• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Swordfishy/ASPEN FETZILLA power amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Dinesh,

I really would discourage you from building a huge power supply. It will not sound so good.

I have done extensive testing with Class AB amps on this issue and found again and again that the very high charge pulses found in high cap power supplies actually militate against good sonic performance. The problem is that these very high charge pulses interfere with the voice coil earth return through the filter caps, causing intermodulation.

I find it's never necessary to go much more than 4,700uF on a 50W amplifier.

The FETzilla uses two 3,300uF caps separated by interelectrode resistors, each of 0.15 ohms. This quarantines the charge pulses principally to the first cap after the rectifier, with the second cap handling much reduced charge pulse and all of the speaker return current. This arrangement gives palpably better sound.

Cheers,

Hugh
 
Hugh

Here is schematics for separate PSU. This is only a general idea if you are planing to use separate power supply instead of the one included on amp PCBs. You can change all the major components to your personal taste. I used here 24VAC 300W transformer. 24VAC is more common than 25VAC and is much easier to get. You don't need 300W transformer for this project, 160W will probably be enough. I used 300W for economical reasons, it costs only $15 more and is bigger by 7mm than 160W one. Soft start circuit is optional in case you plan to use large filtering capacitors to prevent blowing fuses on start-up.

Fetzilla PSU.jpg

Regards
George
 
Lovely work, George, many thanks.

Agree about the 24Vac trafos, plenty around, some at disposals prices.

I do feel you might have too much capacitance there, which incidentally can be mounted onboard at the pcb. I have found that onboard caps improve the sound since the connection to the output devices is not so inductive as a long wire.

Thank you for your excellent contribution!

Hugh
 
put me down for 4+P :)

2+P JoMo Adelaide SA Australia
2+P Ian Finch Coffs Harbour NSW Australia
2+P Gennaji Hyderabad India
2 Ken Lewis LA USA
2 Ron Russell Kent UK
2+P John Kenny Dublin Ireland
2+P Francis Morrin Dublin Ireland
2+P Vitalica Bucharest Romania
2+P Erik Deinum Groningen Netherlands
2+P Keith Correa ? India
2 Igor Polyakov California (city?) USA
2 Steve (aka SRH) Whangarei New Zealand
2+P George (aka JKC) Calgary Canada
2+P Anthony Mills Gold Coast Australia
2+P Dinesh (aka Chlorofille) Christchurch New Zealand
2+P Colin S. (?) (city?) NSW Australia
2 Chris Powell Sydney Australia
2+P Danny (?) Lierde Belgium
2 Sonnya (?) City (?) Denmark
4+P Jozua Cape Town South Africa
2 JT (? aka Sandyhooker) City (?) Arizona USA
2+P Nicola City (?) Vincenza Italy
2+P Sparkey (?) Melbourne Australia
2+P Chalkandtalk (?) City (?) NW England UK
4+P Woody1911a1 Marblehead , Ma USA

cheers Woody
 
You're in, Woody. Now at 54 boards and building!

2+P JoMo Adelaide SA Australia
2+P Ian Finch Coffs Harbour NSW Australia
2+P Gennaji Hyderabad India
2 Ken Lewis LA USA
2 Ron Russell Kent UK
2+P John Kenny Dublin Ireland
2+P Francis Morrin Dublin Ireland
2+P Vitalica Bucharest Romania
2+P Erik Deinum Groningen Netherlands
2+P Keith Correa ? India
2 Igor Polyakov California (city?) USA
2 Steve (aka SRH) Whangarei New Zealand
2+P George (aka JKC) Calgary Canada
2+P Anthony Mills Gold Coast Australia
2+P Dinesh (aka Chlorofille) Christchurch New Zealand
2+P Colin S. (?) (city?) NSW Australia
2 Chris Powell Sydney Australia
2+P Danny (?) Lierde Belgium
2 Sonnya (?) City (?) Denmark
4+P Jozua Cape Town South Africa
2 JT (? aka Sandyhooker) City (?) Arizona USA
2+P Nicola City (?) Vincenza Italy
2+P Sparkey (Brian ?) Melbourne Australia
2+P Chalkandtalk (?) City (?) NW England UK
4+P Woody1911a1 (?) Marblehead , MA USA


Thanks for the support, this is going very well......

Cheers,

Hugh
 
Hugh

As I said before those values can be changed. I think that the best compromise will be to use two big capacitors just after the bridge in the range 6800uF to 10000uF and another pair of smaller capacitors 3300uF to 4700uF directly on the boards. In this case length of the wires you ran from power supply to the amp will be not that critical. If you need to get low ripple voltage on the rails you will need quite big caps. In this case at 50W into 8ohm impedance will produce 3.5A peak current. With four 8200uF you looking at ripple voltage around 0.76Vpp at peak current. In reality it will be little smaller than that because of those series resistors between caps (those resistors should be an inductive wire-wound type for the best filtration). However, if you are looking for the worst case scenario let say 100W into 4ohm impedance load then your ripple voltage on the rails will double to 1.5Vpp at peak current of 7A which is quite high. Because of non-linearity of the speakers you should be expecting that kind of situation on a transient basis.

Regards
George
 
..........at 50W into 8ohm impedance will produce 3.5A peak current.......................100W into 4ohm impedance load ........peak current of 7A which is quite high. Because of non-linearity of the speakers you should be expecting that kind of situation on a transient basis.
Your transient peak currents to the speakers when driven to nearly max Vout, could be much higher than the resistor load predicts. Expect at least two times higher and probably design for 3times higher.
A recent paper suggested that all three speakers tested at ~5times higher when worst music conditions met with worst reactive conditions.
 
Welcome to our little project, George! Now at 58 boards and building!

2+P JoMo Adelaide SA Australia
2+P Ian Finch Coffs Harbour NSW Australia
2+P Gennaji Hyderabad India
2 Ken Lewis LA USA
2 Ron Russell Kent UK
2+P John Kenny Dublin Ireland
2+P Francis Morrin Dublin Ireland
2+P Vitalica Bucharest Romania
2+P Erik Deinum Groningen Netherlands
2+P Keith Correa ? India
2 Igor Polyakov California (city?) USA
2 Steve (aka SRH) Whangarei New Zealand
2+P George (aka JKC) Calgary Canada
2+P Anthony Mills Gold Coast Australia
2+P Dinesh (aka Chlorofille) Christchurch New Zealand
2+P Colin S. (?) (city?) NSW Australia
2 Chris Powell Sydney Australia
2+P Danny (?) Lierde Belgium
2 Sonnya (?) City (?) Denmark
4+P Jozua Cape Town South Africa
2 JT (? aka Sandyhooker) City (?) Arizona USA
2+P Nicola City (?) Vincenza Italy
2+P Sparkey (Brian ?) Melbourne Australia
2+P Chalkandtalk (?) City (?) NW England UK
4+P Woody1911a1 (?) Marblehead , MA USA
4+P George Cherhit Sacramento CA USA

We have one week to Greg returning from Europe, and two days or so after that I will pull the lever on the board manufacture.

Cheers,

Hugh
 
PCB REFINEMENT

Folks,

Here is the advantage to taking your time with the pcb layout; insurance against mistakes, and refinement.

I have added the output inductor and 10R resistor to the output buss. This saves time as the inductor can be soldered directly to the pcb.

Cheers,

Hugh
 

Attachments

  • FetzillaPCB.gif
    FetzillaPCB.gif
    49.1 KB · Views: 651
Last edited:
Steve,

That would be optimum, but what is important is that the feedback is taken from the midpoint. A good compromise would be to thicken the track between the fb takeoff and the left side of L1.

Thanks for the tip.

Nice resistor, Sonnya, thank you. Planned resistance is 0.15R, in fact. The 25 5W series will fit fine; the default size is the stock sandcast ceramic which is not a wonderful component but is at least ubiquitous. I really think 10W at 47.2mm long is too large, however.

Hugh
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.