Just a few newbie questions about rail voltage

[MikeHunt79]

Ok, I'm looking at the possibility of building one of those 41hz AMP8 kits. The only (low-power) amps I have built so far have been run from batteries or pre-made PSU's.

After reading one of the threads over at 41Hz on power supply requirements, it looks like I'm going to need 60v + and - rails, with a toroid capable of at least 1800VA, as this amp may pull 15A when running 4 ohm loads.

Now, here's my questions:

1) Instead of getting one huge 2000VA toroid, could I get 4 x 500VA toroids instead, and wore them in parallel? I can't find toroids over 500VA in the UK.

2) For 60V rails, I would need a toroid that generates + and - 45v... is this correct? Does anyone know of a handy RMS to peak calc at all, as I'd like to be sure... I can't find square root on Windows Calc.

3) I already have some rubycon 10000uF caps that are rated at 63V. Would these be suitable or is 63V a little too close? I could hunt down some 80V caps, but I'd like to use the 63V ones if I could.

4) Is it ok to use Diodes in series to drop the voltage on the rails if they are a little high? Say to get from 65V to 60V?

EDIT: Is it essential to use toroids? Could I get away with non-toriod trafo's instead? This will be a sub amp, and I'm also thinking about having the trafo's and caps in a separate chassis to the amp.

[Iain McNeill]

1. yes, as long as you use separate bridge rectifiers for each transformer to prevent circulating currents.

2. you'd get 63.6V peak from a 45Vrms winding with full wave rectifier. You'd need 2 such windings to get ±63VDC or a center tapped 90V winding. Or you could half wave rectify the single 45V winding into ±63VDC with more ripple. This assumes zero transformer resistance, zero diode drop and zero power drawn.

3. You need to add a good 25% to the peak voltage in case of mains supply variations. Also consider surge and transients. A zener is fast but can't handle too much power. MOV's are good but react slowly. I'd over-rate my caps more than that.

4. Diodes in series are OK. Just check power rating.

5. Don't electrocute yourself and don't work with metal tools on charged caps

[NanoFrog]

For calculating peak, just use 1.414.

[AndrewT]

Hi Mike,
what power are you expecting from your finished amplifier?

You can use a transformer VA rating that is between 1*max power to 2*max power.

+-63V supply rails could supply upto 300W to 400W to a 4r0 load.
This max power could be run from a 300VA to 800VA transformer.

I got 1000VA transformers from Farnell @ £60 each and they are available from Rapid as well as RS.

The disadvantage of using smaller transformers is the increased regulation, but once you get to 600VA and above the improvement in regulation is quite small.

[MikeHunt79]

Originally posted by Iain McNeill
1. yes, as long as you use separate bridge rectifiers for each transformer to prevent circulating currents.

I'll remember that, I'll order loads of bridge rectifiers.

Originally posted by Iain McNeill
2. you'd get 63.6V peak from a 45Vrms winding with full wave rectifier. You'd need 2 such windings to get ±63VDC or a center tapped 90V winding. Or you could half wave rectify the single 45V winding into ±63VDC with more ripple. This assumes zero transformer resistance, zero diode drop and zero power drawn.

Here is info on the amp: http://www.41hz.com/main.aspx?pageID=87

By look looks of it, 60V is the maximum, so I may add a few extra (high current) diodes to be on the safe side, and I'll also test with a multimeter to check it doesn't go above 60V.

Originally posted by Iain McNeill
3. You need to add a good 25% to the peak voltage in case of mains supply variations. Also consider surge and transients. A zener is fast but can't handle too much power. MOV's are good but react slowly. I'd over-rate my caps more than that.

Ok, cool... I'll pick up some 80v or 100v caps then.

Originally posted by Iain McNeill
5. Don't electrocute yourself and don't work with metal tools on charged caps

Been there, done that.

Originally posted by NanoFrog
For calculating peak, just use 1.414.

Ooo, nice, makes things a lot easier.

Originally posted by AndrewT
Hi Mike,
what power are you expecting from your finished amplifier?

You can use a transformer VA rating that is between 1*max power to 2*max power.

+-63V supply rails could supply upto 300W to 400W to a 4r0 load.
This max power could be run from a 300VA to 800VA transformer.

I got 1000VA transformers from Farnell @ £60 each and they are available from Rapid as well as RS.

The disadvantage of using smaller transformers is the increased regulation, but once you get to 600VA and above the improvement in regulation is quite small.

I'm expecting around 1kw, or maybe 1.2kw... it will be powering 4 x 8 ohm subs, so each sub will get around 250 or 300W.

Amp info:
http://www.41hz.com/main.aspx?pageID=87

Bear in mind it is 2 channel, and I guess 60V @ 15amps (per chan) = 900W peak... but I need twice that (2 channels) hence where I got 1800VA from.

By the looks of it, I could get away with 1500VA or 1000VA, then that would be good... I guess it would be worth adding a thermometer to the toroids if I went for the lower powered toroids?

Also, thank for the info on the 1000VA toroids, it will probably be easier to go for them.... but, when hunting around, I managed to find a very good UK source for toroids: http://www.airlinktransformers.com/t...ansformers.asp

They look ideal, and good value, I'm tempted to just get the 1500VA one and be done with it... Anyone used this company, and are toroids good quality?

[AndrewT]

Hi,
there is no way to get 1000W from +-60Vdc.
If your amp can only use upto +-60Vdc then you must ensure that this voltage is not exceeded when the mains is running at maximum tolerance. which in the UK is 254Vac.
For a nominal +-58Vdc use a 40Vac transformer but this can go as high as 63Vdc and could make your amp fail.

I suggest you lower your expectations to 38Vac if you can find someone to wind custom secondaries for you.

[MikeHunt79]

Originally posted by AndrewT
Hi,
there is no way to get 1000W from +-60Vdc.
If your amp can only use upto +-60Vdc then you must ensure that this voltage is not exceeded when the mains is running at maximum tolerance. which in the UK is 254Vac.
For a nominal +-58Vdc use a 40Vac transformer but this can go as high as 63Vdc and could make your amp fail.

I suggest you lower your expectations to 38Vac if you can find someone to wind custom secondaries for you.

Sorry, I've just had a look at the datasheet for the TK2350 chip that the amp is based on:

www.tripath.com/downloads/TK2350.pdf

It states the the absolute maximum that the chip can take is 70v. I should have checked sooner...

The 1500VA toroid that Airlink sell does 45v AC at 230v AC

so at 254v, it will give ~50v AC (45/230*254 rounded up)

50V ac = 70.7v DC peak (50*1.414)- damn that's close!

Still over a volt should be lost thru the bridge rectifier (shouldn't it)?

I can always add a few extra diodes to be on the safe side... and hope that the mains never goes above 255V. Is there some sort of circuit that will warn you if the voltage goes over 250v?

also, I don't see why 1kw is not possible with 60v rails....

assuming 4 ohm load on each channel...

v=ir
60v = I * 4
I = 15A

2 channels, so total current = 30A

Power = VA
Power = 60*30 = 1800W

This is peak power however

1800 / 1.414 = 1273W RMS

Have I calculated something wrong here?

[AndrewT]

Originally posted by MikeHunt79 It states the the absolute maximum that the chip can take is 70v.

absolute max supply is often when the amp is not outputting any current or the output is open circuit. When operating there may be a lesser working maximum voltage.

Originally posted by MikeHunt79

The 1500VA toroid that Airlink sell does 45v AC at 230v AC

so at 254v, it will give ~50v AC (45/230*254 rounded up)

50V ac = 70.7v DC peak (50*1.414)- damn that's close!

You have omitted to take account of transformer regulation. 1500VA could be between 3% and 4.5% depending on design/construction.

Originally posted by MikeHunt79

Still over a volt should be lost thru the bridge rectifier (shouldn't it)?

If you use a single bridge rectifier the Vdrop of one diode is subtracted from the raw DC peak. your 45+45Vac transformer will give about +-73Vdc on no load.

Originally posted by MikeHunt79
also, I don't see why 1kw is not possible with 60v rails....

assuming 4 ohm load on each channel...

v=ir
60v = I * 4
I = 15A

2 channels, so total current = 30A

Power = VA
Power = 60*30 = 1800W

This is peak power however

1800 / 1.414 = 1273W RMS

No,
sinewave voltage has that ratio of 1.414 (sqrt2) that you've used already for peak to effective power voltage. 1000W into 4r0 requires sqrt(4000)V =63.25Vac and this is sqrt(8000)Vpk=89.4Vpk. You cannot get 89Vpk out of +-60Vdc supply. Worst case impedance for normal speakers can be as low as 35% of the nominal impedance.
Your 1000W amplifier feeding 4ohm reactive loads could send 89.4Vpk into 0.35*4ohms. output current can approach 89.4/4/0.35=64Apk. Your 1kW amplifier should be capable of meeting this transient loading condition 89.4Vpk and 64Apk.

A 400W into 4r0 amp requires to meet 56.6Vpk and 40Apk to drive a severe speaker load to specification.
As I said earlier your target power for +-60Vdc supply rails should be about 300W to 400W into 4r0.