Is Higher VA always better? and Avel close or Antek exact?
Are there any disadvantages to using a 400VA transformer over 300VA for a chipamp with one power supply for two chips?
Put another way, if you want 2x 22V (or 18V or 25V or whatever) is it always better to have 400VA than 300VA?
Does the implication hold for 500VA over 400VA, and where is the limit where things get complicated (with the soft start I've read about?)
Part II of this question:
Suppose you want to drive 8 ohm speakers also let's take arguendo that 22V is just right for a chip like LM3875 in a stereo (two chips one transformer) setup.
Suppose next that each of the following have the same price. Are any of these transformers be unambigiously better?
a: Avel 330VA 2x25
b: Avel 250VA 2x18
c: Antek 300VA 2x22
The subtlety of this question is that there is the additional factor of make.
Avel has a pretty good reputation as a plitron substitute, but not exactly the right specs. Antek might have a lesser reputation but has a bigger menu of output voltages (compared to those from avel at partsexpress for example).
I'm trying to figure out what is thought about these trade-offs.
The broader implication is that if the answer to part I finds that if bigger is better (at least to a point) then a 500VA 2x22V might be better than a 330VA 2x25V even if Avel transformers are unambiguously better than Antek by construction.
Better? This is a very nebulous term. There are tradeoffs like price, weight, etc. So better is a subjective term and you need to weight the pros and cons.
In general, I don't think that you can do any "harm" by using a transformer that has a higher VA rating, but the higher VA will cost more. You may need to take some care in the turn on surge when you get above 300VA, and that is an added cost and complication. So, the question is "do you really need 300VA or 400VA?". Forget about 500VA!
I happen to be developing a spreadsheet that let's one evaluate the performance of a real transformer in a linear, unregulated power supply typically used in audio amps. It models the voltage sag in the transformer secondary as a function of current draw. It also takes into account power losses - dissipated power for a class AB amplifier, lost voltage "headroom" because of ripple on the rails, diode drops, and the drop out voltage of the output devices in the amplifier. Copper and core losses in the transformer are also included. The data needed to model the transformer can be obtained from an Antek datasheet, when these are available for the transformer you are interested in, or can be obtained by measurements done on a transformer in your possession.
I plugged in some numbers for the Antek AN-3222 (22VAC secondary, 300VA), which seems to be one of the units you are considering. I find that, regardless of the chip amp you use, the power supply can support for stereo operation about 63WPC into 4 ohm loads and about 33WPC into 8 ohm loads. The power is limited by the maximum rail voltage that can be generated by the PS. The VA rating on the transformer can support a much higher peak power, including the crest factor, but the rail voltage is limiting.
The above limitations are purely for how much power the power supply can deliver to the amplifier. There may be addition limitations posed by the amplifier itself, for instance it may not be able to support the rail voltage or output current required to drive these loads. Nonetheless, the data gives you some direction in your choice of a transformer. From the attached screen shots, you can see that the transformer is over dimensioned in terms of the VA specification and the amplifier output power is limited by the rail voltage. This indicates you can significantly reduce the VA rating of the transformer if it will be used to power a 2-channel 35WPC@8ohm/65WPC@4ohm stereo amp. Something on the order of 150VA should be sufficient to meet the needs of these power levels.
I ran the numbers for the Antek AN-1222 and AN-1224 and both come up a little short voltage wise (because of sag). I suggest you go for the AN-2222 or, better yet, the AN-2224 (as long as the chip amp you choose can handle the rail voltages of +/-31V DC that will be generated under no-load conditions. Output power will still be limited by rail voltage but you won't be wasting money on more VA than you really need.
You can always waste money on an amplifier spending more to make
it "better" with bigger transformers, bigger/more expensive caps etc.
Fact is though a carefully designed and optimised for cost amplifier, for
the same price, will walk all over something your chucking money at.
The VA numbers you are talking are way over sensible, halve them
at least, in reality about 1/3 to 1/4 would be fine for an amplifier.
(Its a music amplifier, not a sine wave generator, and its not run
at full volume most of the time, you need a good perspective.)
Chip amps and big transformers is a design oxymoron.
Seems like 300VA is the upper limit before worrying about turn-on surge. I'm going to take that as the highest I want to go.
The issue to me seems to be that the difference between the 300VA and 200VA is about $5.00. (At least on Anteks).
That's less than half the fixed cost of shipping. I'm not a high roller, and I've decided I'm not buying a pintron for my first go at this, but I'm not worried about the price difference between Avel and Antek. I'm also not worried about size or weight for my first build.
I just want to buy something that is as right as I can get. Seems that the 22V is what Peter D. prescribed for the LM3875 and it is in a sweet spot as somewhat universal for speaker resistance.
I regret that I don't really get the whole picture with your spreadsheet. I think it is time I just get cracking. I have a feeling this won't be the last transformer I ever buy, so maybe next time I'll try something a little different.
To sreten: the problem is that I'm not really chucking money. The difference between 160VA and 300VA is negligible. I think that's why I'm not sure how to optimize. I really don't care if the transformer is $30 or $60 or even $80. The shipping is the same on either and either 1/3 or 1/6 of the price.
Sreten, I don't want to come off the wrong way here. I really don't want to spend more than I have to either, but it is the design part I'm concerned about.
If higher VA isn't worse in any way, why not?
If design oxymoron means worse performance I would like to know why.
If it doesn't, I'm OK with having something that is overkill since maybe I can put it to a different use one day.
It is exactly this that is my question:
>a carefully designed and optimized for cost amplifier, for
>the same price, will walk all over something your chucking money at.
If there is really an optimum, I'm all for using it. It seems 22V seems to be a good spot on the graph for LM3875. Higher starts to get close to the limits, and I haven't seen a lot of advocacy for lower. Maybe it is out there? I'm not going in to this with anything but ambition to make a nice machine.
Deciding on the VA rating seems simple. More doesn't cost that much, I don't want to deal with this soft start business so 300VA seems to be a kind of sweet spot.
If there are clear advantages to going lower--other than price--- I would. I don't yet understand what those advantages are.
Without disadvantages of going larger, the disadvantage of going lower than necessary is constraints which become binding for unknown future preferences.
Fair enough. Such is DIY that you can make choices that would make
no commercial sense. There is nothing I know that investigates the
consequences of overdoing it in a commercial context, most are
about what you actually need for a given context.
Having said that, if you've got a big power supply by far the most cost
effective thing you can do to a chip amp is parallel two of them with
the same supply and heatsinking, pretty much a no-brainer IMO.
Agreed. That's what I'm going for. I have yet to hear anyone knock Antek in any serious way. Seems that they are making coils of wires that are doing the trick.
You wrote AN- models from Antek. Seems that the AS designation is specifically designed for audio, and cost about $5.00 more than the AN versions.
Not sure if this was an intentional distinction.
A fascinating thread.
One minor correction: the difference between Antek's AS-3222 and AN-3222 toroids is that the AS-3222 is shielded, whereas the AN-3222 is not. Antek doesn't offer shielding on all its transformers, but I find the $5 they charge to be a very inexpensive way to combat hum.
Also, I've happily used Plitron toroids when I could afford them, though I wonder why they are roughly double the price of similarly-sized Antek products. My experience with Antek toroids has been very positive and I wouldn't hesitate to use them, especially in a chip amp.
|All times are GMT. The time now is 08:21 AM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio