Thought: Using headphone dongles as speaker amplifiers to make high quality audio with nothing else but a phone?

[RogersCompactMonitorFan]

Hello,

I've been reading the thread "JCMALLY JM20MAX Headphone Dongle Review" - https://www.audiosciencereview.com/...cally-jm20-max-headphone-dongle-review.62260/

The specs on the box state output power of 192mW into 32 ohms. I was curious - would it be feasible to use this as an audio amplifier to drive a speaker? Not a super loud party amplifier, but perhaps one for listening to music relatively quietly but at a very high quality?

My calculations suggest it could drive my speakers to produce an output level of 77.6dB at 1 meter. This sounds loud enough.

I saw reports on the thread of the unit running hot. Maybe it would need physical modifications/extra heatsinks.

Question: Have I gone horribly wrong in my calculations, am missing some real world amplifier design constraint (e.g. heat dissipation) - and is this a silly idea? Would it be worth purchasing one and soldering up a 3.5mm to banana plug speaker cable lead to try? It would be cool to have a high quality HiFi from nothing but a phone and a dongle. If not this dongle - perhaps there are others?

My calculations:

1) The rated power into 32ohms is 192mW. Find the supply voltage:

P=(V^2)/R
=> 192=(V^2)/32
=> V=sqrt(P/R)
=> V=sqrt(0.192/32)

So the supply voltage equals 2.45v. This is about half USB supply voltage, so sounds feasible.

2) Given a supply voltage what is the maximum output power?
Max power at 2.45v into 8 ohms:

P=(V^2)/R
= (2.45**2)/8
= 750mW

3) But is the power limited by the current supply?

The box says "Op-amp: SGM8262"

The data sheet: https://www.sg-micro.com/rect/assets/e637497c-405d-45a6-88dd-21c16c22188f/SGM8262-2.pdf says:

peak current into 1ohm at 2.25v is 200mA
Power=I^2*R
=0.2*0.2*8
=320mW

So I believe this amplifier will give 320mW into 8 ohms.

4) How loud is this power?

My Focal 906 speaker sensitivity is 89dB at 1W at 1m

Calculate the ratio in dB of 1W to 320mW

sensitivity_measured_at=1000
output_power=320
ratio=10*math.log(sensitivity_measured_at/output_power)
=10*math.log(1000/320)
=11.4 dB

sensitivity=89
output_level=sensitivity-ratio
=89-11.4

Sound output level will be 77.6 dB at 1 meter

 

[amirm]

Focal's minimum impedance is 4.5 ohm, not 8:

The dongle will either get very hot, shut down or both. I have a USB powered sound bar. It has tiny 3 to 4 inch speakers. It will play reasonably well but turn it up and it will shut down.

There is no harm in trying but I suspect it will not be usable.

 

[DVDdoug]

There are lots of powered monitors that will work from a line-output or headphone output, but they are not USB powered. Maybe some have USB inputs. You can also find some with USB or Bluetooth. Or there are powered PA speakers that you could drive with just a phone.

 

[RogersCompactMonitorFan]

Thank you for the replies.

I originally thought that this experiment could be an cheap+easy way to hear what a "very high performance" amplifier would sound like - and see if I can hear any difference compared to the TPA3255 Aiyima A07 amplifier I'm using. This would be an easy way to see if an amplifier upgrade would be worthwhile.

I forgot 8 ohms was "nominal" so thanks for pointing this out. Doh. I suppose audio science is all about having theories and then having them disproven after analysis/experiment!

So into 4.5 ohms max power will I think be 180mW.

I made a second mistake (doh again) in that output current is 200mA peak rather than constant. So to derate this to RMS I assume would be 200/sqrt(2) = 141mA

This (I think) into 4.5 ohms will be 89mW.

I thought perhaps other "power op-amps" could provide similar performance, but this is essentially what the LM1875 or LM3886 amplifiers are.

I'm still tempted to research single ended class A amplifier designs to try to see what performance can be achieved very cheaply but for low power output. I don't know what single ended designs with minimal transistor count can do.

The designs I found online don't have accurate measurements so it is hard to know if they are worthwhile vs the amplifier I have. Most measurements of distortion just state "below 0.1%".

Perhaps the TPA3255 already exceeds the threshold of hearing (at least for low output volumes such as I'm talking about) making this irrelevant however.

 

[Rja4000]

Focal's minimum impedance is 4.5 ohm, not 8:

The dongle will either get very hot, shut down or both. I have a USB powered sound bar. It has tiny 3 to 4 inch speakers. It will play reasonably well but turn it up and it will shut down.

There is no harm in trying but I suspect it will not be usable.

Maybe by using 2 of them in serie...

 

[Marc v E]

Thank you for the replies.

I originally thought that this experiment could be an cheap+easy way to hear what a "very high performance" amplifier would sound like - and see if I can hear any difference compared to the TPA3255 Aiyima A07 amplifier I'm using. This would be an easy way to see if an amplifier upgrade would be worthwhile.

(....)

Perhaps the TPA3255 already exceeds the threshold of hearing (at least for low output volumes such as I'm talking about) making this irrelevant however.

Yes, it's beyond the threshold of hearing ime and imo.

What is worthwhile is using eq to let your speakers sound good in your room. See REW and try something relatively cheap like eq on your phone or a cheap dsp to see the results.

If you want to know what hi-end sounds like I would recommend to buy a cheap but well measuring in ear and use it with your phone.

Imo in the end it's mostly the speakers, sub and eq that counts. The last because the room interacts with the sound coming from the speakers. Just my 2 cents.

 

[olieb]

There are several things with this calculation.

1) The rated power into 32ohms is 192mW. Find the supply voltage:

P=(V^2)/R
=> 192=(V^2)/32
=> V=sqrt(P/R)
=> V=sqrt(0.192/32)

So the supply voltage equals 2.45v. This is about half USB supply voltage, so sounds feasible.

But this will give you RMS voltage. The peak voltage (which the supply must deliver) is 1.41 times this value.

For the efficiency calculation you use the natural logarithm but the formula is for log to base 10. So the "11.4 dB" result is not correct.

It is easier to skip 2) 3) and 4).
Instead:
192mW with 32Ω gives 2.45Vrms max output voltage.

The "efficiency" of 89dBspl at 1W power (8Ω) is in fact a sensitivity of 89dBspl at 2.83Vrms voltage.
Or 88dBspl at 2.5Vrms. (Or 80dBspl at 1Vrms)
So the dongle can play even louder (88dB peak with 2.5V). But do not forget this is the peak SPL with 0dBFS. Average loudness (that is what your SPLmeter will show you) will be much lower, how much so will depend on programme.
But it will work, well until it gets hot and/or protection sets in.

An open question is at what point distortion/clipping/protection will set in with such a low impedance.
If the impedance (for a specific music impulse) is actually more in the range 4.5Ω then the 200mA peak current will only produce 900mV peak or 630mVrms.
This is about 4 times lower than calculated before and this would reduce the SPLpeak by additional ≈12dB (if the impulse energy is centred in the regions of low impedance). Non resistive load is another story still.

 

[Rja4000]

By filtering hard below, say, 50Hz, you may still keep impedance at a reasonable level with the Aria 906, I suppose.
I'm curious, so I'll give it a try

 

[RogersCompactMonitorFan]

Thank you for the correction of peak vs RMS. I'll stick to DC only in future... No peak/RMS to get wrong ;-)

Rja4000 - I'm interested to hear how you get on.

I have been enjoying the sound of the JM20 dongle with my Sennheiser HD25 headphones. I've no idea how these headphones measure, but they sound good to my ears.

Room correction is on my list of things to try, so sounds like I should prioritise this rather than messing about with amplifiers - thanks for the tip.

 

[Rja4000]

So I did.
OK, not with the one you mentioned but with what I have at hand.

I did drive, for 1/2h now, one of my Aria 906 from an E1DA #9039S.

This small dongle Ivan (@IVX) sent to me for measurements - sorry, it's long overdue- was measured by Archimago at THD+N below -124dB at 16.2 ohm 3.3Vrms (around 0.7W).
I personnaly measured mine at over 125dB SINAD.

And it worked. Even without filtering out lowest frequencies.
I don't have any way to measure SPL at hand, but from one single speaker, the level was certainly enough to listen at 1m-1.5m.
That is with keeping the output level around -3dB.
From time to time, I pushed it to 0dB.
All this from my Samsung S23 ultra as sole power and music feed.

Yes, the device is warm, but it's been worse during measurements (which is normal - we measure by pushing the limits)

 

[boxerfan88]

The specs on the box state output power of 192mW into 32 ohms.

That pretty much says it’s rated to 32ohms only, anything lower isn’t guaranteed.

If you’re assuming the power will scale linearly down to 16ohm / 8ohm / 4ohm, me thinks impossible.

Many proper big case amplifier rated at 8ohm doesn’t scale linearly down to 4ohm/2ohm. Expecting this tiny device to scale linearly down to 8/4ohm is wishful thinking. In any case, no harm buying one to try … it’ll probably clip quite easily.

 

[RogersCompactMonitorFan]

Rja4000 - Interesting - thanks for trying. Was there a reason you used only one speaker rather than two? It shows at least my idea wasn't completely silly. It looks like the E1DA device you used has a higher output power than the JC20 Max I was thinking about.

boxerfan88 - Do you know what the limitation is that prevents a design scaling down to a lower impedance? Is it a current limit?

 

[Rja4000]

Was there a reason you used only one speaker rather than two?

Yes. Lazyness.
You know, this mindset... which lead the man to invent the wheel...

 

[RogersCompactMonitorFan]

That's a fair reason ;-)