Help building a META42
Dec 3, 2002 at 5:02 AM Thread Starter Post #1 of 38

Kal525

100+ Head-Fier
Joined
Dec 1, 2002
Posts
147
Likes
1
after some research and posts, i have decided to go ahead and build my own meta42. I just ordered the sennheiser hd600s for $250, which will become $200 after the rebate and want to get started on this amp. I have browsed tangent's website and found a lot of useful information, i am however still confused about many of the optional parts and parts in which i have a choice ( i don't like choices since i have no idea what the impact will be when it comes to this sort of thing). If anyone is willing to help me please let me know, i really don't want to mess anything up. I guess i'll start with a pretty important question........what op-amp should i use? I have no clue on this one, i just want something that sounds good, and will i need the C5 capacitor. I of course have a lot more questions, but this is a start, and i do not want to bore you all with my long list of questions.
 
Dec 3, 2002 at 6:55 AM Post #2 of 38
If your sources don't have DC offset, dump C1.
Use buffers, EL2002's are nice if you are willing to pay the extra for them. Stack them if you want.
If you get buffers, use the Multiloop topology. ie. R3 to R6 are all occupied. R7 is also occupied.
Populate R8 or R9 if you get a hiss on your headphones and it is caused by the amp. To check if it's the amp, unplug all the source from the amp and turn the amp up. No hiss == jumper across both.
C5 isn't actually needed.
Using R10 with cascoded jfets (Q1 & Q2) will improve sound (YMMV depending on the opamp used).

Opamps:
Good ones are OPA627's (with adaptor) or AD8620's.
If batteries are used, consider the AD823.
 
Dec 3, 2002 at 7:19 AM Post #3 of 38
ok, i was under the impression that i needed C1 in order to prevent DC damage to my headphones. I want to be able to use this with batteries and a wall outlet, so i'm guessing that means i will need C1. As for the buffers, i'm not sure what stacking is exactly, i'm assuming that it means that i buy 4 of them instead of 2, and solder them on top of each other. As for R3 to R6, should i stick with the resistances in the schematic, or should i pick different ones? And what resistance should R7 be, it says 47-100 ohms, but not sure what exactly would be best. I don't think i will need R8 or R9 since the hd600s have 300 ohm impedence and so i should not hear the hiss unless i'm using headphones with less impedence. I'm still not really understanding the part with biasing the op-amp into class A, which is what the cascoded jfets have to do with i'm assuming.

Now for the opamps, i am not sure which will sound best, is there one that is known for overall good quality? There are a bunch on the parts list, and yet another one on tangent's parts order form. And as for C5 i thought i had to populate it if if the op-amp oscillates, so since you said i don't need it, i'm guessing that a good op-amp doesn't oscillate.......sorry if what i'm saying sounds a little ignorant, but this is really confusing and i'm trying my best to absorb as much as i can, thanks for your help.
 
Dec 3, 2002 at 7:38 AM Post #4 of 38
Just my two cents, if you're going to be using the amp with the HD600, go for the AD8620 opamp. (Unless you need a great deal of portability, of course, then you'll need to go with something that works great at lower voltages.) It probably won't oscillate, so in all likelihood you won't need C5. You might get a little hiss if you're using a wall power supply and your house's power is fairly dirty, so there's a possibility you'll want to populate either R8 or R9.

Also, stacking the output buffers just means putting one buffer on top of another, then soldering the pins together so the top one doesn't fall off. This will give your amp a bit better control over the bass, especially with low impedance headphones, though you'll probably notice some difference with Sennheiser HD600s too.
 
Dec 3, 2002 at 7:40 AM Post #5 of 38
Quote:

Originally posted by Kal525
ok, i was under the impression that i needed C1 in order to prevent DC damage to my headphones. I want to be able to use this with batteries and a wall outlet, so i'm guessing that means i will need C1.


No. C1 prevents damage to your headphones by blocking DC voltage on the signal line (the audio input of the amp). This will prevent the DC voltage from being amplified by the opamps (hence, frying the phones).
If your source doesn't have DC offset, C1 is not needed. This has got nothing to do with the power source (batteries or wall warts alike).

Quote:

Originally posted by Kal525
As for the buffers, i'm not sure what stacking is exactly, i'm assuming that it means that i buy 4 of them instead of 2, and solder them on top of each other.


Yes, that is correct.

Quote:

Originally posted by Kal525
As for R3 to R6, should i stick with the resistances in the schematic, or should i pick different ones? And what resistance should R7 be, it says 47-100 ohms, but not sure what exactly would be best.


You can stick with the values for R3 to R6 in the schematic. They work fine and especially so since you've high impedance headphones.

Quote:

Originally posted by Kal525
I'm still not really understanding the part with biasing the op-amp into class A, which is what the cascoded jfets have to do with i'm assuming.


Yes, the cascode jfets help to bias the amp into class A. Stick a resistor (between 100 to 1K ohm) in R10 as well.

Quote:

Originally posted by Kal525
Now for the opamps, i am not sure which will sound best, is there one that is known for overall good quality?


OPA627 or AD8620 or AD843 would be your best bet. OPA637's can also be used if you're setting the gain at least above 5. The default values in the schematic works with the OPA637 just fine. (That's what I've tried)
Note: The OPA627 & 637 are single opamps, you'll need a Browndog single-to-dual op-amp adaptor to use them.

Quote:

Originally posted by Kal525
And as for C5 i thought i had to populate it if if the op-amp oscillates, so since you said i don't need it, i'm guessing that a good op-amp doesn't oscillate.......


It's not that a good opamp doesn't oscillate. However, the multiloop topology (using R3 through R6) should be enough to prevent the opamp from oscillating.

Quote:

Originally posted by Kal525
sorry if what i'm saying sounds a little ignorant, but this is really confusing and i'm trying my best to absorb as much as i can, thanks for your help.


It's ok. Everyone has a to start somewhere. Maybe some other forum members can correct me if I'm wrong just as well.
 
Dec 3, 2002 at 3:34 PM Post #6 of 38
Quote:

Originally posted by Dreamslacker

OPA627 or AD8620 or AD843 would be your best bet. OPA637's can also be used if you're setting the gain at least above 5. The default values in the schematic works with the OPA637 just fine. (That's what I've tried)
Note: The OPA627 & 637 are single opamps, you'll need a Browndog single-to-dual op-amp adaptor to use them.



wait a minute... does the overall gain of the amp have to be >5, or just the gain of the inner loop?
 
Dec 3, 2002 at 4:28 PM Post #7 of 38
alright, thanks a lot for your help, but that just opened up some more questions
smily_headphones1.gif


I really don't want to mess my phones up, and presumably i'll be using this thing at home and with a portable, so i'm not sure if my source will have DC offset or not. In the parts advice list it says to go with C1 if i'm gonna be using it with a wide variety of equipment. And if i use a Wima MKS-2 10µF/16V capacitor, i should have minimal distortion, is this correct? Or would you really recommend just leaving it out?


Now for the buffers, i might as well stack them since it's not more money so i will probably do that. I think i am going to use the AD620 op-amp as it seems that a lot of people have been using them in recent builds of the meta42. Now in deciding between the EL2001 vs the EL2002, it says i should consider the speed of the op-amp since i want a significantly faster buffer. Knowing this, which buffer would be better suited to match the op-amp i am using.


The next thing i'm confused about also has to do with buffers. In the "removing the buffers" section of tweaks, it says there are 3 buffers, and since you either have 2 or 4 EL buffers, i'm not sure what 3 buffers he is talking about, since the only buffers i see in the diagram are 2 EL2001 buffers. He also mentions a "power supply buffer" which again i don't know what it is.


As for R3 - R6, i will stick with the values in the schematic as you said.


Now, moving onto biasing the op-amp into class A: Am i correct in stating that by populating Q1 and Q2 i am putting a current source from the output to the V-rail? Tangent talks a lot about current levels and such. Do i have to play around with anything, or do i just populate Q1 and Q2 as instructed. Also, he initially recommends that i do not add a current source, since some op-amps are themselves are sometimes "biased to operate in class A at low current levels" So would it make an improvement if i did use cascoded Jfets with the ad620? Assuming i did do this, how would you recommend i obtain 2 JFETs with two diffrent Idss values?

The last point discussed so far is the C5 capacitor. you said that R3-R6 prevents the need for this capacitor depending on the op-amp that i use, knowing that i will be using the ad620, what would you recommend now?

That covers pretty much what's been discussed so far, i still have other questions about the capacitors in the power section of the board, but this is enough for now. Thank you very much for your help.
 
Dec 3, 2002 at 5:37 PM Post #8 of 38
Quote:

does the overall gain of the amp have to be >5, or just the gain of the inner loop?


Just the inner loop. However, a high inner loop gain isn't a cure-all with fast op-amps like the 637. Bypass caps, fast buffers, and other tweaks may still be necessary. So, newbies shouldn't use chips like this. Kal525, you should use a tolerant chip until you're ready to start gaining deep troubleshooting skills. (Which is to say, you're ready to start getting into deep trouble.
smily_headphones1.gif
)

Quote:

And if i use a Wima MKS-2 10µF/16V capacitor, i should have minimal distortion, is this correct?


C4 doesn't have anything to do with distortion. It has to do with op-amp stability and transient handling. To put it another way, it has to do with the amp's 'agility', both helping the amp to be more agile, and also to ensure that the op-amp's speed doesn't translate into instability.

Quote:

Now for the buffers, i might as well stack them since it's not more money so i will probably do that.


Ummmm... the extra buffers do cost money. Also, they add to the current drain of the amp, so your battery life will drop.

Quote:

Now in deciding between the EL2001 vs the EL2002, it says i should consider the speed of the op-amp since i want a significantly faster buffer. Knowing this, which buffer would be better suited to match the op-amp i am using.


For battery use, use the 2001. For a wall-powered gung-ho cost-no-object amp, what the hey, get 2002s.

Quote:

i'm not sure what 3 buffers he is talking about, since the only buffers i see in the diagram are 2 EL2001 buffers.


The schematic only shows one amplifier channel, since the other channel is identical. The 3 buffers are for left output, right output, and virtual ground.

Quote:

He also mentions a "power supply buffer" which again i don't know what it is.


That's the virtual ground buffer, labelled 2001G on the board.

Quote:

Do i have to play around with anything, or do i just populate Q1 and Q2 as instructed


The simplest way is to just order a "cascode set" with your META42 board. I match the transistors for you and provide 4 that will work together. You're welcome to buy a bag of transistors and match them yourself, but it sounds like that doesn't interest you.

Quote:

would it make an improvement if i did use cascoded Jfets with the ad620?


Yes.

Quote:

The last point discussed so far is the C5 capacitor. you said that R3-R6 prevents the need for this capacitor depending on the op-amp that i use, knowing that i will be using the ad620, what would you recommend now?


The multiloop topology reduces the need for this cap, but it doesn't eliminate it. However, that cap was added mainly as a sop to silence certain critics. I have personally never had to populate C5. The critics have a valid theoretical point, but in practice C5 doesn't seem to be necessary in most amps. Leave C5 out until you know for certain you have to have it. Just like R8 and R9, IMHO.
 
Dec 3, 2002 at 8:24 PM Post #10 of 38
Sorry to hijack your thread Kal
smily_headphones1.gif
I do have related questions though I believe...

Quote:

C1 prevents damage to your headphones by blocking DC voltage on the signal line (the audio input of the amp). This will prevent the DC voltage from being amplified by the opamps (hence, frying the phones). If your source doesn't have DC offset, C1 is not needed.


So... if I were interested in building a meta42 for use with my soundcard would I use a DMM set to measure DC voltage coming from the line-out on my soundcard while playing music? If so, I'm getting readings of under+/- 1 mV. Does this sound right? Would this be enough DC voltage to require C1? I plan on adding a decent source eventually (CDP/SACDP) and am wondering if many of those output any DC voltages with the signal.

Additionally, would anyone know the impedance of the Sony V6? If there are no capacitive/inductive elements in a pair of phones am I right to say that measured resistance of 130ohms at the connect would also be the impedance? I'm curious because I'd like my meta42 to be able to handle both a V6 and eventually HD-600 (300ohm impedance).
 
Dec 3, 2002 at 9:04 PM Post #11 of 38
ok, i did not realize that there were two amplifier sections, but that has me kind of confused. If there's a right and a left channel, then wouldn't you have two output and two input jacks as well, or is there still only 1 input and 1 outupt jack?

So basically the ground buffer is also an EL2001.....so when and if i stack the buffers, i will only stack the right and left channel, but not the "2001G" is that correct? So i would have either 5 EL2001, or 1 EL2001 and 4 EL2002 buffers assuming that i stacked them.

I don't even know what gain means so i hope the op-amp i picked (AD620) is a "tolerant chip"

As for C4, is the best part for my setup the one that you had listed with the 10 micro F capacitance, you have other parts listed in your opptional parts list, but none are in bold so i think that means you haven't used them.


I think ordering the cascode set from you is a very good idea
smily_headphones1.gif


I will also leave out R8 R9 and C5 for the time being, unless i have hissing problems.

In choosing the buffers, what would you go with if you were designing this thing for portable and home use? The EL2001 or EL2002? And if i went with EL2002 would "2001G" also be populated with EL2002?

Lastly, do you think i should include C1, and if so, what capacitance should i choose?

Thanks
 
Dec 3, 2002 at 9:30 PM Post #12 of 38
Quote:

...then wouldn't you have two output and two input jacks as well, or is there still only 1 input and 1 outupt jack?


Depending on the type of input jack you choose (RCA or mini) will determine how many. RCAs you will need two, minis you will need one. You will only need one headphone jack, just make sure it's a stereo jack (3 conductor). Unless, of course, you want to build it with both a 1/4" and a 1/8" jack for flexibility. Quote:

In choosing the buffers, what would you go with if you were designing this thing for portable and home use? The EL2001 or EL2002? And if i went with EL2002 would "2001G" also be populated with EL2002?


Personally, I would go with the 2001's if you are going to use both wall and battery power. You can populate the VG buffer with the 2001 and the output with 2002s, I have done it with no ill effects. Quote:

Lastly, do you think i should include C1, and if so, what capacitance should i choose?


The higher the better. I typically use a 1.0µF poly cap in that position.
 
Dec 3, 2002 at 9:35 PM Post #13 of 38
Quote:

So... if I were interested in building a meta42 for use with my soundcard would I use a DMM set to measure DC voltage coming from the line-out on my soundcard while playing music? If so, I'm getting readings of under+/- 1 mV. Does this sound right? Would this be enough DC voltage to require C1?


It's quite common for audio outputs to have near-zero DC offset. As long as total DC offset through the system is under roughly 20 mV, you're okay. Since the average headphone amp will have a total system gain of 1-2x, between the attenuation of the pot and the 5-10x gain gain following it, 1 mV of offset isn't nearly enough of matter.

C1 is paranoia. We put it in because we don't want people whining to us if they happen to have a source with significant DC offset and the amp fries their cans. Chances are excellent that jumpering C1 is safe, but that's a bad gamble when people are plugging $300 headphones into this amp.

Quote:

would anyone know the impedance of the Sony V6? If there are no capacitive/inductive elements in a pair of phones am I right to say that measured resistance of 130ohms at the connect would also be the impedance? I'm curious because I'd like my meta42 to be able to handle both a V6 and eventually HD-600 (300ohm impedance).


Yes, you can just measure the resistance of the headphone to get a rough idea of working impedance, but it's not the right number to answer the question you're asking. My Koss KSC-35s get loud even with weak sources; my Senn HD-570s require a pretty strong source to get loud, and their impedance is 64 ohms. See what I mean about this being the wrong number?

The one you want is "sensitivity". This tells you how loud the headphones will get with a particular signal. Usually it's given in terms of a 1 kHz signal with a 1 V RMS amplitude. You also need to know the full line level voltage of your source, and then take into account pot attenuation. All together this will tell you where gain should be.

That's the scientific way. The practical way is to just try various values for R3 until you get the right balance between gain and "touchiness" on the volume knob.

Quote:

If there's a right and a left channel, then wouldn't you have two output and two input jacks as well, or is there still only 1 input and 1 outupt jack?


The META42 is a stereo amplifier -- 2 channels. Your headphones and source are also stereo components. An 1/8" and 1/4" jack are a single connector for stereo sound, but you need a pair of RCA jacks for stereo. So the answer is either, "yes you need 2 jacks", or "yes you need just one jack". Both are correct.

Quote:

So when and if i stack the buffers, i will only stack the right and left channel, but not the "2001G" is that correct? So i would have either 5 EL2001, or 1 EL2001 and 4 EL2002 buffers assuming that i stacked them.


You can also stack the ground buffer, if you want. This is discussed in the META42 docs.

Quote:

is the best part for my setup the one that you had listed with the 10 micro F capacitance, you have other parts listed in your opptional parts list, but none are in bold so i think that means you haven't used them.


The documentation tells you whether it's better to get the 10uF part or the 6.8uF part. See the "Capacitor Voltage Tolerance" section on the "Assembly Tips and Hints" page.

Quote:

I will also leave out R8 R9 and C5 for the time being, unless i have hissing problems.


You jumper R8 and R9 -- don't leave them out. And C5 doesn't fix "hissing problems". It's to fix oscillation, which can sound hissy, but most often it doesn't. Even if it does sound hissy, it's usually loud, whereas R8 and R9 are there to fix very low hiss levels with certain setups. Entirely different symptoms.

Adding C5 should never be your first troubleshooting step in the META42.

Quote:

In choosing the buffers, what would you go with if you were designing this thing for portable and home use?


2001s.

Quote:

astly, do you think i should include C1, and if so, what capacitance should i choose?


See the docs.
 
Dec 3, 2002 at 9:40 PM Post #14 of 38
Quote:

Originally posted by tangent
Just the inner loop. However, a high inner loop gain isn't a cure-all with fast op-amps like the 637. Bypass caps, fast buffers, and other tweaks may still be necessary. So, newbies shouldn't use chips like this. Kal525, you should use a tolerant chip until you're ready to start gaining deep troubleshooting skills. (Which is to say, you're ready to start getting into deep trouble.
smily_headphones1.gif
)


Yeah, I think I've figured that out the hard way. I have a pair of OPA637s in my multiloop amp and they aren't happy unless I keep R3 down below 500ohms (total gain=6). I like my total gain to be around 1.5-2 (I listen low). Inside loop gain is 101, and it still is prone to occilation. One possible source of problems may come from the PS, as I have many electrolytics across both rails, but only some smallish film caps between the Vground ourtput and the rails. I have tons of board space though so that may change very soon. Suggestions on type and values of caps to add would be welcome. I already have about 7,000uF worth of caps in the PS, but they are mostly before the final inductor and just help smooth things out.

I have another question that is somewhat related to this thread, I am using BUF634 buffers in all positions of my amp and currently have all of them in high-bandwidth mode because this is a wall powered amp. Is this a good idea, or should I leave the virtual ground driver in a lower bandwidth mode?

You mention a capacitor to control the opamps' occilation, could you elaborate? I can get opa228s to work just fine, but these 637s are tough. I plan to do a few reviews of different TI chips if I ever get them all working.
 
Dec 3, 2002 at 10:20 PM Post #15 of 38
alright, i'm getting pretty overwhelmed now. There's just so much to think about and i really don't understand half of what's going on, i'm just picking parts based on what is recommeneded. I really don't understand the technical explanations of how to pick parts that well. I do think it would be cool to be able to put this thing together myself though

JMT, i saw a picture of your grey meta42 amp you built and thought it looked pretty cool, i was wondering if you could give me a parts list or something like that, because now i am confused in more than one area and i don't feel like bothering all of you with what are probably simple questions to 95% of you. At first i didn't even know i would need an amp for the hd600s, but after i learned i needed one, i went from wanting to purchase an airhead, to stumbling upon these forums and seeing that the meta42 is "easy" to build and outperforms the airhead by a lot.

I'm sure i will be able to put the meta42 once i have all the parts, it's just that i do not understand some of the technical explanations of how to choose the parts and, i just want something that sounds good with my headphones without paying $200 for a flimsy total airhead.
 

Users who are viewing this thread

Back
Top