F5 Turbo Builders Thread

[BigE]

oops, counting bias for fets that are off. That's 150 VA less

Around 1450 va

[BigE]

Check 7.2A x 4 fets x 50 Volts = 1440 VA total. This is a fine estimate

[a.wayne]

Originally Posted by BigE

If each Fet can handle 12 amps full load, then derating to 60% for longevity gives you a max of 7.2 amps/FET. That's and additional 1.2 amps x 4 fets *the other 4 are off*. Which is 4.8 amps or an extra 240 VA.

70% is 8.4 amps/FET => 2.4 * 4 = 9.6 amps or an extra 480 VA. That is a 1080 VA transformer.

Derating to 80 % gives your 9.6 amps/FET, or an extra 4 x 3.6 = 14.4 amps or 720 VA for a total of 1320 VA.

Looking at this backwards, if you have a 1500 va transformer, you've got 900 VA above 150 watt class A. 900 / 4 = 225 VA /FET = 4.5 amps. Which means that each FET can push 10.5 amps steady state, which only derates to 10.5/12 = 87.5 percent.

I'm not so sure this is a good idea. It is looking like a 1KVA transformer is about the max you want to supply to this sort of monoblock.

I would use 4 amp/ fet for SOA ....

[AndrewT]

Originally Posted by BigE

Sanity check time for me....

V3, 4 pairs per channel, biased to 3,0 amps per polarity, 50 volt rails. That gives 0,75 watts/FET, or 0,35 Volts. Total bias current is 6 amps.

Now total dissipation is 8 * 50 * 0.75 = 300 watts. That can easily be spread over 4 regular sized F5 heatsinks at 75 watts per heat sink -- each heatsink carries one pair of FETS.

Efficiency is around 33% at full class A power. Does that sound about right?


As for Transformer VA, clearly 300 VA is needed just to run the bias. Now, add the load requirements. P=150 rms = 300 peak.

So, P=IV , V = 50 says I = 6 amps peak. That is another 300 VA to power class A operation. (That checks with doubling the bias current to come out of class A).

So far, the transformer has to be 600 VA *bare minimum*.

What do you spec for Class A/B overhead? Do you limit the VA to provide 50% of the steady state operating current of the output transistors?

Also, how do the diodes impact the heatsink size?

BigE,
I don't know where you found your references, but most of what you have posted is plain wrong.
The total bias current is 3A. The rail to rail voltage is 100Vdc. The dissipation is 300W. You got the correct answer using rubbish data.
A 300VA 37+37Vac will get your +-50Vdc supply rails when supplying about 2Adc. Let's check: 300VA & 35+35Vac results in a rated current of 4.05Aac.
This will give a maximum continuous dc current from the capacitor input filter of ~2Adc. You cannot and must not draw 3Adc from a 300VA transformer.

For that continuous current draw of 3Adc, you must use a transformer rated @ 6Aac i.e. 444VA.

Now to your next fallacy.
A +-50Vdc 3A bias ClassA amplifier when delivering zero Watts to the load consumes 300W as shown earlier. When it delivers ClassA current to the load, the total load seen by the pSU does not change. If 20W goes to the load the amp dissipates the remaining 280W. If 50W goes to the load, the amp dissipates the remaining 250W.
if 100W goes to the load, then 200W is dissipated by the amplifier.

Why is this so?
Because the rail currents are not constant. as one rail current increases the other rail current decreases by exactly the same amount. The difference in rail currents goes to the load. The total load seen by the PSU does not change.

Now that I have settled that, lets go back to transformer VA.
A 444VA transformer delivering 3Adc continuously will run hot. As hot as the manufacturer allows for their product.
For cooler running and good reliability it is generally recommended that the continuous DC current drawn be approximately 50% of the manufacturer's specified maximum continuous DC current when feeding a capacitor input filter.
This would require a transformer rated @ 800VA to 900VA.
This has nothing to do with delivering ClassA power to the load. This is simply to allow reliable and cooler operation than the manufacturer's maximum rating while the amplifier is idling.

As stated earlier when the amplifier is delivering ClassA power the total load seen by the transformer stays the same. You don't need any extra VA for ClassA power delivery.

If you want more power than ClassA then the amplifier transitions to ClassAB and the power drawn by the amplifier becomes the average for A ClassAB with high bias when delivering high output currents.
The usual factor to apply is roughly 1 to 2 times the maximum output power as a VA value.
Lets suppose your ClassA is 140W and you decide that the maximum ClassAB should be double this, i.e. 280W. This would require a 280VA to 560VA transformer. But we need to make an adjustment for the high bias. 444VA for high bias + 1.5*280W for ClassAB on a continuous basis is an overestimate for required VA. This indicates that a 864VA would be the upper limit for needed VA for the ClassAB amplifier. Same as for a cooler running transformer.

Now look at the normal range of VA suggested by Pass and others. Use about 6times to 10times the maximum ClassA output power. For a 140W amplifier 6 times comes out at 840VA, and 10times indicates 1400VA.

Your

the transformer has to be 600 VA *bare minimum*.

does not give much assistance to any builder.
Even the 6 to 10 times is a better guide.

The logic you followed in later posts is also flawed. By luck you appear to be getting closer to a real transformer rating
Peak current to the load comes from the capacitors.
It has very little to do with average currents supplied to keep the capacitors charged.

[BigE]

Thanks for sorting that out for me Andrew. I can see the flaws in my understanding and am a little less ignorant now. I dare not say more knowledgeable....

[Blink_PT]

I'll build the PSU to F5T 2 in a board made ​​by me, but there is a doubt on the links of MUR3020W!

Thanks to confirm that the connections are correct. thank you

[carlomar]

Originally Posted by Blink_PT

I'll build the PSU to F5T 2 in a board made ​​by me, but there is a doubt on the links of MUR3020W!

Thanks to confirm that the connections are correct. thank you

Hi Blink,

This is the correct connetions

[Blink_PT]

[QUOTE=carlomar;3282447]Hi Blink,

This is the correct connetions

.......
Hello Carlomar,

It will be more explicit way. The problem will be passing it to the PCB. I'll wear clothes iron, and paper photocopy laser... (half dozen copies)

[omishra]

Originally Posted by carlomar

Hi Blink,

This is the correct connetions

.......
Hello Carlomar,

It will be more explicit way. The problem will be passing it to the PCB. I'll wear clothes iron, and paper photocopy laser... (half dozen copies)

It is easy with 2 layer PCB. See how I made it for 4 blocks, you are referring one of them.

PS_Rectifier_Ver1.png

PS_Rectifier_Top_Ver1.png

PS_Rectifier_Bottom_Ver1.png

[Blink_PT]

Originally Posted by omishra

It is easy with 2 layer PCB. See how I made it for 4 blocks, you are referring one of them.

Attachment 317360

Attachment 317361

Attachment 317362

They are all the same, change the color? The rectifiers in this scheme are separated from PSU?

What is the value of the capacitors connected will bridge?

Saw its construction in Hifivision and really liked...