F5 power amplifier

[sekess]

Hey Andrew,
Couldn't you slow charge the caps from the primary or the secondary side of the transformer? What is the advantage of doing it from the secondary side?

Thankls,

[sekess]

Hey Andrew,
Plus, it's a bit of a pain slow starting from the secondary side. With two power supplies (for dual-mono), you need 4 resistors (one on each of the secondaries), which all have to be bypassed with either a 4-way relay or multiple relays.
If it's doable from the primary side, you can do the whole thing with one resistor (or combination of resistors) and one relay as long as the ratings are high enough.

Now for that matter, if you plan on slow starting from the secondary side, wouldn't it be even better to slow start from just after the rectifiers. this way the rectifiers never see a shhort at start-up?

Thanks,
Steve

[jackinnj]

Originally Posted by bear

jackinnj, seems like a whole lot of folks have smoked these resistors due to operator error... so IF the mosfet survived, either they are good fuses or else they are maybe on the edge...

If I wasn't lazy and unable to calculate accurately in my head (what is in my head?) I'd figure the peak current through them and so the peak watts... if it was within 50% of the resistor's rating, I'd double it up myself... Ymmv.

_-_-

I've had folks with issues regarding the 10 ohm source resistors -- but not the 0.47R Panasonics, Vishay or Dales. At peak the JFET source resistors dissipate 0.57W, definitely not a place for a 1/4 or 1/2 W resistor.

The machine inserted Panasonics are specified to be mounted 16mm (about 5/8inch) above the PCB.

The Ohmite TWW5's are very efficient at getting the heat off the board. They aren't quite as flat as the Panasonics or Dales, but are rated at 5W. There was a brief period of time when Ohmite switched over to ROHS and they were out of stock at all the usual suspects. Janneman appears to have used them in one of his designs for Elektor.

I have gotten my 576 curve tracer to smoke on occasion. sometimes you can make the current limiting resistors unhappy.

[bear]

I was going to bring up the 10 ohm resistors... hadn't gotten there yet.

So, maybe 2 x 20ohm 1/2w then?


sekess, if you put a resistor in series with the primary (the usual way to go) for a soft start the caps will never fully charge because of the vdrop of the resistor...fwiw.

How long before you throw the relay shorting out the series resistance depends on the resistance and the amount of capacitance you have to charge up in the real world. A little bit of experimentation will be sufficient. Watch an ammeter in line or else look to see if the house lights blink?

Of the many ways of doing the switching are the RC + transistor, 555 timer, manual switch, and comparator that looks at the rail voltage...

_-_-bear

[sekess]

Hey Bear,
Yea, I know while the resistor is in place the rail voltage won't reach 24 volts. But, the caps will stop charging at some lower voltage. then, when the resistor is switched out, the rails will reach about 24 volts and the caps will finish their charging.

Basically, I am trying to figure out the following (for soft-starting and slow-starting):

-- What size resistor(s) to use - both voltage and wattage
-- Where to place them - primary, secondary, or at the rails
-- How much time before I bypass them with the relay

If practical, I would like to soft start as well as slow charge the caps (as Andrew T has alluded to).

I got a decent deal on some very large caps. So, I wanted to try those guys. I am going to end up with 280 000uF per power supply (dual-mono -- so 560 000uF total). With that amount of capacitance, I thought it would be best if I could slow-charge them at start-up.

Thanks,
Steve

[jackinnj]

Originally Posted by bear

I was going to bring up the 10 ohm resistors... hadn't gotten there yet.
So, maybe 2 x 20ohm 1/2w then?

Smaller Panasonics -- 2 W -- but 2x 0.5W should work as well.

[bear]

Right, absolutely necessary.
I'd suggest 50watt wirewound ceramic resistors at minimum.
You can get away with less, but in the event something goes wrong or delays your relay you will smoke smaller resistors.
A value around 8-10 ohms is likely to be good.
With the lower voltages maybe a higher value will suffice...
I would only put this before the transformer at the primary.
Also, do NOT use a single relay contact. Use multiple high current contacts, imo. If you just have to use a medium size "cube" relay, at least get one with 4 contacts...

My Symphony No.1 amp has 500,000ufd @65vdc rail voltage. That and >2KVa of iron... so I use very large P&B relays. You can see a pic on my website if you hunt down in the amplifier section a bit...

As I said you can figure out the time thing after you put the supply together and have it running... just look at the pulse when you bypass it... that and the time it takes for the rails to come up. On my amp I used 10 seconds... more than sufficient to bring up the rails and let the amp stabilize (no speaker relays, so don't want thumps)

I'm thinking that a soft recovery type rectifier will be best in consideration of the current waveform that will be charging the cap banks...

How many caps to achieve this total capacitance??

_-_-bear

[Snokker]

Originally Posted by sekess

I got a decent deal on some very large caps. So, I wanted to try those guys. I am going to end up with 280 000uF per power supply (dual-mono -- so 560 000uF total). With that amount of capacitance, I thought it would be best if I could slow-charge them at start-up.

I'm very curious how your transformers/mains will behave with this amount of capacitance. I often experienced lots of humming and buzzing in the toroid due to the very, very bad power factor (cos phi). One of my F5's has 130.000uf per channel on a Amplimo 2x18V, 625VA and I can't go further without having the buzz.....

[sekess]

Hey Snokker,
I did not realize that too much capacitance could result in buzzing. I just thought it creates start-up issues.

Thanks,
Steve

[a.wayne]

Originally Posted by bear

Right, absolutely necessary.
I'd suggest 50watt wirewound ceramic resistors at minimum.
You can get away with less, but in the event something goes wrong or delays your relay you will smoke smaller resistors.
A value around 8-10 ohms is likely to be good.
With the lower voltages maybe a higher value will suffice...
I would only put this before the transformer at the primary.
Also, do NOT use a single relay contact. Use multiple high current contacts, imo. If you just have to use a medium size "cube" relay, at least get one with 4 contacts...

My Symphony No.1 amp has 500,000ufd @65vdc rail voltage. That and >2KVa of iron... so I use very large P&B relays. You can see a pic on my website if you hunt down in the amplifier section a bit...

As I said you can figure out the time thing after you put the supply together and have it running... just look at the pulse when you bypass it... that and the time it takes for the rails to come up. On my amp I used 10 seconds... more than sufficient to bring up the rails and let the amp stabilize (no speaker relays, so don't want thumps)

I'm thinking that a soft recovery type rectifier will be best in consideration of the current waveform that will be charging the cap banks...

How many caps to achieve this total capacitance??

_-_-bear

Bear I'm prepared ... tell me more ! how low can we go .......