Gustard A18 Balanced Desktop DAC Review

[sq225917]

The LME49860 has better figures for imd and thd+ noise at 5v output level, then it goes off a cliff in either direction and performance is worse that the OPA1612. if you don't sit with the volume at 0db the whole time the OPA1612 might be a better choice AKM certainly thought so.

I have no idea what access Gustard have to local markets, but globally the OPA1612 looks like the more expensive chip from any 3PD

 

[Dannemand]

Hi
Thanks for your question, it's interesting

1. In fact, the purchase price of LME49860 is more expensive than OPA1612A. LME49860 has better rail-to-rail performance. The requirements for SINAD performance are met in A18.

2. It is safe to use powerstrip instead of the switch. The power switch of the A18 body can be kept on. If the machine is not used for a long time, it is recommended to turn off the body switch.

3. The A18 MQA version still maintains the output level of 6V rms, and the measurement level of about 5V is used because APX555 will switch gears near 5V. The performance of 6Vrms cannot be accurately evaluated. The attenuation of 2dB digital volume output measured in the 5V gear of the AP will have less performance loss than the 6V output and the 10V gear of the AP.

Thanks a lot for your response. Even in the middle of New Year's holidays. I much appreciate that!

1. Understood, thank you. Curious then, why AKM chose OPA1612 for their reference design. I have no personal experience with either OpAmp, but I am glad to hear you chose it for a design reason. A18 does sound great, so no concerns there!

2. Thank you, perfect! I'll just use the powerstrip switch then. What made me ask was the note in the AK4499 data sheet about the importance of power on/off sequence (between DAC chip and output stage) to prevent IC damage.

3. Great explanation, thank you. That's a rather important piece of added context, since both Amir's and Wolf's measurements point to better performance at (or just below) 5V output than 6V. I do remember @JohnYang's post about it (here) but your explanation clarifies perfectly. I still wish A18's output level was closer to the nominal 4V, but fortunately my Bryston amp has gain pots, so I can compensate.

Thanks again for responding. Your activity on ASR does Gustard credit!

 

[Dannemand]

The LME49860 has better figures for imd and thd+ noise at 5v output level, then it goes off a cliff in either direction and performance is worse that the OPA1612. if you don't sit with the volume at 0db the whole time the OPA1612 might be a better choice AKM certainly thought so.

I have no idea what access Gustard have to local markets, but globally the OPA1612 looks like the more expensive chip from any 3PD

Looking at Mouser I also found OPA1612 to be more expensive. But maybe it's different in volume.

I see your point about the OPA1612 vs LME49860 performance depending on output level. Interesting. Gustard has been very adamant about keeping output at 0dB. The newer X16 uses OPA1612A, but of course ES9068AS is a different chip. I understand AK4499 is hard to tame.

 

[sq225917]

The opa1612 is a better choice for high frequency noise too it's not as affected by rf.

I checked all the global distributors the opa1612 is more expensive from all of them, even in multi thousands. I suspect the lme was available direct from the fab.

It's all down to specific use case of course and the numbers on the Gustard are exrmplary, but the smsl400 is better, with its opa1612.

 

[ArturoKiwi]

The opa1612 is a better choice for high frequency noise too it's not as affected by rf.

I checked all the global distributors the opa1612 is more expensive from all of them, even in multi thousands. I suspect the lme was available direct from the fab.

It's all down to specific use case of course and the numbers on the Gustard are exrmplary, but the smsl400 is better, with its opa1612.

And the D90 too

 

[Dannemand]

@sq225917 @ArturoKiwi:

Thank you, both.

I am aware that both D90 and M400 use OPA1612. And they do measure better. Both of them are also more expensive than A18 and outside the range of what I could afford.

Back in November when I bought A18, it was the more expensive option compared to my alternative, SMSL M200. I ended up keeping the A18 because it sounded better to my ears and those of my wife.

(In case the Nobody-Can-Hear-Any-Difference police is watching: Yes, we did perform a series of level matched, filter matched ABX blind listening tests. And no, we didn't construct a special room for it, nor did we bring in 20 other listeners. We did it for our own sake, not to prove anything to anybody. )

So I am still very happy with the A18. It sounds wonderful to me, not too bright and analytical, not too bassy. It just effortlessly brings out the best in any music we throw at it.

That said, I too am wondering if OPA1612 might have performed better, particularly as we mostly use A18 with slight attenuation (-10db - 0dB range). But I understand OpAmps should be matched to caps, so just replacing the LME49860s isn't an option for me -- even if I were good with a soldering iron, and even if I were willing to sacrifice my warranty.

I have confidence in Gustard. No doubt they would recommend A22 instead with its discrete class-A output stage, and would say this is their low cost AKM option. But for me, being on a budget, A18 is the higher end choice, and one I hope to keep for a long time.

And I do appreciate Gustard's responsiveness and willingness to answer questions here on the forum!

 

[Veri]

And I do appreciate Gustard's responsiveness and willingness to answer questions here on the forum!

Agreed

 

[ArturoKiwi]

Agreed

Me too and until now I'm really happy to read a brand that follow their buyers and is so transparent.

But now:
"1. In fact, the purchase price of LME49860 is more expensive than OPA1612A. LME49860 has better rail-to-rail performance. The requirements for SINAD performance are met in A18."

TThat's not true :-(

 

[deafenears]

Me too and until now I'm really happy to read a brand that follow their buyers and is so transparent.

But now:
"1. In fact, the purchase price of LME49860 is more expensive than OPA1612A. LME49860 has better rail-to-rail performance. The requirements for SINAD performance are met in A18."

TThat's not true :-(

Which bit is not true?

"The purchase price of LME49860 is more expensive than OPA1612A" ? How do you know how much it costs Gustard to source these components?

"LME49860 has better rail-to-rail performance" ?

"The requirements for SINAD performance are met in A18" ? Front page says it's true - the A18 is right up there in terms of SINAD.

 

[ArturoKiwi]

Which bit is not true?

"The purchase price of LME49860 is more expensive than OPA1612A" ? How do you know how much it costs Gustard to source these components?

"LME49860 has better rail-to-rail performance" ?

"The requirements for SINAD performance are met in A18" ? Front page says it's true - the A18 is right up there in terms of SINAD.

If you read previous messages, from more than one user, they all wrote that OPA1612A are more expensive

 

[gustard]

Hi everyone

Our electronic components are purchased from local distributors. In short, based on their quotation, the price of opa1612A can indeed be lower than that of LME49860. Perhaps because of the difference in sales volume, because opa1612A is more popular, dealers can get lower prices from Texas Instruments.
We don’t deny that OPA1612A is the more expensive one on Mouser and Digikey.

"LME49860 has better rail-to-rail performance" ------ practice brings true knowledge. It does show this on the A18 circuit board

A18 should be better, but because of the accident at the AKM factory. We can only say that it is very regrettable.

 

[ArturoKiwi]

Hi everyone

Our electronic components are purchased from local distributors. In short, based on their quotation, the price of opa1612A can indeed be lower than that of LME49860. Perhaps because of the difference in sales volume, because opa1612A is more popular, dealers can get lower prices from Texas Instruments.
We don’t deny that OPA1612A is the more expensive one on Mouser and Digikey.

"LME49860 has better rail-to-rail performance" ------ practice brings true knowledge. It does show this on the A18 circuit board

A18 should be better, but because of the accident at the AKM factory. We can only say that it is very regrettable.

Thanks for your kind reply

 

[Dannemand]

Our electronic components are purchased from local distributors. In short, based on their quotation, the price of opa1612A can indeed be lower than that of LME49860. Perhaps because of the difference in sales volume, because opa1612A is more popular, dealers can get lower prices from Texas Instruments.
We don’t deny that OPA1612A is the more expensive one on Mouser and Digikey.

Once again thank you for your response and your input. If there is one thing everybody here seems to agree upon (not just this thread) it is how much we appreciate your engagement on the forum!

I do understand about the price variations, depending on your supply channel. And as I posted in my original question to you, I would even be OK if your answer had been that LME49860 was chosen to shave cost on the BOM, given that A18 is a more affordable AK4499 implementation.

"LME49860 has better rail-to-rail performance" ------ practice brings true knowledge. It does show this on the A18 circuit board

A18 should be better, but because of the accident at the AKM factory. We can only say that it is very regrettable.

I don't think I fully understood this part. Maybe something got lost in translation, or maybe I am just dense. "A18 should be better, but because of the accident at the AKM factory". Are you saying that you intend to improve A18 once AKM is up and running again? Or that it should be better thanks to LME49860's rail-to-rail performance (compared to using OPA1612A)?

Any additional clarification/elaboration would be great.

Thanks again!

 

[Veri]

I don't think I fully understood this part. Maybe something got lost in translation, or maybe I am just dense. "A18 should be better, but because of the accident at the AKM factory". Are you saying that you intend to improve A18 once AKM is up and running again? Or that it should be better thanks to LME49860's rail-to-rail performance (compared to using OPA1612A)?

Yeah I think he means that, due to the incident they can not really make any improvement anymore, no further R&D since the availability is gone.

 

[sq225917]

That's what I thought, as I said price on parts varies with supply, especially local markets and bulk buying.

 

[Nejiro]

I am the owner of an A18 that I am very satisfied with, I misunderstood or maybe this product will no longer be supported? Even if it is fine with me already .......

 

[Veri]

I am the owner of an A18 that I am very satisfied with, I misunderstood or maybe this product will no longer be supported? Even if it is fine with me already .......

It is and will be supported, but they will not spend any more time on improved or new AKM chip models, is how I understand it.. because they won't be able to keep building them :/

 

[ArturoKiwi]

Is there any news about AKM? They are rebuilding the factory or they will completely stop the chip production?

 

[ArturoKiwi]

I don't tknow what to think. Gustard recently released the A22 with dual AK4499EQ, but says the A18 is a dead project due to inability to improve it.
So why would anyone buy any AKM chip dac right now?

 

[sq225917]

They don't have supply to build it, thats why it's dead. Its a complete finished product, it works. If one likes the price, sound and features then buy it, if you can find one.