I'm starting a new thread on this because it was going off topic in the thread where it started.

I was impressed by the little panel mounted mp3 player that ARF member Easy had built into his AM transmitter:
viewtopic.php?f=12&t=132867&start=185



A couple of weeks ago, I was on the DealExtreme.com site looking for some other electronics, and ran across some similar units. I ended up buying this one for $7.80 and free shipping:
http://www.dealextreme.com/p/1-5-led-mp ... 3-7v-94814

This one has a red LED display, an FM tuner built in, and plays mp3 files from either SD memory cards or USB memory sticks. It also has a power amp built in and can drive speakers directly with reasonable volume. Also, it comes with a remote control.

Here is a picture of it running off of three AAA cells, and driving a pair of Minimus 7 speakers:


And a rear view showing the connections.

I added a connector for the FM antenna input, and soldered the audio line-out cables directly to the board. The power input and speaker outputs had factory installed connectors.



Dimensions are as follows:
Panel cutout: 75mm (2.95") wide, 27mm (1.06") high
Bezel: 93mm (3.66") wide, 35mm (1.38") high
Depth: 40mm (1.57")

As for audio quality, I noticed no distortion.
For noise, if I put my ear right up against the speaker (with no audio playing) I could hear a very faint amount of hiss. In other words, no audible noise under normal conditions.
The only fault I could find in the audio quality is some bass rolloff below 100Hz. I'll need to hook up a signal generator and scope to find out exactly what frequency where the rolloff starts. For audio going into a transmitter, this rolloff may be a good thing.

It does have six selectable equalization settings as follows:
1. Flat...ish
2. Tinny
3. Tinnier
4. Tinniest
5. Muffled
6. Rock

I added the "...ish" to the flat setting due to the bass rolloff. All settings seem to have the same bass rolloff regardless of the rest of the response curve. Because of the small power supply, it's probably not capable of delivering any bass even without the deliberate rolloff, and since it was likely intended to be matched up with cheap tiny speakers, deep bass, if present, would just rattle the speaker cones. Unfortunately, the same equalization applies to the line outputs, and so the bass rolls off there too.

On the subject of the line outputs, these are considerably lower than normal line level, and so would require some additional amplification before going into a power amp. A small opamp signal booster with some bass boost would likely cure all the shortcomings.

As for the comments by other users on the sound quality, it's hard to tell without knowing exactly what they were doing with the unit. For example, the power amplifier uses a bridge output. If someone tried to connect one side of the speaker connection to a common ground (eg., 3-wire stereo headphones) there's no telling what it would sound like.

This unit came with zero documentation, but the support forum on the DealExtreme site had enough useful info to get me started, and I've been documenting things as I make new discoveries. Mainly, I've been sorting out the remote control and all the operational features. With no documentation, it takes a bit of time to figure out how to navigate the tunes on the memory card if they're organized into sub-folders.

I'll measure the current draw and post the results later.

Super. Thanks for the info.

It has a power amp? You'd think the Chinese might find that worth mentioning. What is it, 100 mW?

I imagine the line outs are like cell phone outs, say, .5V because of the "3.7V" spec (sounds like catering to a lithium Ion battery).

I was thinking about a Granco UHF converter 'dial window' and that 1.38" might make it, although it may be .05 too big. Could probably shave one or the other down a hair.

Power draw varies with whichever input mode is selected:
Soft power off (Standby) mode: 16mA
AUX input: 53mA
FM tuner: 60mA
SD memory: 60mA
USB memory: 120mA
This is with speakers disconnected, or volume muted.

When driving speakers at max output, the current consumption is over 250mA. This is probably limited by my AAA battery supply.

It appears the power amplifier is rated at 3 watts when driving a 4 ohm speaker from a supply voltage of 5V. (See datasheet below.)

Some part numbers:
Main controller & MP3 decoder: GPDS209A-001D (possibly Amtel clone)
On board EEPROM memory: 24C02
FM tuner: Beken BK1080
Audio Power Amp (2 ea.): Power Analog PAM8002

Datasheets:
FM tuner - http://www.datasheet4u.net/download.php?id=688279
Audio Amp - http://www.poweranalog.com/pdf/PAM8002.pdf

Schematic diagram for a virtually identical MP3 player:
http://www.chipkingdom.com/upfile/20111211712780010.pdf
I say virtually identical, because I did a random check of wiring and they match in every case. Only difference is that some component numbering is different, and they label the controller chip part number as CECL08D.

It appears from the diagram, that many additional function buttons can be added to the main unit simply by adding pushbuttons and the correct resistor value. I tried this and it worked!

Unfortunately, the only thing that appears to be missing, functionally, is a shuffle play mode.

Terrific stuff.

Well of course, a class D power amp. The current being pulled must reflect an average program level since it's only about 1 1/8 watt total. I suppose one could put some mini speakers in the transmitter and use it as a sort of 'local monitor'.

Using the FM tuner and rebroadcasting to AM radios is kind of an amusing thought.

Which 'function button' did you add?

I added the "Play/Mode" function button. Since there's already a Play button and a Mode button, I thought Play/Mode may be something that allows you to select shuffle play. But no, it just combines the existing Play and Mode functions into a single button. Pressing it briefly gives the play function, and pressing and holding for a couple of seconds gives Mode function (input source selection).

Meanwhile I confirmed with scope and signal generator what my ears already told me. There is definitely bass rolloff and the corner frequency is about 120Hz. It's not very steep, so a simple bass boost network in the add-on preamp should fix it. I'll attack that tonight.

And, you were correct about the line level voltage. It appears to be about 0.3 Volts. So, it won't take much gain to fix that. In fact this may be enough signal to drive my transmitter directly.

I perused that schematic you posted and don't see any 'circuit' reason why it would roll off. I mean, line in is 10uF into a series 200k, ridiculously low, and line out is 10uF, which would be fine into a 10k load. Were you measuring line out or speaker out?

It kinda bugs me needing to 'fix' what shouldn't be broke but you're probably right that they 'expect' these things to be mated up to dinky speakers. But it would have been nicer if they'd just lowered the 'power amp' blocking caps rather than ruin line out with what must be a 'firmware' roll off.

Is that nominal or peak? If nominal it sounds like too much as nominal for 'consumer' 2V is 316mV but, of course, that depends on crest factor. I mean, 2V comes to 5.7 Vp-p and you just can't do it single ended with a 3.7V battery, or 5V either. Was volume maybe up too much and clipping? Or was that the 'most' you could get out driving it with a tone generator?

By comparison, I have a notebook which seems to be 1V because it goes into clipping right about 1.6 Vp (3.2 Vp-p), or so, just a bit over the 1.414 Vp 1V would max out at, so it's probably 5V powered inside. That'll drive my transmitters but cell phones are noticeably weaker, which is why I presume they're more like .5V, and that seems to hold for small MP3 players too, like the Nano. That sort of makes sense because discharge volts on a 3.7 V Lithium Ion is roughly 2.7 V and you couldn't get 3.2 Vp-p out of it, especially since there has to be some headroom for voltage drop across the output drivers, so I'd expect nominal to run, let's be generous, around 100 mV or less.

Btw, they have another unit that looks identical for 50 cents more. I wonder what you get for the extra 4 bits. Speaking of which, the Chinese don't seem to have very strong grasp of 'marketing' because they tend to leave out little things like what the heck the product has and does.

Component-wise, it costs nothing to do the roll-off in firmware. And, it's trivial programming, because MP3 files (and I assume WMA files too) are stored as frequency domain samples rather than time domain. So, it's just trivial arithmetic to equalize/roll-off any arbitrary frequency. However, since it has six different equalization settings, it would have been nice for one of them to be totally flat.

The 0.3 volts was peak, and it was a rough estimation from the scope, so don't rely on it too heavily. Also, I used Audacity to create the test frequency track, and save to an MP3 file, which I then played back on the MP3 player. With Audacity, I used the default amplitude setting of 1 (no units given), which appears to provide maximum output level.



Well, if you have really deep pockets go for it. Yeah the marketing leaves a lot to be desired. I get the feeling they don't like to write anything down where it may come back to bite them. Just try to find a manufacturer's datasheet for any of these things.

Sure, it costs nothing in firmware but, in this case, it also costs nothing in hardware. 'Small' caps don't cost more than 'big' ones.

I don't know if it's a case of ignorant engineering, "everybody uses 10 uF so, duh," or 'firmware syndrome', "yippee, anything can be done in firmware." Maybe a bit of both because it's mind boggling to screw up line outs for the sake of speakers on the power amp. I mean, really now, did someone think it a 'feature' you can connect an external power amp, or anything else for that matter, and have it sound just as tinny as the dinky built in speakers?

I wonder if the 'car' modules are any better since maybe they figure there will be better speakers.

Ah, ok. Yes, if you did a "generate tone" in Audacity then the "1" means full scale, 0 dBfs. If you then tell it to 'normalize' the file, since default is 3 dB under peak, you'll see the thing snap to .707. The actual "volts" is whatever the DAC does. I.E. "1" will be 2 V for a 2 V DAC. Well, if mixer volume is also set to 0 dBfs.

Now, at 300 mV max, it sounds too low

I think you may have more trouble with that amplitude than seeing "300 mV" suggests because program nominal could easily be -12 dB, or less. Back in the day when 'Hi-Fi' still meant something I think it was usually around -18 dBfs, which came from .316V/2V being -16 dB plus, say, 2 dB headroom under the digital 0 dBfs brick wall. It depends on the material, of course, but you probably know this stuff better than I do.

Oh I know. I say the Chinese will never 'conquer the world' until they learn to write spec sheets, and there's a serious side to the joke. I mean, you can't make satellites, rockets, jet fighters, or any other complex system without well defined specs so somebody over there must know how. Maybe it's the same guy that can translate English because he sure as heck isn't working in the consumer end either.

But maybe we've discovered the 'secret' to their high volume: you have to buy one to find out if it's what you want to buy

Interestingly, if you restrict your search to "DIY Parts & Components" only two show up. I wonder if that indicates "not made for dinky speaker stuff."

Can you use a playlist file with this one?
That's the one feature I'd like to see in the one I used, an automated way of sequencing the MP3 files.

-Sam T.

I don't think so. I've never used an MP3 player that used playlist files, so I'm not really familiar with them.

It appears that the file functions in this unit are the bare minimum.

I built a preamp with a gain of 5 which seems to be plenty. I discovered after my previous testing, that my stereo audio patch cable apparently has a pair of 180 ohm load resistors built in, and this would have seriously affected the measurements. Anyway, the preamp has the needed bass boost network which fixes the bass rolloff.


I tested it with my current transmitter project, and discovered that whoever designed this MP3 player really liked to monkey with the equalization. The "Flat...ish" equalization setting is obviously not as flat as I'd thought. It now appears that there's some serious treble boost going on somewhere above ~6kHz, which won't be audible on most AM radios, but it's overloading my transmitter. When someone hits a cymbal or sings something with an "essss" it's like someone turned on a buzz saw. I tried different equalization settings, and naturally all of the various "Tinny" settings are bad. The "Muffled" setting gets rid of the problem, but not surprisingly, sounds muffled.

I dug out an old Radio Shack multiband equalizer, and hooked it up between the above preamp, and the transmitter. I set it to chop off everything above about 7500Hz. That fixed the "ssss" problem. So, it looks like I'll have to add a high filter to the preamp now. Am I thinking that this was a waste of $7.80? No, not yet, but getting closer.

On the subject of playlist files. I looked it up on Wiki, and figured out how to create one. I did this and tested it. As I suspected, it didn't do anything.

Well, fixing the hot high end was pretty simple. I just added a 560pF capacitor in parallel with the 47k feedback resistor, and things sound much better.

Weird. I'm trying to imagine what the heck they were thinking and all I can come up with is maybe it's designed to equalize some specific mini speakers?

It's interesting there didn't seem to be an audible high frequency problem running into your speakers.

I suppose that it probably was noticeable, except that I was blaming it on the base roll-off. It was easier to spot when I did an A-B comparison of a tune being played on the MP3 player and the original from the CD.

I was too lazy to find exactly where the treble is starting to boost, but just did a quick estimate based on the fact that the test receiver has an 8kHz bandwidth and the boosted treble was already noticeable. I should probably make another test recording with Audacity, and put it on the scope again, to check out the whole audio range. However, as it is now, the audio that ultimately comes out of the receiver is pretty good. So I'm happy with the results.

Makes sense.

I'm glad you got it working well enough and muchas gracias for all the great information on it.

Can you suggest an alternate IC for the LM3900. I think they are getting harder to find.

Ed.

Well, Mouser seems to have plenty.

As for something else, it depends on how fancy you want to get. A lot of folks use the LM358 but, IMO, it's not really suitable for 'Hi-Fi'. On the other hand, it's probably fine for a 5V powered MP3 player going into an AM transmitter, because you're not trying to swing 2V+ line levels out to 50 KHz, and they're cheap. Same reason I was going to use it in that solid state AM transmitter I never built. Btw, a gimmick I heard is to put a DC load on the output forcing the totem to operate Class A but I don't know if it really makes an audible difference.

On the 'high performance' end Burr Brown (owed by T.I.) makes the "Sound Plus" OPA2134 (dual), which will operate down to 5V, but those suckers are 3 bucks.

There are others, of course, and those are just two I knew of offhand.

The LM3900 is also sold under the LM2900 number which may be more common nowadays. Digi-key still lists them under both numbers I think. I bought a bunch just a few months ago. However, any regular op amp will work too. It just requires that the non-inverting input would be biased with a voltage divider to half the supply voltage, instead of the current bias that the LM3900 uses.

The LM2900/LM3900 are just different temperature ranges. LM2900 is 'industrial' and LM3900 is 'consumer'. That doesn't matter for your MP3 player and Mouser has 10 times as many of the LM3900.

Yeah, if it can operate down to 5V with still enough output swing. That's pretty common these days with so many battery powered devices but not all can operate that low.

How about the very well respected op-275
http://www.analog.com/static/imported-f ... /OP275.pdf
http://search.digikey.com/us/en/product ... -ND/820338

I'm currently using the op-275 here in my little 3 watt, home-brew (iPod/Mp3 player) PP stereo tube amp.
The one that's using 5902 sub-mini tubes.
Here's the thread on it:
viewtopic.php?f=6&t=182783