"The Wire AMP" Class A/AB Power Amplifier based on the LME49830 with Lateral Mosfets

hey Cristi, just reaffirming my interest, if all goes well i will take 2-4 (assuming one powers one balanced channel? or 2?) you could also look at a website called zero, who do really nice deep drawn enclosures, they fully anneal them and even have mumetal options. they have like 30000 already tooled shapes so the likelihood of needing a custom design is slim.

on that note would it be possible to get without your enclosure as an option if you don't go for the above? because i will probably go for one from there and would rather not pay for an enclosure i won't use. they will do further tooling as needed too and i expect bulk pricing.

i would be looking at 60-70 and 45-55vdc voltages. with fully regulated preferred

the enclosure does look good though, i guess i'll see how it pans out
 
Last edited:
Looks great Cristi!

I see a nice PFC and a pretty hefty looking transformer (ETD35 or ETD39?). The PFC will be nice to have, especially for those looking at multi-channel setups where you can quickly max out a 15A circuit without a good PFC. With the PFC you should be able to draw a full 1800W from the wall all day long on a NA 120V/15A circuit.

I can't see where you're taking the feedback for regulation, but I'm guessing it's off the high current rails to maintain regulation under load? If so, then the low current supply will probably wander upwards under heavy load, so it might be advisable to post-regulate those supplies. The current draw is so low that even an LM317/337 would easily be up for the task as long as the outputs don't get shorted. I don't see any post regulation on board, but nothing is stopping people from adding their own in between.

I really like that style of case as well. Cheap and functional. I've use that style of supply before for 36VDC motor drivers, and they worked admirably. You can mount them pretty much anywhere, and the shielding should help keep radiated noise to a minimum providing the outputs are well filtered before they leave the chassis.

If you need some assistance in testing these, I'd be happy to give it a go on one of my prototypes here. If you ship it here, then I'll cover return shipping or I'll just buy it if it works. Let me know what would work best for you!

Regards,
Owen
 
At the high current output GND I would think. It might be a benefit to have a few extra solder points for speaker GND and input GND so everything can be cleanly returned to the supply. I have all of these points on the board, but it's usually better to return everything to the supply.

Cheers,
Owen
 
cool, as i thought, yeah i agree it would be nice to have multiple copies of the high current output ground, or just 2 of a heavier duty screw term so that heavy duty ring terminals could be stacked on top of each other and bolted down.

or was the picture of the enclosure just illustrative of the type of enclosure and there isn't going to be any high current screw terms on the output? i would prefer that at least for this purpose as it would make for much easier connection of more outputs, or more channels, as desoldering a 12awg wire from a star ground of this scale would not be fun.

either that, or i have seen some very neat board mount amphenol ground buss bar connectors that would do nicely.
 
custom toroidal transformer
Is that for audio or general purpose?

Are you having:
a 30% over wind to reduce magnetostriction noise,
an inter winding screen,
GOSS wrapping,
potting,
insulative over wrapping,
M0 core (or at least M3),
silicone core impregnation.

Instead of 55 volt windings you can have a pair of 12, or so, volt windings and join them to the main windings to get the extra 12 volts.
 
Hi Ian,

Sorry, I did forget to add you again! I'll put you on the next update, and you can definitely order just PCBs. I'll add you to the list for two more.

As for mounting the LME directly on the same heatsink as the transistors, National specifically recommends againstdoing this, but I think it's more for thermal reasons than HF coupling. Either way, if the manufacturer tells you not to implement it a certain way, then I'm sure they have a good reason for it.

If you want to add external protection diodes for use with the Toshiba parts, I would suggest soldering them directly to the pins of the transistor before you mount it.

As for the transformer, your list is more of a "nice to have" than "absolutely necessary". It will likely drive the cost up to $250 and depending on how and where you mount it, it might not provide much benefit overall. I think if you want something that extreme, you might need to organize it yourself. In terms of winding arrangement, you are correct, it would probably be better to just add a few turns to the existing high current windings to get the extra voltage.

Cheers,
Owen
 
I expect these will be fine for many people but for those wanting to draw more power and needing to better get the heat out of the devices the 6mm thick 'chassis' of these may be less that adequate and 10mm thick sinks would be more suitable.

Bear in mind that mosfets 'r on' increases as the devices get hotter. So getting the heat out and running them cooler is a good thing. You might like to aim for 45°C which is perhaps only a 20°C rise above room temperature.

For 80mA per pair, 3 pairs of 2SK1058 2SJ162, we run a 300mm x 150mm 40mm comb section sink with the 10mm 'chassis' thickness. That's for about 220W rms mono block amplifier and about 40°C sink temp at quiescent current at about 23°C room temp.
 
Last edited:
I'll put you on the next update, and you can definitely order just PCBs. I'll add you to the list for two more.
Thanks very much 🙂

mounting the LME directly on the same heatsink...more for thermal reasons
Yes I forgot about thermal coupling. Thanks.

external protection diodes....directly to the pins of the transistor
I was just concerned about excess current through them in the over voltage condition, and also if there was any possibility of them failing closed circuit and damaging the gate resistor and maybe then the LME chip.

As for the transformer...your list is more of a "nice to have" than "absolutely necessary".
Whilst that can be true if you get 'lucky' with a general purpose one (as we have with some 500VA and 600VA, but then we tried many from probably all the manufacturers in the UK and got those off the shelf ones from one we thought had done a generally good job with the Audio ones) I was trying to point out that specific audio ones are available and can be better.

I certainly can't tolerate a badly humming transformer, even a quietly humming 500VA sets a background noise level that eventually can become a nuisance. And some of them electrically radiate quite badly. An associate had a 500VA sample, the field from which was detected meters away using a simple sensor he made. Not acceptable is when such a field intersects a phono stage and causes hum and noise. We've had all this.

I currently have left in stock 10 x 1000VA, 2 x 600VA and 6 x 500VA Tx's with various features from the above list and more. In one case the manufacturer was able to supply me a special shaped stack developed during research at a UK university. The lowest mechanical noise and best sounding had the primary on the outside. They have all the features in the list except GOSS but aren't intended to be used near a phono stage.

I've had them a for a while but the price was about £80 each for 10 ish delivered to the same address. I don't know what that would be now but I expect a lot higher.

I've not really tested the big ones against smaller ones in lower power amplifiers and it may have been folly to some extent to get them. Except that they are 'voltaged' for the 300W mono blocks I was making. Those amps can deliver cleanly 600W into 4 ohms and 1000W into 2 ohms or so.

If I recall correctly, the 1000VA make about ±90Volts DC at quiescent load at 240volts mains ac. The 600's about ±80V and the 500 make about ±70V. They all have a pair of about 12v windings generally with about 0.56mm wire.

I think that now out of curiosity I would be interested in trying an R core to see what that did.
 
Last edited:
The conviction that stock toroidal power supply transformers are "best" for audio is rather persistent (to say the least...).
When you (Ian) live in the London area problems with toroidals do not surprise me.
I know of one person in your neighbourhood who discovered that (in this case primarily because of line overvoltage) stock 240V toroidals are useless because core saturation caused them to hum and rattle "all over the place"...
Lower T transformers, preferably with a very small airgap, might solve your problems.
C-cores are very well suited (R cores "might" be better but these are, like toroidals, also closed magnetic loops with their inherent sensitivity for line overvoltage).
 
I expect these will be fine for many people but for those wanting to draw more power and needing to better get the heat out of the devices the 6mm thick 'chassis' of these may be less that adequate and 10mm thick sinks would be more suitable.

Bear in mind that mosfets 'r on' increases as the devices get hotter. So getting the heat out and running them cooler is a good thing. You might like to aim for 45°C which is perhaps only a 20°C rise above room temperature.

For 80mA per pair, 3 pairs of 2SK1058 2SJ162, we run a 300mm x 150mm 40mm comb section sink with the 10mm 'chassis' thickness. That's for about 220W rms mono block amplifier and about 40°C sink temp at quiescent current at about 23°C room temp.

and your point is? i already mentioned in that very post (or the one before) that those who have very high power needs maybe need to look elsewhere; it was a suggestion of a group buy for 2 of the more common power requisites, group buys are not intended to cater for the extremes. those people should look after themselves for their particular needs
 
Last edited:
That wasn't intended for you to answer QUSP, I used your link to just generally point out something to the readers 🙂

The point is that people who want higher power might be interested in knowing and welcome that tidbit of data 🙂

220W is by no means extreme 🙂 Fairly normal really amongst all my friends with not particularly efficient floor standers in say a 10 foot by 14 foot room. It does no harm to have some headroom and a bigger better PSU often confers better transient response.

Our 100W per channel integrated amp has 80,000uF and a 500VA. It beats 200W per channel amps. A pal with one recently took it to a high end shop to evaluate another amp and speakers he was looking to buy. The dealer commented how it beat that amp retailing at £2500 (or maybe it was £5000, I've forgotten, but it's equalled £5000 amps before). The decent PSU is in no small part responsible for this and it's why higher power amps often sound better than lower power ones.
 
Last edited:
probably those looking to build 1kw+ amplifiers already know that they have rather more special needs WRT heatsinks, not generally the type of thing you take on without knowing something about it, but i take your point.
 
I have just pulled out a stock 100VA 230:9+9Vac toroid.

I wired it up as an auto transformer.
I connected the 245Vac mains to the middle of the 9V+9V winding and the neutral to the far end of the 230V winding.

Measuring the voltages on open circuit load I get.
245Vac input, and outputs of 235Vac @ 5Aac, 245Vac @ 13Aac and 255Vac @ 5Aac.

This little 100VA transformer which is very cheap could be used to ensure that UK mains supply voltages of 235Vac and upwards cannot overvoltage a 230Vac rated toroids and loads not exceeding 1200VA.

Simply swapping the 240V input tap from the middle of the 9+9 tapping to the far end would give outputs of 220Vac @ 5Aac, 230Vac @ 5Aac and 240Vac @ 13Aac when fed with nominal 240Vac mains power.
 
Last edited:
Hi,

Owen, earlier in this thread you mentioned a three layer PCB including a ground plane. Is this ground plane divided in more than one area like small signal ground and decoupling ground? If so I'm delighted 🙂
 
C-cores are very well suited
Thanks for that. It's interesting as I had what I was told was custom designed 100VA winding on a 300VA C core with some super secret spec that I wasn't told, and on my CD player it produced the sweetest treble of all the transformers I tried there. I happen to have it again now. It's potted and canned so I can't easily reverse engineer it without damage?

Our audio transformers don't hum any more or any less with changing mains voltage. From 215 to 270V AC I've measured and they are quiet. What does set them off is the likes of a running hair dryer plugged into the house sockets anywhere. The ones with the outside primary, which are normally completely silent, become the loudest, making a clacking sound!
 
Last edited:
This little 100VA transformer which is very cheap could be used to ensure that UK mains supply voltages of 235Vac and upwards cannot overvoltage a 230Vac rated toroids and loads not exceeding 1200VA.
I think you might find that the high source impedance from that will kill the sound quality.

I expect it's been known for ages but I recently discovered the difference it can make when juxtaposing the mains supply L and N wires around with respect to the transformer. Done it on 500VA Toroids and a small R core. All with a surprising size of change. There's each time been an easily to hear best way around. You just have to try it yourself. (I'm assuming I don't have to say in this thread that you should get a qualified electrician to do it if you're not sure 🙂 )
 
Last edited: