Looking for a "fix" for my iPod + HA-2 DAC/amp stack

[entsnack]

Hello Scientists,

One of my portable rigs is an iPod Classic gen 6 stacked with an Oppo HA-2 DAC/amp. The HA-2 is an old "Made for iPod" device that extracts 24/48 digital audio from the iPod, which it then amplifies. With the HA-2, my iPod sounds on-par with my other portable rig – an iPhone SE gen 1 + an Apple dongle or an iFi Go Blu DAC/amp (they both sound almost the same to me). I use Sennheiser IE 200s with both stacks.

I am looking to build up my iPod stack to outperform my iPhone stack (ideally not not necessarily in an audible manner). The HA-2 has a line-out so I can bypass the amp. I know the HA-2 measures poorly (here are measurements from its successor, the HA-2 SE).

What can I add to my iPod stack to "fix" this? Any oldheads out here with advice on 2007-era audio tech?

I can't change out the HA-2 (the intersection of made-for-iPod DACs with available and user-replaceable batteries is...the HA-2). I don't want to change out the iPod (I have multiple iPhones if I want a touchscreen). I don't want to switch to headphones (I already have a desktop setup with Arya Stealths), but am open to getting other IEMs instead if the money is better spent there.

- ent

 

[Sauerkrautonaut]

Any oldheads out here with advice on 2007-era audio tech?

Not really, since I myself can be considered 2007-era
I'm feeling benevolent today though, so I've decided to give you advice anyway. For better or worse.

If I've understood this correctly, all you need is a good, pocket-sized headphone amp that can connect to the HA-2 and - ideally - outperform its built-in amp.

There are some products that can do this, such as the Topping NX3s and the Topping NX4, and I don't see where the difficulty in finding them would be. It took me about three minutes of searching the Audio Science Review index to find these two, and I can guarantee you there are countless more on the market.

That said, as Amir found out in his testing, the headphone amp is not the issue with the HA-2:
"[...] Its DAC has some warts such as excessive noise and jitter and anemic output level. Its headphone amplifier is clean with low impedance but does not have competitive amount of power [...]"

...and replacing the amp won't solve the issues of the DAC.

The lack of iPod-compatible DAC/Amps is for a good reason, namely that audiophiles who want better performance have devices that can supply them with better performance, both for desktop and for portable use. This was as true then as it is now. Therefore, I'd simply accept this as a limitation of the stack. I'm not sure it's even possible improve this stack any further without switching to a completely different platform or DAC.

 

[entsnack]

Thank you! I know a lot more now than I did when I posted this, and now more after reading your message. Amir's review is an amazing resource too.

I have also come to the realization that the MFI authentication needs to be bypassed for this stack to improve meaningfully, so I'm now focusing my effort there with some co-conspirators.

 

[Sauerkrautonaut]

[...] that the MFI authentication needs to be bypassed [...]

That's certainly an interesting approach. So you intend to use a different DAC over USB by bypassing the "made for iPod" authentication thing?
I'm not sure if that will work as intended. If I remember correctly, the iPod 30-Pin dock connector does have USB data lines, but I don't think the software side of the iPod will recognise a modern USB DAC. Not without major engineering wizardry at least.

Modern DACs work on modern apple devices, because iOS comes with UAC (USB Audio Class) drivers.
The iPod Classic runs a rudimentary operating system called Pixo OS, which does not contain UAC drivers.

And I'm not sure if the iPod will recognise anything at all without giving the correct resistance to the accessory indicator pin (21) on the 30-Pin connector. Problem is, the accessory indicator pin doesn't have a resistance for generic USB accessories. There is a resistance for serial communication, which is too slow to send audio over. There is also a resistance for docking station use, but this is useless too, because it sends audio over the line out pins on the 30-Pin connector, which use the iPod's internal DAC/Amp.

I read rumours about using the iPod camera connection adaptor to use a DAC, but this is impossible, because to my knowledge, it strictly uses PTP, which you cannot transmit audio over.

I found this old thread from someone who has the same issue as you do, might be worth giving it a read if you haven't already.

 

[entsnack]

not without major engineering wizardry

Yes, that's the plan. This document has more details, but the idea is to do the MFI handshake with a donor accessory that can handle the authentication challenge, and then switch to the DAC.

We are not yet sure separating the two stages is possible (due to logical checks or timing) but that's our current hypothesis.

Some recent software work handling iAP in Rockbox makes it possible to prototype and capture data more easily now, so we are not limited by what was available at the time of the old 2021 thread. Specifically we are not limited to the stock iPod OS.

Once we figure out auth, handling audio output is the next challenge.

All my past projects started with someone telling me why it is not possible, so I'm happy to see the streak continue.

 

[Sauerkrautonaut]

Fascinating. This is far more fleshed out than I thought it was. I had a look through the references of that document, one of which conveniently links to the original iAP specification. I had a quick read through the relevant bits. While it was definitely one of the more unusual novels I've read, it was certainly funny to see the lengths to which Apple goes in order to prevent their customers from enjoying their products, which they paid for, however they want.

Anyway, one of the commenters (you?) mentioned Rockpod. According to the Rockpod documentation, Rockpod implements a custom protocol stack that, as far as I understand it, works completely using the USB data lines:

[...] a custom iAP-over-USB-HID transport layer [...]

Therefore, some of the biggest issues in the GitHub doc:

[...] Auth chip I²C bus is internal to the accessory — does NOT appear on 30-pin connector pins [...]

and

Does auth happen over dedicated pins (25/26) or over iAP serial/USB transport? [...]

and

Is the auth challenge-response proxied by the accessory MCU (over iAP), or does the iPod talk directly to the auth chip (over dedicated pins)? [...]

are already solved. They are also answered in the specification:

Accessory identification and authentication may take place over any of the iPod’s transport links—UART, USB, or Bluetooth. [...]
(iAP Spec Page 48)

Further questions are answered in the iAP specification, which I bothered to read, because whatever AI that wrote the GitHub doc seemingly didn't:

Is there a timing constraint between auth completion and lingo declaration? [...]

-> This question is nonsense, because the lingo declaration happens *before* the auth. This misconception is spread throughout the entire GitHub doc. This is explicitly noted numerous times in the specification:

IMPORTANT: To identify an accessory to an attached iPod or iPhone, before authentication, new accessory designs must use the Identify Device Preferences and Settings (IDPS) process described in Appendix B [...]
(iAP Spec Page 48)

That said, there are timing restrictions within the IDPS process; see pages 47, 49, 108, 109, 296, 533, 535, 540, 543, 545, 548, and many others of the iAP specification. There are too many to quote, and too many to list comprehensively.

Does the iPod re-challenge the auth chip periodically during audio streaming? [...]

-> No. Not at random, and not periodically. However, the host (iPod) can issue a full re-identification and re-authorization at any time, and it may do so in some situations:

When an iPod transitions from sleep to power on, it sends a RequestIdentify command to the accessory. The accessory must respond with StartIDPS and go through the IDPS process [...]
(iAP Spec Page 109)

Does pin 21 resistor value affect auth behavior? In dock mode (1kΩ), does the iPod initiate auth automatically? Does 68kΩ (line-out mode) skip auth entirely?

-> No. The resistor on its own does not alter or remove auth requirements.
-> No. The iPod never automatically initiates auth upon sensing a resistor.
-> No. Line-out mode does not skip auth, because it does not require auth. Auth is only required for accessories that utilize features that require iAP auth, and line out is no such feature. You can connect something to the line-out pins of the 30-Pin connector anytime. No auth is required for this.

I am definitely excited to see how this project develops going forward. I wish you and everyone else working on it good luck with this project.
Feel free to contact me, should you require the assistance of my reading skills again.

Hope this helps.

 

[entsnack]

Thank you! This is super useful, I think we will crack this soon, and I just found a cheap HA-2 on eBay for sacrificial parts. The journey is quite thrilling and may be an end in itself.

And once we crack it we'll send the whole stack over to Amir to measure.

 

[Sauerkrautonaut]

Has there been any development in this project? It's been three days since I commented a link to this thread under the GitHub doc, and there has been no reaction or acknowledgement whatsoever. I'm sad to see this project die, because it will definitely bring some fresh air into the stale and boring iPod community.
If I/we/you manage to defeat iAP, that would be an absolute game changer. Oh, to think of all of the things you could do with unrestricted USB connectivity and freely programmable software like Rockbox!