power supply for a big bass amp-> sound ok?

[Steve Conner]

Well, let's do some math:

300W output at 60% efficiency = 300/0.6 = 500W of HT

6 tube heaters at 1.6A each = 1.6*6.3*6 = 60W

for a total of 560W.

However, it's highly unlikely that the amp will be called to produce a continuous tone at 300W for the whole duration of the gig. So we can use the old rule of thumb that allows us to overload transformers to twice their rating in intermittent service. (But not for the heater draw: they're on all the time.)

So we have (500/2)+60 = 310W.

I ignored the "utilization factor" of the rectifier which is something like 1.2 or 1.4. This means that it takes 1.2 or 1.4VA of transformer capacity to produce 1W of DC power. If it were 1.4:

(500*1.4/2)+60 = 410VA.

So, if we used the 400VA tranny, our amp might post 300W RMS on the bench, but if we left it running like that for an hour, the transformer would burn. Most amps are designed this way, and that's why transformers have thermal fuses inside nowadays.

[black_labb]

Well, let's do some math:

300W output at 60% efficiency = 300/0.6 = 500W of HT

6 tube heaters at 1.6A each = 1.6*6.3*6 = 60W

for a total of 560W.

However, it's highly unlikely that the amp will be called to produce a continuous tone at 300W for the whole duration of the gig. So we can use the old rule of thumb that allows us to overload transformers to twice their rating in intermittent service. (But not for the heater draw: they're on all the time.)

So we have (500/2)+60 = 310W.

I ignored the "utilization factor" of the rectifier which is something like 1.2 or 1.4. This means that it takes 1.2 or 1.4VA of transformer capacity to produce 1W of DC power. If it were 1.4:

(500*1.4/2)+60 = 410VA.

hmm, i didnt factor in the heaters into the VA rating. i was just looking at the current itself. i'll ask them about the inclusion of the heaters and how conservatively rated it is.

and thanks for bringing up the issue loudthud, i'll do some research. may end up going with the 450v tranny and getting a bit less power if neccesary.

[black_labb]

oh, and thanks for that breakdown steve. didnt know anything about the utilisation factor of rectification. i always thought that the only power they take up is the 0.7v they drop. is this related to the tranny putting out alot of current half of the cycle then 0 current at the other half?

[Steve Conner]

It's because the current drawn through the rectifier isn't a sine wave, so it has a higher RMS-to-average ratio. This heats the windings more than a sinusoidal current would. It's not like the diodes burn up the extra power, it's just that you can get less power than you'd expect before the transformer overheats.

I can't remember where I read it, but I think it's 1.4 for a FWCT and 1.1 for a bridge. Maybe someone else can comment.

[loudthud]

A view from PSUD

There is a thread on TalkBass where a guy scratch built a 300W amp. You might have to register to view. I like that forum because most of the members are not technical, they just play bass. One of the heavy hitters there is Bill Fitzmaurice, a guy that builds horn speaker enclosures and publishes in Audio Express magazine.

http://www.talkbass.com/forum/showthread.php?t=465327

The amp evolves in the thread. Started with 4 KT90s and ended up with 8 KT88s. Uses the driver transformer from a Fender 400PS and a couple of chokes. The thing must weigh over 100Lbs.

Rectifier losses: The 0.7V per diode is not realistic in most circuits. When a diode is near it's max current, 1.0V or 1.1V is closer to reality. Not very important in a high voltage supply but when you try to build a DC heater supply it's significant. The thing that gets you in a high voltage supply is the transformer starts to distort (shown below). Remember that in a full wave bridge you effectively have two diodes in series.

Here's some numbers from an amp I have yet to build. The OT is a hammond 1650R. The PT is a 300VA torroid that I got custom wound from a guy on ebay. My intent was to have a 600V B+ and a 300V screen supply running a single pair of 6550s. IIRC there is a 6.3V@6A heater winding and 60V@100ma bias winding. The secondary specified at 420VCT (under load). To get accurate numbers out of PSUD you need to plug in the primary and secondary resistances and the open circuit secondary voltage. I did this and PSUD came up with a lumped losses of 41 ohms with 120V 60Hz primary. Seems high to me but I did it several times and got the same numbers.

Simulations were run at 90mA and 333mA. The filtering is a pair of 220uF caps in series so the screen supply comes from the center tap which PSUD does not simulate. I use the delay function to remove the startup surge so I can get accurate peak and RMS current numbers for secondary current. If you care, you can also get numbers for the filter's ripple current. Simulations are for 1000mS after a delay of 1 second. (Edit: looking closely it looks like the sim ran at 50Hz but I don't want to do it over to correct if possible)

90mA load current
IT1 223mA RMS 689mA peak
Ripple voltage 7V p-p
Min Val 595V (bottom voltage of the ripple)

333mA load current
IT1 661mA RMS 1.64A peak
Ripple voltage 22V
Min Val 546V

In the pictures, the red trace is the rectifier current (you only see half wave), the yellow is the transformer's secondary voltage. Note how on the rising portion of the ripple where current is flowing to charge up the cap, the transformer's output really distorts. And that limits the peak voltage available to the cap.

[Steve Conner]

Hey, that's cool. I have the same transformer set as that guy, and I got mine from Bob Rissi too. I was going to build it up with a 500VA toroidal PT that I pulled from some piece of medical equipment, and six KT88s. Unfortunately I quit my band, and even if I hadn't, they weren't loud enough to need such firepower

I wonder where he got a chassis strong enough to support it all.

[black_labb]

i've seen that thread before loudthud, didnt get very far in it as there wasnt much technical detail after the initial description of the amp and i stopped looking. now that you brought it up i'm looking a bit further into it. thanks for that. i'm a member at TB but dont always check it regularly, i sometimes feel there is too much going on to keep up with with so many active members.

as for my design, i'm condidering going for the 450v tranny as opposed to the 500v one, and thinking of using 2, and putting in 8 sockets. would put out around 350-400w with 6550's at ~610v or so.
from what people have said about the hammond power rating it could probaby put out the 400w. if it doesnt put out the full power at the lowest frequencies thats not bad either, as most bass speaker cabinets (atleast commercial ones that are ported) would be well past their xmax with 400w of fundamental. im thinking with ~600v i could run 8 jj e34l's in it for ~300w. probably a fair bit cheaper than 6550's, especially considering i have 4 of them already.

might as well get 1 more transformer, make it more reliable, more powerful and i was going to get an extra 6v transformer anyway for the preamp filiaments

[black_labb]

just for people's info, i got this from a representative at antek reguarding hpw conservatively rated the trannies are.
"Hi,
Our transformer can run 120% continuous. the core and copper that we use are always bigger than required from calculation. You can run the full power on both the high voltage and the 6.3V. I am sure that the temperature will not be more than 50 degree C."

seems like therit pretty conservatively rated. still leaning towards getting 2, as it would drop less voltage at full power, which seems to be why many big amps dont reach the power that the tubes are capable of.

i dont really know why i need the power though, i would be fine with 200w. guess i want to get the most out of the big tranny, and i already have 2 200 watt valve monsters.