Review and Measurements of E1DA 9038S BAL Portable DAC & Amp

[raanany]

Ohh, thank you for the review. Regarding the SNR, it depends on USB 5V rail quality and current loops around it. I took the video with #9038S tied to the USB OTG Android, powered from power-bank i.e. completely isolated from 230VAC outlet, and you can see the SNR is 122.9-123.8dbA, next I moved my wireless 2.4GHz mouse(DUT AP2702 and mouse on the same table) to stop video record and this action reduced SNR to 121dbA. I submit 1dbA just in case and published -122dbA as specs. -118dbA I never have seen yet but I believe it could be in some particular setup.

Hi IVX, I just ordered your Dac from Aliexpress. I have a few questions please:

1. Can I be sure that it is your dac that they are selling? https://a.aliexpress.com/lELDcRAs

2. Are they (who ever is manufacturing it in China) performing any validation to make sure the DAC meets the specifications?

Thanks,
Raanan

 

[IVX]

It is our store, I do 3 times the test of each unit. General + S/N before assembling, and one more general test before packaging.

 

[mgmate]

It is our store, I do 3 times the test of each unit. General + S/N before assembling, and one more general test before packaging.

IVX, Any Estimated time for 9038D?. It is not for putting presure, it's just to have an estimated date.

Regards,

Manuel

 

[raanany]

It is our store, I do 3 times the test of each unit. General + S/N before assembling, and one more general test before packaging.

Thanks. I’m trying to use the discord invites you gave but they are all expired. The most recent one is https://discord.gg/vdXUZ4

Also, before I dive into the discord discussion and learn about how to control volume on a mac - let me get this straight - “software volume” is practically going to increase the digital noise floor and eat away the E1DA beautiful SNR figure to death, isn’t it? Electrical engineer here, please be as technical as you can get with the answer.

 

[IVX]

raanany
In IEM mode and 24bit, the SW volume should be under -50dbfs(at 1% position of the Windows volume slider) to affect 9038S resolution.

 

[raanany]

raanany
In IEM mode and 24bit, the SW volume should be under -50dbfs(at 1% position of the Windows volume slider) to affect 9038S resolution.

Thanks @IVX. I saw on discord (managed to join eventually using: https://discord.gg/nkUBYSv) that a gen 3 device is in the works which will have HW volume control. Will this gen 3 also have the 2mm balanced plug? (I just ordered $200 matching balanced cable from campfire audio for my Andromeda gold to pair with your DAC... )? Will the HW volume use attenuation of the D2A output or will it simply reduce the digital signal volume as the SW volume does? Will the Mac volume buttons be able to control the volume?

sorry for having so many questions. I was really excited to order your DAC and the cables and was disappointed to learn that I’d have to use iTunes volume slider (rather than the convenient volume buttons) and eat away the resolution.
Thanks,

raanany

 

[Veri]

I just ordered $200 matching balanced cable from campfire audio for my Andromeda gold to pair with your DAC...

Oh no.

 

[raanany]

Oh no.

I know it sounds crazy but considering the Andromeda Gold IEMs have an impedance of as low as 3ohms (at low frequencies), you should strive to get the lowest cable resistance and amp output impedance that you can find. Otherwise, Bass is negatively affected (high frequency impedance and therefor gain is up to 12 dB higher). I ordered the E1DA because of the low impedance and the excellent SNR. If you control volume digitally you eat this wonderful SNR. Instead, If you attenuate after the D/A conversion (before the AMP section) you lower both signal and DAC noise and retain (true, up to a certain AMP noise floor) the original SNR. Having no volume control (= resorting to SW volume control) is not a good solution for a portable DAC+AMP. It could be acceptable for a desktop DAC connected to an amplifier which has its own volume control. A portable DAC+AMP must employ HW volume control after the D/A conversion. Do you agree @IVX or am I missing something?

 

[Veri]

I know it sounds crazy but considering the Andromeda Gold IEMs have an impedance of as low as 3ohms (at low frequencies), you should strive to get the lowest cable resistance and amp output impedance that you can find.

Just one thing to take into mind. Campfire Audio is an offspring of ALO audio, an expensive overpriced cable manufacturer.

So yes, they stuff a bunch of sensitive BA drivers into an IEM that might just be sensitive enough to give minor changes swapping cables. Who is the number one entity to profit from manufacturing and marketing these cables for the Andromeda et al?

It's just one big scheme similar 8-core and 16-core cables are available in $20-$50 price ranges from China/aliexpress etc.

 

[IVX]

raanany
can I ask you, how do you think HW volume works, what kinda magic there help to win SW volume?

 

[raanany]

raanany
can I ask you, how do you think HW volume works, what kinda magic there help to win SW volume?

Assuming the AMP section (programmable attenuator + current driver) of the DAC+AMP combo has a lower noise floor compared to the DAC output, then if we attenuate the signal coming out of the DAC, we attenuate both signal and DAC noise. Theoretically, given no AMP noise at all, we could retain the DAC output SNR regardless of attenuation. In practice, the more we attenuate, the signal will go down but the noise will eventually hit the AMP noise floor and the signal to noise will start to show impact. Do you agree until now?

How to design the programmable attenuator - I think if you take a cheap 8 bit DAC (let's call it volume_dac), drive the main DAC output into the volume_dac reference voltage input and connect the volume command to the volume_dac digital input (after some simple db-->lin look-up-table conversion) then you get digital HW volume control that is better than SW attenuation of the PCM samples. An 8 bit dac can be very cheap and will easily beat the SNR noise floor of the 32 bit main DAC so the overall effect will be better SNR at low volumes. Do you agree? If so, I'd like a discount for the next version

 

[IVX]

I mean HW volume control implemented in a DAC, like 9038, it is simple multiplier 0..1 integer 32 bit. If the volume control on the host side implemented without stupid mistake(as I remember, tidal has noise problem with its volume control at -30-50dbfs), no difference where is such multiplier is implemented.

 

[raanany]

Just one thing to take into mind. Campfire Audio is an offspring of ALO audio, an expensive overpriced cable manufacturer.

So yes, they stuff a bunch of sensitive BA drivers into an IEM that might just be sensitive enough to give minor changes swapping cables. Who is the number one entity to profit from manufacturing and marketing these cables for the Andromeda et al?

It's just one big scheme similar 8-core and 16-core cables are available in $20-$50 price ranges from China/aliexpress etc.

Veri, I agree that brand names charge an extra. This wonderful site shows us how much they are sometimes not worth the money they charge. Unfortunately, out of possibly 100 vendors of cables in aliexpress, a certain, unknown percentage will be "not as advertised". Other's will meet the spec but will break after a month. With others, instead of silver on copper you'll get something else and instead of 10 milliOhms you'll get 1 Ohm. Even the good manufacturers will exhibit variance due to cutting the cost of quality control. If audiosciencereview recommends something (like the E1DA), I'll happily buy it (as I did) but without this confidence, I'm worried I'll never be sure what I bought.

I'm one of those people (I hate the term "audiophile" but I admit it applies here) who are obsessed with quality. I know that my ears don't hear that well anymore but still I can't live with the notion that I don't have the best I can buy. I think most people who come here, are like me.

 

[raanany]

I mean HW volume control implemented in a DAC, like 9038, it is simple multiplier 0..1 integer 32 bit. If the volume control on the host side implemented without stupid mistake(as I remember, tidal has noise problem with its volume control at -30-50dbfs), no difference where is such multiplier is implemented.

Yes, such multiplication of the PCM sample with a programmable gain is the same as SW gain. Don't you agree the other approach is better?

 

[IVX]

No in portable world I believe.

 

[Veri]

I'm one of those people (I hate the term "audiophile" but I admit it applies here) who are obsessed with quality. I know that my ears don't hear that well anymore but still I can't live with the notion that I don't have the best I can buy. I think most people who come here, are like me.

I respect that I simply would not buy from Campfire or ALO due to personal reasons. I don't like the way they work. They sell the kind of pricy cables that should not make an audible difference. They sell a flagship (Solaris) that uses counterfeit knock-off Bellsing drivers intead of real Knowles drivers. When you look at the price their stuff sells at, that is unbelievable.

 

[raanany]

I respect that I simply would not buy from Campfire or ALO due to personal reasons. I don't like the way they work. They sell the kind of pricy cables that should not make an audible difference. They sell a flagship (Solaris) that uses counterfeit knock-off Bellsing drivers intead of real Knowles drivers. When you look at the price their stuff sells at, that is unbelievable.

I had no idea. I really like the sound of the Andromeda Gold and I don't regret buying it. As for stealing secrets... I'm not sure how I feel about it. I'm part of this industry (I'm en electrical engineer), admittedly not doing audio but the concept of IP theft is the same. There are only a few ways to implement something correctly. When you leave a company (call it company B) and move to another (company C) you naturally take with you what you've learned. Several years back you brought to company B some background experience from company A and company B didn't seem to mind then but all of a sudden, all your knowledge belongs to them. That's bullshit. As I said, there are a limited number of ways to implement something right. Saying you can't work in a competing company is unfair. Many times the patents are actually your brainchild. You invented something and the company you worked at owns the patent rights. You switch jobs and now you can't use your idea. Sometimes you or company C get sued just because they are competitors of company B and therefor the court is in their favor. This is unfair to you as a professional (limits your employment options) and unfair towards consumers who are devoid of the benefits of a healthy normal migration of engineers. Not sure about the Bellsing vs. Knowles case (didn't read the fine print) so maybe in this case some moral injustice was involved.

 

[raanany]

I mean HW volume control implemented in a DAC, like 9038, it is simple multiplier 0..1 integer 32 bit. If the volume control on the host side implemented without stupid mistake(as I remember, tidal has noise problem with its volume control at -30-50dbfs), no difference where is such multiplier is implemented.

Interesting to learn about Tidal. I had a long and frustrating email thread with them trying to convince them that a track I happened to hear had terrible saturation distortion. I tried it with various players, amps and headphones and it was always there at the same portion. The support guy kept insisting that I check my internet connection and DNS settings (who hired this guy ) and insisted that he can't hear the distortion. The same track on iTunes and a CD I later bought was clean. I'm sure there was something broken with their loudness equalisation algorithm but they wouldn't admit it. I stopped my subscription. No point in paying for a premium lossless streaming service if you can't trust their content quality.

 

[Veri]

Not sure about the Bellsing vs. Knowles case (didn't read the fine print) so maybe in this case some moral injustice was involved.

My point was, they sell kilobuck IEMs. They really could have used non-counterfeit drivers instead of the knock-offs lol. It was surely "within budget"

 

[IVX]

raanany
but anyway, the most ridiculous thing is MQA )) I always proud how mankind so much moved forward the last 30 years, at least in marketing tricks. Why the mp3 was so boringly introduced, and for free? Named as a robot's model "mp3" but not more allusive like "Master Quality Authenticated", being the same lossy compression? Marketing Science leading us somewhere what looks like hell.