.

 

[Ordin Aryguy]

Oh, crap. I'm found out.... Ok, I guess it's time to come clean.

You're absolutely right. Never have I designed a Class D amplifier. Not one. Similarly I have never even designed a circuit using a SiC FET.

In my defense, I never claimed to, either. But, you're absolutely correct in your assertions on both accounts.

What I have designed related to this discussion are circuits where the FET functions as a high frequency switch, be it a SMPS, transmission line clock driver in the many 10's of MHz, or just a simple LED on/off driver. No specific Class D stuff, but the fundamentals are the same.

Engineering is little more than making do with a whole bunch of difference constraints. Everything is a compromise. Size, weight, cost, thermal performance, electrical performance, even schedule these days with the current semiconductor supply chain mess that exists.

Si FET versus SiC FET is no different. There are compromises to be made using either.

As you pointed out, I have never designed with a SiC FET, but have designed many with the typical Si FET's. So I did some reading last night, and this morning again. Yep, you're right again that SiC FET's have a significantly higher blocking voltage (VDS), and as you've pointed out most are at least 650V and go up from there. Nowhere did I read that the devices must be operated with a drain voltage anywhere near that high to take advantage of the SiC FET's other positive traits. I could have missed it, and if I did I trust that you'll let me know the errors of my ways.

Again, you're right in that SiC FET's are a little more difficult to drive. The VGS has to get upwards of 30V in some cases to turn on the device to levels where it will have lower RDSon that a typical Si FET. Normal logic levels applied to the gate just aren't going to cut it when designing with SiC FET's if you really want to take advantage of them. Creating a 30V gate drive isn't an impossibility, however, it just adds to the component count, cost, and design schedule. In some cases those compromises are worth it, or SiC FET's would never get designed into products.

... and now for the compromises. If someone happened to be designing a circuit, let's pretend it's a Class D amplifier in this case, and the design constraints happen to include high output power capabilities, very fast PWM fundamental frequency, low output impedance capabilities, small case size (so no large heat sinks), no cooling fans, and cost isn't a huge consideration. Given those constraints I'd give SiC FET's a serious look, and if it made sense with the rest of the design, I might even choose one. Might not, too. It all depends on the rest of the design and the million other compromises that happen along the way in real-world design.

So, you're right on both of the claims you made against me. Admittedly, I have never designed a Class D amplifier, nor have I ever used a SiC FET in a real-world design... At the same time I can read and understand datasheets and can picture use cases where SiC FET's will make sense.

I'll stand with my original assertion that there is a use case for SiC FET's in high power Class D amplifiers, not in every case, but one could certainly be made.

 

[Deleted member 21219]

Admittedly, I have never designed a Class D amplifier, nor have I ever used a SiC FET in a real-world design... At the same time I can read and understand datasheets and can picture use cases where SiC FET's will make sense.

Always remember: In theory there's no difference between theory and reality. In reality, there is. Jim

 

[audio2design]

So, you're right on both of the claims you made against me. Admittedly, I have never designed a Class D amplifier, nor have I ever used a SiC FET in a real-world design... At the same time I can read and understand datasheets and can picture use cases where SiC FET's will make sense.

I'll stand with my original assertion that there is a use case for SiC FET's in high power Class D amplifiers, not in every case, but one could certainly be made.

Well your initial assertion did not specify high power. However, even with that caveat, unless you are >1000 watts into 4 ohms, there is no justification and I expect that extends up to at least 2000 watts.

If you are splitting bass/mid-high, then there is even less justification as you don't need that level of speed for good low frequency performance.

The combination of RDSon and speed, even if the Si MOSFET heats up some does not justify using SiC even for fairly high power. From a cost/performance standpoint, you GaN makes far more sense as a next step. When you get up to the voltage of the power supply where you need 400+ volts plus margin, then SiC makes sense.

The advantage of SiC wrt thermal is not in getting the heat out, it is that the performance is relatively consistent over typical operating temperatures with RDSon dropping initially with temperature increase before increasing, and higher maximum junction temperatures.

 

[whislai]

I'm have been using a ClassDAudio SDS-470C power amp for years, just got the Mini Gan 5 yesterday, first impression sound is boomier and less sparkling on treble, most probably due to the "balance" output for both bass & treble frequency? Then today I heard it differently, it sounded more transparent and though I thought the bass just could be a bit more than I want, but I like the sound, it still impress me on details, compare to previous SDS amp, but there's one more problem, when I crank the volume near max, for some songs the amp stop the sound, so wonder is it a protection feature to the speakers? This is not happen from the SDS amp though.

 

[antcollinet]

The "today I heard it differently" is a perfect illustration of subjective hearing. I am certain the performance of your system wasn't different today.

I hear music differently if I am tired, or stressed, or warm, or cold - or if I've had a drink etc etc. We will also hear a difference if we are listening for one.

State of mind effects what you hear far more than miniscule differences in Amp characteristics. We hear with our brains, not with our ears.

 

[Schollaudio]

I'm have been using a ClassDAudio SDS-470C power amp for years, just got the Mini Gan 5 yesterday, first impression sound is boomier and less sparkling on treble, most probably due to the "balance" output for both bass & treble frequency? Then today I heard it differently, it sounded more transparent and though I thought the bass just could be a bit more than I want, but I like the sound, it still impress me on details, compare to previous SDS amp, but there's one more problem, when I crank the volume near max, for some songs the amp stop the sound, so wonder is it a protection feature to the speakers? This is not happen from the SDS amp though.

What power supply voltage is used on the SDS-470c user. The GAN5 uses 34V in a balanced design for a peak of 288 watts at 8 ohms not counting any rail sag. That should be fairly loud with most speakers.

 

[whislai]

What power supply voltage is used on the SDS-470c user. The GAN5 uses 34V in a balanced design for a peak of 288 watts at 8 ohms not counting any rail sag. That should be fairly loud with most speakers.

aha now it's not happen, suspect is because of my speakers banana metal part touch some other metal part from my monitor, I'm from Malaysia where electric shock can causing death btw haha. Also this amp is not running cool at all, living on 33 level condo without aircon, I feel the massive heat from this amp lol. The song playing is Inva Mula - Lucia Di Lammermoor + The Diva Dance

 

[whislai]

The "today I heard it differently" is a perfect illustration of subjective hearing. I am certain the performance of your system wasn't different today.

I hear music differently if I am tired, or stressed, or warm, or cold - or if I've had a drink etc etc. We will also hear a difference if we are listening for one.

State of mind effects what you hear far more than miniscule differences in Amp characteristics. We hear with our brains, not with our ears.

yeah, I'm posting on behalf of that I'm human and express my feeling subjectively haha. Anyway my further subjective experience after switching back and forth to compare with my previous SDS, I think I like this GAN amp better, my previous SDS has heat issue which solved by removing the top case and putting few coins on the heat sink.

 

[daniboun]

yeah, I'm posting on behalf of that I'm human and express my feeling subjectively haha. Anyway my further subjective experience after switching back and forth to compare with my previous SDS, I think I like this GAN amp better, my previous SDS has heat issue which solved by removing the top case and putting few coins on the heat sink.

Thanks for sharing your impressions.
Question : Is The SDS-470c based on the IRS2092 chip ? Do you own other Class D amps to compare ?

 

[daniboun]

Another question for the experts .... the manufacturer shared this measurement in relation to the headroom of the mini Gan ... I have trouble understanding what he wants us to understand....

I have to admit that ClassDaudio communicates little and badly .... Any idea ?

Here the answer :

"Hi,

This graph was from a discussion I was having with someone and it was regarding the amps headroom. This little amp puts out lots of power and lots of very clean headroom. Some of the components of the amp are not made to handle this king of power for long periods so it is rated at 200W RMS per channel into 8 ohms.

Tom"

 

[whislai]

Thanks for sharing your impressions.
Question : Is The SDS-470c based on the IRS2092 chip ? Do you own other Class D amps to compare ?

I owned 3 SDS amps, same chip I suppose. Yes I have another NAD master series M22 v2 power amp, Hypex NCore, which I think is neutral with excellent distortion control and more neutral to the sense as pointed by most reviewers. Though overall I prefer smoother sound with just a little bit of warm & fast bass control amp.

 

[daniboun]

I owned 3 SDS amps, same chip I suppose. Yes I have another NAD master series M22 v2 power amp, Hypex NCore, which I think is neutral with excellent distortion control and more neutral to the sense as pointed by most reviewers. Though overall I prefer smoother sound with just a little bit of warm & fast bass control amp.

OK thanks. You mean : you prefer the Mini Gan 5 vs the M22 ?

 

[pma]

@daniboun : It is really difficult to separate passion from listening impressions and to get independent and neutral view. It is not impossible however difficult. I know it from my own experience. At least you need to start with level matched listening within 0.5 dB.

 

[daniboun]

@daniboun : It is really difficult to separate passion from listening impressions and to get independent and neutral view. It is not impossible however difficult. I know it from my own experience. At least you need to start with level matched listening within 0.5 dB.

Relevant and you re right )