Shouldn't need to! It is not even needed to have the P and N mosfets matched between them.
You need all matched P's and N's in each amp matched, but not across mono blocks. The bias procedure should make the output current equal.
But having matched devices across the mono blocks could in theory contribute to having the mosfets in the same spot on the transfer curve and yield better sonic results.
I disagree. You need the P-type closely matched and the N-type also closely matched in each CHANNEL. In other words, in the case of the F4, matched triplets all around.
Grimberg is correct - you don't need the N matched to the P. (Of course, it certainly doesn't hurt...)
The F4 has an adjustment for DC offset for exactly this reason - a strict N-P match is hard to get.
The output devices that share current are when needs to be matched, so one device doesn't current hog and burn itself up. So in each channel, the N need to matched to each other, and the P need to be matched to each other. The actual Vgs of the N in one channel doesn't need to be the same as the P.
The F4 has an adjustment for DC offset for exactly this reason - a strict N-P match is hard to get.
The output devices that share current are when needs to be matched, so one device doesn't current hog and burn itself up. So in each channel, the N need to matched to each other, and the P need to be matched to each other. The actual Vgs of the N in one channel doesn't need to be the same as the P.
What is the purpose of the 33pf ceramic capacitor across the AC line?
I seem to have lost the one I ordered from mouser, and my local shop only has a 2.2nF one, which is quite a bigger value than 0.0033nf on the schematic.
Can I use the bigger one?
I seem to have lost the one I ordered from mouser, and my local shop only has a 2.2nF one, which is quite a bigger value than 0.0033nf on the schematic.
Can I use the bigger one?
The line cap is 0.0033uF, which is 3.3nF, and 3300pF
2.2nF is actually smaller. It will still be fine 🙂
The cap is there to stop the switch from sparking on power up as well as a bit of filtering of the AC mains. MAKE SURE that the cap is X1 or X2 rated for across-the-mains use, as the cap must fail open if there is a problem.
2.2nF is actually smaller. It will still be fine 🙂
The cap is there to stop the switch from sparking on power up as well as a bit of filtering of the AC mains. MAKE SURE that the cap is X1 or X2 rated for across-the-mains use, as the cap must fail open if there is a problem.
They are X2/Y2. I have 2.2 and 4.7 available, should I go with the larger one since I have large 600VA transformer?
X2 rated capacitor between Live and Neutral. This can be relatively large without deleterious effect on other equipment.
Y2 rated capacitor between Live and Earth and/or between Neutral and Earth.
The current passing through the Y2 capacitor is equivalent to Earth Leakage and is usually minimised to have little effect on other equipment.
Y2 rated capacitor between Live and Earth and/or between Neutral and Earth.
The current passing through the Y2 capacitor is equivalent to Earth Leakage and is usually minimised to have little effect on other equipment.
I want to get less power from my F4 built, and lower the temperature of the heatsink by 10F.
To get 16Wrms from your F4 I am thinking of two ways
1. keep 3 pairs of Mosfets per channel and reduce the bias to 330mA per Mosfet
2. disconnect one Mosfet pair and increase the bias to 500mA.
Question: Which would be an overall better solution for keeping the distorsions low and soundstage wide?
Cheers,
Radu
To get 16Wrms from your F4 I am thinking of two ways
1. keep 3 pairs of Mosfets per channel and reduce the bias to 330mA per Mosfet
2. disconnect one Mosfet pair and increase the bias to 500mA.
Question: Which would be an overall better solution for keeping the distorsions low and soundstage wide?
Cheers,
Radu
I'd keep all three pairs and just decrease the bias. Removing a pair would increase the output impedance.
It will sound fine at decreased bias and still give full power. 🙂
The other cooling trick is to make sure the heatsinks have 2" below them and 6"+ above. This will increase the chimney effect and keep the sinks cooler. The total heat does not change, but it increases the exchange away from the amp.
The other cooling trick is to make sure the heatsinks have 2" below them and 6"+ above. This will increase the chimney effect and keep the sinks cooler. The total heat does not change, but it increases the exchange away from the amp.
Thanks 6L6.
I remember I read in one of Nelson's responses that he calculated the output power by multipleying 8 ohms by the square of 2.5A and divide by 2. ( I thought 2.5 A he used is the quiescent total current). Same way if bias goes down to 2A then total power is only 16W. Is anything wrong with this calculation?
Thanks,
Radu
I remember I read in one of Nelson's responses that he calculated the output power by multipleying 8 ohms by the square of 2.5A and divide by 2. ( I thought 2.5 A he used is the quiescent total current). Same way if bias goes down to 2A then total power is only 16W. Is anything wrong with this calculation?
Thanks,
Radu
You need 2.5A to get 25WRMS in Class A (not accounting for any inefficiencies). Since the output stage is push-pull, the actual idle bias need only be half this figure. With an idle bias of 1A, you'll get 16WRMS as you claim in Class A. The amp will still run to higher wattage in Class AB (though possibly with higher distortion).
Your calculations are correct, the class-A power will be 16W. But it's not going to clip until it hits the voltage rail, and that's going to be the full 25W.
(There is no Klunk)
(There is no Klunk)
Here's an update on my build:
Power supply wiring is done, I'll fire it up tomorrow evening if everything goes well. Mosfets and amplifier PCB's are installed and if PSU tests well I should be ready to finish cabling by the weekend and fire it up! Can't wait!!
An externally hosted image should be here but it was not working when we last tested it.
Power supply wiring is done, I'll fire it up tomorrow evening if everything goes well. Mosfets and amplifier PCB's are installed and if PSU tests well I should be ready to finish cabling by the weekend and fire it up! Can't wait!!
Is that a transforemer screen wire going across the mounting panel?
The screen connection to chassis should be low impedance to work properly.
i.e. it must be VERY SHORT.
A 25mm long lead bolted directly to the front panel would be 10times less inductance than your long lead.
The screen connection to chassis should be low impedance to work properly.
i.e. it must be VERY SHORT.
A 25mm long lead bolted directly to the front panel would be 10times less inductance than your long lead.
That is indeed a transformer screen wire. I thought it would be best to have just one central ground connection to the chassis.
Won't I run into ground loop issues if I have the screen connected on another place than the central ground connection?