CRCRC :
First C 2 x 18,000 uF
First R 4 x 1R0
Second C 1 X 18,000 uF
Second R 4 x 1R0
Third C 7 x 2200 uF.
The last C should be high quality low ESR caps.
Simulation looks good with ripple below 100 mV and thats with RMS output current of around 5 A. Not bad at all.
Considering I will never really need 5 A RMS continously it should work very well with more "normal" loads where you only really need alot of current for transients.
I think Im done now. 😀
First C 2 x 18,000 uF
First R 4 x 1R0
Second C 1 X 18,000 uF
Second R 4 x 1R0
Third C 7 x 2200 uF.
The last C should be high quality low ESR caps.
Simulation looks good with ripple below 100 mV and thats with RMS output current of around 5 A. Not bad at all.
Considering I will never really need 5 A RMS continously it should work very well with more "normal" loads where you only really need alot of current for transients.
I think Im done now. 😀
Then you can start with a 20 to 22 volts transfo.
Though, I am not shure the second RC cell is necessary.
Generally, three cells are used when the voltage of the transfo is too high
Why not to follow the Nelson's scheme (increasing C values if you like). He really knows what to do.
Or, if you want a very low ripple, you can also build a cap multiplier.
Though, I am not shure the second RC cell is necessary.
Generally, three cells are used when the voltage of the transfo is too high
Why not to follow the Nelson's scheme (increasing C values if you like). He really knows what to do.
Or, if you want a very low ripple, you can also build a cap multiplier.
Last edited:
30 V transformer, going for cascode F5.
Supply would give me around 37 Volts out with RMS load current of 4,6 A and only 66mV ripple. Load current of 8,3 A gets me 33,3 volts and 115 mV ripple. But since these numbers are continous output current, in a real world scenario with a "normal" speaker load with varying current demands the performance should be much, much better.
Could follow nelsons schematic 100% but I like to do stuff my own way. Anyone can copy a schematic and make it work. Which is why I also made layout for that 4 x output pair cascoded F5, although that has dropped to 3 x output pairs in a revised design.
Last edited:
68W/8 ohm yes. 🙂 And something like 120W / 4 ohm.
Now that I have reduced the number of output pairs to 3 and going with 500 mA pr fet we are talking around 110 W. Heatsinks will have to be big, no doubt about it.
My speakers are 8 ohm but they drop to 4 ohm in the bass region. 3 x output pairs with a supply of 37 volts should give me around 18 A current limit from my amplifier. I need around 4.2 A in 8 ohm to reach 68W and double that for 136 W in 4 Ohm, so 8.4 A in 4 ohm. Well within the current limit and with power to spare. With weird phase angles and similar, having current to spare is a good idea.
Last edited:
I am well on my way to completing my F5 using cvillers mk2 amp boards and p/s boards and wondered if it was advisable/necessary to use lead offcuts to solder in the three plated through holes near the middle of the p/s pcb? They are where the 0.47R resistors fit. TIA
So, with 1.5A bias that's 36w classA /8ohms.
As i use an active crossover, the amp direct coupled to my 8 ohms speakers, i have used 33v rails, 1.7A bias to get 50w classA.( from 200 to 2500hz)
But i already said that.
I'm planning on doing something like that myself or running a jumper wire. Maybe would have been better to have solder pads instead of the three vias.
Dear All,
I will probably put the last connector today, and the project will be complete. (That's strange it never happens). Anyway, I noticed that the bias is not super stable, mainly related to temperature. I have a Shredinger cat effect. The bias change with open and close chassis, but I need to open it to change it. How do you do it ?
I am speaking about the voltage around the 0.47 ohm resistors varying 15-20 mV. Could it be that the thermistors are too close to the MOSFET ?
Thanks,
Davide
I will probably put the last connector today, and the project will be complete. (That's strange it never happens). Anyway, I noticed that the bias is not super stable, mainly related to temperature. I have a Shredinger cat effect. The bias change with open and close chassis, but I need to open it to change it. How do you do it ?
I am speaking about the voltage around the 0.47 ohm resistors varying 15-20 mV. Could it be that the thermistors are too close to the MOSFET ?
Thanks,
Davide
Probably not. I glue mine right to the case of the Mosfet,
which minimizes the open case effect.
😎
Neutraity
couple of questions as I am working on similar set up.
Is the las cap 2200uf 2200 microfarads?is this corect.
at 37 volts 1.4 A device is about 50W disipation each correct?
what are you doing for heat sink and thermal interface?
Tanks
couple of questions as I am working on similar set up.
Is the las cap 2200uf 2200 microfarads?is this corect.
at 37 volts 1.4 A device is about 50W disipation each correct?
what are you doing for heat sink and thermal interface?
Tanks
I am even more puzzled, I thought that the increase in temperature had the effect of lowering the current. After few hours of use and test I had 630 mV, while yesterday I set it around 590 mV. The DC offset is stable.
I am trying to understand if I am figuring an issue or not. The heatsink is not too hot, I can put my hand on it indefinitely, but inside the chassis I have 50 C.
Thanks,
D.
Last caps are 7 x 2200 uF caps, thats 2200 microfarads.
37 volts at 500mA pr FET is 18,7 W dissipation pr FET.
That is my current plan, bias pr FET might be a little more or a little less, depending on cooling performance of the case Im planning on using.
Will be a case from modu
Last edited:
I would probably increase the final capacitor bank - when F5 really pulls, it would be a shame if it had to pull a peak current all the way from first bank.
Last edited: