| theAnonymous1 |
Before I ask my questions, let me explain that I built this amp for my girlfriend who is hearing impaired. The amp was built to drive a pair of custom molded earbuds.
http://earplugstore.stores.yahoo.net/chmisearph.html
Specs: 25 ohm 115db/1mw (they don't seem that sensitive to me though)
The amp is the basic CMOY circuit using an OPA2134UA and +-9v supply. All components are SMD.
Pic
The amp works perfectly, but I am disappointed with the output level. I was hoping that with a +-9v supply and 25 ohm phones that I would get some serious output, but that just isn't the case. It plays loud, but not uncomfortably loud; which is what I was shooting for. I tried one of her hearing aids on and they get uncomfortably loud, so I know the amp needs more output.
So I guess my question is what can I do? Is there something that can be optimized in the circuit, or do I just simply need more voltage? This amp needs to be as small as possible so trying to get close to the +-18v limit of the opamp with batteries gets REALLY bulky.
Any thoughts or suggestions is greatly appreciated. |
|
|
| teemuk |
For a somekind of earphone system the power to 25 ohms obtainable from a +-9V supply should be more than enough. Have you measured how much power the amplifier really gives out? With that supply voltage an ideal amplifier should give out about 1,62 watts to 25 ohms. I'm not familiar with the chip you used but I gather that it's not actually a low power opamp. I checked out the datasheet quickly and it seems that 600 ohms is the lowest load that the concerned chip can cope with.
I know from experience that LM386 low voltage power amplifier chip running from +9V rail gives out about half watt of power to 8 ohms. With sensitive earpiece this would be enough to cause a series hearing loss for an average person. Same chip running from +16V rail would give about 1W to 32 ohms. The issue is definitely not about the rail voltage. How much power does your amplifier really put out? Does it collapse under a small load such as 25 ohms?
Anyway, for an earpiece system even less than hundred mW should be plenty.
If the amplifier is fine and can drive a 25 ohm load then I would suspect that the earpiece is faulty. |
|
|
| BWRX |
What is the maximum level of the input signal you're using and what is the gain of each channel?
The 2134 can do 35-40mA of output current, so it should be able to swing about 1Vp-p with that 25ohm load before giving up. At those levels distortion will probably be higher than you'd like. |
|
|
| theAnonymous1 |
The amp technically sees a 76R load, as there is a 51R resistor in series with the output.
The earbuds are brand new and there is nothing wrong with them.
Here is the design I followed......
Here is a link to the design page......
http://headwize.com/projects/showfi...e=cmoy2_prj.htm
I used the same values as above except R5 is 51R, C1 is 2uf, and I omitted R1. The gain of the amp is 11.
I'm not sure of the output level of the mp3 player I used, but I'm sure it's the same as the majority of players on the market. I can make the output of the amp clip, so input signal and gain seem plenty high enough.
This amp is definitely louder than an amp I recently built based on a TPA6111A2DR amp IC powered by a single 3.7v lithium cell. That amp was louder than the mp3 player, but not by much.
You have to remember that what you would consider excessive volume is what my gf would consider acceptable volume.
I don't understand how that little hearing aid of hers is so darn loud from that little tiny battery.:confused:
EDIT: Ok, I just measured the impedance of her phones and they are 155 ohm, not 25 ohm like the specs say. That is kind of frustrating, as I probably wouldn't have bought them if I knew that before hand. I will definitely be contacting the earplug store and asking why there is such a discrepancy in the specs.:mad: |
|
|
| raromachine |
I'm not sure, but I don't think DC impedance (measured with a DMM) will match the impedance for AC (your audio signal) 1 for 1. And the 25ohms will probably be between certain frequencies.
I'm suprised that with a gain of 11 you're not getting loud, whether it's into 25ohms or 115ohms.
Try removing R5 (it's optional anyhow) and seeing what happens?
Try some other headphones (i.e. 32ohm earbuds). |
|
|
| teemuk |
If something is loud that is usually due to efficiency of the speaker system not the power of the amplifier.
The chip you use has internal current limit of 40 mA so your amplifier can barely drive the 155 ohm earpiece. You don't need to have a load that is any lower value than this - your amp couldn't drive it.
To load of 155+51 ohms the amplifier's output power should be about 155 mW RMS. Have you verified this? If the earpiece's efficiency is indeed 115 dB @ 1mW (I doubt it) this should be uncomfortably loud. Note that an average human has a high risk of getting permanent hearing damage when exposed to sound pressures greater than 85 dB.
With all the math considered, something in your system does not work the way you want it to... Trying another earpiece might give you better results. I suspect that the SPL ratings of the one you tried and bought are bloated. |
|
|
| Nordic |
What source are you useing... the output impendance may also affect the output, but I think more likely you woud do good to add another 9V battery, and turn up the gain a bit..
I mention the impendance, because with that circuit, I have a portable FM radio that plays at lower volume trough it than without... Also an Ipod shuffle... healthy volume gain... Ipod nano smaller volume gain...
Those chips can do about 40mA output per channel (I doubt at 9V though)....
Rod Elliot has a nice headphone amp design with crossfeed... now that plays plenty loud.....
http://headwize.com/projects/showfi...e=opamp_prj.htm
Have a look at the high current buffers for ideas on modifying your cmoy to add more current... |
|
|
| carstereogeek |
I have used this circuit before to drive a pair of 16 ohm shure earbuds I have and I also had the same problem.
Do not remove R5,the circuit will go into oscillation with the low impedence phones(I used 47 ohms).
Replace R3 with a 1k trimpot(10 turn works best) so you can adjust the gain.When you find the gain you need,measure the trimpot and replace with a fixed value.
Hope this helps! |
|
|
| Pit Hinder |
Sounds sensible to me. Carrying a bleeping rucksack full of batteries might get her off music. And Anonymous - stand by your girl! Having people think she is stupid, not realizing that she just can't hear them...sheer hell. But ain't life hell for you, too? Kitten used to shout at me, and before I could shout back she took her hearies out.
Pit |
|
|
| theAnonymous1 |
| Wow, whats goin on with the forum lately. This is the second time I have lost a post from the forum being restored from before I posted. I don't really feel llike typing it all over again. |
|
|
| Pit Hinder |
| Yeh, problem here, too. My email tells me there are new posts, but you ain't there buddy! |
|
|
| theAnonymous1 |
Well I just got done talking to the earplug store and they are willing to remake the earbuds with the correct impedance. They offer a 19 ohm version and I think I will go for those. I just have to wait 2 weeks for them to be made.:mad:
I did some crude measurements of the amp. I tried the 155 ohm phones + 51R resistor, 155 ohm phones without the 51R resistor, my 35 ohms phones + 51R resistor, and my 35 ohm phones without 51R resistor.
No matter what the load was I could only get around 0.75v output with music before clipping. I could get almost 4.5v output with a 1k sine, but It's hard to tell when a sine is clipping without a scope. I think music is a more realistic test in this case anyway. I tried both the mp3 player and my PC as a source and got the same results.
I have some quad OPA4134UA opamps. Should I try the high current buffer setup shown in Fig. (a) below, or should I try something completely different.
Thanks again for everyones help. It's kind of hard being an audiophile with a hard of hearing girlfriend. I feel guilty that I spend so much time and money on these things and she can barely hear at all. I was hoping to bring her a bit closer to my world by sharing the joy of music through something other than her horrible sounding hearing aids.
To listen to music with her hearing aids on she has to turn them on "telephone" which picks up all sorts of hum and distortion. Awful. |
|
|
| BWRX |
| quote: | Originally posted by theAnonymous1
No matter what the load was I could only get around 0.75v output with music before clipping. I could get almost 4.5v output with a 1k sine, but It's hard to tell when a sine is clipping without a scope. |
Since you said you don't have a scope I guess it's safe to assume you used a multimeter set to AC volts to measure the 0.75V output while the amp was reproducing the 1khz sine wave. Don't forget that the AC volts setting on a multimeter measures rms voltage. So that 0.75Vrms you measured means you were seeing about a 2Vp-p sine wave from the 2134. That's seems lower than it should be if your load is actually 150ohms. Are you using batteries for the supply rails? If so, are they fully charged? |
|
|
| teemuk |
| quote: | Originally posted by theAnonymous1
I did some crude measurements of the amp. I tried the 155 ohm phones + 51R resistor, 155 ohm phones without the 51R resistor, my 35 ohms phones + 51R resistor, and my 35 ohm phones without 51R resistor.
No matter what the load was I could only get around 0.75v output with music before clipping. I could get almost 4.5v output with a 1k sine, but It's hard to tell when a sine is clipping without a scope. I think music is a more realistic test in this case anyway. I tried both the mp3 player and my PC as a source and got the same results.
|
Increasing the load should decrease the current demand and allow the amplifier to swing closer to rails. 40 mA max with 200 ohms should allow the output to swing to about max. +-8 Vpeak. If your signal clips as early as 0,75V even with this load then the circuit is faulty. Period. I still suspect that the 600 ohm load is the minimum for the concerned chip.
Did you feed the sine wave from the same source as the music thus overruling the possibility that there is somekind of an mismatch problem between the source and the amp? How did you measure the output amplitude of music without a scope? Can you trust these measurements to be exact?
You wanted to know if you should built something different. I believe you should. The way I see it, the whole CMOY design is horrible: Current limited opamp that is (in theory) cabable to drive a minimum of 200 ohms load. Simple LM386 chips can deliver higher currents to lower loads, work from single supply rail and likely fit to smaller space as well. |
|
|
| theAnonymous1 |
I measured the voltage with a Fluke 111 true RMS meter. The 1k sine and the music tested were both played from my laptop. The 1k sine produced 4.5Vrms from the headphone amp with the laptops output at near full volume. When I played music the amp would clip at anything over 0.75Vrms, which was at an input level far lower than the 1k sines max level. I don't have a scope so when I say it clipped; I mean audible clipping heard through the phones.
The batteries are new and fully charged. Here is a pic of the supply connected......
http://i19.tinypic.com/2afeu1j.jpg
Here is a hi-res close-up......
http://i18.photobucket.com/albums/b...1/cmoy004-1.jpg
The LM386 looks like it would work well with 2 9v batteries, but it requires large output capacitors and I would like to stay away from that option. |
|
|
| Nordic |
I'm bussy with the (a) one... just took board from etchant a few minutes ago... it is the datasheet headphone amp circuit for the opa2227... will let you know how it went...
It is basicaly the same, except the resistor to ground on the positive leg is removed...(internal) the 2 resistors on the output side is 200 ohms the gain setting resistors are both 1k.
Depending on the impendance of the source, I would use opa2227 for up to 20k, opa2277 for 20k - 100k and opa2132 for 100k+ sources...
Eagle files....
4cm x 8cm
 |
|
|
| dsavitsk |
You might try looking over at Headwize.com. There are lots of people with a lot of experience in driving headphones. But, the short answer is that I'd use something with higher current capability instead of the OPA. Try a NJR4556 or a LM6172, or try putting a BUF634 (or two in parallel) in the feedback loop of the opamp. The OPA simply has too little current to drive a low impedance load. I'd also remove the output resistor as it is not doing you any good.
If you want a readymade PCB, try the Pimeta at http://tangentsoft.net/shop/ It is not the end all of amplifiers, but it is a step up from a cmoy and will at least give you some ideas on how to drive these phones a little better. |
|
|
| theAnonymous1 |
I think I'll give the current buffer circuit a try after Nordic lets me know how it works for him. I already have some OPA4134UA so the only thing I'll be wasting if it doesn't work out is my time.
This CMOY doesn't sound bad, it's just not loud enough for this application. Maybe I'll throw it in a case and use it myself. It's definitely more powerful than my mp3 players output. |
|
|
| paulb |
| quote: | Originally posted by theAnonymous1
Well I just got done talking to the earplug store and they are willing to remake the earbuds with the correct impedance. They offer a 19 ohm version and I think I will go for those. I just have to wait 2 weeks for them to be made.
...
The LM386 looks like it would work well with 2 9v batteries, but it requires large output capacitors and I would like to stay away from that option. |
I would seriously consider using the LM386 or one of its newer cousins. Opamps (except some high power ones) will not drive a 19-ohm load properly. The parallelled version was published at headwize as intended for 68-ohm Grado headphones.
I've seen the LM386 used for a headphone guitar amp and results with a music source were quite acceptable. Not the ultimate in high-fi but still quite good. Not sure why you have a problem with large output capacitors, is it the physical size? If you can get to the 4 individual wires on the earbuds you could use a BTL arrangement. See, for example
http://www.national.com/pf/LM/LM4950.html |
|
|
| dsavitsk |
| quote: | Originally posted by dsavitsk
I'd also remove the output resistor as it is not doing you any good. |
I wanted to reiterate this. It is a simple fix, and I'd bet it will solve the problem. There is no reason to have a resistor in series with the output. People use it to remove some hiss that can occur with some chips, but a little hiss is likely not an issue for you.
Really, any of these chips will drive low impedance phones to literally dangerous volumes. If they don't seem ear splitting to you, then there is something not working right in the circuit, and I think it is that resistor -- or else you set the gain wrong.
Another chip to try is the opa551. It is a high current opamp (~200mA I think) commonly used as a unity gain buffer and is very capable of driving low impedance phones. With a little gain it should work quite well.
Oh, and grados are 32ohm, not 68. |
|
|
| Nordic |
OK, I'm listening to mine, and sounds downright awesome... none of that japanese amp overated symbals and highs... much more natural proportions to highs and lows than standard opa2132 cmoy... I think it may be ok to leave out the 10uf caps, as bass is abundant, and it may just make it even better... Listeing to Norah Jones - Not too late CD now... those acoustical instruments sound scary realistic... and that off one half flat 9v cell...
Very smoothe with a Z.... best out of the 15 or so headphone variations I built so far...background should be called blackground... it is but pitch black silent... even with ipod phones.
Re the PCB ... change package for R4 and R8 to one size smaller, or you wil have to drill a hole in the trace like me...
But guess what? it seems to be an impendance mismatch for my ipod, just like the cmoy is for my portable fm-radio, so there is a good chance it will mate perfectly with your source.... tommorow I will try to double the gain :) |
|
|
| paulb |
| quote: | Originally posted by dsavitsk
Oh, and grados are 32ohm, not 68. |
Right, I stand corrected. But still think a small power amp is a better choice for this application. |
|
|
| Nordic |
Paul Although you are not wrong... take for instance the rod elliot headphone amp with crossfeed... that thing plays phenominaly loudly... but is to heavy on the juice for one 9v cell. But it is why I will look at upping the gain... as long as I stay under 5 It will be great.
I have been so impressed I spent a few minutes fixing the files while listening to music... will test with opa2132 in the morning but that one resistor may cause mean offset, which at the moment measures 0 mv on one channel and 1mv on the other.. which explains the silent background. |
|
|
| dsavitsk |
| quote: | Originally posted by Nordic
one resistor may cause mean offset, which at the moment measures 0 mv on one channel and 1mv on the other.. which explains the silent background. |
Sound is AC, offset is DC. Otherwise every device with a cap on the output would be silent, which is clearly not the case. |
|
|
| Nordic |
Off course it was a dc measurement... but I see what you are saying, which now gets me another step to understanding biasing.... I couldn't understand how an amp could be silent if it had a constant voltage added to the signal... dc has no sound, unless you count the crackle from the flames...
One channel lacks the small bypass caps... just got it going so long... this circuit is completely DC coupled with no caps appart from those in the PSU... I.e. you put DC in, you throw headset away...
I know you have alot of experience.. I so enjoy going over your dac circuits etc... but don't discount this one... try it... I got the chips for free as samples... I guess you could do the same... |
|
|
| theAnonymous1 |
| quote: | Originally posted by dsavitsk
Another chip to try is the opa551. It is a high current opamp (~200mA I think) commonly used as a unity gain buffer and is very capable of driving low impedance phones. With a little gain it should work quite well. |
Thank you dsavitsk , I think I will try this opamp. If I use 2 in parallel that gives me 400ma of current to work with. I will power it with +-7.4v lithium cells. |
|
|
| dsavitsk |
| quote: | Originally posted by Nordic
but don't discount this one... try it... I got the chips for free as samples... I guess you could do the same... |
This looks like what is known in the headphone circles as the A47 Amp (http://www.diyfactory.com/projects/...d_schematic.pdf)
I have built it, and most other opamp headphone amp variations. It is a pretty good performer for low impedance phones considering its low cost. I do think a single LM6172 sounds a little better, but it is a chip prone to oscillations and needs some care in layout.
| quote: | | If I use 2 in parallel that gives me 400ma of current to work with. I will power it with +-7.4v lithium cells. |
I can't imagine you'll need two.
Also, the 551 is a mono chip. There is the 552 which I think is dual, but you should check the datasheet. |
|
|
| theAnonymous1 |
The 552 is a single also. It has a 24V/µs slew rate and 12mhz bandwidth vs. 15V/µs and 3mhz for the 551. The 552 also requires a gain of at least 5.
I removed the 51R resistor in series with the output on my amp and it makes no difference. I also tried adjusting the gain and the amp just gets louder with a lower input signal and still distorts at the same output level. I also tried a +-14v 5A supply and while the bass seemed a lot more full it didn't really play any louder. I think it's just as simple as not enough current output. |
|
|
| Nordic |
Naah the Grado one is significantly diffirent...
non-inverted input, no protection from over current on the input... lower resistors on the output... mine has 220R (closest i had)
See a couple of things I will steal from there... |
|
|
| Nordic |
OK, up to 2k now, so 2x gain... sounding just as good, only gets louder. I thought at first it was missing some top frequencies, but after a few hours of listeing (its very non-fatigueing), I would say its just because it is so much cleaner...
Spent lots of time listening to jazz etc, a few minutes ago I thought, heck lets try some digital music... synth friggen sounds almost analog.
I plan to build an all film caps version, and getting precision resistors, once I settle on a gain level.... with the low parts count, it is worthwile getting GOOOD components. |
|
|
| Pit Hinder |
You are halfway there!
Non fatiguing means your CPU (spell it "brain") has less overtime to do to disencrypt the music - solder on and tell us about the results.
Pit |
|
|
| Nordic |
Stunning noise rejection used as a preamp on PC output...
I modified, went to 330uf for the electrolytics, and 22nf on the bypass caps (all I had enough to make sets of after measuring)
latest version pcb with fixes |
|
|
| theAnonymous1 |
I finished drawing up the board for the OPA551. I have to connect two nets with jumper wire, but I think it came out pretty good considering it's single sided with the components so close together. Size is approximately 25mm x 25mm.
The OPA551 will be delivered on Thursday. I will post my results after it's all put together and tested. I think I'm also going to order an AD8397. It will fit on the board I already made and has pretty high output current.
 |
|
|
| theAnonymous1 |
I'm a bit of a nub when it comes to making boards. Eagle is a bit too much for me so I just use Pad2Pad. I connect all the nets manually, then route, then fix the traces manually.:xeye:
Works fine for small boards like this.
I was still working on the trace placement in the above pic. Here is the final layout ready for laser printing......
 |
|
|
| BWRX |
| quote: | Originally posted by theAnonymous1
I'm a bit of a nub when it comes to making boards. |
Hey, you gotta start somewhere! I would change the routing to use 45 degree bends instead of 90 degree bends. Can you post a schematic? |
|
|
| theAnonymous1 |
| quote: | Originally posted by BWRX
Hey, you gotta start somewhere! I would change the routing to use 45 degree bends instead of 90 degree bends. Can you post a schematic? |
Ermmmm.....:ashamed:
Pad2Pad actually makes all the traces bend at 45's when it can. I manually made them all 90's:blush:. The 90's allow for better clearance between traces and components in some instances, and I like the way it looks for some reason. I know it technically makes the traces shorter, but the board is so small I'm not worried about it. Is there other reasons why 90's are no good?
As far as a schematic, it's just the basic CMOY using single opamps and without the output resistor. I connected it all by hand using what's left of my brain. I use double sided PCB even thous it's a single sided layout. The other side will be used as a ground plain and 6 of the square pad will be connected to it through drilled holes.
Told you I'm nub. |
|
|
| BWRX |
90 degree bends in traces are not optimal from a manufacturing standpoint. It's just harder to maintain the integrity of a 90 degree bend than it is a 45 degree bend.
If you're going to make it on a double sided pcb you could use the other side to run a couple traces instead of using jumpers. |
|
|
| theAnonymous1 |
I use the "toner transfer" method of etching and putting traces on both sides of the board is a royal PITA. If I try to do both sides with one etch I have trouble ironing on the second side without messing up the first.
I could iron on the top, mask the bottom with tape, etch, iron on the bottom, mask the top, and then etch again. That's a whole lot of work. It's easier to just use two jumper wires this time around.
You are the board making King in my eyes Brian. I envy your skills.:D |
|
|
| BWRX |
I'm a noob compared to a lot of PCB gurus out there :cannotbe: but I like to think I half know what I'm doing ;)
A good way to make sure both sides a DIY double sided PCB line up is to put a little cross in opposite corners of the layout. Mark where the center of the crosses are and drill through the board. Now you have references to line up both sides of your board to. As for transferring one side without messing up the other, you just have to be careful I guess. |
|
|
| dsavitsk |
| quote: | Originally posted by theAnonymous1
I finished drawing up the board for the OPA551. I have to connect two nets with jumper wire, but I think it came out pretty good considering it's single sided with the components so close together. Size is approximately 25mm x 25mm. |
Looks pretty nice. I think if you futz with it a lot you can eliminate those jumpers, but it is probably not worth the time. Jumpers should be fine.
As for the 45 degree thing, this only really matters for super high frequencies. You should be fine here.
| quote: | Originally posted by theAnonymous1
I think I'm also going to order an AD8397. It will fit on the board I already made and has pretty high output current.
|
Be forwarned that this chip is a beast to work with. Without a lot of attention to a lot of details (decoupling, ferrites on the output, caps in the feedback loop, etc.), it will oscillate and destroy itself, and perhaps your headphones. Even when not oscillating, it is prone to a lot of offset. Look at Headwize for discussions of the Mini3 and the Pint, and good luck. People do seem to like it when they get it to function. I still think the LM6172 sounds a little better, but a lot of people disagree. Oh, and the NJR4556 is another high current opamp to try. It is used in the Grado RA-1, and only costs like $0.50 so is worth the experiment. All of these are prone to oscillation and offset, but the SAD8397 is the worst of the bunch. |
|
|
| theAnonymous1 |
I can't let my idol down.:D
Top.....
Bottom....
 |
|
|
| BWRX |
| quote: | Originally posted by dsavitsk
As for the 45 degree thing, this only really matters for super high frequencies. You should be fine here. |
A layout that works well at high frequencies will also work well at low frequencies, but not vice versa. |
|
|
| theAnonymous1 |
Here is a pic with the silk layer so you can better see what everything is......
 |
|
|
| Fenris |
Anonymous -
I've got some spare through-hole stuff laying around. Would you like me to put together a small OPA2132 and BUF634 board for you? |
|
|
| BWRX |
| Anonymous1, I whipped up a quick schematic in Eagle and routed a board as well. There is no easy way to do a one sided layout without jumpers when two separate chips share supply rails. I liked your idea of a solid ground plane on the back and ran with that. For this layout you'd have to drill out the 4 ground pads and solder your wires on both sides to connect them to the ground plane. You'd also have to jumper each supply rail over to the chip on the right. |
|
|
| BWRX |
Feel free to use it if you'd like. The main change I made was to add supply rail decoupling caps for each chip. It is best to place them as close to the chip's power pins as possible.
I can send anyone the eagle schematic and board files if they want to play around.
The actual dimensions of the board outline are 1.375in by 1.1125in. |
|
|
| Fenris |
Here's an example of a headphone board I've built, probably way overbuilt with bypass capacitors. It's about 1.4 x 2.3 inches, but could be made alot smaller by eliminating caps. It's single sided with one short jumper.
 |
|
|
| theAnonymous1 |
Fenris, thank you for the offer. I am in the middle of etching my board and would like to try and see how the OPA551/2 work out first.
Brian, thanks for making up that board. If the one I'm etching now doesn't turn out maybe I'll give it a try.
Doing double sided toner transfer is pure hell. I had to redo it 4 times already using acetone to clean the toner off.
I'll post a pic of the board when it's done as long as it's not too horrible looking. It's taking forever to etch double sided 2oz. with spent ferric chloride. |
|
|
| Nordic |
often there is much less traces on one side... drill the holes and manualy draw one side with a marker.... I do it all the time
I leave my FeCl in an open bowl... most of the water evaporates... when I need it I just pour some boiling water in and stir the crud up.. the heat helps much.
Get those decoupling caps even closer if you can.. |
|
|
| theAnonymous1 |
| quote: | Originally posted by Nordic
often there is much less traces on one side... drill the holes and manualy draw one side with a marker.... I do it all the time
I leave my FeCl in an open bowl... most of the water evaporates... when I need it I just pour some boiling water in and stir the crud up.. the heat helps much. |
Not a bad idea Nordic. Drawing some line with a sharpie would have been a lot easier.
Board is done. Now I just have to wait for the FedEx guy to show up tomorrow.;)
Wish I has some tin plating solution.
Top.....
Bottom....
 |
|
|
| BWRX |
| Very nice job on something that small! |
|
|
| theAnonymous1 |
Thanks Brian. Now I just have to find where I put my tiny drill bit.:xeye:
I have another question about this project. I plan on using two 7.4v lithium batteries with a real ground point. I noticed that when doing it this way you have to disconnect both + and - at the same time or you get huge DC offset.
The pot I plan on using has a built in switch, but it's only a single pole. How can I disconnect both the + and - supply leads with a single switch? I know the obvious answer would be a relay, but they are too current hungry for a portable amp. Is there any other way? |
|
|
| Fenris |
| Nice job on the board. I use HCl and H2O2 to etch (and toner transfer) - I think it's easier than FeCl. Also, to tin I've stopped using Tinnit, it was more trouble than it was worth. I now just wipe a very thin layer of flux on and then use my iron to cover the traces with solder. It's actually faster than Tinnit (unless doing several square feet at a time). I don't think there's a way to switch on the amp with only a SPST switch without using a relay. A small 12v relay will have about 1500 ohm resistance. Use a 350 ohm resistor in series and use both batteries to switch it on/off. Current draw should be about 8ma. If you're already having problems sourcing 40ma of current from each opamp and expect to probably be drawing 100+ma from each, 8ma is negligable. Plus it will draw equally from each battery, so they won't drain unevenly. |
|
|
| theAnonymous1 |
Oh wow, nevermind about the relay current issue. I work midnights and still havn't gone to bed yet.:xeye:
I was looking at the spec sheet for a 12v DIP reed relay and I got mW confused as mA. I read it as 144mA, not 144mW. It would actually be 12mA @ 12v.:smash:
Time for bed. |
|
|
| Nordic |
| quote: | Originally posted by Fenris
I now just wipe a very thin layer of flux on and then use my iron to cover the traces with solder. It's actually faster than Tinnit |
I totaly do the same...
If you don't own flux, go and buy some before you solder another dot. You get nice big jars of it at the hardware shop, plumbers use a tone of the stuff to solder pipes.. |
|
|
| theAnonymous1 |
I just finished the board, and it works. I soldered all the resistors and caps while I was waiting for FedEx to deliver the OPA. Right when I was about to unplug my iron I heard they guy put the box at my door. FedEx is awesome.
This little amp absolutely ROCKS!! It's insanely loud and sounds excellent.
It plays loud enough that I could actually blow the earbuds I tested it with if I tried. Any loud bass notes in a song makes my earbuds distort and crackle from over excursion. I have the gain at 11 like the last amp and even with my mp3 players volume all the way up; I don't hear any clipping or distortion from the amp itself. I am very satisfied with how it came out.
I can't test it for it's intended purpose yet because I still have to wait to get the new pair of lower impedance custom buds back.
I would like to thank everyone who help out. I love this place.
 |
|
|
| Nordic |
Operator: good morning this is Fedex, how can I assist you...
Caller: I have a package I need delivered to Idaho...
Operator: Its allready delivered
Caller: No it isn't its right here on the kitchen tabl... hey!!! wait where is it now?
Operator: Your dad is trying on the shirt right now....
Nicely done mate
Here is a picture of my near completed build... still have to attach one screw mount, but it all works and everything.. went for gain of 4.7 eventualy and used some Oscon caps I got from Franz G.
I'm giving it to a friend I had since school for his birthday tommorrow. |
|
|
| Pit Hinder |
A small step in electronics, a huge leap in quality of life for your girlfriend! Big handshake from across the ocean. Well done.
Pit |
|
|
| PB2 |
Nice work, looks like you've got a good solution.
I would have suggested a simple output buffer like Figure 14
here on Self's page:
http://www.dself.dsl.pipex.com/ampins/webbop/opamp.htm
It should be possible to do this without large caps, and at lower standby current than with the OPA551. A low power OP amp, driving a very low bias AB output stage for example.
CMOY burns power just in producing the split supply, and then
it requires large caps.
You should be able to use series pass BJT or FETs on the +/- supply, and have the power switch just bias both on, rather
than a relay. Or you could just use a double pole switch.
Pete B. |
|
|
| theAnonymous1 |
I'm sure there are a lot of better solutions than the one I chose, but small size and simplicity where my main concerns. I'm pleased with how it turned out.
Thank you for the link to the opamp page. I think I will spend some time reading through it. |
|
|
| dsavitsk |
| quote: | Originally posted by theAnonymous1
[B]This little amp absolutely ROCKS!! It's insanely loud and sounds excellent. |
I'm glad the 551's worked out. :) |
|
|
| theAnonymous1 |
Yes, thanks again for recommending it. I like it so much I'm in the process of making a dual mono DIP version for myself to use at home. Each channel will have separate transformers, rectifiers, etc.
I'm kind of stuck on how I'm going to power my gf's portable now. The 7.4v 400mah batteries I was going to use are sold out and prices from other places are just too high. I need a supply with a profile smaller than two 9v batteries.
Anyone know of a small boost converter circuit that can take 3.7v in and output +-8v-10v or single 16v-20v? Prefferably using an 8 pin SOIC chip. |
|
|
| dsavitsk |
| quote: | Originally posted by theAnonymous1
Anyone know of a small boost converter circuit that can take 3.7v in and output +-8v-10v or single 16v-20v? Prefferably using an 8 pin SOIC chip. |
You can try to copy the one off of my HPDAC. It uses a TPS61040 which should do what you want, though it only has 5 pins. Just be aware that layout is important to keeping it stable, and ferrites and re-regulation are important to keeping it quiet.
http://www.ecp.cc/HPDAC.html |
|
|
| theAnonymous1 |
| quote: | Originally posted by dsavitsk
You can try to copy the one off of my HPDAC. It uses a TPS61040 which should do what you want, though it only has 5 pins. Just be aware that layout is important to keeping it stable, and ferrites and re-regulation are important to keeping it quiet.
http://www.ecp.cc/HPDAC.html |
Thank you again. The TPS61040 looks like a good candidate.
Is there a way to make a dual supply with it using just one battery? Preferably without having to add a separate rail splitter circuit. |
|
|
| dsavitsk |
| No, I think you'll need a rail splitter. |
|
|
| BWRX |
| If you use two batteries you can make two boost regulators to get your +/- rails just like you would with two positive linear regulators. Don't forget that with a DC-DC convertor Pout=Pin*efficiency. For a boost regulator that means that since the output voltage will be higher than the input voltage the input current will be higher than the output current. Something to keep in mind since you'll be using batteries. |
|
|
| theAnonymous1 |
| quote: | Originally posted by BWRX
If you use two batteries you can make two boost regulators to get your +/- rails just like you would with two positive linear regulators. Don't forget that with a DC-DC convertor Pout=Pin*efficiency. For a boost regulator that means that since the output voltage will be higher than the input voltage the input current will be higher than the output current. Something to keep in mind since you'll be using batteries. |
Yeah I was thinking about two boost circuits in series but 2 boost circuits + 2 batteries = 2 big.
I realize the current draw from the 3.7v battery will be quite a bit. I have a few 3.7v 800mah lithium cells. She will only ever use it for a couple hours at a time so that should be enough.
Maybe just doing 3.7v to 18v and then using a resistor divider for the rails is the simplest smallest solution. But then there's the issue of the divider resistors eating up current.:cannotbe: |
|
|
| Nordic |
Lol time to make a cap with solar cells...
Makes mental note to get some of the strips I saw the other day 9V and about 13mA |
|
|
| PB2 |
| quote: | Originally posted by theAnonymous1
Yeah I was thinking about two boost circuits in series but 2 boost circuits + 2 batteries = 2 big.
I realize the current draw from the 3.7v battery will be quite a bit. I have a few 3.7v 800mah lithium cells. She will only ever use it for a couple hours at a time so that should be enough.
Maybe just doing 3.7v to 18v and then using a resistor divider for the rails is the simplest smallest solution. But then there's the issue of the divider resistors eating up current.:cannotbe: |
I have to wonder if it would be simpler to design a very simple power amp that can go close to the rails say 3-4 Vpp off of +/- 3.6V (7.2V total) batteries, than to include boost circuits and so on. A very low power rail to rail OP amp with an output buffer is another possibility.
Wondering if it can be made to sound good with very low idle currents leads to the thinking that it would be interesting to have it turn off the current sources and output stage with the lack of an input signal, and power up at higher idle currents with an input signal. This could have a very long on dwell time so that on/off switching is not an issue from cut to cut.
Just offering the ideas since this thread has me thinking about the problem.
Pete B. |
|
|
| lineup |
| quote: | Originally posted by PB2
.I have to wonder if it would be simpler to design a very simple power amp that can go close to the rails say 3-4 Vpp off of +/- 3.6V (7.2V total) batteries, than to include boost circuits and so on. A very low power rail to rail OP amp with an output buffer is another possibility.
. |
If you need a
- very low drop of voltage ( like Input Voltage -0.1 Volt worst case )
- extremely low supply current ( around 0.100 mA consumption)
- precision
Discrete Regulator - 3 small transistors
I have such a circuit!
Actually I have one working. set for 5.00 Volt output
*******************************************
More information on FEATURES of my regulator circuit:
Microphone Amplifier 9 Volt Battery Discrete Regulator
See attached picture of my little great regulator!
As most voltage regulators chips it has 3 pins: in, out, gnd
For size compare I have used an Audio Op-Amp IC in my image
Regards
lineup
Lineup Audio Circuits :cool: |
|
|
| lineup |
:cool:
this is the bottom view
of my small and very own regulator idea and design
using only 3 transistors TO92
enjoy!
Lineup Audio constructions
:) |
|
|
| theAnonymous1 |
Ok, it seems like every time I solve one problem it creates another.
I'm going to go a completely different route now I think. I have some TPA6111A2 150mw stereo headphone IC's left over from my first attempt at a portable amp. It worked good, but just lacked some power. I will admit that when I tested the TPA6111A2 with the custom earbuds I hadn't discovered they were 150 ohm yet. I was under the assumption they were 25 ohm, so results should be better with the new phones when they arrive.
http://focus.ti.com/lit/ds/symlink/tpa6111a2.pdf
The idea I have is to use two stereo IC's each in a BTL config. I will then power each circuit with it's own 3.7v lithium battery. By using seperate batteries I can then tie one output from each circuit together so that a standard stereo connector can be used.
Below is a pic of what I have in mind. I'm not sure if this is the correct way to do it, but it seems logical to me.
I got the idea from the similarity with a Tripath amp bridging diagram.
 |
|
|
| martin clark |
I've got nothing to contribute to the thread directly , just posting to applaud your sentiment generally. My (ex) S/O had significant hearing difficulties, yet loved music... Iunderstand where you are coming from.
(funnily enough she loved my (large by UK standards) horns but hated the ESLs - which I sold...) |
|
|
| PB2 |
Any update with the new low impedance headphones?
As far as the bridged configuration you show there, why not try it first unbridged? With Low Z phones you'll need a lot of current.
The requirements for the amp depend on the load impedance, and since some of these OP amps are marginal driving low Z loads it is good to first look at your requirements. How many volts out do you need to the phones? Then calculate the current. I'm sure you know this, but it is good to step back and first determine what you need.
I tried a TLO82 with a transistor output stage, +/- 9V supply and found that the OP amp limited the swing to about 6 Vpp. My mp3 player puts out about .5 Vpp and even just 1-2 Vpp is very loud, bass comes almost as wind out of my Grado phones. Yet it does not seem to sound so good, don't know, perhaps it is too much for the phones, or some filtering is required. I should probably try a better front end, as it just sounds a bit rough and grainy. I have found that it has a tendency to oscillate, then it sounds real bad. I just wired it on a proto board so that is probably part of the problem.
I'm very tempted to do a simple discrete design but I don't really need a headphone amp. Also, rail to rail video OP amps look good, on paper anyway, since they're designed to drive 75 ohm loads. 3 in a package are available (AD813 http://www.analog.com/en/prod/0,2877,AD813,00.html ) and they could be paralleled for even lower impedance loads.
Are you planning to fit it into an Altoids case? |
|
|
| theAnonymous1 |
Hi PB2,
I'm still waiting for the new phones to show up. I don't know whats taking so long. I will have to call and check up on their status Monday.
I've already built the bridged circuit I proposed above. The output power seems on par, if not a bit better than the OPA551 board @ +-9v. The OPA551 board sounds better in my opinion, but sound quality isn't the most important thing here.
http://i16.tinypic.com/33af5ug.jpg
I think I will end up using the OPA551 board though. I found some lithium cells the size of AAA batteries. Four of these in series should work out well and still be relatively small.
http://www.batteryspace.com/index.a...ROD&ProdID=2598
I'm trying to aim for something a bit smaller than an altoids container. |
|
|
| PB2 |
I forgot to mention, you can't bridge since there is a common lead for right and left? Unless you're running it mono?
I was going to suggest AAA lithiums but I wasn't sure of how much space you had.
Pete B. |
|
|
| theAnonymous1 |
| quote: | Originally posted by PB2
I forgot to mention, you can't bridge since there is a common lead for right and left? |
True, but I found a way around this problem by accident. I tied one output from from each chip together to great the common "ground" for the headphones. Then, instead of referencing the input signal to the normal input ground(battery -), I referenced it to the new common ground on the outputs.
I know, crazy, but it worked.:xeye: |
|
|
| PB2 |
| quote: | Originally posted by theAnonymous1
True, but I found a way around this problem by accident. I tied one output from from each chip together to great the common "ground" for the headphones. Then, instead of referencing the input signal to the normal input ground(battery -), I referenced it to the new common ground on the outputs.
I know, crazy, but it worked.:xeye: |
Hi, well it's interesting how things technically incorrect can sort of work. If it works at all it is because most source material has much common mono content. The common bridged signal is now mono. If you drive one side only the common connection will have the two amplifiers fighting. There will be very high cross currents if you've wired it as I think you have.
Pete B. |
|
|
| PB2 |
| quote: | Originally posted by PB2
I tried a TLO82 with a transistor output stage, +/- 9V supply and found that the OP amp limited the swing to about 6 Vpp. My mp3 player puts out about .5 Vpp and even just 1-2 Vpp is very loud, bass comes almost as wind out of my Grado phones. Yet it does not seem to sound so good, don't know, perhaps it is too much for the phones, or some filtering is required. I should probably try a better front end, as it just sounds a bit rough and grainy. I have found that it has a tendency to oscillate, then it sounds real bad. I just wired it on a proto board so that is probably part of the problem.
|
The problem was probably a combination of poor connections (old proto board) and the compensation cap being off by a factor of 10. There was not enough major loop gain to cover the output stage, the frequency response was not even flat. It's all fixed and sounds very good, a lot of headroom! Still it is not optimal since it cannot even come close to the rails which would allow lower voltage batteries to be used. This is easy to fix with a rail to rail OP amp.
OK folks, any favorite rail-to-rail OP amps that are low power, will run on say +/- 3 V, and sound good? Output current drive is not important as I'm using transistor buffers. AD822? or MAX4494?
Pete B. |
|
|
| Nordic |
 |
|
|
| PB2 |
Thanks for the info, OPA132 looks good however it is not a low power device, the AD820 also looks very good.
I made a measurement error, I'm getting about 12 Vpp out from the now +/- 8V supply. Much better.
Pete B. |
|
|
| theAnonymous1 |
Here is a low voltage rail-to-rail opamp you could try. Maybe a little too low voltage.
http://focus.ti.com/lit/ds/symlink/opa365.pdf
In other news, I just received some MAX629 500ma DC-DC converter chips. I'm going to build up a circuit to take a single 3.7v lithium battery up to around 20v to use with the OPA551 board I made. I have everything I need except the 47uH SMD inductors.
Still waiting impatiently for the new earbuds to arrive. |
|
|
| theAnonymous1 |
I have a question about this virtual ground circuit.....
Can I replace Q1 with a ZXMN3A01F MOSFET and Q2 with ZXMP3A13F ? If so, what value should R3 and R4 be? |
|
|
| PB2 |
| quote: | Originally posted by theAnonymous1
I have a question about this virtual ground circuit.....
Can I replace Q1 with a ZXMN3A01F MOSFET and Q2 with ZXMP3A13F ? If so, what value should R3 and R4 be? |
That is nearly the output stage that I'm using, interesting coincidence. That is in the signal path even if it's just producing the ground point.
Using those big caps, why not just put them in series with the output and run the OP amps single ended off the supply? The CMOY caps that produce the ground reference are in the signal path anyway.
Just offering some different perspectives.
Pete B. |
|
|
| theAnonymous1 |
I won't be using 220uf caps due to space. I plan on using 100uf tants. The caps ground point won't be connected to the diodes like they are in the diagram from Tangents site. Sijosae's original schematic does not have the caps connected this way (see fig I below).
Going single ended with 100uf tants into 25 ohm phones would pretty much kill all the bass. This wouldn't be such a bad thing for my girlfriend as she can't hear that low anyways, but I might want to use the amp occasionally. |
|
|
| PB2 |
Yes, 100 uF is a bit low.
I believe that now the Grado phones are the headroom limit as I can hear significant distortion during heavy bass passages at loud listening levels. Phones are obviously going to have excursion limits in the same way as a woofer.
Filtering below 20 Hz might make sense with most phones. You might want to filter higher for your girlfriend if you need very high midrange levels so that the midrange SPL will not be limited by bass excursion.
Pete B. |
|
|
| theAnonymous1 |
:mad: :mad: :mad: :mad: :mad: :mad: :mad: :mad: :mad:
OK, I finally received the new pair of custom buds and I am steaming @#%$ing mad. The new buds were supposed to be 25 ohm and they measure 6.6R. The sick thing is the companies name is Precision Laboratories INC. Not so @#$%ing precise are they.
The proof is in the pudding so they say.......
http://i15.tinypic.com/4c2y3k0.jpg
I'm going to try my best monday to keep my head from exploding when I call Tammy at the earplugstore. She is just the middle man(women) who sells Precision Labs service. I did find it a bit discerning when I first called her with my original complaint and she didn't even know what "impedance" is or why it is an important spec. She sells headphones for a living.:rolleyes:
I tried the buds on the OPA551 amp I built and it distorts almost right away. The MP3 player that feeds the headphone amp drives them better, but it also distorts before it plays at a decent volume.
I just purchased $40 worth of parts from digi-key to make a dc-dc converter and rail splitter for the cmoy and now I can't even use any of it.
These were supposed to be a xmas present...... it's @#*$ing March. All I can do is *sigh* at this point.:( |
|
|
| BWRX |
| That's a very low impedance for any style of headphones/earbuds. What kind of driver is used in them? |
|
|
| theAnonymous1 |
The specs say "Balanced Armature single driver ", but their specs obviously can't be trusted. It could be a piezo in parallel with a 6.6R resistor for all I know.
My dear friend Tammy told me they are able to make the drivers in any impedance. The only problem seems to be they can't read. How does 25 look anything like 6.6? |
|
|
| theAnonymous1 |
Well, good news and bad news.
The good news first......
I finished the headphone amp and it works like a charm. I used it with a pair of cheap 30 ohm headphones and it reaches ear splitting levels.
The tiny amp board is only 1"x2" and contains a DC-DC converter, a discrete rail splitter, and the amp circuit with two OPA551. The DC-DC converter takes a 3.7v lithium cell and converts it to about 23.5v (odd number, don't ask). At full volume into my 30 ohm phones the rails never deviate more than 0.5v from each other and only sag to about +-10.8v. This thing is crazy powerful for it's size and sounds good to boot.
Here is a pic of it powered up and singin' some songs......
http://i15.tinypic.com/2ewhybm.jpg
http://i18.tinypic.com/2ufc11x.jpg
Now the bad news.......
I called my good friend Tammy Bashir of the earplugstore to complain about the second round of screw up. She seemed very aggravated that I expected what I paid for and immediately referred me to a guy named Bill Jr. at Precision Labs. She said she didn't have time to play the middle man. That was nice of her to say considering she had the time to process my $220 credit card order.
So anyway, I call this Bill Jr. guy and explained to him about how imprecise his operation is. He too seemed aggravated at my request to get what I paid for. He said that if they aren't what I want then he can't help me. How nice of him huh? He explained that he just purchases the drivers and they are what they are. It's nice that on his own site he advertises their custom molded buds come with 19 ohm drivers, yet he has no 19 ohm drivers. After a few choice words he said I can mail them back to him AGAIN at my own cost and he will take a look at them, but won't promise he can solve the problem.
At this point I am extremely exhausted and frustrated. I'm not sure if I should send them back again, or go through the painful process of my credit cards buyer protection plan.
*sigh*
|
|
|
| Zigis |
Hi, theAnonymous1 !
Nice thiny Amp !
I am thinking to build something similar, without large 9v batterry.
How long can You ran this amp with 3.7V lithium, compared standart 9V?
How large (mA) are lithium cell?
do You use MAX629 ?
Zigis. |
|
|
| theAnonymous1 |
Hi Zigis,
Yes, I am using the MAX629. I have the output of the boost circuit set at around 23v. This is a little high for the tantalum caps I'm using, but they don't seem to mind. There was a little audible switching noise from the supply at first, but a 0.1uf cap across pins 3 & 4 completely got rid of it and now the amp is dead silent.
The lithium cell is 900mah. The amp will run at very high output levels for around 5-6hrs per charge. The nice thing about the cell phone battery I am using is that it will automatically cut off the power when the battery reaches 3v to prevent damage to the cell.
I was in a rush to put the thing together and forgot to measure the quiescent current draw of the whole circuit.
In the picture below I outlined each circuit with a colored box. Red is the MAX629, Blue is the discrete rail splitter, and Green is the amplifier.
 |
|
|
| AndrewT |
Hi,
since you used your credit card (not debit card) you have the power of them behind you.
Write to the supplier (the sales woman that can't be bothered) and ask for a refund of the cost (both if both don't meet the intended specifications) and copy the letter to your credit card company. Make sure the retailer knows you have copied to the credit card by saying so on the letter. The retailer will take heed.
The credit card company will put the sale in dispute and freeze the funds.
If the dispute is not resolved for the wrong goods supplied problem you can use the credit card company to demand your money back for all the goods falsely described.
at least the 6r6 phones are due for refund.
what about the 25r/19r phones? did they meet the original specification?
Now back to the current drive problem way back in posts 1 to 20.
I haven't read those in between 21 and latest.
Power =V^2/R =I^2*R
You know the current limit of the opamp. Therefore use the current power rating.
When using peak currents rather than rms values remember to divide by 2 to obtain the rms based power.
The 2134 has a 40mA current limit.
P=0.04^2*25=0.04Wpk=0.02Wac i.e.20mW.
for 115db/mW phones that equates to a maximum SPL of 115+13=128db.
The average level probably nearer 110 to 115db to avoid clipping.
If the 2134 had been operating correctly it should have been capable of average levels of 110db before clipping.
Reducing the phones to 19r (if they had been manufactured to correct spec) would have REDUCED the power and SPL delivered cf the 25r versions. |
|
|
| Zigis |
Hi !
You use cell phone charger.
How does it work, automatic who turn out charge wen batterry is full, is it in charger, or in phone, or in batterry itself?
What purpose is 3 pins in cell phone batterry ?
Zigis. |
|
|
| theAnonymous1 |
I am still waiting for the 3rd set of earbuds to show up. I should probably call Precision Labs and bug them, or they might never show up. If the new phones aren't to spec or if they give me any trouble when I call, then I will go ahead with legal actions.
Zigis,
I bought the battery off ebay and it came with a charger.
http://cgi.ebay.com/ws/eBayISAPI.dl...em=160059864568
I should note I bought 2 of the above items and both of the chargers did not work. There is a small transformer inside to provide 5v or so to the charging circuit; this transformer has a manufacturing flaw and does not work. I had to provide a separate 5v supply. Still not a bad deal for the money considering the charging circuit part does work.
I'm not too sure what the other pin on most batteries is for. I think after the battery drains past around 3v that pin might output a small voltage signal. |
|
|
|
