Yes, output of 2700 has transformer. It is pretty obvious they didn't just twist it up and insert it into circuit without extensively testing its performance, and providing results as performance specification statistics.
Suggesting addition of an output transformer to amplifier for suspected power supply issue is backwards. If power supply is limiting performance of amplifier, improve the power supply.
Sure, for audio listening is final arbiter for perceived sound quality, but is no way to design and build, especially for novice.
Good place for novice to start is with proven design. Sure proof of design could be by totally subjective listening, but by many experienced listeners that share same impressions across a spectrum of diverse designs.
A single builder's subjective results have little merit. They all turn into "It sounds great to me, I think it will sound great for you" conjecture. If somebody builds one of your designs, and on listening finds serious flaw, can they fault your design? Not without measurements that rule out construction error, and show consistency with similar metrics for design reference, or sending unit to you for evaluation and then trusting your opinion.
Suggesting addition of an output transformer to amplifier for suspected power supply issue is backwards. If power supply is limiting performance of amplifier, improve the power supply.
Sure, for audio listening is final arbiter for perceived sound quality, but is no way to design and build, especially for novice.
Good place for novice to start is with proven design. Sure proof of design could be by totally subjective listening, but by many experienced listeners that share same impressions across a spectrum of diverse designs.
A single builder's subjective results have little merit. They all turn into "It sounds great to me, I think it will sound great for you" conjecture. If somebody builds one of your designs, and on listening finds serious flaw, can they fault your design? Not without measurements that rule out construction error, and show consistency with similar metrics for design reference, or sending unit to you for evaluation and then trusting your opinion.
How does a novice go about improving the power supply on an existing amp? Any guidelines for the OP?
OP hasn't stated what current amplifier is, or it's power supply, but has stated desire to build starting with an existing PCB. This implies a design for which interest induced a production run of PCB, for which interested builders used, and likely created feedback for.
OP didn't claim desire to modify existing setup.
OP didn't claim desire to modify existing setup.
You brought up poor power supply in #2:
Doesn't adding 5:1 output transformer require higher voltage swing from amplifier? Does this help if amplifier has inadequate power supply?
Yes, it does.
It helps because by 'inadequate power supply' I didn't mean it didn't swing sufficient volts, rather that it sagged whilst driving the 'phones due to insufficiently low output impedance at LF. A typical headphone amp needs to cater for a much wider impedance range than a typical amp driving a loudspeaker, so it'll often have enough voltage swing to drive a 600ohm load because this is about the highest impedance seen in 'phones. Yet many headphones nowadays are 16ohms - the excess voltage swing isn't required (it'll overload the drivers) so it makes sense (at least to me) to transform the 16ohms to something higher and better utilize the available power.
So then - back to my earlier question. What's the thinking behind suggesting improving the power supply?
Last edited:
It all goes back to your claim that "typical headphone amps lack in dynamics due to inadequate power supplies for driving low impedances"
Which, I've to say, is a rather vague claim and probably flat out wrong for all the designs suggested in this thread. Specify what you mean as "adequate power supplies for driving low impedances" and you might get a meaningful answer.
How so? I can't see - do please add some meat to the bones of this.
Meaning lots of capacitance on the rails (10mF would be the bare minimum) and full regulation, low output impedance rails. If you know of such, do flag them up. It could well be that I just need to get out more
Why would it be flat out wrong, even if only 'probably' ? Any explanation on offer or is this just an unsupported claim?
You shouldn't lump the O2 headamp and a CMOY together, either.
The O2 uses a separate chip for gain and two paralleled high current chips for output that can source/sink up to 120mA or so of current and hit jcx's number. A single battery CMOY with virtual ground may not even be able to hit jcx's 3.2Vpk swing number either because you are dividing the 9V battery in half, minus chip overhead. The O2 uses two batteries giving around 7V peak swing. A CMOY with two batteries (they exist) still likely can't push even half of 100mA if using just one chip.
I say "likely" because the Grado RA1 headamp is essentially a single chip CMOY that uses one of the two output chips paralleled in the O2, so it can hit 60-70mA. Most CMOYs use chips that are only good for 20-30mA or so of output current.
Never mind agdr. Lets see what a decent headphone amp looks like
Regards
Really ?
I couldn't be more obvious. You claim that the power supply is the weak link in many headamps and suggest an unpractical solution (the transformers). It seems rather logical to point out that an inadequate power supply is something that can be fixed and that it is often easier/cheaper than using transformers.So the goal is :
- 10mf. Why 10mf rather than 4.7mf or 22mf ? You ask for supported claims, support yours too.
- full regulation (whatever "full regulation" means). Even the "lowly" lm317 have very low output impedance at LF. You can quite easily get down to 10mR. Is this enough for you or not ? If not, why ?
Really ?
No, its only your claim that transformers are impractical. You've not yet demonstrated that they are.
But when its noted (as Barleywater has) that the OP shows no signs of wanting to do that, that rather makes it an impractical suggestion. What am I missing? From my pov the opposite to your claims is true - modding a power supply is very much an individual thing because all headamps will need different solutions. A naive DIYer stands a very big chance of screwing up such a mod without a lot of hand-holding about what caps/regs to choose, how to mount them, whether the existing case will accommodate them physically etc etc.
By contrast a trafo is an external device which can work its stuff on a wide variety of headamps and can be wound according to the particular model of phones one has.
Which amps have you found using LM317s?
The jury's out on that one at present. I'm still researching audibility of regs.
If power supply is sagging, it is fault. Demands on power supply do not change with output transformer. Getting same power out of headphones/speakers places same demands on power supply.
You meant its at fault? The sag needn't be great - a few mV.
Nonsense - the ripple (sag) on the power supply is a function of the current drawn. Higher impedance loads draw a lower current for the same power output.
Back once more to my earlier question - improving the power supply. Why?
Last edited:
Trivial:
- the only existing commercial good ones are very expensive (Sowter's)
- there is no good explanation to be found anywhere on how a beginner could properly wind one. Where is your explanation ? Where are the measurements proving your claims ?
Hence, saying to a beginer to use transformers is completely useless if you're not willing to hold his hand through the process of building and measuring transformers. Good luck with that.
The OP hasn't asked for much: a design that works well, with a pcb available. And he said explicitly he didn't want to wind transformers.
Restraining myself to actual suggestions made in this thread:
The altronix kit doesn't have an integrated PS, so one can get a good one.
The minimax uses one lm317.
The CKIII I suggested uses lm7815/7915 which have similar output impedances at LF.
The Gilmore JCX and I pointed out has an even better PS (probably completely overkill).
Irrelevant as I didn't suggest buying off the shelf.
Should the OP have been interested enough to follow up my suggestion, I'd have given him that. As its fairly simple, no problem. But as he said he wasn't at all interested I offered no explanation.
Why would measurements prove something about a subjective experience?
I was willing so your remark is irrelevant. That makes two irrelevant remarks and one irrelevant question so far and I'm only half-way through replying
Yes he did - after my suggestion. Which is fair dos. Yet after that, the idiotic remarks about transformers began to appear so I returned to the thread to explain precisely why they were idiotic.
What counts as 'good' in your estimation? Got an example?
Thanks.
Ugh... We're speaking amplifiers design, that's not a purely subjective experience. Measurements are needed to known that you're not simply building a sweet distortion machine. I'm not asking for 0.000001% thd but a flat frequency response and some idea of the added THD is a bare minimum.
Furthermore, with transformers, you need to have at least an idea of the inductance and DCR ranges, in order to properly interface them with the amplifier.
Overkill but good: The σ22 Regulated Power Supply
Overkill but good: http://www.diyaudio.com/forums/diyaudio-store/247281-super-regulator.html
Good: what was used for the "wire" headphones amp
Probably sufficient: Alimentation linéaires - Module d' Alimentation linéaire régulé Double DC LT1084 12V 1.5A
The whole debate is pretty much academic anyway. Many headphones amp will indeed have problems with a low impedance load because they were not designed for it, neither the power supply nor the output stage. But those designed for low impedance loads will most of the time have a decent power supply. Most of the amps suggested here are fine with the 32r grado.
Ironically, I would agree that a transformer is not a bad idea, it fixes a lot of problems (gain structure, output stage limitations, etc). I actually use a pair at the output of a tube amp. But if the only problem is the power supply, transformers are overkill.
Hold on a second here - I just backed up to see how we reached this point. You'd asked about 'measurements to prove my claims' but in fact I was talking about dynamics which are subjectively perceived, not measured. I said that in my experience the dynamics improved with a suitable trafo. If I'd claimed something objective then perhaps measurements would be relevant. The context most definitely wasn't 'amplifier design'.
I'd agree but we'd not got to that part because the OP didn't express an interest. Had he done so then I'd have talked about numbers of turns and the like.
Fine as in produce adequate SQ? I wouldn't disagree - but I got the impression that the OP was hankering for a little more - he did after all mention 'Class A'.as a reference point for 'decent quality'. How many of the amps you've mentioned are that?
Do explain why - a better solution (bearing in mind that the OP doesn't want to mod his power supply) is? A trafo does also have secondary effects - distortion from the output stage (nasty crossover spikes) tends to be considerably lower with higher impedance loads. There's less modulation of transistor beta with output current (because the output currents are lower). The transistors undergo less heating (again because output currents are lower) so thermally induced problems (bias shift for example) are reduced.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.