[OCC7N]

A stable continous 3.7V signal would still better a lipo battery?

 

[Thorsten Loesch]

No matter how less I know about electronics, I have an assumption the battery inside diablo is weak for current hungry headphones

It SHOULD not be, but it is possible. Batteries are protected and have a circuit that drops the battery off line if they detect excessively high current.

I dont believe a bypass happens when charger is connected…

Well, your believe is incorrect.

The charger IC used in the Diablo is a "battery bank" IC, meaning if external power is attached, it is supplied to the main power bus and the same power is used to charge the battery. The battery does not participate at all in the system, it is being charged via a switching charger util full, after that the switching charger stops.

1. I believe 3.7V psu is better and it eliminates the USBC port.

First, 3.7V will be the "wrong" Voltage. It is equivalent to a 50% full/empty battery. A full battery is 4.2V or 4.3V.

Second, when operating from battery, a switcher will operate to make a constant 5V from the Battery.

Thus using a high quality supply via USB-C and letting the battery charge and enter "desktop mode" will likely be a better choice than supplying an equal quality power via the battery connection.

Additionally, to supply power via the battery connection, the unit needs to be opened up, voiding warranty and there will be no easy way to route the wires out of the unit looking nicely.

2. Idefender will be used instead of purifier3 for the purpose of an less altering adapter.

iDefender is useful only if there is a ground loop in the system, involving the source. Otherwise iPurifier is preferred, as it is a full USB signal repeater.

The idefender is designed the way it is as it should be placed on the source side of the system. It is not meant to plug into the "sink" device like an OTG adapter.

A really "extreme" system would use the iDefender on the source side with a 5V/1A+ linear PSU (or iFI iPower) providing power to the downstream iPurifier that is plugged directly into the audio device, which in turn is powered by a linear PSU (or iFi iPower).

3. Mod a very durable plastic cover to replace the backplate, so not only the idefender will fit but a socket for 3.7V DC power supply.

That is what I remarked before. That seems a lot of work, for what is likely the worse quality solution compared to using USB-C.

The other purpose of the idefender is to connect it straight to ideon 3R Renaissance mk2 reclocker - and yeah it beats the ipurifer3 which is included in diablo.

Hmm, I find that interesting. The 3R is functionally identical to iPu but with a separate power supply, based on this:

https://www.avmentor.net/reviews/2017/ideon_audio_3r_usb_renaissance_1.shtml

Feeding a separate 5V line to the iPu via iDef may even out the results. Maybe not? Would need to be tried.

But to start and test before all modding I will use it normally and see how it is.

I strongly suggest you try a high quality 5V supply via USB-C. And try a separate power line via iDef to iPu BEFORE you make any changes that are not reversible, carry the risk of damaging things or voids warranty.

Thor

 

[Thorsten Loesch]

Hmm i recall he did not. He only designed the ZENs..?

Well, I did design pretty much anything iFi sells, excluding some most decent products, though the current trend I notice at iFi seems to re-use my designs and "copy-paste" merge them to make "new" products such as the xSDSD Gryphon" and "iCAN Phantom".

The products I really had nothing to do with are the Uno and both stream's, though the DAC inside of these and a number of other design elements are based on products I designed previously, such as the USB signal regeneration and SPDIF reclocking.

And some recent releases such as the "go" devices were finalised and completed after I left and I only did the initial electronics design. Just how much was changed between the initial design and final release I do not know, though the basics (chipsets etc.) seem identical top what I originally designed. Yes, there is a "Team" at iFi, until relatively late before my departure this team was very small and most ifi products from concept via electronics design and PCB design or direct oversight during PCB design to final testing was me.

Thor

 

[Thorsten Loesch]

A stable continous 3.7V signal would still better a lipo battery?

Probably not. It would depend on the precise battery fitted (there are different factories that offer equivalent batteries and may have been used by iFi, original details of the first generation batteries I posted before) and it's build in current limit in the electronics vs the external power supply current limit.

The battery fitted is 4,800 mAh meaning it is typically designed for discharge at 1C (4.8A) or even 2C (9.6A) but the electronics fitted to the battery may limit current.

If you ask me, 5V @ >4A via USB-C (the connector is rated at 5A max) would be the best choice. Don't use a USB Cable, use DC to USB-C adapter:

I would suggest a 5V/4A linear power supply, if you asked me. I linked the item I would buy personally before.

And if you ABSOLUTELY INSIST on using the "feed external power to the battery connector after modifying the case to have a socket and then allow the power to be converted to 5V via a switching regulator", make sure not to use 3.7V but 4.2V which is equivalent to a fully charged 3.7V LiPo battery.

Thor

 

[OCC7N]

Probably not. It would depend on the precise battery fitted and it's build in current limit vs the external power supply.

The battery fitted is 4,800 mAh meaning it is typically designed for discharge at 1C (4.8A) or even 2C (9.6A) but the electronics fitted to the battery may limit current.

If you ask me, 5V @ >4A via USB-C (the connector is rated at 5A max) would be the best choice. Don't use a USB Cable, use DC to USB-C adapter:



I would suggest a 5V/4A linear power supply, if you asked me. I linked the item I would buy personally before.

And if you ABSOLUTELY INSIST on using the "feed external power to the battery connector after modifying the case to have a socket and then allow the power to be converted to 5V via a switching regulator", make sure not to use 3.7V but 4.2V which is equivalent to a fully charged 3.7V LiPo battery.

Thor

Thank you very much for all the useful info.

Help me try to understand why LiPO is rated at 3.7V. Google says it may be 4.2V at fuld charge but it drops to the nominal voltage 3.7V.

Also this is the battery inside diablo. I see no 4.2V

As for ipurifier3 vs ideon 3R Renaissance mk2 I tested them both. No competition Ideon slays the ipurifer, which is not a full unaltered signal for my ears.

 

[Thorsten Loesch]

Help me try to understand why LiPO is rated at 3.7V. Google says it may be 4.2V at fuld charge but it drops to the nominal voltage 3.7V.

Please peruse this:

https://batteryuniversity.com/article/bu-501a-discharge-characteristics-of-li-ion

LiIon/LiPoly batteries typically get charged top 4.2V (some are 4.3V) with 4.25V (4.35V) being "ABSOLUTE MAX" at room temperature.

Voltages higher than that risk what is euphemistically called "rapid unintended disassembly" and / or fire.

These batteries drop voltage as result of discharge. Typically 3.2V is considered a safe discharge limit.

So a fully charged LiIon/LiPoly battery will have 4.2V and will drop to ~ 3.2V at 0%. The midpoint between full and empty is 3.7V, so the battery is rated as "3.7V".

Also this is the battery inside diablo. I see no 4.2V

Did you charge the battery to full and measure the voltage?

As for ipurifier3 vs ideon 3R Renaissance mk2 I tested them both. No competition Ideon slays the ipurifer, which is not a full unaltered signal for my ears.

I am unsure what "not a full unaltered signal" means.

The iPu is in effect a single port USB hub with a relatively high quality "Phy" which dramatically reduces the risk of errors in the USB Audio stream.

The 3R is acting in exactly the same way. It uses a different brand USB Hub Chip and has a separate power supply, while iPu (intentionally) run's the Bus power.

A very superficial and further dumbed down from my original version coverage of the subject is here:

https://www.audiocircle.com/index.php?topic=148763.0

It was originally published on "Michael Lavorgnia's "Audiostreams".

A very simple one liner is that USB Audio is like CD. Audio is a continuous stream, data is either correct and flagged as "good" or not. Any bad data is first attempted to be restored from the very limited redundancy and then error concealment is used, that is missing data is "straight line" interpolated, if enough data is lost the audio stream mutes.

This mechanism is the main cause for differences in perceived sound quality from different USB cables etc. The CD original is also behind sonic differences from sometimes questionable CD treatments. In both cases the "fancy" items make the situation objectively worse, but have been preferred sonically by some.

Thor

 

[OCC7N]

Please peruse this:

https://batteryuniversity.com/article/bu-501a-discharge-characteristics-of-li-ion

LiIon/LiPoly batteries typically get charged top 4.2V (some are 4.3V) with 4.25V (4.35V) being "ABSOLUTE MAX" at room temperature.

Voltages higher than that risk what is euphemistically called "rapid unintended disassembly" and / or fire.

These batteries drop voltage as result of discharge. Typically 3.2V is considered a safe discharge limit.

So a fully charged LiIon/LiPoly battery will have 4.2V and will drop to ~ 3.2V at 0%. The midpoint between full and empty is 3.7V, so the battery is rated as "3.7V".



Did you charge the battery to full and measure the voltage?



I am unsure what "not a full unaltered signal" means.

The iPu is in effect a single port USB hub with a relatively high quality "Phy" which dramatically reduces the risk of errors in the USB Audio stream.

The 3R is acting in exactly the same way. It uses a different brand USB Hub Chip and has a separate power supply, while iPu (intentionally) run's the Bus power.

A very superficial and further dumbed down from my original version coverage of the subject is here:

https://www.audiocircle.com/index.php?topic=148763.0

It was originally published on "Michael Lavorgnia's "Audiostreams".

A very simple one liner is that USB Audio is like CD. Audio is a continuous stream, data is either correct and flagged as "good" or not. Any bad data is first attempted to be restored from the very limited redundancy and then error concealment is used, that is missing data is "straight line" interpolated, if enough data is lost the audio stream mutes.

This mechanism is the main cause for differences in perceived sound quality from different USB cables etc. The CD original is also behind sonic differences from sometimes questionable CD treatments. In both cases the "fancy" items make the situation objectively worse, but have been preferred sonically by some.

Thor

Ok Thanks! I got it now! - Im wondering what Diablo sees the full charged states at before it bypass...4.2V?

When the Ideon mk2 is paired with 9V psu it also blows the ER isoregen out of the water, confirmed in audio session with friends.

-

I don´t think you are following me on this one.

I want to use 3.7V as nominal voltage. Im eliminating the USBC 5V input. The unit is not going to reverse current/charge back into the 3.7PSU.

Which means there are not point of 4.2V or 4.3V...as its operating at the nominal 3.7Voltage. I could use whatever psu between 3.2 - 4.2 and the diablo would not start to charge because that port is not connected

 

[Thorsten Loesch]

Ok Thanks! I got it now! - Im wondering what Diablo sees the full charged states at before it bypass...4.2V?

No, not how it works.

The system has three mosfet switches. Here in red (this is not the actual chip used, but a similar one that has good illustrations):

Power to the system is from PMID (after C2).

VIN has the USB-C power on it.

You see two switching transistors connected to "SW" and then the inductor L1 which is the switching inductor which acts as temporary energy storage for the switching system. The battery is connected on the "far" side of L1.

Whenever the system is on and external power is applied, the left most switch is on and power flows from VIN to PMID and power the system. The voltage is 5V nominal (actual 4.7-5.2V depending on load and quality of external PSU).

If the system determines that the battery needs to be charged, the two mosfet switches on the right start switching alternately at around 500kHz -1.5MHz switching frequency, sending energy into the inductor and then allowing the energy to pass into the battery. The precise switching timing of the two transistors determines how much current flows.

Whenever the system determines that the battery is in the required charge state (this is a complex logic that includes the "desktop mode" and also depends on temperature, so there is no simple single answer), both of the transistors on the right are turned off. The battery is completely disconnected from PMID and is in effect completely removed from the system.

If no external power is provided and the system is on, the transistors on the right start switching and are controlled to give 5.1V on PMID. This is a switching system that creates noise not present if the battery charged and external 5V is supplied. The battery power is never used directly, the battery is always connected to the system through a switching DC-DC converter.

When the Ideon mk2 is paired with 9V psu it also blows the ER isoregen out of the water, confirmed in audio session with friends.

I am always somewhat leery of excessively casual listening tests. If you want to compare things it is always important to exclude other variables. One of these variables mis expectation bias.

In the end, as long as the result makes you happy, your listening session is valid and the outcome is fine. If you want to extrapolate to anybody else you have to operate to higher levels of evidence.

I don´t think you are following me on this one.

There is nothing to follow.

I want to use 3.7V as nominal voltage.

And I am telling that if you want this to work properly you should use 4.2V, not 3.7V and EVEN BETTER, if you want the optimum result, you should NOT supply power via the battery connection and use the USB-C input instead, as this will create a system with less noise than using the battery connection to provide power.

Im eliminating the USBC 5V input.

Bad idea.

The unit is not going to reverse current/charge back into the 3.7PSU.

Ok. And you have a current limit on the switcher that takes the 3.7V (or 4.2V) and turns them into 5V, if you provide higher input voltage than 3.7V, you will have more power output on the 5V system feed.

But the best solution is to leave the battery alone and supply 5V to USB-C as this eliminates the switching mode converter linked to the battery entierly.

Which means there are not point of 4.2V or 4.3V...as its operating at the nominal 3.7Voltage. I could use whatever psu between 3.2 - 4.2 and the diablo would not start to charge because that port is not connected

Correct. And if you use 3.2V you would have the lowest amount of power actually available at the 5V system feed, 3.7V would give more power and 4.2V would be the most power.

Using 5V would be even better, however that would trigger battery overvoltage protection.

BUT you can use 5V (which will give the most available power at given ampere rating AND normally have the switching circuitry linked to the battery OFF and thus less noise) by supplying 5V on the USB-C connector with the adapter I showed.

Your choice of 3.7V via the battery connector is based on reasoning that is not correct in reality.

Now you insist to do what you want to do and you want to convince me that it is a good idea. As the guy who designed the product you are talking about I am telling:

1) This is a BAD idea from any angle
2) It will not actually produce the result you expect
3) YOU WILL get the result you expect (maximum output power, lowest noise etc.) if you instead supply 5V via USB-C

I will leave it at that.

Thor

 

[OCC7N]

No, not how it works.

The system has three mosfet switches. Here in red (this is not the actual chip used, but a similar one that has good illustrations):


You see two switching transistors connected to "SW" and then the inductor L1 which is the switching inductor which acts as temporary energy storage for the switching system. The battery is connected on the "far" side of L1.

Whenever the system is on and external power is applied, the left most switch is on and power flows from VIN to PMID and power the system. The voltage is 5V nominal (actual 4.7-5.2V depending on load and quality of external PSU).

If the system determines that the battery needs to be charged, the two mosfet switches on the right start switching alternately at around 500kHz -1.5MHz switching frequency, sending energy into the inductor and then allowing the energy to pass into the battery. The precise switching timing of the two transistors determines how much current flows.

Whenever the system determines that the battery is in the required charge state (this is a complex logic that includes the "desktop mode" and also depends on temperature, so there is no simple single answer), both of the transistors on the right are turned off. The battery is completely disconnected from PMID and is in effect completely removed from the system.

If no external power is provided and the system is on, the transistors on the right start switching and are controlled to give 5.1V on PMID. This is a switching system that creates noise not present if the battery charged and external 5V is supplied. The battery power is never used directly, the battery is always connected to the system through a switching DC-DC converter.



I am always somewhat leery of excessively casual listening tests. If you want to compare things it is always important to exclude other variables. One of these variables mis expectation bias.

In the end, as long as the result makes you happy, your listening session is valid and the outcome is fine. If you want to extrapolate to anybody else you have to operate to higher levels of evidence.



There is nothing to follow.



And I am telling that if you want this to work properly you should use 4.2V, not 3.7V and EVEN BETTER, if you want the optimum result, you should NOT supply power via the battery connection and use the USB-C input instead, as this will create a system with less noise than using the battery connection to provide power.



Bad idea.



Ok. And you have a current limit on the switcher that takes the 3.7V (or 4.2V) and turns them into 5V, if you provide higher input voltage than 3.7V, you will have more power output on the 5V system feed.

But the best solution is to leave the battery alone and supply 5V to USB-C as this eliminates the switching mode converter linked to the battery entierly.



Correct. And if you use 3.2V you would have the lowest amount of power actually available at the 5V system feed, 3.7V would give more power and 4.2V would be the most power.

Using 5V would be even better, however that would trigger battery overvoltage protection.

BUT you can use 5V (which will give the most available power at given ampere rating AND normally have the switching circuitry linked to the battery OFF and thus less noise) by supplying 5V on the USB-C connector with the adapter I showed.

Your choice of 3.7V via the battery connector is based on reasoning that is not correct in reality.

Now you insist to do what you want to do and you want to convince me that it is a good idea. As the guy who designed the product you are talking about I am telling:

1) This is a BAD idea from any angle
2) It will not actually produce the result you expect
3) YOU WILL get the result you expect (maximum output power, lowest noise etc.) if you instead supply 5V via USB-C

I will leave it at that.

Thor

Let me ask this shortly wihtout Im running in circles with questions.

how is it possible to make diablo work without battery?

EDIT: iFi has no intention obviously releasing a pure dac in the mid-fi.

 

[Thorsten Loesch]

Let me ask this shortly wihtout Im running in circles with questions.

how is it possible to make diablo work without battery?

It is not possible. The way the Diabolo is designed, it needs the battery because without a battery present it will presume a damaged battery and keep the system off (correct behaviour for product safety).

BUT, if you take advantage of what is called "desktop mode", the battery will sit idle charged to around 70% and this is equivalent to sitting on the shelf. The "nominal" battery life is 70% capacity, which is 5 years at 25-30 degrees. Note, at this point the battery still works fine, just capacity is lost.

We have not seen LiIon / LiPoly batteries for long enough to get an idea when the battery becomes completely unusable, if not cycled. So in practice, just make sure the 5V on the USB-C are always on and turn the Red Label / Diablo off/on on the front and you can forget about the battery. By the time you need to think about it again, I doubt you will still use the Red Label / Diablo.

EDIT: iFi has no intention obviously releasing a pure dac in the mid-fi.

I cannot comment on iFi's intentions. The Diablo is NOT a pure DAC. If you need a line out DAC, then the Diabolo is not a great choice, IMHO.

I have commented before on the Zen DAC Sig and Zen CAN Sig. With better power supplies these make a good desktop system. real output power is comparable to the Red Label / Diablo (with enough power from the power supply), most other elements are also similar/identical.

Thor

 

[OCC7N]

It is not possible. The way the Diabolo is designed, it needs the battery because without a battery present it will presume a damaged battery and keep the system off (correct behaviour for product safety).

BUT, if you take advantage of what is called "desktop mode", the battery will sit idle charged to around 70% and this is equivalent to sitting on the shelf. The "nominal" battery life is 70% capacity, which is 5 years at 25-30 degrees. Note, at this point the battery still works fine, just capacity is lost.

We have not seen LiIon / LiPoly batteries for long enough to get an idea when the battery becomes completely unusable, if not cycled. So in practice, just make sure the 5V on the USB-C are always on and turn the Red Label / Diablo off/on on the front and you can forget about the battery. By the time you need to think about it again, I doubt you will still use the Red Label / Diablo.



I cannot comment on iFi's intentions. The Diablo is NOT a pure DAC. If you need a line out DAC, then the Diabolo is not a great choice, IMHO.

I have commented before on the Zen DAC Sig and Zen CAN Sig. With better power supplies these make a good desktop system. real output power is comparable to the Red Label / Diablo (with enough power from the power supply), most other elements are also similar/identical.

Thor

The battery is the LPSU. How can the Diablo see its missing its battery when its feed the necessary voltage?

 

[Thorsten Loesch]

how is it possible to make diablo work without battery?

Actually, there is a way. You need to in effect "fake" a "full battery". So you need a PCB PCB that holds a regulator that provides 4.2V from the 5V input and 5V input via USB-C.

The battery is the LPSU.

The LPSU is NOT a battery.

Yes, feeding 3.7V to the battery connection will kind of work. Kind of, because you should use 4.2V at least, to maximise the system's power. There is also a thermistor in the battery that you would have to emulate with a fixed resistor of the correct value.

By doing this you force the circuit to run the switching DC/DC converter to take your 4.2V (or 3.7V) to turn it into 5V. So using a high quality LPS is pointless.

If you want to power the system from the LPS and not from a switcher, you need to do it from the 5V USB-C charging port and either a have a charged battery in the Diablo or a "fake" battery.

Maybe I misunderstand your goal.

My presumption is that you want maximise sound quality and power available by optimising the power supply.

If you just want to somehow make it work without battery, with no consideration of sound quality and without maximising output power, what you propose will work (with the Thermistor Caveat) to make the unit power up and operate. Will this deliver the maximum sound quality and power? No.

Thor

 

[OCC7N]

Actually, there is a way. You need to in effect "fake" a "full battery". So you need a PCB PCB that holds a regulator that provides 4.2V from the 5V input and 5V input via USB-C.



The LPSU is NOT a battery.

Yes, feeding 3.7V to the battery connection will kind of work. Kind of, because you should use 4.2V at least, to maximise the system's power. There is also a thermistor in the battery that you would have to emulate with a fixed resistor of the correct value.

By doing this you force the circuit to run the switching DC/DC converter to take your 4.2V (or 3.7V) to turn it into 5V. So using a high quality LPS is pointless.

If you want to power the system from the LPS and not from a switcher, you need to do it from the 5V USB-C charging port and either a have a charged battery in the Diablo or a "fake" battery.

Maybe I misunderstand your goal.

My presumption is that you want maximise sound quality and power available by optimising the power supply.

If you just want to somehow make it work without battery, with no consideration of sound quality and without maximising output power, what you propose will work (with the Thermistor Caveat) to make the unit power up and operate. Will this deliver the maximum sound quality and power? No.

Thor

Look @Thorsten Loesch I respect if the Diablo design is 100% yours, why did you use 2 Burr browns?

Every other dacs are 1 BB up to the iDSD Pro and Signature......and...the Diablo sounds much better than iDSD Pro(non signature) to my ears, so I dont understand why you also recommend me the ZENs

 

[Thorsten Loesch]

Look @Thorsten Loesch I respect if the Diablo design is 100% yours, why did you use 2 Burr browns?

Because that was used in the unit the Diablo was based upon, the original iDSD.

Every other dacs are 1 BB up to the iDSD Pro and Signature......

The iDSD Pro actually uses four DAC's. The iDSD micro/Black Label/Signature/Finale use two DAC's.

There is no real change in sound quality, going from one to two and four DAC's in parallel, the numbers get better.

and...the Diablo sounds much better than iDSD Pro(non signature) to my ears

To my ears it is the product I like the least out of all I designed. Perhaps we have differing tastes?

so I dont understand why you also recommend me the ZENs

I recommend to have a listen with improved power supplies.

The Zen DAC shares the analogue stage with Diablo and iDSD Neo but as you noted has only a single DAC Chip.

The Zen CAN is closer to the Diablo Headphone Amp than to the one in iDSD micro. It's probably half way in between iDSD Pro headphone out with more power and the Diablo.

Otherwise the Zen SIGNATURE series has the same level of passive parts and the same circuitry, active parts etc that you fin in the Black Label / Diablo.

Of course, no need to listen, if you already hate the idea. I linked this before, it is VERY important reading:

https://tu-dresden.de/mn/psychologie/ifap/kknw/die-professur/news/we-hear-what-we-expect-to-hear

So in any way, suit yourself, do as you please. I have given you the best advice I can give. Take it for exactly what you paid for it.

Thor

PS, a drool old story goes like this....

A long time in India, an enlightened Guru tried to teach the people. He was giving his teachings for free, begging for food, his clothes were cheap and in a bad state. Almost nobody listened to him. Of the few that did listen, even fewer put what they had been thought into practice.

Only one of the Guru's students did follow everything and was well on the way to enlightenment. This student had been a con man before and had grown rich, but he no longer wanted material things. He begged the Guru to take his advice. He told the Guru - "Baba, you teach for free. People do not value that which they get for free. People listen to you but do not follow your wisdom, for they do not value it. They look at you, they see you poor and think 'What could this poor beggar possibly offer me?' So people do not value you and your gifts."

The Guru asked: "What should I do?"

The student, the former con man told the Guru: "You need dress richly, wear much gold, you must build a richly decorated stand that will make you look rich and important. And then you must charge those you teach a lot of money to teach them. Charge each according to their means, but make sure you charge enough that person feels pain when they hand over the money. Then they will value your counsel and wisdom and will implement your teachings. I have much money I will give you, I now just want to reach enlightenment."

The Guru agreed and both disappeared for a few days. When the Guru returned, he was dressed more richly and wore more expensive jewels the even the Raja. Before teaching, he would ask people to pay him, upfront and so much that they felt the pain handing over the money. The Guru became greatly famous, many learned from him and took his wisdom to heart. The Guru also became fabulously wealthy and build a huge golden temple for his many disciples to worship in, after contributing gold to make the temple even larger. Pilgrim's valued greatly the opportunity to worship in the golden temple and gave much.

 

[OCC7N]

Well you made your point and made me cancel my orders on every iFi devices. Thanks