The 3875 GC kit guide recommends a 160VA or higher transformer. However, these are really big (toroidals) and I'm building a mono amp with just one 3875 (both channels somehow summed before getting to the 3875) and I want to drive this amp to no higher than 30-40W...maybe rail DC voltage of around 18V or so.
Is it OK to use a smaller toroid? What are the disadvantages of using a smaller toroid...let's say 30 or 60 VA?
I've been checking out the general purpose ones put out by Bicron.
http://www.bcrn.com/Pages/AUBUt-TPTrev.htm
Thanks a ton!
Is it OK to use a smaller toroid? What are the disadvantages of using a smaller toroid...let's say 30 or 60 VA?
I've been checking out the general purpose ones put out by Bicron.
http://www.bcrn.com/Pages/AUBUt-TPTrev.htm
Thanks a ton!
Calculate the minimum power of transfo with this formula :
Pamplifier * 1.5 = VA transfo
examle :
Amplifier 30WRMS = 30*1.5 = 45 VA (minimum) trasformer power
Amplifier stereo 40+40W = 80*1.5 = 120 VA
This formula is valid for normal class AB without other device.
If you use a smaller trasfo then at max volume amplifier plays
really distorted and trasfo can overheat or to burn
bye
Pamplifier * 1.5 = VA transfo
examle :
Amplifier 30WRMS = 30*1.5 = 45 VA (minimum) trasformer power
Amplifier stereo 40+40W = 80*1.5 = 120 VA
This formula is valid for normal class AB without other device.
If you use a smaller trasfo then at max volume amplifier plays
really distorted and trasfo can overheat or to burn
bye
Hi,
the big disadvantage of using a small transformer is the high regulation that comes with it.
A 12-0-12Vac transformer is rated at rated current. So 60VA will give 2 times 2.5Arms @ 12Vac.
However on open circuit the voltage will rise by the regulation percentage about (10 to 15% or worse) so the peak voltage at the smoothing caps becomes 12*1.15*1.414=19.5Vpk.
and when the mains voltage rises by another 6% that peak becomes 20.7Vpk.
Take off the diode drop for the rectifier and you have a maximum of 20V on the smoothing caps and that is what your amplifier will feel when idle (your Iq may pull this down slightly).
That high voltage will drop when the amp starts delivering power. This usually shows as poor bass and wandering stereo image, but you only plan one channel so that may not be an issue.
the big disadvantage of using a small transformer is the high regulation that comes with it.
A 12-0-12Vac transformer is rated at rated current. So 60VA will give 2 times 2.5Arms @ 12Vac.
However on open circuit the voltage will rise by the regulation percentage about (10 to 15% or worse) so the peak voltage at the smoothing caps becomes 12*1.15*1.414=19.5Vpk.
and when the mains voltage rises by another 6% that peak becomes 20.7Vpk.
Take off the diode drop for the rectifier and you have a maximum of 20V on the smoothing caps and that is what your amplifier will feel when idle (your Iq may pull this down slightly).
That high voltage will drop when the amp starts delivering power. This usually shows as poor bass and wandering stereo image, but you only plan one channel so that may not be an issue.
Thanks! I see the problem.
Two related questions...
1- The toroids I'm looking at have dual secondary windings. How does this work out with a single LM3875 kit/single rectifier board scenario? Does having dual secondaries help any of the low VA issues brought up?
2- Would going with higher capacity or non-polarized caps help the situation?
Two related questions...
1- The toroids I'm looking at have dual secondary windings. How does this work out with a single LM3875 kit/single rectifier board scenario? Does having dual secondaries help any of the low VA issues brought up?
2- Would going with higher capacity or non-polarized caps help the situation?
Here is some advice for noobs everywhere: when you have a question about an IC or a transformer, the FIRST place you should go is the data sheets and ap notes provided by the manufacturers. You come to places like this only after you have exhausted all other credible sources of information because you can never be sure of the quality of info you'll get here.
Transformer sizing is not a matter of opinion, it is a matter of numbers. The math is easy. The formulas are in the app notes. Can't find an ap note at your transfomer manufacturer's site? Go to another transformer manufacturer's site and use their formula. Power transformers are power transformers. The same formula will apply no matter who makes the thing.
Go to the source!
I_F
Transformer sizing is not a matter of opinion, it is a matter of numbers. The math is easy. The formulas are in the app notes. Can't find an ap note at your transfomer manufacturer's site? Go to another transformer manufacturer's site and use their formula. Power transformers are power transformers. The same formula will apply no matter who makes the thing.
Go to the source!
I_F
Have a look at this:
http://www.hammondmfg.com/pdf/5c007.pdf
See the capacitor input filter and load near the bottom on the right side? Notice that the DC output current is only 0.62 x the secondary AC current rating. That means that if you need 1 A DC out, your secondary current should be spec'd at 1 A/0.62 = 1.6 Arms.
The DC voltage out is 0.9 x the secondary rms rated voltage. If you want 10VDC, you better have a secondary rated voltage of 10 V/0.9 = 11.1 Vrms.
If you want 10VDC at 1A out, and you don't know better, you calculate that you need 10V x 1 A = 10VA rated transformer, but what you really need is 11.1V x 1.6 A = 18 VA.
You may want to use one of the other configurations shown. Make the appropriate changes to the calcs...
You have to decide what your load is before you calculate the required transformer size. If you were going into production and needed to keep costs down to be competitive, you'd want to be pretty accurate about the calculation so you could use the absolute minimum sized transformer to do the job. Since you are making one unit, and the price difference between a 50VA and a 100VA transformer is pretty small, why not just put in the bigger transformer?
In the above calculation you see that a 10 watt DC load requires about a 18VA transformer. That's about double the DC power. So you can estimate the required power then just double it to size the transformer and you should be in pretty good shape. That will take care of core losses and power dissipated in the rectifiers.
For regulated supplies, transformer selection is more critical. See here: http://tangentsoft.net/elec/ps-est.html
I_F
http://www.hammondmfg.com/pdf/5c007.pdf
See the capacitor input filter and load near the bottom on the right side? Notice that the DC output current is only 0.62 x the secondary AC current rating. That means that if you need 1 A DC out, your secondary current should be spec'd at 1 A/0.62 = 1.6 Arms.
The DC voltage out is 0.9 x the secondary rms rated voltage. If you want 10VDC, you better have a secondary rated voltage of 10 V/0.9 = 11.1 Vrms.
If you want 10VDC at 1A out, and you don't know better, you calculate that you need 10V x 1 A = 10VA rated transformer, but what you really need is 11.1V x 1.6 A = 18 VA.
You may want to use one of the other configurations shown. Make the appropriate changes to the calcs...
You have to decide what your load is before you calculate the required transformer size. If you were going into production and needed to keep costs down to be competitive, you'd want to be pretty accurate about the calculation so you could use the absolute minimum sized transformer to do the job. Since you are making one unit, and the price difference between a 50VA and a 100VA transformer is pretty small, why not just put in the bigger transformer?
In the above calculation you see that a 10 watt DC load requires about a 18VA transformer. That's about double the DC power. So you can estimate the required power then just double it to size the transformer and you should be in pretty good shape. That will take care of core losses and power dissipated in the rectifiers.
For regulated supplies, transformer selection is more critical. See here: http://tangentsoft.net/elec/ps-est.html
I_F
Hi,
Its as I stated, for music VA transformer ~= clipping RMS power output.
As music has a peak to mean ratio much higher than sine waves so
unless your listening to heavily compressed lift music or death metal
the above rule specifies the minimum you need.
Extra capacitance above say 4,700uF (8R) to 6,800uF (4R) won't help much.
/sreten.
Its as I stated, for music VA transformer ~= clipping RMS power output.
As music has a peak to mean ratio much higher than sine waves so
unless your listening to heavily compressed lift music or death metal
the above rule specifies the minimum you need.
Extra capacitance above say 4,700uF (8R) to 6,800uF (4R) won't help much.
/sreten.I_Forgot said:
Thanks, IF. I'm actually making a few (lotsa kids in the family and they may be Xmas gifts
) and it's a question of packaging...it's all gotta fit under (inside) an appropriate speaker cabinet for just a single 5" woofer and leave reasonable space for the LM3875 and PSU. Thanks to all you guys that responded.
Again...basic idea is just one LM3875 being driven to put out about 30-40W to a single 5" driver. I'm shooting for low voltage...maybe like 18-25V DC to the rails. What is the minimum VA that'll work OK? It seems that the answer is about 60VA or so. The Bicron units of this VA rating have a regulation % of 12-13%.
Hi,
All the following apply to ClassAB and ClassB amplifiers i.e. the vast majority.
The MINIMUM VA that will work is probably of the order of half the maximum output power. The performance is likely to be terrible.
A sensible minimum would be VA = maximum output power but the performance will be compromised. This will give a "cheap" amplifier.
The optimum VA ~=1.5 times maximum output power. This gives good performance and good value for money.
VA>1.6times maximum output power generally gives only small or very small gain in performance and value for money decreases rapidly unless a cheap or secondhand transformer becomes available.
Be carefull to check the full range of voltages that can come out of your PSU, particularly if you hang another voltage limited IC across the rails.
All the following apply to ClassAB and ClassB amplifiers i.e. the vast majority.
The MINIMUM VA that will work is probably of the order of half the maximum output power. The performance is likely to be terrible.
A sensible minimum would be VA = maximum output power but the performance will be compromised. This will give a "cheap" amplifier.
The optimum VA ~=1.5 times maximum output power. This gives good performance and good value for money.
VA>1.6times maximum output power generally gives only small or very small gain in performance and value for money decreases rapidly unless a cheap or secondhand transformer becomes available.
Be carefull to check the full range of voltages that can come out of your PSU, particularly if you hang another voltage limited IC across the rails.
AndrewT said:
Guys...guys...your inputs are greatly appreciated and actually I thought that you're both converging around the same answer...1.5 times max power output.
I found a nice 90VA 20V+20V with 7% regulation Avel Lindberg on the cheap and that seems to meet the 1.5 times criteria for a single LM3875 driven moderately. Do you agree?
If the 3875 GC kit guide recommends a 160VA or higher transformer, it was based not on numbers but actual listening tests. PS is very critical to this amp design and the recommendation was given for a reason. Personally, I wouldn't go less than 200VA per channel.
Of course, the amp will work with 40VA transformer, but I thought you approaching "this GC LM3875 as art..."
Of course, the amp will work with 40VA transformer, but I thought you approaching "this GC LM3875 as art..."
AndrewT said:
Hi AT,
I thought you'd read Self's Power Amplifier Design book ?
If you have/do, then you'd understand why I'm not attempting
to go through Selfs reasoning other than stating that for music
replay as opposed to steady state sine waves at full power,
VA = clipping power level is a good commercial choice.
(That is : instead of oversizing the power supply build a more
powerful amplifier with that supply simply because you can.)
For active speakers in particular you don't have to generalise
the load, and say always design for 4ohm power output. If
the drivers are 8 ohm then VA = 8 ohm power RMS clipping.
In the context of this thread where in the application space is
tight I'm just being insistent on what is engineeringly required.
The VA needed without major compromise is the designed power
output, in this case 30VA to 40VA, with a decent capacitance level.
On music signals the continuous average power output even
with excessive levels of clipping should not exceed 10W, which
is not a major strain on a 30VA supply.
30W continuous sine waves for hours are a very different matter.
/sreten.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.