| pjanda1 |
I'm beginning an LM3875 project. I've read over many of the archieve threads on tranformer selection. It looks to me like 160VA should be more than adaquate for a stereo amp, yet many choose a larger unit. Is there any empirical reason to go larger? I'm specifically looking at the 22v 160va Avel from PE. I also want it small because I want to squeeze it into the smallest possible chassis.
thanks! |
|
|
| Nuuk |
160va is enough (IMHO) for a good stereo Gainclone.
Larger transformers may have an advantage for reasons that I will not state here.
BUT, and I welcome comments from others here, I have found larger transformers to be a little noisier (ie louder hum). |
|
|
| peranders |
| You could manage with a smaller transformer. Think average power outtake. I think down to 80 VA isn't too wrong. |
|
|
| pinkmouse |
| I have a mono LM3875 running from a 30VA 18-0-18. It's used as a loudspeaker test amp, but goes more than loud enough for most uses. |
|
|
| chris ma |
| quote: | Originally posted by pjanda1
I'm beginning an LM3875 project. I've read over many of the archieve threads on tranformer selection. It looks to me like 160VA should be more than adaquate for a stereo amp, yet many choose a larger unit. Is there any empirical reason to go larger? I'm specifically looking at the 22v 160va Avel from PE. I also want it small because I want to squeeze it into the smallest possible chassis.
thanks! |
For mee.-self anyway
I picked larger ones so that it can be recycled for other projects in the future. There is not much different in the $ amount for higher VA, but the trouble of looking/buying one and wait for it.. the delivery time....
The Butcher:D |
|
|
| mod_evil |
As i transformer VA to Ampers?
I have 127v in the frist and 15-0-15 em second.... of transformer.
Thanks. |
|
|
| sam9 |
You would think VA=Watts but it doesn't. On the other hand that's a way of thinking about it that may be crudely useful. I think it is more accurately considered a thermal rating -- the higher the VA the cooler it remains. Not surprisingly there is a relationship between weight and VA.
While you could determine the VA rating needed to drive the amplifier with a continuous (i.e., day after day non-stop) sine wave at max power as a way to ensure the VA rating is totally adequate-- as a practical matter, normal (or even vigorous) home use requires rather less than that.
Dopes the above sound vague? Good choice of transformer VA rating is a bit vague by nature. |
|
|
| merlinx76 |
| quote: | Originally posted by sam9
While you could determine the VA rating needed to drive the amplifier with a continuous (i.e., day after day non-stop) sine wave at max power as a way to ensure the VA rating is totally adequate-- as a practical matter, normal (or even vigorous) home use requires rather less than that.
|
A lower VA rated transformer will have a lower voltage at the demanding peaks of the audio signal which is when the voltage is needed most. While I agree that a transformer wouldn't need to push that current continuously but it still has to maintain desired voltage when the peak current is required.
When you think about it there is not much difference in price of transformers to go up a size or 2 just to be safe. Personally, I don't want to skimp here. |
|
|
| kmj |
| quote: | | I picked larger ones so that it can be recycled for other projects in the future |
A wise choise, for as someone in the tube-forum said "You don't just build ONE amplifier":angel: |
|
|
| djk |
I used an 80VA 18-0-18 to run a stereo amp with 8 ohm speakers. It sounded a bit 'thin' when driven hard.
Changing to a 160VA 20-0-20 gave a much better sound.
Even though the AC voltage was higher, the DC voltage was about the same with the 160VA unit because it had better regulation. |
|
|
| grapetonix |
| I'm also thinking of doing an LM3875 stereo design. Possibly using 80VA 24-0-24 toroid. Is this just too weak? I'm not going to suck more than 30 or possibly at the very highest 40 watts per channel out of it. |
|
|
| djk |
| It will make noise (ie: work), but don't expect the best sound quality. |
|
|
| Nuuk |
| Grapetonix, it sounds like you need to read this !;) |
|
|
| DDF |
Hi,
I've been going through this design exercise with the LM3876, which has similar (if not identical) drive current limitations to the 3875. The choice of appropriate transformer for the 3875 depends on the speaker impedance.
This part has serious difficulty driving 4 ohm loads, and is much more comfortable at 6 ohms and above. Under the premise that Spike sounds bad and you want to avoid it, if your speakers dip below 4 ohms, the supply rails must be below 25V to avoid spike. This means no more than an 18V secondary. If the speaker is 6 ohms and above, you can go to higher rails, and the 22V secondary, without triggerring Spike. At 22V rails, and speaker impedance above 6 ohms, 160VA is more than enough as the bridge load is less than 110 watts into two speakers.
I design speakers and don't like being restricted to 6 ohms. I'm making my 3786 work without Spike into 4 ohms loads, by using 18V secondaries. I've designed mine so that unloaded, it results in +/-26V at the rails (my line voltage is 120V, not 115V). When loaded, it dips down to just below 25V, allowing the supply to just dip below the Spike threshold (~25V) on hard transients, but stay above it for very short transients which won't kick in Spike.
In summary, for the 3875/3876, into 4 ohm loads, I would argue that they benefit from LESS stiff supplies for this reason: more voltage drive into 6 ohms and above, and avoiding spike at lower impedances as the supply sags with current draw.
On the flip side, I would recommend spending the extra $10 for a pair of LM3886s. Its behaviour into low impedance loads is far superior, and you can then go to higher rails, with bigger trannies.
I'm sticking with the 3876 for now, just as I'm making this amp 100% from the parts bin and don't want to buy a larger transformer. If I like the sound, I'll retrofit in 3886s and a larger tranny.
Just need to make sure the heat sinks and cap voltage ratings are appropriate for the larger amp, day 1. |
|
|
| sam9 |
| "Soggy" as opossed to "stiff" isn't necessarily bad. The story, possably apocryphal, is that Carver noticed very early that some supossedly modest performance amplifiers sounded much better than they should. Deciding that the their soggy supplies must let them reproduce transienr peaks greater than their formal ratings would indicate the idea behind the M-400 and it's sucessors was born. |
|
|
| pjanda1 |
Thanks for the advice everyone. As I said, I was hoping to keep it small to fit in a little chassis. The other reason was that I hadn't seen an easily available cheap 22v transformer larger than 160VA. It turns out that Avel is offering 330 and 500VA 22+22V units, they just aren't listed in any catalogs yet. I'm now leaning towards getting the 330VA directly from them.
I'm also still considering the 18+18V 250 VA from PE. I'm going to build some new high efficiency speakers in the very near future, but for now the amp will be driving some 89.5 db, 4 ohms nominal and minimum speakers. Hopefully the higher rails will be fine with those until I get something else constructed. |
|
|
| DDF |
Hi,
As I mentioned, the higher rails and 4 ohms will not play nicely together. Go for the LM3886, I think you'll regret it if you don't, for your application.
Sam, many prefer unregulated for amplifiers for the reason that regulated supplies can strangle transients. |
|
|
|
