Hello,
I've been repairing amplifiers for a long time now and they always seem to run from -+50 Volt or so power rails
I also built amplifiers from kits and use +-25 as a test supply, but to my surprise i don't feel much need for higher voltages.
If i really needed more power i would just resort to bi-amping or tri-amping, which is actually more efficient than one big powerful amplifier.
At lower voltages:
Am i overlooking something or are higher voltage rails only used for max power?
I've been repairing amplifiers for a long time now and they always seem to run from -+50 Volt or so power rails
I also built amplifiers from kits and use +-25 as a test supply, but to my surprise i don't feel much need for higher voltages.
If i really needed more power i would just resort to bi-amping or tri-amping, which is actually more efficient than one big powerful amplifier.
At lower voltages:
- Choice of transistors increases
- Capacitor cost could be reduced
- Less power draw at idle
Am i overlooking something or are higher voltage rails only used for max power?
Member
Joined 2009
Paid Member
The obvious comment is that many people don’t want to bi or tri amp. A single high power amp is then preferred.
Even at lower powers using high voltage rails is still a good idea, the only tradeoff is extra cost and heat. The advantages depend on details but may include
Better linearity for the active devices
Higher voltage caps canhave thicker oxides and be of better quality
More headroom so less risk of clipping on dynamic passages
Even at lower powers using high voltage rails is still a good idea, the only tradeoff is extra cost and heat. The advantages depend on details but may include
Better linearity for the active devices
Higher voltage caps canhave thicker oxides and be of better quality
More headroom so less risk of clipping on dynamic passages
So it seems, but the system as a whole would greatly benefit from that. One could choose the optimum amplifier for the woofer and perhaps a small class A for the high frequencies
Isn't this only true if you really go low on the supply voltage? As in without modification of the circuit.
Higher voltage caps tend to store more energy per unit of volume than lower voltage ones. I think ideally amps would run direct from rectified mains (power factor corrected at higher powers) and use an output transformer to transform impedance and provide safety isolation. An input trafo would be needed too which met mains isolation requirements.
Here are some 400V/68uF caps which store ~5J each, they're 10mm*50mm high. To store the same 5J but at 50V needs 4700uF which is 22mm*35mm, around 3X greater in volume.
????LED???? ???????? 400V68UF 68UF400V 10X50-???
Here are some 400V/68uF caps which store ~5J each, they're 10mm*50mm high. To store the same 5J but at 50V needs 4700uF which is 22mm*35mm, around 3X greater in volume.
????LED???? ???????? 400V68UF 68UF400V 10X50-???
Electrolytic capacitor volume is proportional to the product of capacitance and voltage, CV.
The energy stored is 1/2(CV^2).
Then the energy per volume that the capacitor can store is proportional to the applied voltage,
since (CV^2)/CV = V
Last edited:
The public, in general, doesn't deal with exponential scales well. If amps were rated in dBm instead of Watts, people would take a 20Watt amp (+43dBm) seriously. Most people expect 100 Watts (+50dBm), but as you can see 7dB is not a big difference. But to get 100 Watts, you need +/-43VDC ( 8 Ohms), probably +/-50VDC if the output saturation voltage is not well managed. To get 20W@8 Ohms, you only need +/- 18VDC. High voltage has been a problem for BJT amps because their SOA falls off sharply above about 15VDC, resulting in secondary breakdown. There are other problems with old BJTs which meant the failure rate was high, but today we have MOS and better BJTs. But too little too late since consumer products are moving to very cheap class-D on a chip products.
Exactly my point, and if that wasn't a problem: What about the loudspeaker itself? I've been building speakers long enough to know that up to a few watts speakers are reasonably easy too handle, but once the power increases to the speaker (and therefore cone excursion) control with just one amp becomes messy.
Overall i feel that just increasing power is not the way forward, and meanwhile class-d seems to have filled in the multi-amp scenario, but i am not happy with class-d amplifiers at all.
I find it strange that solid state amplifiers have not adapted more to the current possibilities and technology. Looking at class D, they mostly run at low voltages.
The only multi-amp scenario consumers got was surround amplifiers, hardly a benefit for most who just want to enjoy music.
I've converted quite a few of those surround amplifiers into tri-amp systems but the bias was always too low (because of poor cooling) and that is where i thought, why have a 50V supply rails anyway?
Last edited:
Some music has very high transients.
To stop them getting clipped at high volume requires a high voltage supply.
My first amp was 30 volts and clipped very quickly and badly.
The next amp was +/-55VDC and was very good even ta high volumes.
To stop them getting clipped at high volume requires a high voltage supply.
My first amp was 30 volts and clipped very quickly and badly.
The next amp was +/-55VDC and was very good even ta high volumes.
Speaker technology improvement has also made lower rail voltage amps viable, but they may not have sold. The standard speaker impedance 1960-1980 was 8 ohms, which at 200 W required 40 v out. To get that reliably from junction transistors took ~80 v rails.
Improvement in magnet material made 4 ohm speakers quite acceptable in the eighties. 200 W speaker at 4 ohm is 28 v requiring ~56 v rail. However to maintain compatibility with existing speakers amps needed a high wattage 8 ohm rating, until very recently. I'm still using 8 ohm speakers.
Anybody that has put a music system in a bar or auditorium knows that 30 W systems don't cover crowd noise. Bogen sold a lot of 30 W systems that were suitable only in a school environment where quiet is enforced by the faculty. The B3 organ tone cabinet and the Leslie rotary speaker were about 40 W, and both were often miked and relayed through the house sound system.
100 w is Especially needed in a room full of coats in winter. 150-200 W/channel is about the next sales point up the curve. That power is where low end powered mixer's sell quite well. Hammond went to 70 W amps in 1968 with the H model, with a second and third "tone cabinet" available with also 70 w for auditoriums holding 100 or more. Rodgers & Allen organ builders went to 2 channels of 100 W amps about 1968 as their minimum system. These were for quiet church crowds.
There is another sales point at 500-800 w/ch, for audiences in the hundreds. With the inflation in wattage rating common now, with many 2000 w amps blowing up if delivering 500 w rms 8 hours at a time, the sale of 4000 w amps has become quite common. But they are class D or T.
Improvement in magnet material made 4 ohm speakers quite acceptable in the eighties. 200 W speaker at 4 ohm is 28 v requiring ~56 v rail. However to maintain compatibility with existing speakers amps needed a high wattage 8 ohm rating, until very recently. I'm still using 8 ohm speakers.
Anybody that has put a music system in a bar or auditorium knows that 30 W systems don't cover crowd noise. Bogen sold a lot of 30 W systems that were suitable only in a school environment where quiet is enforced by the faculty. The B3 organ tone cabinet and the Leslie rotary speaker were about 40 W, and both were often miked and relayed through the house sound system.
100 w is Especially needed in a room full of coats in winter. 150-200 W/channel is about the next sales point up the curve. That power is where low end powered mixer's sell quite well. Hammond went to 70 W amps in 1968 with the H model, with a second and third "tone cabinet" available with also 70 w for auditoriums holding 100 or more. Rodgers & Allen organ builders went to 2 channels of 100 W amps about 1968 as their minimum system. These were for quiet church crowds.
There is another sales point at 500-800 w/ch, for audiences in the hundreds. With the inflation in wattage rating common now, with many 2000 w amps blowing up if delivering 500 w rms 8 hours at a time, the sale of 4000 w amps has become quite common. But they are class D or T.
Last edited:
One way to add inherent protection to amp is to use higher voltage supply, with less reservoir capacity and appropriate transformer, and to expect the rail voltage to sag under load. This allows for large fast transients but limits the SOAR stress under heavy load or when bogging down the output stage.
Very good point, that is definitely something commercial manufacturers would have to take into consideration since the end user might have a big room and/or low efficiency speakers.
Out of curiosity, what sort of size room and what music are we talking about?
Years ago I worked in pro audio, and one thing we noticed was that different amps sounded very different when over driven. At the time we called it "overload recovery". I have seen it call "rail sticking" here in DIYA. I had a modification that I applied to the Yamaha P2200 on request that reduced the transient that appeared following positive clipping. In retrospect, I probably should have done it differently, but it worked well enough.
But many amps had worse problems, especially with VI output protection, which busts into oscillations when the amp protection is triggered. My point is that clipping transients on a good amp is almost imperceptible, and only get nasty when seriously over driven. If you focus on the THD at 10 Watts, it's easy to overlook some very bad behavior. Every amp has a limit and the difference between them is only a few dB. But the sound at high levels can be very different, and the lower-power amp may have a higher useful output capability. Whenever I simulate an amp, I always spend a lot of time looking at the clipping behavior. In most cases, clamping diodes in the right place prevents early stages from swinging wildly out of range when the following stages reach their limit first, cutting off feedback. When the input stage looses feedback, it swings towards it's own clipping value and takes time (sticking) to return to the linear range when feedback returns.
But many amps had worse problems, especially with VI output protection, which busts into oscillations when the amp protection is triggered. My point is that clipping transients on a good amp is almost imperceptible, and only get nasty when seriously over driven. If you focus on the THD at 10 Watts, it's easy to overlook some very bad behavior. Every amp has a limit and the difference between them is only a few dB. But the sound at high levels can be very different, and the lower-power amp may have a higher useful output capability. Whenever I simulate an amp, I always spend a lot of time looking at the clipping behavior. In most cases, clamping diodes in the right place prevents early stages from swinging wildly out of range when the following stages reach their limit first, cutting off feedback. When the input stage looses feedback, it swings towards it's own clipping value and takes time (sticking) to return to the linear range when feedback returns.
Some very accomplished r&d folks/engineers I know just swear by higher voltage, common pa amps for home use, sneer at the audiophool stuff. For the most part I can’t argue with them!
This thread is making me more motivated to resurrect my hafler amp that uses 93v rails.
This thread is making me more motivated to resurrect my hafler amp that uses 93v rails.
- Status
- Not open for further replies.